Altered dietary salt intake for preventing diabetic kidney disease and its progression

Abstract

Background

There is strong evidence that our current consumption of salt is a major factor in the development of increased blood pressure (BP) and that a reduction in our salt intake lowers BP, whether BP levels are normal or raised initially. Effective control of BP in people with diabetes lowers the risk of strokes, heart attacks and heart failure and slows the progression of chronic kidney disease (CKD) in people with diabetes. This is an update of a review first published in 2010.

Objectives

To evaluate the effect of altered salt intake on BP and markers of cardiovascular disease and of CKD in people with diabetes.

Search methods

We searched the Cochrane Kidney and Transplant Register of Studies up to 31 March 2022 through contact with the Information Specialist using search terms relevant to this review. Studies in the Register were identified through searches of CENTRAL, MEDLINE, and EMBASE, conference proceedings, the International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.

Selection criteria

We included randomised controlled trials (RCTs) of altered salt intake in individuals with type 1 and type 2 diabetes. Studies were included when there was a difference between low and high sodium intakes of at least 34 mmol/day.

Data collection and analysis

Two authors independently assessed studies and resolved differences by discussion. We calculated mean effect sizes as mean difference (MD) and 95% confidence intervals (CI) using the random‐effects model. Confidence in the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

Main results

Thirteen RCTs (313 participants), including 21 comparisons (studies), met our inclusion criteria. One RCT (two studies) was added to this review update. Participants included 99 individuals with type 1 diabetes and 214 individuals with type 2 diabetes. Two RCTs (four studies) included some participants with reduced overall kidney function. The remaining studies either reported that participants with reduced glomerular filtration rate (GFR) were excluded from the study or only included participants with microalbuminuria and normal GFR. Five studies used a parallel study design, and 16 used a cross‐over design. Studies were at high risk of bias for most criteria. Random sequence generation and allocation concealment were adequate in only three and two studies, respectively. One study was at low risk of bias for blinding of participants and outcome assessment, but no studies were at low risk for selective reporting. Twelve studies reported non‐commercial funding sources, three reported conflicts of interest, and eight reported adequate washout between interventions in cross‐over studies.

The median net reduction in 24‐hour urine sodium excretion (24‐hour UNa) in seven long‐term studies (treatment duration four to 12 weeks) was 76 mmol (range 51 to 124 mmol), and in 10 short‐term studies (treatment duration five to seven days) was 187 mmol (range 86 to 337 mmol). Data were only available graphically in four studies. In long‐term studies, reduced sodium intake may lower systolic BP (SBP) by 6.15 mm Hg (7 studies: 95% CI ‐9.27 to ‐3.03; I² = 12%), diastolic BP (DBP) by 3.41 mm Hg (7 studies: 95% CI ‐5.56 to ‐1.27; I² = 41%) and mean arterial pressure (MAP) by 4.60 mm Hg (4 studies: 95% CI ‐7.26 to ‐1.94; I² = 28%). In short‐term studies, low sodium intake may reduce SBP by 8.43 mm Hg (5 studies: 95% CI ‐14.37 to ‐2.48; I² = 88%), DBP by 2.95 mm Hg (5 studies: 95% CI ‐4.96 to ‐0.94; I² = 70%) and MAP by 2.37 mm Hg (9 studies: 95% CI ‐4.75 to ‐0.01; I² = 65%). There was considerable heterogeneity in most analyses but particularly among short‐term studies. All analyses were considered to be of low certainty evidence.

SBP, DBP and MAP reductions may not differ between hypertensive and normotensive participants or between individuals with type 1 or type 2 diabetes. In hypertensive participants, SBP, DBP and MAP may be reduced by 6.45, 3.15 and 4.88 mm Hg, respectively, while in normotensive participants, they may be reduced by 8.43, 2.95 and 2.15 mm Hg, respectively (all low certainty evidence). SBP, DBP and MAP may be reduced by 7.35, 3.04 and 4.30 mm Hg, respectively, in participants with type 2 diabetes and by 7.35, 3.20, and 0.08 mm Hg, respectively, in participants with type 1 diabetes (all low certainty evidence).

Eight studies provided measures of urinary protein excretion before and after salt restriction; four reported a reduction in urinary albumin excretion with salt restriction. Pooled analyses showed no changes in GFR (12 studies: MD ‐1.87 mL/min/1.73 m², 95% CI ‐5.05 to 1.31; I² = 32%) or HbA1c (6 studies: MD ‐0.62, 95% CI ‐1.49 to 0.26; I² = 95%) with salt restriction (low certainty evidence). Body weight was reduced in studies lasting one to two weeks but not in studies lasting for longer periods (low certainty evidence). Adverse effects were reported in only one study; 11% and 21% developed postural hypotension on the low‐salt diet and the low‐salt diet combined with hydrochlorothiazide, respectively.

Authors' conclusions

This systematic review shows an important reduction in SBP and DBP in people with diabetes with normal GFR during short periods of salt restriction, similar to that obtained with single drug therapy for hypertension. These data support the international recommendations that people with diabetes with or without hypertension or evidence of kidney disease should reduce salt intake to less than 5 g/day (2 g sodium).

Plain language summary

Does reducing your salt intake help to prevent and treat chronic kidney disease in people with diabetes?

What is the issue?

There is strong evidence that we all eat too much salt, which increases our risk of high blood pressure (BP). This is particularly important in people with diabetes as diabetes increases the risk of stroke, heart attack and kidney failure, and also having high BP will further increase these risks. Reducing salt intake could help by reducing BP and thus reducing the risk of heart attacks and of worsening kidney function.

What did we do?

We searched the Cochrane Kidney and Transplant Register of Studies up to 31 March 2022 for randomised controlled trials which compared low and high levels of salt intake in people with diabetes. We calculated the average level of reduction in systolic BP (the "top" level of BP measured) and in diastolic BP (the "bottom" level of BP measured) in diabetics when they received a high salt diet and when they received a low salt diet. We also looked at whether the amount of protein in the urine (a marker of kidney damage) was reduced in diabetics receiving a low salt diet.

What did we find?

We found 13 studies, including 313 people with type 1 or 2 diabetes. We found that reducing salt intake by an average of 5 g/day lowered BP, with systolic BP reduced by 7 mm of mercury (Hg) and diastolic BP reduced by 3 mm Hg. We found that the amount of protein in the urine was reduced in four of the eight studies that reported this outcome. Only one study reported side effects with low BP when standing up with low salt diets reported in a quarter of the participants.

Conclusions

Lowering dietary salt intake to the recommended levels of 5 g/day or less would benefit people with diabetes by lowering BP by similar amounts to a single BP medication.

Summary of findings

Summary of findings 1. Reduced salt intake versus usual or high salt intake for preventing diabetic kidney disease and its progression.

Reduced salt intake versus usual or high salt intake for preventing diabetic kidney disease and its progression
Patient or population: patients with diabetes with or without CKD Setting: inpatients and outpatients Intervention: reduced salt intake Comparison: usual or high salt intake
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (RCTs)	Certainty of the evidence (GRADE)
	Risk with usual or high salt intake	Risk with reduced salt intake
Systolic BP: long‐term studies Duration: up to 12 weeks	The mean systolic BP was 6.15 mm Hg lower with a reduced salt intake (9.27 lower to 3.03 lower) compared to a usual or high salt intake		‐	288 (7)	⊕⊕⊝⊝ LOW 1 2
Systolic BP: short‐term studies Duration: 1 week	The mean systolic BP was 8.43 mm Hg lower with a reduced salt intake (9.6 lower to 5.81 lower) compared to a usual or high salt intake		‐	112 (5)	⊕⊕⊝⊝ LOW 1 2
Diastolic BP: long‐term studies Duration: up to 12 weeks	The mean diastolic BP was 3.41 mm Hg lower with a reduced salt intake (5.56 lower to 1.27 lower) compared to a usual or high salt intake		‐	288 (7)	⊕⊕⊝⊝ LOW 1 2
Diastolic BP: short‐term studies Duration: 1 week	The mean diastolic BP was 2.95 mm Hg lower with a reduced salt intake (4.96 lower to 0.94 lower) compared to a usual or high salt intake		‐	112 (5)	⊕⊕⊝⊝ LOW 1 2
MAP: all studies Duration: up to 12 weeks	The mean MAP was 3.01 mm Hg lower with a reduced salt intake (4.95 lower to 1.07 lower) compared to a usual or high salt intake		‐	421 (13)	⊕⊕⊝⊝ LOW 1 2
GFF: all studies Duration: up to 12 weeks	The mean GFR was 1.78 mL/min/1.73 m² lower with a reduced salt intake (4.21 lower to 0.65 higher) compared to a usual or high salt intake		‐	392 (12)	⊕⊕⊝⊝ LOW 1 2
Body weight: all studies Duration: up to 12 weeks	The mean body weight was 1.21 kg lower with a reduced salt intake (1.73 lower to 0.68 lower) compared to a usual or high salt intake		‐	454 (12)	⊕⊕⊝⊝ LOW 1 2
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI) CKD: Chronic kidney disease; CI: Confidence interval; RCT: Randomised controlled trial; BP: Blood pressure; MD: Mean difference; MAP: Mean arterial pressure; GFR: Glomerular filtration rate
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Heterogeneity between studies

² High or unclear risk of bias for allocation concealment and blinding due to study design (cross‐over studies) with small numbers of enrolled participants

Summary of findings 2. Net change in blood pressure in hypertensive and normotensive participants with a reduced salt intake versus usual or high salt intake for preventing diabetic kidney disease and its progression.

Net change in BP in hypertensive and normotensive participants with a reduced salt intake versus usual or high salt intake for preventing diabetic kidney disease and its progression
Patient or population: patients with diabetes with or without CKD Setting: inpatients and outpatients Intervention: reduced salt intake Comparison: usual or high salt intake
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (RCTs)	Certainty of the evidence (GRADE)
	Risk with usual or high salt intake	Risk with reduced salt intake
Systolic BP Duration: up to 12 weeks	The mean systolic BP in hypertensive patients was 6.45 mm Hg lower with a reduced salt intake (11.69 lower to 3.61 lower) compared to a usual or high salt intake		‐	108 (5)	⊕⊕⊝⊝ LOW 1 2
	The mean systolic BP in normotensive patients was 8.43 mm Hg lower with a reduced salt intake (9.6 lower to 5.81 lower) compared to a usual or high salt intake		‐	108 (5)	⊕⊕⊝⊝ LOW 1 2
Diastolic BP Duration: up to 12 weeks	The mean diastolic BP in hypertensive patients was 3.15 mm Hg lower with a reduced salt intake (6.49 lower to 0.18 lower) compared to a usual or high salt intake		‐	108 (5)	⊕⊕⊝⊝ LOW 1 2
	The mean diastolic BP in normotensive patients was 2.95 mm Hg lower with a reduced salt intake (4.11 lower to 2 lower) compared to a usual or high salt intake		‐	108 (5)	⊕⊕⊝⊝ LOW 1 2
MAP Duration: up to 12 weeks	The mean MAP in hypertensive patients was 4.88 mm Hg lower with a reduced salt intake (10.39 lower to 0.63 lower) compared to a usual or high salt intake		‐	59 (3)	⊕⊕⊝⊝ LOW 1 2
	The MAP in normotensive patients was 2.15 mm Hg lower with a reduced salt intake (4.56 lower to 0.26 lower) compared to a usual or high salt intake		‐	182 (8)	⊕⊕⊝⊝ LOW 1 2
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BP: Blood pressure; CKD: Chronic kidney disease; CI: Confidence interval; MD: Mean difference; MAP: Mean arterial pressure
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Heterogeneity between studies

² High or unclear risk of bias for allocation concealment and blinding due to study design (cross‐over studies) with small numbers of enrolled participants

Summary of findings 3. Net change in blood pressure in participants with type 1 or type 2 diabetes with a reduced salt diet versus usual or high salt intake for preventing diabetic kidney disease and its progression.

Net change in BP in participants with type 1 and type 2 diabetes with a reduced salt intake versus usual or high salt intake for preventing diabetic kidney disease and its progression
Patient or population: patients with diabetes with or without CKD Setting: inpatients and outpatients Intervention: reduced salt intake Comparison: usual or high salt intake
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No. of participants (RCTs)	Certainty of the evidence (GRADE)
	Risk with usual or high salt intake	Risk with reduced salt intake
Systolic BP Duration: up to 12 weeks	The mean systolic BP in patients with type 1 diabetes was 7.35 mm Hg lower with a reduced salt intake (14.49 lower to 0.21 lower) compared to a usual or high salt intake		‐	96 (4)	⊕⊕⊝⊝ LOW 1 2
	The mean systolic BP in patients with type 2 diabetes was 7.35 mm Hg lower with a reduced salt intake (10.32 lower to 4.38 lower) compared to a usual or high salt intake		‐	304 (8)	⊕⊕⊝⊝ LOW 1 2
Diastolic BP: Duration: up to 12 weeks	The mean diastolic BP in patients with type 1 diabetes was 3.20 mm Hg lower with a reduced salt intake (5.16 lower to 1.23 lower) compared to a usual or high salt intake		‐	96 (4)	⊕⊕⊝⊝ LOW 1 2
	The mean diastolic BP in patients with type 2 diabetes was 3.04 mm Hg lower with a reduced salt intake (5.20 lower to 0.89 lower) compared to a usual or high salt intake		‐	304 (8)	⊕⊕⊝⊝ LOW 1 2
MAP Duration: up to 12 weeks	The mean MAP in patients with type 1 diabetes was 0.08 mm Hg higher with a reduced salt intake (1.92 lower to 2.08 higher) compared to a usual or high salt intake		‐	62 (3)	⊕⊕⊝⊝ LOW 1 2
	The mean MAP in patients with type 2 diabetes was 4.03 mm Hg lower with a reduced salt intake (6.54 lower to 2.05 lower) compared to a usual or high salt intake		‐	359 (10)	⊕⊕⊝⊝ LOW 1 2
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BP: Blood pressure; CKD: Chronic kidney disease; CI: Confidence interval; MD: Mean difference; MAP: Mean arterial pressure
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

¹ Heterogeneity between studies

² High or unclear risk of bias for allocation concealment and blinding due to study design (cross‐over studies) with small numbers of enrolled participants

Background

Description of the condition

Approximately 75% of cardiovascular disease (CVD) in people with diabetes can be attributed to raised blood pressure (BP), and it is the largest direct cause of death due to stroke, heart attack and heart failure in people with diabetes (Adler 2000; Mozaffarian 2014; Sowers 2001; Strain 2018). Chronic kidney disease (CKD) associated with diabetes is responsible for the largest group on dialysis programmes, with a significant burden on healthcare resources (Harvey 2003). Lowering BP reduces the progression of CKD in people with diabetes, and diabetic patients with BP less than 130/80 mm Hg have the same reduction in glomerular filtration rate (GFR) due to age as healthy individuals (Giunti 2006; Schrier 2002).

Description of the intervention

In an analysis for the Global Burden of Disease Study (GBD 2017), a high salt intake was determined to be one of the main factors leading to increased death and morbidity among people worldwide due largely to CVD. Reducing salt intake reduces systolic BP (SBP) and diastolic BP (DBP) in adults and children (WHO 2012). Therefore, WHO recommends a reduction in salt intake to below 5 g/day (< 2 g/day sodium) in adults to reduce BP and the risk of CVD, stroke and coronary heart disease (WHO 2012). In agreement with the WHO, the KDIGO guidelines (KDIGO 2020) on diabetes management in CKD recommend salt intake should be below 5 g/day (< 2 g of sodium) in people with diabetes and CKD to potentiate the suppression of the renin‐angiotensin system (RAS) by angiotensin‐converting enzyme inhibitors (ACEi) or angiotensin receptors blockers (ARB). RAS blockade is considered essential to delay the progression of CKD in people with diabetes. However, cardiovascular and kidney protection by RAS blockade may not be complete. Therefore, additional strategies to improve the efficacy of RAS blockade, including dietary salt restriction, may be useful in people with diabetes to potentiate RAS blockade efficacy.

How the intervention might work

Diabetic patients differ from the nondiabetic population in having an increase in total body sodium, an increase in renal tubular sodium reabsorption, and an impaired ability to excrete a sodium load. These factors suggest that reducing dietary salt intake may play an important role in the management of hypertension in the diabetic population (Houlihan Losartan 2002; Houlihan Placebo 2002). The activity of the sodium‐glucose co‐transporter increases proximal tubular reabsorption of sodium, and this reduces distal sodium delivery (Woods 1987). Insulin directly increases distal tubular sodium reabsorption (DeFronzo 1981). Furthermore, increased angiotensin II in the tubular fluid activates sodium channels in the collecting duct (Nishiyama 2002). People with diabetes have a lower incidence of low renin state and a higher plasma renin activity on a high salt diet than volunteers. This lack of suppression of the RAS with a high salt diet may play a part in the development of hypertension (De'Oliveira 1997; Price 1999) in people with diabetes. Lowering salt intake also reduces urinary protein excretion in several populations, and increased urinary protein excretion is an important marker of CKD and CVD in people with diabetes (Allen 1997; Cianciaruso 1998; He 2009a; Swift 2005).

Why it is important to do this review

There is strong evidence that our current consumption of salt is a major factor in increasing BP (GBD 2017; He 2009b). Despite the high cardiovascular risk and theoretical reasons for increased salt sensitivity in people with diabetes, the current knowledge of the role of salt in regulating BP and its impact on people with diabetes is limited to small studies. Therefore, we carried out a systematic review of all studies where the salt intake was altered in people with diabetes. There are no standard criteria for high and low salt diets in people with diabetes. As with previous meta‐analyses of salt reduction, this review included studies where the difference in sodium intake between periods of high and low sodium intakes was at least 34 mmol/day (He 2002; He 2003; Law 1991). This is an update of a review first published in 2010.

Objectives

We aimed to evaluate the effects of salt restriction on the prevention and progression of CKD in people with diabetes by studying the effects of changing salt intake on BP, urinary protein excretion and surrogate markers of CVD and CKD.

Methods

Criteria for considering studies for this review

Types of studies

For inclusion, studies needed to satisfy the following inclusion criteria. 

• Randomised controlled trials (RCT) with random allocation to either a low or a high salt intake.

• Sodium intake was estimated by 24‐hour urinary sodium excretion (24‐hour UNa), and studies achieved a minimum difference in 24‐hour UNa of 34 mmol (i.e. 2 g salt/day). The reduction in 24‐hour UNa was calculated as UNa (high) ‐ UNa (low) in cross‐over studies and was calculated as (UNa (high) ‐ UNa (low)) low salt group ‐ (UNa (high) ‐ UNa (low)) control group in parallel studies. For cross‐over studies, UNa (low) was the 24‐hour UNa at the end of the reduced salt period, and the UNa (high) was the 24‐hour UNa at the end of the high salt period. For parallel studies, UNa (high) and UNa (low) were the 24‐hour UNa measurements at baseline and at the end of the intervention for each group.

• Studies where concomitant interventions, such as antihypertensive medication or other dietary modifications, were used during the study period were included providing that these interventions were constant throughout the low and high salt period.

Types of participants

We included all RCTs of adults (18 years or older) with type 1 or type 2 diabetes mellitus, irrespective of ethnicity or gender. We excluded studies of children and pregnant women.

Types of interventions

We included studies where a low salt intake was compared to a high salt intake with a minimum difference in 24‐hour UNa, as described above.

Types of outcome measures

The updated primary and secondary outcomes aimed to cover all clinically relevant and critical outcomes according to the recent KDIGO guidelines (KDIGO 2020).

Primary outcomes

1. Net change in 24‐hour UNa

2. Net change in 24‐hour BP (SBP, DBP, mean arterial pressure (MAP))

3. Death (any cause)

Secondary outcomes

1. Functional kidney measurements: urinary albumin excretion (UAE), proteinuria, serum creatinine (SCr), creatinine clearance (CrCl), estimated (e) GFR, effective renal plasma flow (ERPF)

2. Glycated haemoglobin (HbA1c), hypoglycaemia

3. Body weight: body weight, body mass index (BMI)

4. CVD events: stroke, heart failure, myocardial infarction

5. Adverse events

Search methods for identification of studies

Electronic searches

We searched the Cochrane Kidney and Transplant Register of Studies up to 31 March 2022 through contact with the Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources:

1. Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)

2. Weekly searches of MEDLINE OVID SP

3. Handsearching of kidney‐related journals and the proceedings of major kidney conferences

4. Searching the current year of EMBASE OVID SP

5. Weekly current awareness alerts for selected kidney and transplant journals

6. Searches of the International Clinical Trials Register (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of search strategies, as well as a list of handsearched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website under CKT Register of Studies.

See Appendix 1 for search terms used in strategies for this review.

Searching other resources

1. Reference lists of review articles and relevant studies.

2. Letters seeking information about unpublished or incomplete studies to investigators known to be involved in previous studies.

Data collection and analysis

Selection of studies

Two review authors independently undertook an assessment of citations. We retrieved full articles if the initial assessment suggested that studies or reviews might include relevant data. We independently assessed abstracts and, if necessary, the full text to identify suitable studies. We resolved disagreements through discussion between the authors.

Data extraction and management

We extracted data using standard data extraction forms. We recorded relevant data, including characteristics of the study, design (parallel or cross‐over), type of study (open, single‐blind, double‐blind), method of blinding, study duration and pre‐ and post‐intervention results. Where more than one publication of a study existed, we used the publication with the most complete data.

Assessment of risk of bias in included studies

The following items were assessed independently by two authors using the risk of bias assessment tool (Higgins 2022) (see Appendix 2).

• Was there adequate sequence generation (selection bias)?

• Was allocation adequately concealed (selection bias)?

• Was knowledge of the allocated interventions adequately prevented during the study?

  • Participants and personnel (performance bias)

  • Outcome assessors (detection bias)

• Were incomplete outcome data adequately addressed (attrition bias)?

• Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?

• Was the study apparently free of other problems that could put it at risk of bias?

Measures of treatment effect

We calculated treatment effects for each study. We calculated the mean difference (MD) for outcome measurements with 95% confidence intervals (CI). In cross‐over studies, we calculated the MD in outcomes as the difference between the end of the low salt period and the end of the high salt period. For parallel studies, we calculated the treatment effect as the difference between the two treatment groups in the change in outcomes from baseline.

Unit of analysis issues

The majority of included studies were cross‐over studies, providing separate data for each treatment period's end. We calculated the treatment effect for cross‐over studies as the MD between the results at the end of the low salt period and those at the end of the high salt period. 

Dealing with missing data

We contacted the original authors when relevant data were not reported in the publication.

Assessment of heterogeneity

We first assessed the heterogeneity by visual inspection of the forest plot. We then quantified statistical heterogeneity using the I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than sampling error (Higgins 2003). A guide to the interpretation of I² values was as follows.

• 0% to 40%: might not be important

• 30% to 60%: may represent moderate heterogeneity

• 50% to 90%: may represent substantial heterogeneity

• 75% to 100%: considerable heterogeneity.

The importance of the observed value of I² depends on the magnitude and direction of treatment effects and the strength of evidence for heterogeneity (e.g. P value from the Chi² test or a CI for I²) (Higgins 2022).

Assessment of reporting biases

We planned to use funnel plots to assess the potential existence of small study bias (Higgins 2022). However, there were too few studies included in the review to allow assessment of reporting bias. 

Data synthesis

We pooled data using generic inverse variance with both fixed‐effect and random‐effects models to ensure the robustness of the model chosen and susceptibility to outliers. Data are presented using the random effects model.

Subgroup analysis and investigation of heterogeneity

Subgroup analyses were used to explore possible sources of heterogeneity.

• Long‐term (four to 12 weeks) versus short‐term studies (up to seven days)

• Studies of hypertensive versus normotensive participants

• Studies including type 1 versus type 2 diabetes participants

Heterogeneity between studies may be related to the types of participants. Heterogeneity between studies could be related to variations in the degree of sodium reduction, the different durations of the studies, differences between parallel group and cross‐over designs and the presence or absence of a washout period between interventions in cross‐over studies. 

Sensitivity analysis

We had planned to perform sensitivity analyses to explore the influence of the following factors on effect size. However, we could not undertake most of these analyses as outlined below.

• Repeating the analysis, excluding unpublished studies

• Repeating the analysis taking account of the risk of bias, as specified

• Repeating the analysis, excluding any very long or large studies, to establish how much they dominate the results.

Summary of findings and assessment of the certainty of the evidence

We presented the main results of the review in 'Summary of findings' tables. These tables present key information concerning the quality of the evidence, the magnitude of the effects of the interventions examined, and the sum of the available data for the main outcomes (Schunemann 2022a). The 'Summary of findings' tables also include an overall grading of the evidence related to each of the main outcomes using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (GRADE 2008; GRADE 2011). The GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. The quality of a body of evidence involves consideration of the within‐trial risk of bias (methodological quality), directness of evidence, heterogeneity, the precision of effect estimates and risk of publication bias (Schunemann 2022b). We presented the following outcomes in the 'Summary of findings' tables:

• Net changes in SBP, DBP, MAP, GFR and weight with reduced dietary salt in all participants

• Net changes in SBP, DBP and MAP with reduced dietary salt in hypertensive and normotensive participants

• Net changes in SBP, DBP and MAP with reduced dietary salt in type 1 and type 2 diabetes participants.

We had planned to include data on death and cardiovascular outcomes in the Summary of Findings Tables, but no data were available for these outcomes from included studies.

Results

Description of studies

Results of the search

The initial search in 2010 identified 914 citations, of which 72 reports were potentially eligible. After full‐text analysis, 13 RCTs (16 reports, 254 participants) were included, and 53 studies (56 reports) were excluded.

For this 2022 update, two additional RCTs were identified, of which one was excluded (Ushigome 2019) and one was included (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). Petrie 1998, which was included in the 2010 review, was reviewed and was excluded from this update as frusemide was only given to the high salt intake group and would have increased sodium excretion. Therefore, for this update, we included 13 studies (20 reports) with 313 participants were included in this review (Figure 1).

1.

Study flow diagram

Included studies

To avoid confusion, we used the term "study" for each entry throughout this meta‐analysis because some RCTs had two or more comparisons and have been split and entered more than once for the purpose of the meta‐analyses. There were 21 separate studies from 13 RCTs with 313 included participants.

• One article presented two separate studies (one with a cross‐over design and the other with a parallel design); we entered the data for these as separate studies in this meta‐analysis (Dodson_P 1989; Dodson_X 1989).

• Six studies (113 participants) used parallel design (De'Oliveira 1997; Dodson_P 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Miller 1997; Mulhauser 1996), and 15 studies (200 participants) used a cross‐over design.

• Seven studies (Lopes De Faria 1997; Luik 2002; Miller 1995; Miller 1997; Mulhauser 1996; Trevisan Micro 1998; Trevisan Normo 1998) enrolled 99 participants with type 1 diabetes, while 14 studies enrolled 214 participants with type 2 diabetes.

• Four studies (Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014) included some participants with reduced overall kidney function. The remaining 17 studies either reported that participants with reduced GFR were excluded from the study or only included participants with microalbuminuria and normal GFR (Appendix 3).

• Six studies enrolled participants with normal levels of albuminuria (Dodson_P 1989; Dodson_X 1989; Lopes De Faria 1997; Luik 2002; Miller 1995; Miller 1997). Three studies included only participants with microalbuminuria (Houlihan Losartan 2002; Houlihan Placebo 2002; Mulhauser 1996), and two studies included participants with micro‐ or macroalbuminuria (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). Nine studies presented data separately for participants with normal levels of albuminuria or with albuminuria (Imanishi Micro 2001; Imanishi Normo 2001; Trevisan Micro 1998; Trevisan Normo 1998; Vedovato Micro 2004; Vedovato Normo 2004; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

• Four studies compared salt reduction with normal salt intake in two groups of participants with type 2 diabetes and presented data for both categories separately.

  • Two studies (Houlihan Losartan 2002; Houlihan Placebo 2002) compared salt restriction with normal salt intake in two groups, one receiving losartan and one receiving a placebo.

  • Two studies (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014) compared salt restriction with normal salt intake in two groups, one receiving a placebo and one receiving hydrochlorothiazide, with all participants receiving maximal doses of lisinopril.

• The duration ranged from four to 12 weeks in seven studies and were considered to be long‐term studies (Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Mulhauser 1996). The remaining 14 studies were short‐term studies with treatment durations of less than two weeks.

• Six studies included 102 participants with high BP (> 140/90 mm Hg) (De'Oliveira 1997; Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Mulhauser 1996), and 13 studies included 166 participants with normal BP (< 140/90 mm Hg). Two studies (45 participants) included participants with BP ≤ 180/110 mm Hg (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014).

• UNa excretion was only available graphically in four studies (Imanishi Micro 2001; Imanishi Normo 2001; Vedovato Micro 2004; Vedovato Normo 2004).

• The age of participants ranged from 18 to 65 years.

We have summarised the studies' characteristics in the review in Appendix 3.

Outcome reporting

All studies reported BP measurements and 24‐hour UNa. There was much variability in reporting of measurements of kidney function and UAE. Studies used different isotopes to measure GFR and other haemodynamic measurements. UAE was reported in ways that could not be entered into a pooled analysis: two studies reported data as geometric mean and percentage change (Houlihan Losartan 2002; Houlihan Placebo 2002); two studies reported data as mean albumin excretion and as albumin:creatinine ratios (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014); one study reported data as a change in 24‐hour albumin excretion (Luik 2002); one study reported the mean level and range of urinary albumin (Mulhauser 1996); and two studies presented data as the median and interquartile range of albumin excretion (Vedovato Micro 2004; Vedovato Normo 2004). We have summarised the outcomes reported in each study in Appendix 4.

Excluded studies

Fourteen studies in people with diabetes were excluded. Six studies were excluded because they studied multiple interventions (DNETT Japan 2010; Dodson 1984; Ekinci 2009; Gilleran 1996; Helou 2016; HHK 2018) rather than altered salt intake. Three studies included some participants without diabetes, and the data for participants with or without diabetes could not be separated (LowSALT CKD 2012; Suckling 2016; ViRTUE‐CKD 2016). Two studies did not measure urinary sodium levels (Imanishi 1997; Ushigome 2019). One study included in the initial review was excluded in this update because additional frusemide was given to the high‐salt group but not the low‐salt group (Petrie 1998). PROCEED 2018 was excluded because the difference in 24‐hour UNa between high and low sodium intakes was less than 34 mmol/L.

Risk of bias in included studies

All included studies were small and of variable methodology resulting generally in high risk of bias (Figure 2; Figure 3)

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

Three studies were at low risk of bias for sequence generation. Low risk of bias was judged when the method of randomisation was reported as random number allocation in Luik 2002 and by computer randomisation in blocks of two (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). Sixteen studies were judged to be at unclear risk of bias as they provided no details about how the methods of randomisation were performed. Two studies were judged as having a high risk of bias because the dietician randomised the patients upon enrolment (Miller 1995; Miller 1997).

Allocation concealment

Two studies provided a description of allocation concealment by reporting that the randomisation code remained a secret during the entire trial and was judged to be at low risk of bias (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). One study was judged to be a high risk of bias (Lopes De Faria 1997). The remaining 18 studies did not provide information about allocation concealment and were judged to have an unclear risk of bias.

Blinding

Blinding of participants and personnel (performance bias)

Only Mulhauser 1996 adequately described the methods used for double‐blinding of both the patients and investigators and was judged to be at low risk of bias.

Three studies were stated to be double‐blind trials but failed to provide any further details about how the participants or personnel were kept blind to treatments, and so were judged to be an unclear risk of bias (Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002).

The remaining 17 studies were unblinded, open‐label studies and judged to be at high risk of bias. In Kwakernaak HTZ 2014 and Kwakernaak Placebo 2014, the dietary component was open‐label, but the hydrochlorothiazide and placebo were double‐blind, and for this reason, the study was judged at a high risk of bias.

Blinding of outcome assessors (detection bias)

Only Mulhauser 1996 adequately described the methods used to keep all investigators, and outcome assessors blind until the last patient had completed the study and was judged to be at low risk of bias.

Miller 1995 reported the investigators remained blind to the diets, but no further details were provided about how this was carried out, so this study was judged as at unclear risk of bias.

The remaining 19 studies were either unblinded (De'Oliveira 1997; Dodson_P 1989; Imanishi Micro 2001; Imanishi Normo 2001; Lopes De Faria 1997; Luik 2002) or stated to be double‐blind studies but provided no information regarding the blinding of outcome assessors (Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Miller 1997; Trevisan Micro 1998; Trevisan Normo 1998; Vedovato Micro 2004; Vedovato Normo 2004; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998). In Kwakernaak HTZ 2014 and Kwakernaak Placebo 2014, the dietary component was open‐label, but the hydrochlorothiazide and placebo were double‐blind, and for this reason, the study was judged at high risk of bias.
 

Incomplete outcome data

In two studies, all participants were accounted for from the start to the end of the study, with only two participants withdrawing per study (attrition rate = 4%). Intention‐to‐treat (ITT) analysis was performed for the original participant samples, including those two withdrawals (ITT; N = 45), so the studies were judged to be low risk of bias (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014).

In three studies, all participants were accounted for from the start to the end of the study. No participants dropped out. However, it was unclear whether ITT analysis was performed; therefore, the studies were judged to be at unclear risk of bias (Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

In the remaining 16 studies, either withdrawals or data were not complete; study attrition was not clearly reported, or an ITT analysis was not reported or performed, and therefore the studies were judged to be at high risk of bias.

Selective reporting

All 21 studies were judged to be at high risk of bias. In all studies, the outcomes planned in the methods were reported in the results. However, none of the studies had published protocols or trial registrations to provide evidence to show that reporting bias was not present. In four studies, the protocol was provided as supplementary materials. either online or in appendices. However, all were published at the same time as the article, which was considered insufficient, so they were judged to be at high risk of bias (Imanishi Micro 2001; Imanishi Normo 2001; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014).

Other potential sources of bias

Funding declaration

Seventeen studies declared their funding sources. Twelve were not industry funding and were judged to be at low risk of bias. Five declared funding from the pharmaceutical industry and were judged to be at high risk of bias (Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Mulhauser 1996). The remaining four studies failed to declare any sources of funding and were judged to be at unclear risk of bias (Lopes De Faria 1997; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

Conflicts of interest

Only three studies reported conflicts of interest, and all three declared no competing interests (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Lopes De Faria 1997). Two studies were judged to be at high risk of bias (Houlihan Losartan 2002; Houlihan Placebo 2002). The remaining 16 studies did not report any details about whether they had conflicts or not and were judged to be at unclear risk of bias.

Other information

Eight studies were judged to be at low risk of other potential biases because they either did not require a washout period (parallel design) or a sufficient washout period of seven days or more was applied (De'Oliveira 1997; Dodson_P 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Lopes De Faria 1997; Luik 2002; Miller 1997; Mulhauser 1996).

In one cross‐over study, insufficient details were reported on whether a washout period was applied between treatment arms and was therefore judged to have an unclear risk of bias (Dodson_X 1989).

In the remaining 12 studies, no washout period was applied between cross‐over phases resulting in possible carryover effects, and therefore they were judged to be at high risk of bias (Imanishi Micro 2001; Imanishi Normo 2001; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Miller 1995; Trevisan Micro 1998; Trevisan Normo 1998; Vedovato Micro 2004; Vedovato Normo 2004; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

No other potential sources of bias were identified.

Effects of interventions

See: Table 1; Table 2; Table 3

As studies were not of sufficient duration to test the efficacy of salt reduction on outcome measurements such as heart attacks, end‐stage kidney disease (ESKD) or death, changes in salt intake on BP and surrogate markers of CKD in participants were considered in evaluating the evidence for this review.

Urinary sodium excretion

In 17 studies (Appendix 5), the median 24‐hour UNa on the usual or high salt intake was 203 mmol (4.7 g sodium/day) with a range of 159 to 361 mmol/24 hours (3.7 to 8.3 g sodium/day). The median 24‐hour UNa on a low salt diet was 73 mmol (1.7 g sodium/day) with a range of 14 to 164 mmol (0.6 to 7.1 g sodium/day). The median net change in 24‐hour UNa was 123 mmol/24 hours (5.4 g sodium/day) with a range of 51 to 337 mmol (2.2 to 14.7 g sodium/day). Four studies (Imanishi Micro 2001; Imanishi Normo 2001; Vedovato Micro 2004; Vedovato Normo 2004) were excluded from these calculations as they only provided graphical data of urinary sodium concentrations.

In the seven long‐term studies (Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Mulhauser 1996), the median 24‐hour UNa on the high salt diet was 202 mmol, with a range of 188 to 224 mmol and on the low salt diet was 123 mmol, ranging from 80 to 164 mmol. The median net change in 24‐hour UNa was 76 mmol, ranging from 51 to 124 mmol.

In short‐term studies, data were available from 10 studies (De'Oliveira 1997; Lopes De Faria 1997; Luik 2002; Miller 1995; Miller 1997; Trevisan Micro 1998; Trevisan Normo 1998; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998). Median 24‐hour UNa on the usual or high salt diet was 224 mmol, with a range of 159 to 361 mmol and on the low salt diet was 32 mmol, with a range of 14 to 99 mmol. The median net change in 24‐hour UNa was 187 mmol, with a range of 86 to 337 mmol.

In 10 studies, which included participants with hypertension (De'Oliveira 1997; Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998), the median net change in 24‐hour UNa was 89 mmol, with a range of 51 to 181 mmol. In seven studies (Lopes De Faria 1997; Luik 2002; Miller 1995; Miller 1997; Mulhauser 1996; Trevisan Micro 1998; Trevisan Normo 1998) which included participants with normal BP levels, the median net change in 24‐hour UNa was 211 mmol with a range of 107 to 337 mmol.

In seven studies of participants with type 1 diabetes, the median net change in 24‐hour UNa was 211 mmol with a range of 107 to 337 mmol (Lopes De Faria 1997; Luik 2002; Miller 1995; Miller 1997; Mulhauser 1996; Trevisan Micro 1998; Trevisan Normo 1998). In 10 studies of participants with type 2 diabetes, the median net change in 24‐hour UNa was 89 mmol with a range of 51 to 181 mmol (De'Oliveira 1997; Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

Effects of different sodium intakes on blood pressure

All studies that could be included in the meta‐analyses reported measurements of BP as SBP and DBP or MAP (defined as DBP + (SBP‐DBP)/3).

Systolic blood pressure

Data from 12 studies were available for analysis of SBP. Combining data on longer‐term (up to 12 weeks) with short‐term use (one week) of low sodium intake demonstrated that a low sodium intake diet may reduce SBP by 7.36 mm Hg (Analysis 1.1 (12 studies): MD ‐7.36 mm Hg, 95% CI ‐10.75 to ‐3.98; I² = 74%). In longer studies, low sodium intake may reduce SBP by 6.15 mm Hg (Analysis 1.1.1 (7 studies): MD ‐6.12 mm Hg, 95% CI ‐9.27 to ‐3.03; I² = 12%) while in one‐week studies, low sodium intake may reduce SBP by 8.43 mm Hg (Analysis 1.1.2 (5 studies): MD ‐8.43 mm Hg, 95% CI ‐14.37 to ‐2.48; I² = 88%). There was considerable heterogeneity in the analysis of one‐week studies (I² = 88%) but not in the analysis of longer studies (I² = 12%).

1.1. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 1: Systolic BP

The reduction in SBP with a low sodium diet may not differ between studies of participants with (Analysis 1.6.2) or without albuminuria (Analysis 1.6.1) at study enrolment.

1.6. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 6: Systolic BP according to presence/absence of albuminuria at enrolment

Diastolic blood pressure

Data from 12 studies were available for analysis of DBP (Analysis 1.3). Combining longer‐term and short‐term use of low sodium intake demonstrated that a low sodium intake diet may reduce DBP by 3.17 mm Hg (Analysis 1.3 (12 studies): MD ‐3.17 mm Hg, 95% CI ‐4.58 to ‐1.76; I² = 54%). In longer studies, low sodium intake may reduce DBP by 3.41 mm Hg (Analysis 1.3.1 (7 studies): MD ‐3.41 mm Hg, 95% CI ‐5.56 to ‐1.27; I² = 41%) while in one‐week studies, low sodium intake also may reduce DBP by 2.95 mm Hg (Analysis 1.3.2 (5 studies): MD ‐2.95 mm Hg, 95% CI ‐4.96 to ‐0.94; I² = 70%). There was substantial heterogeneity in the analysis of one‐week studies (I² = 70%), but this was lower in the analysis of longer studies (I² = 41%).

1.3. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 3: Diastolic BP

The reduction in DBP with a low sodium diet may not differ between studies of participants with (Analysis 1.7.2) or without albuminuria (Analysis 1.7.1) at study enrolment.

1.7. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 7: Diastolic BP according to presence/absence of albuminuria at enrolment

Mean arterial pressure

Data from 13 studies were available for analysis of MAP. Both long‐term and short‐term use of low sodium intake may reduce MAP by 3.01 mm Hg (Analysis 1.5 (13 studies): MD ‐3.01 mm Hg, 95% CI ‐4.95 to ‐1.07; I² = 63%). In longer studies, low sodium intake may reduce MAP by 4.60 mm Hg (Analysis 1.5.1 (4 studies): MD ‐4.60 mm Hg, 95% CI ‐7.26 to ‐1.94; I² = 28%). In one‐week studies, low sodium intake may reduce MAP by 2.37 mm Hg (Analysis 1.5.2 (9 studies): MD ‐2.37 mm Hg, 95% CI ‐4.75 to ‐0.01; I² = 65%). There was substantial heterogeneity in the analysis of one‐week studies (I = 65%), but this was not seen in the analysis of longer studies (I² = 28%).

1.5. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 5: MAP

The reduction in MAP with a low sodium intake may be greater in studies of participants with albuminuria at study enrolment (Analysis 1.8.2 (7 studies): MD ‐5.40 mm Hg, 95% CI ‐7.72 to ‐3.08) compared with those without albuminuria (Analysis 1.8.1 (3 studies): MD ‐0.11 mm Hg, 95% CI ‐2.27 to 2.05) at study enrolment.

1.8. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 8: MAP according to the presence/absence of albuminuria at enrolment

Certainty of the evidence

The certainty of the evidence is summarised in Table 1. All results were considered to be at low certainty evidence because of the small numbers of included participants, heterogeneity between study results and increased risk of bias.

Effects of different sodium intakes on blood pressure in hypertensive and normotensive participants

Hypertensive participants

Five studies evaluated the effects of salt restriction on SBP and DBP in hypertensive participants (Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Mulhauser 1996), and three studies evaluated MAP (De'Oliveira 1997; Houlihan Losartan 2002; Houlihan Placebo 2002).

In hypertensive participants, salt restriction may reduce SBP by 6.45 mm Hg (Analysis 2.1.1 (5 studies): MD ‐6.45 mm Hg, 95% CI ‐11.47 to ‐1.42; I² = 41%), DBP by 3.15 mm Hg (Analysis 2.2.1 (5 studies): MD ‐3.15 mm Hg, 95% CI ‐6.49 to ‐0.18; I ² = 60%), and MAP by 4.88 mm Hg (Analysis 2.3.1 (3 studies): MD ‐4.88 mm Hg, 95% CI ‐10.39 to ‐0.63; I² = 54%). There was moderate heterogeneity in the analyses.

2.1. Analysis.

Comparison 2: Net change in BP in hypertensive and normotensive participants, Outcome 1: Systolic BP

2.2. Analysis.

Comparison 2: Net change in BP in hypertensive and normotensive participants, Outcome 2: Diastolic BP

2.3. Analysis.

Comparison 2: Net change in BP in hypertensive and normotensive participants, Outcome 3: MAP

Normotensive participants

Five studies evaluated the effects of salt restriction in normotensive participants (Imanishi Micro 2001; Imanishi Normo 2001; Luik 2002; Trevisan Micro 1998; Trevisan Normo 1998), and eight studies evaluated MAP (Lopes De Faria 1997; Miller 1995; Miller 1997; Vedovato Micro 2004; Vedovato Normo 2004; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

In normotensive participants, salt restriction may reduce SBP by 8.43 mm Hg (Analysis 2.1.2 (5 studies): MD ‐8.43 mm Hg, 95% CI ‐14.37 to ‐2.48; I² = 88%), reduce DBP by 2.95 mm Hg (Analysis 2.2.2 (5 studies): MD ‐2.95 mm Hg, 95% CI ‐4.96 to ‐0.94; I² = 70%), and MAP by 2.15 mm Hg (Analysis 2.3.2 (8 studies): MD ‐2.15 mm Hg, 95% CI ‐4.56 to ‐0.26; I² = 68%). There was substantial heterogeneity in the analyses.

Certainty of the evidence

The certainty of the evidence is summarised in Table 2. All results were considered to be of low certainty evidence because of the small numbers of included participants, heterogeneity between study results and increased risk of bias.

Effects of different sodium intakes on blood pressure in type 1 and type 2 diabetes

Type 1 diabetes

Four studies evaluated the effects of salt restriction on SBP and DBP in participants with type 1 diabetes (Luik 2002; Mulhauser 1996; Trevisan Micro 1998; Trevisan Normo 1998), and three studies reported the effects on MAP (Lopes De Faria 1997; Miller 1995; Miller 1997).

Salt restriction may reduce SBP by 7.35 mm Hg (Analysis 3.1.1 (4 studies): MD ‐7.35 mm Hg, 95% CI ‐14.49 to ‐0.21; I² = 91%) and DBP by 3.20 mm Hg (Analysis 3.2.1 (4 studies): MD ‐3.20 mm Hg, 95% CI ‐5.16 to ‐1.23; I² = 62%), but may make little or no difference to MAP (Analysis 3.3.1 (3 studies): MD 0.08 mm Hg, 95% CI ‐4.95 to 2.08; I² = 8%).

3.1. Analysis.

Comparison 3: Net change in BP in type 1 and type 2 diabetes, Outcome 1: Systolic BP

3.2. Analysis.

Comparison 3: Net change in BP in type 1 and type 2 diabetes, Outcome 2: Diastolic BP

3.3. Analysis.

Comparison 3: Net change in BP in type 1 and type 2 diabetes, Outcome 3: MAP

Type 2 diabetes

Eight studies evaluated the effects of salt restriction on SBP and DBP in participants with type 2 diabetes (Dodson_P 1989; Dodson_X 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Imanishi Micro 2001; Imanishi Normo 2001 Kwakernaak HTZ 2014; Kwakernaak Placebo 2014), and 10 studies evaluated MAP (De'Oliveira 1997; Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Vedovato Micro 2004; Vedovato Normo 2004; Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998).

Salt restriction may reduce SBP by 7.35 mm Hg (Analysis 3.1,2 (8 studies): MD ‐7.35 mm Hg, 95% CI ‐10.32 to ‐4.38; I² = 22%), DBP by 3.04 mm Hg (Analysis 3.2.2 (8 studies): MD ‐3.04 mm Hg, 95% CI ‐5.20 to ‐0.89; I² = 57%), and MAP by 4.30 mm Hg (Analysis 3.3.2 (10 studies): MD ‐4.30 mm Hg, 95% CI ‐6.54 to ‐2.05; I = 56%). There was substantial heterogeneity in the analyses.

Certainty of the evidence

The certainty of the evidence is summarised in Table 3. All results were considered to be of low certainty evidence because of the small numbers of included participants, heterogeneity between study results and increased risk of bias.

Concomitant interventions with renin‐angiotensin system blockers

Salt restriction has been demonstrated to reduce BP when the RAS is blocked (MacGregor 1987). Three of seven studies in the long‐term studies included participants treated with RAS blockers (Houlihan Losartan 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). Therefore we recalculated pooled data, excluding data from these studies. Pooled estimates of the changes in the remaining four studies showed reductions in SBP of 5.55 mm Hg (Analysis 1.2.1 (4 studies): MD ‐5.55 mm Hg, 95% CI ‐11.73 to ‐0.62) and in DBP of 2.26 mm Hg (Analysis 1.4.1 (4 studies): MD ‐2.26m Hg, 95% CI ‐6.54 to 2.02) with 95% CIs that overlapped those of Analysis 1.1 and Analysis 1.3 (which included the studies using ACEi or ARBs), suggesting that RAS blockade may not result in further reductions in SBP and DBP in this small group of studies.

1.2. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 2: Systolic BP (excluding studies using RAS)

1.4. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 4: Diastolic BP (excluding studies using RAS)

Effects of altering salt intake on measurements of kidney function

Pooled analysis of seven studies showed that salt restriction may reduce CrCl by 6.05 mL/min (Analysis 1.9 (7 studies): MD ‐6.05 mL/min, 95% CI ‐10.00 to ‐2.10; I² = 0%). However, pooled analysis of 12 studies of eGFR found that there may make little or no difference in eGFR in these studies (Analysis 1.10 (12 studies): MD ‐1.87 mL/min/1.73 m², 95% CI ‐5.05 to 1.31; I² = 32%).

1.9. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 9: Creatinine clearance

1.10. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 10: Glomerular filtration rate

Eight studies reported change in ERPF. Salt restriction may not alter ERPF in longer studies (Analysis 1.11.1 (8 studies): MD ‐0.73 mL/min, 95% CI ‐2.83 to 1.37; I² = 0%) or short‐term studies (Analysis 1.11.2 (5 studies): MD ‐0.14 mL/min, 95% CI ‐68.26 to 67.98; I² = 82%).

1.11. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 11: Effective renal plasma flow

Effects of altering salt intake on urinary albumin and protein excretion

Eight studies presented data on changes in urinary albumin or urinary protein excretion during salt restriction (Houlihan Losartan 2002; Houlihan Placebo 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Luik 2002; Mulhauser 1996; Vedovato Micro 2004; Vedovato Normo 2004. The data for individual studies are shown in Appendix 6. These studies reported the data in ways which were not suitable for meta‐analysis, so pooled analyses were not performed. One study of type 2 diabetics (Vedovato Micro 2004) reported salt restriction may reduce UAE in patients with microalbuminuria, while two other studies reported no change in UAE in type 1 diabetics with (Mulhauser 1996) or without (Luik 2002) microalbuminuria. Salt restriction may reduce UAE in type 2 diabetics receiving RAS blockade (Houlihan Losartan 2002; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014) but may not reduce UAE in those not receiving RAS blockade (Houlihan Placebo 2002).

Effects of altering salt intake on HbA1c

Six studies reported the results of HbA1c at the beginning and end of the study (Dodson_P 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Lopes De Faria 1997; Luik 2002; Mulhauser 1996).

Pooled analysis showed no difference in HbA1c during salt restriction, suggesting that there were no alterations in diabetic control during the studies (Analysis 1.12 (6 studies): MD ‐0.62, 95% CI ‐1.49 to 0.26; I² = 95%). There was significant heterogeneity between studies; however, the exclusion of one small study (Lopes De Faria 1997) (which showed a small decrease of 3% in HbA1c) removed the heterogeneity between studies.

1.12. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 12: HbA1c

Effects of altering salt intake on body weight

Twelve studies reported data on body weight (Dodson_P 1989; Dodson_X 1989; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014; Lopes De Faria 1997; Luik 2002; Miller 1997; Mulhauser 1996; Trevisan Micro 1998; Trevisan Normo 1998; Vedovato Micro 2004; Vedovato Normo 2004).

Salt restriction lowered body weight by ‐1.21 kg (Analysis 1.13 (12 studies): MD ‐1.21 kg, 95% CI ‐1.73 to ‐0.68; I² = 76%). This effect was primarily due to changes in short‐term studies, with minimal weight change with salt restriction in longer studies (Analysis 1.13.1 (5 studies): MD ‐0.35 kg, 95% CI ‐1.63 to 0.94; I² = 0%) compared with short‐term studies (Analysis 1.13.2 (7 studies): MD ‐1.30 kg, 95% CI ‐1.89 to ‐0.72; I² = 86%) suggesting that these changes are likely to be due to a reduction in extracellular volume.

1.13. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 13: Body weight

Adverse effects

Only two studies reported adverse effects (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). These studies reported orthostatic hypotension was identified in 11% of participants receiving a reduced sodium diet compared with 4% receiving a normal sodium diet. It occurred most frequently (21%) in those receiving a low salt diet together with hydrochlorothiazide.

Fixed effects versus random effects model

To ensure robustness, we performed analyses using both the fixed effect and random effects models.

In all studies (Analysis 1.1 to Analysis 1.5), reductions in SBP, DBP and MAP with salt restrictions were similar using both analyses. Compared with the initial version of this review, the addition of Kwakernaak HTZ 2014 and Kwakernaak Placebo 2014 to the subgroup analysis resulted in the net change in DBP in longer studies no longer differing between the fixed effects (MD ‐3.71 mm Hg, 95% CI ‐5.29 to ‐2.14) and random effects models (MD ‐3.41 mm Hg, 95% CI ‐5.56 to ‐1.27) with both models showing a reduction in DBP with salt restriction.

In studies of hypertensive and normotensive participants, salt restriction reduced SBP, DBP and MAP in both groups (Analysis 2.1; Analysis 2.2; Analysis 2.3). Among the subgroup of hypertensive participants, the net change in DBP did not differ between fixed (MD ‐3.78 mm Hg, 95% CI ‐5.73 to ‐1.84) and random effects models (MD ‐3.15 mm Hg, 95% CI ‐6.49 to 0.18) but the 95% CIs crossed 1 with the random effects model. There was no difference between subgroups with either analysis.

In studies of type 1 and type 2 diabetics, changes in SBP, DBP and MAP with salt restriction were similar between fixed and random effects analyses (Analysis 3.1; Analysis 3.2; Analysis 3.3).

Investigation of sources of heterogeneity

We assessed heterogeneity using the I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than sampling error (Higgins 2003). We investigated levels of heterogeneity considered to be substantial (50% to 90%) or considerable (75% to 100%). For the outcomes of SBP, DBP and MAP, there was considerable heterogeneity among short‐term studies (SBP: 88%; DBP: 70%; MAP: 65%), but this variation was not seen in long‐term studies (SBP: 12%; DBP: 41%; MAP: 28%). While variations in MDs between studies within short‐term studies did not appear to be related to either the reduction in urinary sodium achieved or to the minimum urine sodium recorded (Appendix 7), the range of net change in UNa recorded in short‐term studies (86 to 337 mmol) was much greater than that seen in long‐term studies (51 to 124 mmol) suggesting that differences in study design contributed to heterogeneity between studies.

We also examined whether there were differences in the degree of heterogeneity among cross‐over studies in which there was or was not a washout period between the first and second parts of the cross‐over. For SBP (Analysis 1.14) but not for MAP (Analysis 1.15), there was less heterogeneity among cross‐over studies with washout compared with those without washout. However, the results for SBP largely mirrored those seen overall for SBP among long‐term and short‐term studies (Analysis 1.1). For MAP, there remained heterogeneity of 69% among cross‐over studies with washout (Analysis 1.15.1) and 66% among studies without washout (Analysis 1.15.2). Overall, we were not able to identify specific sources for heterogeneity between studies. However, most studies were at high risk of bias for most criteria (Figure 2).

1.14. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 14: Systolic BP in cross‐over studies with or without washout between interventions

1.15. Analysis.

Comparison 1: Net change with altering salt diet, Outcome 15: MAP in cross‐over studies with or without washout between study periods

Sensitivity analyses

We planned to perform sensitivity analyses to explore the influence of the following factors on effect size.

• Repeating the analysis excluding unpublished studies: This was not done as we did not identify any unpublished studies.

• Repeating the analysis taking account of the risk of bias, as specified: This was not done as studies were at unclear or high risk of bias for most criteria (Figure 2; Figure 4; Figure 5)

• Repeating the analysis excluding any long duration or large studies to establish how much they dominate the results: We re‐analysed the results after excluding three studies with long durations and more than 30 participants (Dodson_P 1989; Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). The results for SBP for long‐term studies with these three studies excluded (MD ‐5.57 mm Hg, 95% CI ‐9.51 to ‐1.62) did not differ from the overall results (MD ‐6.12 m Hg, 95% CI ‐9.02 to ‐3.22). The MD for DBP in long‐term studies (MD ‐5.02 mm Hg, 95% CI ‐9.38 to ‐0.65) was slightly higher than for the overall results (MD ‐3.71 mm Hg, 95% CI ‐5.29 to ‐2.14), but the CIs overlapped. Therefore we did not identify that the exclusion of these three studies made major changes to the results.

4.

Forest plot of comparison: 1 Net change with altering salt diet, outcome: 1.1 Systolic BP.

5.

Forest plot of comparison: 1 Net change with altering salt diet, outcome: 1.5 MAP.

Discussion

Summary of main results

This systematic review identified 21 studies (313 participants) from 13 RCTs, which evaluated the efficacy of reducing salt intake on BP and kidney function in participants with type 1 or type 2 diabetes. This update added two additional studies (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014). All studies were of short duration, and no studies addressed the outcomes of cardiovascular events (stroke, heart failure, myocardial infarction).

• The median net reduction in all studies in 24‐hour UNa during the low sodium intake was 123 mmol/24 hours (5.4 g sodium/day) with a range of 51 to 337 mmol (2.2 to 14.7 g sodium/day). The median net reduction in seven long‐term studies (treatment duration four to 12 weeks) was 76 mmol, ranging from 51 to 124 mmol. The median net reduction in 10 short‐term studies (treatment duration five to seven days) was 187 mmol, with a range of 86 to 337 mmol. Data were only available graphically in four studies.

• Sodium restriction may reduce SBP by 7.36 mm Hg, DBP by 3.17 mm Hg and MAP by 3.01 mm Hg in all participants (low certainty evidence).

• No differences were identified between hypertensive and normotensive participants or between participants with type 1 or type 2 diabetes (low certainty evidence).

• During salt restriction, SBP and DBP may be reduced by similar amounts in studies, including participants with or without microalbuminuria at study enrolment. However, the reduction in MAP may be greater in participants with microalbuminuria at study enrolment compared with those without microalbuminuria.

• UAE was formally assessed in only eight studies. During salt restriction, UAE fell in four studies (Appendix 6).

• There may be a small reduction in body weight during salt restriction, but this was largely seen in studies of one to two weeks rather than in longer studies (low certainty evidence).

• There may be no differences in HbA1c during salt restriction (low certainty evidence).

• There may be no changes to GFR or ERPF though CrCl may be reduced by 6 mL/min during salt restriction (low certainty evidence).

• Adverse effects were only reported in two of 21 studies. Two studies reported that orthostatic hypotension was identified more frequently in participants receiving a reduced sodium diet with or without hydrochlorothiazide compared with those receiving a normal sodium diet (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014).

Overall completeness and applicability of evidence

The available data from this systematic review of RCTs of salt restriction indicate that lowering sodium intake may reduce BP in participants with type 1 and type 2 diabetes, with and without hypertension and with normal or slightly abnormal kidney function (G1, A1 or A2 on the KDIGO classification) (KDIGO 2020). No studies were identified in this patient group which evaluated salt restriction in participants with greater degrees of reduced kidney function. Two additional studies were added in this update (Kwakernaak HTZ 2014; Kwakernaak Placebo 2014), but these data did not change the conclusions of the initial version of the review. The duration of the studies (maximum duration three months) and the small size of the studies (maximum number of 45 participants) did not allow any assessment of important patient‐centred outcomes, including death and CVD. The majority of studies were cross‐over studies, with only one or two weeks in each phase of the study. Seventeen studies provided information on the sodium intakes in each group, with four studies providing graphical information only. Twelve studies provided information on SBP and DBP, while nine provided information only on MAP. Data on kidney function were more limited, with only 12 studies providing information on eGFR and an additional three studies providing information on CrCl. Overall, eGFR was maintained unchanged during the study periods. Limited data were available to assess the effect of salt reduction on UAE. UAE is important in monitoring the progression of kidney disease in participants with diabetes and is an independent predictor of CVD (Hillege 2002). We found only eight studies in this meta‐analysis where UAE and/or protein excretion was measured, and it was not possible to perform pooled analysis due to variations in the way data were reported. Four of the eight studies demonstrated a reduction in UAE with salt restriction. Most studies were performed before 2000, and participants did not routinely receive RAS inhibitors. RAS inhibition with either ACEi or ARBs is now recommended for all diabetic patients with albuminuria since RAS blockade slows the deterioration in kidney function (KDIGO 2020). Since high salt intake reduces the efficacy of these agents on BP (MacGregor 1987), guidelines recommend that salt intake be reduced to 5 g/day (sodium intake < 2 g or < 90 mmol/day) (KDIGO 2020). While such reductions in salt intake may be achievable in the short term, it is hard to maintain long‐term reductions without intensive input from dietitians and easily interpretable information about the salt content of commonly used foods.

Quality of the evidence

There are important limitations to this review. The number of studies on the effects of salt reduction in diabetics was limited, and the majority of available studies were of short duration with few included participants. The results show considerable heterogeneity, particularly among studies lasting only one to two weeks. The small number of studies limits the ability to explore potential sources of this heterogeneity. It is likely that differing methodologies and degrees and duration of salt restriction affect heterogeneity. Studies varied from five days to 12 weeks, and longer studies had more modest changes in salt intake. Several short studies restricted levels of salt intake to as low as 0.6 g/day, followed by loading of up to 20 g/day of salt. Acute, large changes in dietary salt intake such as this will stimulate the RAS and increase sympathetic activity. This will have an impact on the effects on BP and renal haemodynamics. This is demonstrated by the lack of any difference in SBP or DBP between shorter and longer studies, despite a change in median salt intake of almost double that seen in the longer studies. While MDs in MAP did differ slightly between long‐term and short‐term studies, the CIs overlapped.

Outcome measurements were mostly restricted to measurements of BP and surrogate markers of kidney function. We did not find any available data on outcome measurements such as time to ESKD, CVD and death. 24‐hour UNa is a surrogate method of estimating salt intake, but since approximately 90% of the salt we eat is excreted through the kidneys, 24‐hour UNa provides an accurate assessment of sodium intake. Compared with all other methods, e.g. dietary record or recall, measuring 24‐hour UNa excretion is the most accurate and accepted method for assessing dietary salt intake.

GRADE assessment of the included outcomes showed that the certainty of the evidence was low for all outcomes when all studies were included (Table 1), when studies evaluated participants with or without hypertension (Table 2) and when studies evaluated participants with type 1 and type 2 diabetes (Table 3). Studies were downgraded for risk of bias issues (allocation concealment and blinding), for heterogeneity between studies and for imprecision based on small numbers of enrolled participants.

Potential biases in the review process

The search strategy was comprehensive and covers all important and relevant databases. We do not believe we have missed any potentially relevant studies. Of the studies identified, we have endeavoured to obtain all data by contacting the authors, but we were unsuccessful in obtaining additional data. Therefore, there is a small potential for bias in the available outcome data. Data collection was undertaken using a piloted data extraction form. Data analysis and assessment of the quality of studies were all performed in duplicate with a third party for disagreement resolution. We do not believe any other biases could have been introduced.

Agreements and disagreements with other studies or reviews

A recently published systematic review of RCTs has evaluated the effect of reducing salt intake on BP in people with type 2 diabetes (Ren 2021). This review included eight RCTs (10 studies) in adults, who underwent a period of salt restriction of at least one week, the sodium intake was calculated from 24‐hour UNa measurements, and the differences in SBP and DBP between low and usual sodium diets were reported or could be calculated. Six studies (Dodson_P 1989; Houlihan Losartan 2002; Houlihan Placebo 2002; Imanishi Micro 2001; Imanishi Normo 2001; Kwakernaak Placebo 2014) were also included in this Cochrane review. The authors found that salt restriction reduced SBP and DBP by 5.6 mm Hg and 1.7 mm Hg, respectively, compared with reductions of 7.2 mm Hg and 3 mm Hg found in our Cochrane review. Recent systematic reviews of RCTs in people with CKD and not on dialysis found that reducing salt intake lowers SBP and DBP by similar amounts to those seen in our review(Garafalo 2018; McMahon 2021). In systematic reviews of studies in people with normal kidney function, reducing sodium intake consistently reduces BP levels with greater reductions in hypertension participants (Graudal 2020; He 2003). In this review, we found that urinary protein excretion was reduced in four of the eight studies in which this outcome was reported. This outcome was not reported in Ren 2021, which reviewed studies of participants with type 2 diabetes. Systematic reviews in people with CKD (Garafalo 2018; McMahon 2021) have found that reduced salt intake reduces urinary protein excretion by about 30%.

Authors' conclusions

Implications for practice.

This updated systematic review confirms previous findings that salt reduction reduces BP in people with type 1 and type 2 diabetes and can increase pharmacological interventions' efficacy in achieving strict BP control.  These findings, in conjunction with other evidence relating salt intake to BP and albuminuria in hypertensive and normotensive people as well as in people with CKD, suggest that reducing salt intake in people with diabetes to the amounts recommended in public health guidelines for the general population may improve outcomes for people with diabetes.

Implications for research.

The studies included in this review were of short duration, involved large changes in dietary salt intake and required frequent measurements of urinary sodium excretion. These studies demonstrated that short‐term reductions in salt intake result in important reductions in SBP, DBP and MAP in people with diabetes with normal GFR with or without microalbuminuria. However, these short‐term studies cannot assess patient‐level outcomes, including CVD and CKD. Assessing such patient‐level outcomes requires larger and longer‐term studies of reduction in salt intake. A recent study of Chinese people with hypertension and previous strokes (Neal 2021) has demonstrated that a longer‐term reduction in sodium intake with a reduction in cardiovascular and cerebral outcomes could be achieved using salt supplements (75% sodium chloride and 25% potassium chloride) to reduce sodium intake. Similar studies could be considered in people with diabetes without reduced GFR since replacing sodium chloride with potassium chloride is potentially harmful in participants with reduced GFR. However, the opportunity to do such studies is limited now that many people with type 2 diabetes are managed with SGLT2 inhibitors, increasing sodium excretion.

What's new

Date	Event	Description
5 January 2023	New search has been performed	Search strategies for MEDLINE, EMBASE, CENTRAL amended; text updated
5 January 2023	New citation required but conclusions have not changed	New studies added, new outcomes reported, SOF tables added

History

Protocol first published: Issue 4, 2007
Review first published: Issue 12, 2010

Appendices

Appendix 1. Electronic search strategies

Database	Search terms
CENTRAL	MeSH descriptor Sodium, Dietary explode all tree MeSH descriptor Diet, Sodium‐Restricted, this term only ((dietary next salt*) or (dietary sodium)):ti,ab,kw (diet* near/10 (salt* or sodium)):ti,ab,kw in Trials ((salt* or sodium) near/10 (restrict* or intak* or chang* or high or low)):ti,ab,kw in Trials (#1 OR #2 OR #3 OR #4 OR #5) MeSH descriptor Diabetic Nephropathies, this term only MeSH descriptor Diabetes Mellitus, this term only MeSH descriptor Diabetes Mellitus, Type 1 explode all trees MeSH descriptor Diabetes Mellitus, Type 2 explode all trees MeSH descriptor Glucose Intolerance explode all trees MeSH descriptor Insulin Resistance explode all trees ((diabetic next nephropath*) or (diabetic next kidney*) or (diabetic next renal*)):ti,ab,kw in Trials (IDDM or NIDDM) in Trials ((insulin dependent diabetes mellitus) or (non insulin dependent diabetes mellitus)):ti,ab,kw in Trials ((diabet*) near/1 (type 1 or type 2)):ti,ab,kw in Trials ((impaired next glucose next intoleran*) or (glucose next intoleran*)):ti,ab,kw (insulin next resistan*):ti,ab,kw (#7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13 OR #14 OR #15 OR #16 OR #17 in trials (#6 AND #19)
MEDLINE	exp Sodium, Dietary/ Diet, Sodium‐Restricted/ (dietary salt or dietary sodium).tw. (diet$ adj5 (salt$ or sodium)).ti. (diet$ adj10 (salt$ or sodium)).ab. ((salt$ or sodium) adj5 (restrict$ or intak$ or change$ or high or low)).ti ((salt$ or sodium) adj10 (restrict$ or intak$ or change$ or high or low)).ab. or/1‐7 diabetes mellitus/ or exp diabetes mellitus, type 1/ or exp diabetes mellitus, type 2/ (IDDM or NIDDM).tw. Diabetic Nephropathies/ diabetic nephropath$.tw. ((diabetic or diabetes) and (kidney$ or renal$ or nephro$ or nephritis or glomerulo$)).tw. (insulin dependent diabetes or non insulin dependent diabetes).tw. ((type 1 or type 2) adj1 diabet*).tw. Glucose Intolerance/ (impaired glucose toleranc$ or glucose intoleranc$).tw. exp Insulin Resistance/ insulin resistanc$.tw. or/9‐20 and/8,20
EMBASE	Salt Intake/ Sodium Intake/ Sodium Restriction/ (dietary salt or dietary sodium).tw. (diet$ adj5 (salt$ or sodium)).ti. (diet$ adj10 (salt$ or sodium)).ab. ((salt$ or sodium) adj5 (restrict$ or intak$ or change$ or high or low)).ti. ((salt$ or sodium) adj10 (restrict$ or intak$ or change$ or high or low)).ab. or/1‐8 Diabetic Nephropathy/ diabetic nephropath$.tw. ((diabetic or diabetes) and (kidney$ or renal$ or nephro$ or nephritis or glomerulo$)).tw. (IDDM or NIDDM).tw. diabetes mellitus/ or impaired glucose tolerance/ or insulin dependent diabetes mellitus/ or non insulin dependent diabetes mellitus/ (insulin dependent diabetes mellitus or non insulin dependent diabetes mellitus).tw. Glucose Intolerance/ (impaired glucose toleranc$ or glucose intoleranc$).tw. ((type 1 or type 2) adj1 diabet*).tw. Insulin Resistance/ insulin resistanc$.tw. or/9‐20 and/9,21

Appendix 2. Risk of bias assessment tool

Potential source of bias	Assessment criteria
Random sequence generation Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence	Low risk of bias: Random number table; computer random number generator; coin tossing; shuffling cards or envelopes; throwing dice; drawing of lots; minimisation (minimisation may be implemented without a random element, and this is considered to be equivalent to being random).
	High risk of bias: Sequence generated by odd or even date of birth; date (or day) of admission; sequence generated by hospital or clinic record number; allocation by judgement of the clinician; by preference of the participant; based on the results of a laboratory test or a series of tests; by availability of the intervention.
	Unclear: Insufficient information about the sequence generation process to permit judgement.
Allocation concealment Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment	Low risk of bias: Randomisation method described that would not allow investigator/participant to know or influence intervention group before eligible participant entered in the study (e.g. central allocation, including telephone, web‐based, and pharmacy‐controlled, randomisation; sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes).
	High risk of bias: Using an open random allocation schedule (e.g. a list of random numbers); assignment envelopes were used without appropriate safeguards (e.g. if envelopes were unsealed or non‐opaque or not sequentially numbered); alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure.
	Unclear: Randomisation stated but no information on method used is available.
Blinding of participants and personnel Performance bias due to knowledge of the allocated interventions by participants and personnel during the study	Low risk of bias: No blinding or incomplete blinding, but the review authors judge that the outcome is not likely to be influenced by lack of blinding; blinding of participants and key study personnel ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding or incomplete blinding, and the outcome is likely to be influenced by lack of blinding; blinding of key study participants and personnel attempted, but likely that the blinding could have been broken, and the outcome is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Blinding of outcome assessment Detection bias due to knowledge of the allocated interventions by outcome assessors.	Low risk of bias: No blinding of outcome assessment, but the review authors judge that the outcome measurement is not likely to be influenced by lack of blinding; blinding of outcome assessment ensured, and unlikely that the blinding could have been broken.
	High risk of bias: No blinding of outcome assessment, and the outcome measurement is likely to be influenced by lack of blinding; blinding of outcome assessment, but likely that the blinding could have been broken, and the outcome measurement is likely to be influenced by lack of blinding.
	Unclear: Insufficient information to permit judgement
Incomplete outcome data Attrition bias due to amount, nature or handling of incomplete outcome data.	Low risk of bias: No missing outcome data; reasons for missing outcome data unlikely to be related to true outcome (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods.
	High risk of bias: Reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardized difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; ‘as‐treated’ analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation.
	Unclear: Insufficient information to permit judgement
Selective reporting Reporting bias due to selective outcome reporting	Low risk of bias: The study protocol is available and all of the study’s pre‐specified (primary and secondary) outcomes that are of interest in the review have been reported in the pre‐specified way; the study protocol is not available but it is clear that the published reports include all expected outcomes, including those that were pre‐specified (convincing text of this nature may be uncommon).
	High risk of bias: Not all of the study’s pre‐specified primary outcomes have been reported; one or more primary outcomes is reported using measurements, analysis methods or subsets of the data (e.g. sub‐scales) that were not pre‐specified; one or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting is provided, such as an unexpected adverse effect); one or more outcomes of interest in the review are reported incompletely so that they cannot be entered in a meta‐analysis; the study report fails to include results for a key outcome that would be expected to have been reported for such a study.
	Unclear: Insufficient information to permit judgement
Other bias Bias due to problems not covered elsewhere in the table	Low risk of bias: The study appears to be free of other sources of bias.
	High risk of bias: Had a potential source of bias related to the specific study design used; stopped early due to some data‐dependent process (including a formal‐stopping rule); had extreme baseline imbalance; has been claimed to have been fraudulent; had some other problem.
	Unclear: Insufficient information to assess whether an important risk of bias exists; insufficient rationale or evidence that an identified problem will introduce bias.

Appendix 3. Summary table of characteristics of included studies

Study	No. of participants	Age (years)	Males (%)	Diabetes	Study design	Study duration (weeks)	Initial BP on usual salt diet (mean ± SD; mm Hg)	Kidney function at study entry
De'Oliveira 1997	19	55	84	Type 2	P	1	MAP: 102 ± 13	Mean SCr ± SEM Low salt group: 1.16 ± 0.05 mg/dL High salt group: 1.21 ± 0.06 mg/dL
Dodson_P 1989	34	61	68	Type 2	P	12	SBP: 180 ± 18 DBP: 91± 11	Participants with reduced kidney function and/or proteinuria excluded
Dodson_X 1989	9	62	67	Type 2	X No washout	4	SBP: 160 ± 20 DBP: 88 ± 11	Participants with reduced kidney function and/or proteinuria excluded Baseline data not provided
Houlihan Losartan 2002	10	61	100	Type 2	X 4 weeks washout	2	MAP: 114 ± 9	Mean SCr ± SD: 92 ± 2.7 µmol/L UAE rate: 10 to 200 µg/min
Houlihan Placebo 2002	10	63	90	Type 2	X 4 weeks washout	2	MAP: 111 ± 9	Mean SCr ± SD: 92 ± 2.7 µmol/L UAE rate: 10 to 200 µg/min
Imanishi Micro 2001	8	61	50	Type 2	X No washout	1	SBP: 136 ± 9 DBP: 82 ± 6	Mean SCr ± SD: 60.4 ± 13.5 µmol/L Microalbuminuria: 20 to 200 µg/min
Imanishi Normo 2001	8	59	63	Type 2	X No washout	1	SBP: 132 ± 11 DBP: 73 ± 7	Mean SCr ± SD: 60.4 ± 13.5 µmol/L No microalbuminuria
Kwakernaak HTZ 2014	45	65	84	Type 2	X 6 weeks washout	6	SBP: 135 ± 16 DBP: 76 ± 9	Mean eGFR ± SD: 65 ± 25 mL/min/1.73 m² Albuminuria: 711 mg/day (95% CI: 485 to 1043)
Kwakernaak Placebo 2014	45	65	84	Type 2	X 6 weeks washout	6	SBP: 147 ± 16 DBP: 82 ± 10	Mean eGFR SD: 65 ± 25 mL/min/1.73 m² Albuminuria: 711 mg/day (95% CI: 485 to 1043)
Lopes De Faria 1997	10	30	70	Type 1	X No washout	1	MAP: 91 ± 9	Only participants with normal kidney function and no albuminuria included
Luik 2002	24	29	63	Type 1	X 10 to 17 days washout	1	SBP < 140 DBP < 85	Only participants with normal kidney function and no albuminuria included
Miller 1995	9	26	100	Type 1	X No washout	1	High salt MAP: 79 ± 3 Low salt MAP: 78 ± 6	Only participants with normal kidney function and no albuminuria included
Miller 1997	12	23	100	Type 1	X No washout	1	High salt MAP: 80 ± 3 Low Salt MAP: 79 ± 7	Only participants with normal kidney function and no albuminuria included
Mulhauser 1996	14	31	75	Type 1	P	4	Low salt SBP: 134 ± 9 Low salt DBP: 87 ± 4 High salt SBP: 139 ± 7 High salt DBP: 88 ± 4	Normal kidney function Stable increase in proteinuria: > 60 mg/24 hours
Trevisan Micro 1998	7	38	86	Type 1	X No washout	1	SBP: 125 ± 16 DBP: 78 ± 13	Normal kidney function UAE rate (µg/min): 77.3 (range: 39 to 198)
Trevisan Normo 1998	9	42	67	Type 1	X No washout	1	SBP: 120 ± 21 DBP: 72 ± 6	Normal kidney function UAE rate (µg/min): 7.5 (range: 2 to 12)
Vedovato Micro 2004	20	57	75	Type 2	X No washout	1	SBP: 130 ± 9 DBP: 80 ± 9	Baseline kidney function not reported UAE rate (µg/min): 91(range: 43 to 180)
Vedovato Normo 2004	21	60	75	Type 2	X No washout	1	SBP: 125 ± 9 DBP: 79 ± 9	Baseline kidney function not reported UAE rate (µg/min): 10 (range: 3 to 18)
Yoshioka Adva Alb 1998	4	51	58	Type 2	X No washout	1	SBP: 131 ± 8 DBP: 75 ± 8	Normal kidney function Mean UAE ± SD (mg/24 hours): 497 ± 91 (range: 309 to 635)
Yoshioka Micro 1998	7	58	58	Type 2	X No washout	1	SBP: 133 ± 16 DBP: 75 ± 11	Normal kidney function Mean UAE ± SD (mg/24 hours): 65 ± 13 (range: 31 to 122)
Yoshioka Normo 1998	8	60	58	Type 2	X No washout	1	SBP: 134 ± 14 DBP: 78 ± 3	Normal kidney function Mean UAE ± SD (mg/24 hours): 16 ± 7 (range: 8 to 27)
Footnotes BP: blood pressure; CI: confidence interval; DBP: diastolic blood pressure; eGFR: estimated glomerular filtration rate; MAP: mean arterial pressure; P: parallel; SBP: systolic blood pressure; SCr: serum creatinine; SEM: standard error of the mean; SD: standard deviation; UAE: urinary albumin excretion; X: cross‐over

Appendix 4. Outcomes reported in the included studies

Study	24‐hour urinary sodium	SCr	CrCl or GFR	RPF	Quantitative measure of urinary albumin or protein excretion reported	ACR	BP	HbA1c	BMI
De'Oliveira 1997	X	‐	X	X	‐	‐	X	‐	‐
Dodson_P 1989	X	‐	‐	‐	‐	‐	X	X	‐
Dodson_X 1989	X	‐	‐	‐	‐	‐	X	X	‐
Houlihan Losartan 2002; Houlihan Placebo 2002	X	‐	X	X	X	X	X	X	‐
Imanishi Micro 2001; Imanishi Normo 2001	X	‐	X	‐	‐		X	X	‐
Kwakernaak HTZ 2014; Kwakernaak Placebo 2014	X	X	X	‐	X	X	X	‐	‐
Lopes De Faria 1997	X	‐	X	X	‐		X	X	‐
Luik 2002	X	X	X	X	X	X	X	X	X
Miller 1995	X	‐	‐	‐	‐	‐	X	‐	‐
Miller 1997	X	‐	X	X	‐	‐	X	‐	‐
Mulhauser 1996	X	X	X	X	X	‐	X	X	X
Trevisan Micro 1998; Trevisan Normo 1998	X	‐	X	X	‐	‐	X	‐	X
Vedovato Micro 2004; Vedovato Normo 2004	X	‐	‐	‐	X	‐	X	‐	‐
Yoshioka Adva Alb 1998; Yoshioka Micro 1998; Yoshioka Normo 1998	X	‐	X	‐	‐	‐	X	‐	‐
Footnotes ACR: albumin:creatinine ratio; BMI: body mass index; BP: blood pressure; CrCl: creatinine clearance; GFR: glomerular filtration rate; HbbA1c: glycated haemoglobin; RPF: renal plasma flow; SCr: serum creatinine

Appendix 5. Measurements of urinary sodium on high and low sodium diets

Study	24‐hour urinary sodium on low sodium diet (mmol/day ± SD)	24‐hour urinary sodium on high sodium diet (mmol/day ± SD)	Net change in urinary sodium (mmol/day) between high and low sodium periods
De'Oliveira 1997	14 ± 6	195 ± 22	181
Dodson_P 1989	137 ± 38	188 ± 60	51
Dodson_X 1989	123 ± 50	199 ± 37	76
Houlihan Losartan 2002	85 ± 44	208 ± ‐63	123
Houlihan Placebo 2002	80 ± 70	204 ± ‐85	124
Imanishi Micro 2001	About 60 (data from figure)	About 140 (data from figure)	‐
Imanishi Normo 2001	About 60 (data from figure)	About 280 (data from figure)	‐
Kwakernaak HTZ 2014	164 ± 73	224 ± 73	60
Kwakernaak Placebo 2014	148 ± 65	224 ± 73	76
Lopes De Faria 1997	99 ± 27	293 ± 80	194
Luik 2002	38 ± 13	249 ± 71	211
Miller 1995	15 ± 6	247 ± 66	232
Miller 1997	19 ± 7	201 ± 10	182
Mulhauser 1996	92 ± 33	199 ± 52	107
Trevisan Micro 1998	26 ± 21	324 ± 127	298
Trevisan Normo 1998	24 ± 21	361 ± 99	337
Vedovato Micro 2004	About 50 (data from figure)	About 280 (data from figure)	‐
Vedovato Normo 2004	About 50 (data from figure)	About 250 (data from figure)	‐
Yoshioka Adva Alb 1998	78 ± 32	164 ± 42	86
Yoshioka Micro 1998	66 ± 16	159 ± 70	93
Yoshioka Normo 1998	58 ± ‐17	190 ± 76	132
Footnotes SD: standard deviation

Appendix 6. Measurements of urinary albumin and protein excretion in included studies

Study	Urine measurement	Difference between high/normal salt and low salt intakes	Significance
Houlihan Placebo 2002	Reported as % change in geometric mean of albumin excretion (tolerance factor)	+38.5% (CI ‐31.7 to 108.8)	NS
Houlihan Losartan 2002		‐43.5% (CI ‐77.5 to ‐9.6)	P = 0.02
Kwakernaak HTZ 2014	Reported as mean albumin excretion (95% CI) Reported as ACR (95% CI)	434 mg/day (306 to 618) to 306 mg/day (203 to 461) 31 mg/mmol (21 to 44) to 23 mg/mmol (15 to 34)	P = 0.0018 P = 0.012
Kwakernaak Placebo 2014	Reported as mean albumin excretion (95% CI) Reported as ACR (95% CI)	711 mg/day (485 to 1043) to 393 mg/day (258 to 599) 52 mg/mmol (35 to 77) to 29 mg/mmol (19 to 45)	P = 0.0002 P = 0.004
Luik 2002	Reported as change in mg/24 hours of albumin	+1.4 mg/24 hours	NS
Mulhauser 1996	Reported as mean proteinuria (range)	High salt: 1.14 g/day (0.1 to 5.06) Low salt: 0.79 g/day (0.1 to 1.39)	NS
Vedovato Normo 2004	Reported as median of albumin (IQR)	9 (6 to 12) µg/min to 8 (5 to 12) µg/min	NS
Vedovato Micro 2004		108 (84 to 178) µg/min to 80 (37 to 108) µg/min	P < 0.001
Footnotes ACR: albumin‐creatinine ratio; CI: confidence interval; IQR: interquartile range; NS: not significant

Appendix 7. Assessment of heterogeneity in short‐term studies: MD for blood pressure compared with urine sodium

Study	MD for SBP	MD for DBP	MD for MAP	Minimum urine sodium (mmol/day)	Net change in urine sodium (mmol/day)
Imanishi Micro 2001	‐11	‐6	‐	Estimated from graph as 60	Estimated from graph as 80
Imanishi Normo 2001	‐7	0	‐	Estimated from graph as 60	Estimated from graph as 220
Luik 2002	‐3.1	‐1.4	‐	38 ± 13	211
Trevisan Micro 1998	‐4	‐2	‐	26 ± 21	298
Trevisan Normo 1998	‐17	‐5	‐	24 ± 21	337
De'Oliveira 1997	‐	‐	‐9	14 ± 6	181
Lopes De Faria 1997	‐	‐	+2	99 ± 27	194
Miller 1995	‐	‐	‐1	15 ± 6	232
Miller 1997	‐	‐	‐1	19 ± 7	182
Vedovato Micro 2004	‐	‐	‐8	Estimated from graph as 50	Estimated from graph as 230
Vedovato Normo 2004	‐	‐	‐1	Estimated from graph as 50	Estimated from graph as 200
Yoshioka Adva Alb 1998	‐	‐	‐6	78 ± 32	86
Yoshioka Micro 1998	‐	‐	‐3	66 ± 17	93
Yoshioka Normo 1998	‐	‐	‐2	58 ± 17	132
Footnotes DBP: diastolic blood pressure; MAP: mean arterial pressure; MD: mean difference; SBP: systolic blood pressure

Appendix 8. GRADE approach

The GRADE approach assesses the certainty of a body of evidence, rating it in one of four grades (GRADE 2008)

• High: we are very confident that the true effect lies close to that of the estimate of the effect;

• Moderate: we are moderately confident in the effect estimate; the true effect is likely to be close the estimate of effect, but there is a possibility that it is substantially different;

• Low: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect; or

• Very low: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

We decreased the certainty of evidence if there was (Balshem 2011):

• serious (‐1) or very serious (‐2) limitation in the study design or execution (risk of bias);

• important inconsistency of results (‐1);

• some (‐1) or major (‐2) uncertainty about the directness of evidence;

• imprecise or sparse data (‐1) or serious imprecision (‐2); or

• high probability of publication bias (‐1).

We increased the certainty of evidence if there was (GRADE 2011):

• a large magnitude of effect (direct evidence, relative risk (RR) = 2 – 5 or RR = 0.5 – 0.2 with no plausible confounders) (+1); very large with RR > 5 or RR < 0.2 and no serious problems with risk of bias or precision; more likely to rate up if effect is rapid and out of keeping with prior trajectory; usually supported by indirect evidence (+2);

• evidence of a dose response gradient (+1); or

• all plausible residual confounders or biases would reduce a demonstrated effect, or suggest a spurious effect when results show no effect (+1).

Data and analyses

Comparison 1. Net change with altering salt diet.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Systolic BP	12		Mean Difference (IV, Random, 95% CI)	‐7.36 [‐10.75, ‐3.98]
1.1.1 Long‐term studies	7		Mean Difference (IV, Random, 95% CI)	‐6.15 [‐9.27, ‐3.03]
1.1.2 Short‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐8.43 [‐14.37, ‐2.48]
1.2 Systolic BP (excluding studies using RAS)	9		Mean Difference (IV, Random, 95% CI)	‐7.44 [‐11.83, ‐3.04]
1.2.1 Long‐term studies	4		Mean Difference (IV, Random, 95% CI)	‐5.55 [‐11.73, 0.62]
1.2.2 Short‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐8.43 [‐14.37, ‐2.48]
1.3 Diastolic BP	12		Mean Difference (IV, Random, 95% CI)	‐3.17 [‐4.58, ‐1.76]
1.3.1 Long‐term studies	7		Mean Difference (IV, Random, 95% CI)	‐3.41 [‐5.56, ‐1.27]
1.3.2 Short‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐2.95 [‐4.96, ‐0.94]
1.4 Diastolic BP (excluding studies using RAS)	9		Mean Difference (IV, Random, 95% CI)	‐2.76 [‐4.53, ‐0.99]
1.4.1 Long‐term studies	4		Mean Difference (IV, Random, 95% CI)	‐2.26 [‐6.54, 2.02]
1.4.2 Short‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐2.95 [‐4.96, ‐0.94]
1.5 MAP	13		Mean Difference (IV, Random, 95% CI)	‐3.01 [‐4.95, ‐1.07]
1.5.1 Long‐term studies	4		Mean Difference (IV, Random, 95% CI)	‐4.60 [‐7.26, ‐1.94]
1.5.2 Short‐term studies	9		Mean Difference (IV, Random, 95% CI)	‐2.37 [‐4.75, 0.01]
1.6 Systolic BP according to presence/absence of albuminuria at enrolment	12		Mean Difference (IV, Random, 95% CI)	‐6.63 [‐10.19, ‐3.08]
1.6.1 No albuminuria	5		Mean Difference (IV, Random, 95% CI)	‐8.08 [‐15.42, ‐0.73]
1.6.2 Albuminuria	7		Mean Difference (IV, Random, 95% CI)	‐5.85 [‐8.76, ‐2.95]
1.7 Diastolic BP according to presence/absence of albuminuria at enrolment	12		Mean Difference (IV, Random, 95% CI)	‐3.54 [‐4.92, ‐2.16]
1.7.1 No albuminuria	5		Mean Difference (IV, Random, 95% CI)	‐3.58 [‐5.75, ‐1.42]
1.7.2 Albuminuria	7		Mean Difference (IV, Random, 95% CI)	‐3.48 [‐5.45, ‐1.52]
1.8 MAP according to the presence/absence of albuminuria at enrolment	10		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.8.1 No albuminuria	3		Mean Difference (IV, Random, 95% CI)	‐0.11 [‐2.27, 2.05]
1.8.2 Albuminuria	7		Mean Difference (IV, Random, 95% CI)	‐5.40 [‐7.72, ‐3.08]
1.9 Creatinine clearance	7		Mean Difference (IV, Random, 95% CI)	‐6.05 [‐10.00, ‐2.10]
1.9.1 Long‐term studies	3		Mean Difference (IV, Random, 95% CI)	‐6.66 [‐18.55, 5.23]
1.9.2 Short‐term studies	4		Mean Difference (IV, Random, 95% CI)	‐7.23 [‐12.88, ‐1.58]
1.10 Glomerular filtration rate	12		Mean Difference (IV, Random, 95% CI)	‐1.87 [‐5.05, 1.31]
1.10.1 Long‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐2.07 [‐5.29, 1.15]
1.10.2 Short‐term studies	7		Mean Difference (IV, Random, 95% CI)	‐3.26 [‐9.93, 3.42]
1.11 Effective renal plasma flow	8		Mean Difference (IV, Random, 95% CI)	‐6.74 [‐17.08, 3.59]
1.11.1 Long‐term studies	3		Mean Difference (IV, Random, 95% CI)	‐0.73 [‐2.83, 1.37]
1.11.2 Short‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐0.14 [‐68.26, 67.98]
1.12 HbA1c	6		Mean Difference (IV, Random, 95% CI)	‐0.62 [‐1.49, 0.26]
1.12.1 Long‐term studies	4		Mean Difference (IV, Random, 95% CI)	‐0.05 [‐0.35, 0.25]
1.12.2 Short‐term studies	2		Mean Difference (IV, Random, 95% CI)	‐1.44 [‐4.47, 1.60]
1.13 Body weight	12		Mean Difference (IV, Random, 95% CI)	‐1.21 [‐1.73, ‐0.68]
1.13.1 Long‐term studies	5		Mean Difference (IV, Random, 95% CI)	‐0.35 [‐1.63, 0.94]
1.13.2 Short‐term studies	7		Mean Difference (IV, Random, 95% CI)	‐1.30 [‐1.89, ‐0.72]
1.14 Systolic BP in cross‐over studies with or without washout between interventions	10		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.14.1 Studies with washout	5		Mean Difference (IV, Random, 95% CI)	‐4.33 [‐7.32, ‐1.33]
1.14.2 Studies without washout	5		Mean Difference (IV, Random, 95% CI)	‐9.92 [‐15.67, ‐4.16]
1.15 MAP in cross‐over studies with or without washout between study periods	12		Mean Difference (IV, Random, 95% CI)	Subtotals only
1.15.1 Studies with washout	6		Mean Difference (IV, Random, 95% CI)	‐2.52 [‐5.50, 0.46]
1.15.2 Studies without washout	6		Mean Difference (IV, Random, 95% CI)	‐3.32 [‐6.28, ‐0.36]

Comparison 2. Net change in BP in hypertensive and normotensive participants.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Systolic BP	10		Mean Difference (IV, Random, 95% CI)	‐7.65 [‐11.69, ‐3.61]
2.1.1 Hypertensive	5		Mean Difference (IV, Random, 95% CI)	‐6.45 [‐11.47, ‐1.42]
2.1.2 Normotensive	5		Mean Difference (IV, Random, 95% CI)	‐8.43 [‐14.37, ‐2.48]
2.2 Diastolic BP	10		Mean Difference (IV, Random, 95% CI)	‐3.08 [‐4.74, ‐1.43]
2.2.1 Hypertensive	5		Mean Difference (IV, Random, 95% CI)	‐3.15 [‐6.49, 0.18]
2.2.2 Normotensive	5		Mean Difference (IV, Random, 95% CI)	‐2.95 [‐4.96, ‐0.94]
2.3 MAP	11		Mean Difference (IV, Random, 95% CI)	‐2.74 [‐4.97, ‐0.51]
2.3.1 Hypertensive	3		Mean Difference (IV, Random, 95% CI)	‐4.88 [‐10.39, 0.63]
2.3.2 Normotensive	8		Mean Difference (IV, Random, 95% CI)	‐2.15 [‐4.56, 0.26]

Comparison 3. Net change in BP in type 1 and type 2 diabetes.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Systolic BP	12		Mean Difference (IV, Random, 95% CI)	‐7.36 [‐10.75, ‐3.98]
3.1.1 Type 1 diabetes	4		Mean Difference (IV, Random, 95% CI)	‐7.35 [‐14.49, ‐0.21]
3.1.2 Type 2 diabetes	8		Mean Difference (IV, Random, 95% CI)	‐7.35 [‐10.32, ‐4.38]
3.2 Diastolic BP	12		Mean Difference (IV, Random, 95% CI)	‐3.17 [‐4.58, ‐1.76]
3.2.1 Type 1 diabetes	4		Mean Difference (IV, Random, 95% CI)	‐3.20 [‐5.16, ‐1.23]
3.2.2 Type 2 diabetes	8		Mean Difference (IV, Random, 95% CI)	‐3.04 [‐5.20, ‐0.89]
3.3 MAP	13		Mean Difference (IV, Random, 95% CI)	‐3.01 [‐4.95, ‐1.07]
3.3.1 Type 1 diabetes	3		Mean Difference (IV, Random, 95% CI)	0.08 [‐1.92, 2.08]
3.3.2 Type 2 diabetes	10		Mean Difference (IV, Random, 95% CI)	‐4.30 [‐6.54, ‐2.05]
3.4 HbA1c	6		Mean Difference (IV, Random, 95% CI)	‐0.62 [‐1.49, 0.26]
3.4.1 Type 1 diabetes	3		Mean Difference (IV, Random, 95% CI)	‐0.94 [‐2.38, 0.51]
3.4.2 Type 2 diabetes	3		Mean Difference (IV, Random, 95% CI)	‐0.12 [‐0.58, 0.34]

3.4. Analysis.

Comparison 3: Net change in BP in type 1 and type 2 diabetes, Outcome 4: HbA1c

Comparison 4. Adverse events.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Orthostatic hypotension	2	180	Risk Ratio (M‐H, Random, 95% CI)	2.50 [0.81, 7.68]

4.1. Analysis.

Comparison 4: Adverse events, Outcome 1: Orthostatic hypotension

Characteristics of studies

Characteristics of included studies [ordered by study ID]

De'Oliveira 1997.

Study characteristics
Methods	Study design Parallel RCT Duration of study/recruitment Not reported Duration of treatment 5 to 7 days Duration of follow‐up 2 weeks from start of treatment
Participants	Study characteristics Setting: single centre Country: USA Inclusion criteria: ≥ 30 years; NIDDM; onset > 30 years; no diabetic ketoacidosis episodes; fasting blood glucose ≥ 140 mg/dL or 2‐hour blood glucose level in standard GTT ≥ 200 mg/dL; hypertension ≥ 140/90 mm Hg Exclusion criteria: azotaemia or gross proteinuria Baseline characteristics Number: intervention group (9); control group (10) Mean age ± SEM (years): intervention group (56 ± 3); control group (55 ± 2) Sex (F/M): intervention group (7/2); control group (9/1) Ethnicity: Caucasian (17); African American (2) Mean BMI ± SEM ( kg/m²): intervention group (28.9 ± 1.5); control group (31.6 ± 1.7) Mean MAP ± SEM (mm Hg): intervention group (102 ± 3); control group (102 ± 6) Mean cholesterol ± SEM (mg/dL): intervention group (234 ± 17); control group (143 ± 10) Diabetes Mean NIDDM duration ± SEM: 7 ± 1 years Mean fasting blood sugar ± SEM (mg/dL): intervention group (181 ± 23); control group (162 ± 23) Mean 2‐hour oral GTT ± SEM: 319 ± 26 mg/dL Mean SCr ± SEM (mg/dL): intervention group (1.16 ± 0.05; control group (1.21 ± 0.06) Mean uric acid ± SEM (mg/dL): intervention group (5.8 ± 0.7); control group (7.0 ± 0.7) CKD definition: not reported
Interventions	Pre‐randomisation/run‐in period Isocaloric diet: 10 mmol NaCl + 100 mmol K+ + 2000 mL water Duration: 5 to 7 days Intervention group Low salt diet NaCl: 10 mmol/day K+: 100 mmol Water: 2000 mL Duration: 5 to 7 days  Control group High salt diet NaCl: 200 mmol/day K+: 100 mmol Water: 2000 mL Duration: 5 to 7 days  Co‐interventions Anti‐hypertensive agents ceased 2 weeks before the study. Participants did not receive ACEi or ARBs during the study
Outcomes	Outcomes reported / outcomes relevant to this review* MAP* ERPF* eGFR* BMI* Fasting blood glucose Cholesterol Uric acid SCr* 24‐hour UNa excretion* PRA Plasma aldosterone FF
Notes	Additional information Exclusions post‐randomisation but pre‐intervention: none Stop or end point/s: not reported Additional data requested from authors: yes, but no response to our request Completeness of follow‐up Analysed: 100% Per cent followed: 100%
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "9 subjects were randomized to remain in the low‐salt diet ... the other 10 were switched to ... 200 mmol NaCl". Comment: insufficient detail was provided about how the randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding the concealment of allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded study
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: unblinded study
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what the final attrition rate is. ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: provided in appendix but not published earlier
Funding source	Low risk	Quote: "this work was supported in part by the National Institutes of Health grants T32 HL‐07609, NCRR GCRC M01RR026376, P01AC00059916, and 1 P50ML53000‐01 and in part by the Canadian Medical Research Council Fellowship Award (Dr. Donald Allan), and a Northern Ireland Council for Postgraduate Medical Education Award (Dr. John McKnight). Dr. Naomi Fisher and Deborah Price were supported by NIH Clinical Associate Physician Awards" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Low risk	Washout period: not applicable to parallel design Comment: no other potential sources of bias identified

Dodson_P 1989.

Study characteristics
Methods	Study design Parallel RCT (followed by a cross‐over study Dodson_X 1989) Duration of study (recruitment) Not clear Duration of treatment 3 months Duration of follow‐up At 3 months end of treatment
Participants	Study characteristics Setting: single centre Country: UK Inclusion criteria: type 2 DM; mild hypertension (SBP > 160 mm Hg or DBP > 95 mm Hg) Exclusion criteria: past or current treatment with insulin; CKD due to diabetes or hypertensive kidney disease (proteinuria or raised SCr); cardiac failure; pregnancy Baseline characteristics Number: intervention group (17); control group (17) Mean age ± SD (years): intervention group (61.9 ± 7.5); control group (61.1 ± 6.3) Sex (F/M): intervention group (12/5); control group (11/6) Body weight (% of ideal body weight): intervention group (126.6 ± 20.2); control group (127.5 ± 24.9) Mean BP SBP/DBP (mm Hg): intervention group (179/98); control group (174/100) Hypertension Duration of 'mild' hypertension (years): intervention group (4.1 ± 3.4); control group (6.5 ± 8.0) Taking atenolol: intervention group (2/17); control group (2/17) Mean HbA1c ± SD (%): intervention group (10.2 ± 1.95); control group (10.4 ± 2.5) Diabetes Type 2 DM Mean duration of diabetes ± SD (years): intervention group (4.1 ± 5.2); control group (5.1 ± 3.4) Taking hypoglycaemic agents: intervention group (2/17); control group (4/17) Mean 24‐hour UNa excretion (mmol/24 hours): intervention group (198.7 ± 65.9); control group (183.2 ± 62.3) Mean 24‐hour urinary potassium excretion (mmol/24 hours): intervention group (65.9 ± 25.4); control group (71.8 ± 27.0) CKD definition: not reported
Interventions	Pre‐randomisation/run‐in No changes, outcome measurements taken Duration: 7 weeks Intervention group Dietary advice: moderate sodium restriction Duration: 3 months Intervention group No advice: continue on normal diabetic diet (low carbohydrate, high fat, low fibre, added salt) Duration: 3 months Co‐interventions None; participants did not receive ACEi or ARB
Outcomes	Reported outcomes (at 3 months) SBP*, DBP* (mm Hg) Body weight* 24‐hour UNa excretion* 24‐hour urinary potassium excretion Alcohol intake Smoking HbA1c* (%)
Notes	Additional information Stop or end point/s: not reported Per cent followed: 100%
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the dietician randomly allocated patients either to receive advice on reducing sodium intake or to a control group" Comment: insufficient information provided on how the randomisation methods were performed by the dietician
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding the concealment of allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded parallel study
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: unblinded parallel study
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what the final attrition rate is. ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	High risk	Quote: "we are grateful to CIBA (Horsham) for supplies of Slow Sodium and placebo" Comment: pharmaceutical industry funding of study drugs
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Low risk	Washout period for cross‐over trials: not applicable, parallel phase Comment: no other potential sources of bias identified

Dodson_X 1989.

Study characteristics
Methods	Study design Cross‐over study after parallel study (Dodson_P 1989) Duration of study (recruitment) Not clear Duration of treatment 1 month Duration of follow‐up At end of treatment
Participants	Study characteristics Setting: single centre Country: UK Inclusion criteria: type 2 DM; mild hypertension (SBP > 160 mm Hg or DBP > 95 mm Hg) Exclusion criteria: past or current treatment with insulin; CKD due to diabetes or hypertensive kidney disease (proteinuria or raised SCr); cardiac failure; pregnancy Baseline characteristics Number (randomised/analysed): 13/9 Mean age ± SD: 62 ± 6.5 years Sex (M/F): 6/3 Mean body weight ± SD: 77.7 ± 4.8 kg Mean SBP/DBP ± SD (mm Hg): 163.8 ± 15.9/95.3 ± 8.1 Hypertension: 'mild' HbA1c%: not reported Diabetes Type 2 DM Mean duration: 5 years Co‐morbidities: not reported CKD definition: not reported 24‐hour UNa excretion: 141.5 ± 47.7 mmol/24 hours 24‐hour urinary potassium excretion: 72.3 ± 20.6 mmol/24 hours
Interventions	Pre‐randomisation 13 patients who just participated in 3 months of 'moderate dietary salt restriction diet' via dietician advice: from the treatment arm of Dodson_P 1989 Washout period: no washout applied Intervention group Continue on 'moderate dietary salt restriction diet' + placebo oral tablet Duration: daily, for 1 month Control group Continue on 'moderate dietary salt restriction diet' + slow‐release sodium supplement oral tablet 80 mmol/day Duration: daily, for 1 month Co‐interventions or additional treatments None; participants did not receive ACEi or ARB Follow up details No follow‐up past the end of the treatment period (2 months in total)
Outcomes	Reported outcomes (at 3 months) / outcomes relevant to this review* SBP*, DBP* (mm Hg) Body weight* 24‐hour urinary sodium excretion* HbA1c* (%)
Notes	Additional information Stop or end point/s: not reported Per cent followed: 100%
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "six men and three women ... completed the randomised double‐blind crossover trial" Comment: insufficient information provided on how the randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding the concealment of allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "six men and three women ... completed the randomised double‐blind crossover trial" Comment: the study states to be double‐blind but there is insufficient information provided on how the blinding of either the participants or the study personnel was undertaken
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding any blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: all participants were accounted for from start to end of the trial Comment: high attrition rate. 4 participants (attrition = 31%) withdrew due to adverse effects (from both sodium and placebo groups) Comment: unclear if ITT analysis was performed or original participant sample (ITT = 13?) as group totals are not provided for final analysis
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	High risk	Quote: "we are grateful to CIBA (Horsham) for supplies of Slow Sodium and placebo". Comment: pharmaceutical industry funding of study drugs
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Unclear risk	Washout period for cross‐over trials: no washout period Comment: no other potential sources of bias identified

Houlihan Losartan 2002.

Study characteristics
Methods	Study design Parallel RCT, with cross‐over dietary salt Duration of study 3 months Duration of treatment 3 months Duration of follow‐up At end of 3 months of treatment
Participants	Study characteristics Setting: multicentre (Austin and Repatriation Medical Centre diabetes clinic, and surrounding district) Country: Australia Inclusion criteria: 30 to 75 years; type 2 diabetes; hypertension: seated BP > 130/85 mm Hg; albuminuria (AER 10 to 200 µg/min); HbA1c < 11.0%; habitual 24‐hour UNa excretion > 100 mmol/24 hours Exclusion criteria: serious systemic illness; history of substance abuse; seated BP > 165/100 mm Hg; serum potassium > 5.5 mmol/L; SCr > 200 µmol/L; long term use of NSAIDs; recurrent UTIs (> 3/year); BMI > 35 kg/m²; cardiac failure; nitrate therapy; intolerance of ACEi Baseline characteristics Number: 20 Mean age ± SEM (years): losartan group (60.6 ± 3.7); control group (63.1 ± 3.9) Sex (M/F): losartan group (10/0); control group (9/1) Mean BMI ± SEM (kg/m²): losartan group (30.4 ± 2.1); control group (28.1 ± 1.6) Mean MAP ± SEM (mm Hg): losartan group (114 ± 3); control group (111 ± 3) Hypertension: hypertensive > 130/85 mm Hg Mean HbA1c ± SEM (%): 7.4 ± 0.4 losartan group (7.9 ± 0.5); control group (7.4 ± 0.4) Diabetes Type 2 DM Duration of diabetes: median 4.0 years (range 1 to 10) Co‐morbidities: none CKD definition: AER: 10 to 200 µg/min Mean AER ± SEM (µg/min): losartan group (26.6 ± 1.4); control group (32.6 ± 1.3) Mean SCr ± SEM (mmol/L): losartan group (97 ± 6.5); control group (92 ± 2.7)
Interventions	Pre‐randomisation/run‐in period Antihypertensives, including ACEi, ARBs and diuretics, stopped for at least 2 weeks before commencing study Intervention group Losartan (oral): 50 mg/day Low sodium diet: 50 to 70 mmol/day Duration: 2 weeks Control group Losartan (oral): 50 mg/day Regular sodium diet: > 100 mmol/day Duration: 2 weeks Washout period 4 weeks washout between treatments (switch from low to regular and regular to low sodium diet) Co‐interventions Losartan (ARB) for the entire 8 weeks  Follow up details No follow‐up past the end of the treatment period (12 weeks in total)
Outcomes	Reported outcomes (at 2 and 4 weeks) / outcomes relevant to this review* 24‐hour ambulatory BP: SBP* and DBP* MAP* eGFR* ERPF* HbA1c* (%) Body weight* (kg) ACR on 24‐hour urine collection Plasma glucose Electrolytes PRA Angiotensin II Aldosterone Urinary electrolytes Urea SCr CrCl
Notes	Additional information Additional data requested from authors: yes; no response to our request
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "patients were randomly assigned in a double‐blind fashion to receive losartan ... or placebo" Comment: insufficient information provided on how the randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding the concealment of allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "patients were randomly assigned in a double‐blind fashion to receive losartan ... or placebo" Comment: insufficient information provided on how the blinding of either the participants or the study personnel was undertaken
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding any blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition rates are not reported. Unclear from the report what the final attrition rate is. Unclear whether an ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	High risk	Quote: "this work was supported by a Merck medical school grant and an Apex Diabetes Australia Research grant. During the period of this work, CAH was supported by grants from the Austin Hospital Medical Research Foundation and the National Health and Medical Research Council" Comment: pharmaceutical industry funding
Conflicts of interest	High risk	Quote: "MEC is on an advisory panel for Merck on diabetic nephropathy, he has received honoraria for speaking engagements for Merck, and his laboratory has received funding from Merck for studies on a new drug to treat retinopathy" Comment: conflicted interests
Other bias	Low risk	Washout period for cross‐over studies: 4 weeks washout between treatment arms Comment: no other potential sources of bias identified

Houlihan Placebo 2002.

Study characteristics
Methods	Study design Parallel RCT, with cross‐over dietary salt Duration of study 3 months Duration of treatment 3 months Duration of follow‐up At end of 3 months of treatment
Participants	Study characteristics Setting: multicentre (Austin and Repatriation Medical Centre diabetes clinic, and surrounding district) Country: Australia Inclusion criteria: 30 to 75 years; type 2 diabetes; hypertension: seated BP > 130/85 mm Hg; albuminuria (AER 10 to 200 µg/min); HbA1c < 11.0%; habitual 24‐hour UNa excretion > 100 mmol/24 hours Exclusion criteria: serious systemic illness; history of substance abuse; seated BP > 165/100 mm Hg; serum potassium > 5.5 mmol/L; SCr > 200 µmol/L; long term use of NSAIDs; recurrent UTIs (> 3/year); BMI > 35 kg/m²; cardiac failure; nitrate therapy; intolerance of ACEi Baseline characteristics Number: 20 Mean age ± SEM (years): losartan group (60.6 ± 3.7); control group (63.1 ± 3.9) Sex (M/F): losartan group (10/0); control group (9/1) Mean BMI ± SEM (kg/m²): losartan group (30.4 ± 2.1); control group (28.1 ± 1.6) Mean MAP ± SEM (mm Hg): losartan group (114 ± 3); control group (111 ± 3) Hypertension: hypertensive > 130/85 mm Hg Mean HbA1c ± SEM (%): 7.4 ± 0.4 losartan group (7.9 ± 0.5); control group (7.4 ± 0.4) Diabetes Type 2 DM Duration of diabetes: median 4.0 years (range 1 ‐ 10) Co‐morbidities: none CKD definition: AER: 10 to 200 µg/min Mean AER ± SEM (µg/min): losartan group (26.6 ± 1.4); control group (32.6 ± 1.3) Mean SCr ± SEM (mmol/L): losartan group (97 ± 6.5); control group (92 ± 2.7)
Interventions	Pre‐randomisation/run‐in period Antihypertensives, including ACEi, ARBs and diuretics, stopped for at least 2 weeks before commencing study Intervention group Losartan placebo Low sodium diet: 50 to 70 mmol/day Duration: 2 weeks Control group Losartan placebo Regular sodium diet: > 100 mmol/day Duration: 2 weeks Washout period 4 weeks washout between treatments Co‐interventions Placebo oral tablet for the entire 8 weeks Follow up details No follow‐up past the end of the treatment period (12 weeks in total)
Outcomes	Reported outcomes (at 2 and 4 weeks) / outcomes relevant to this review* 24‐hour ambulatory BP: SBP* and DBP* MAP* eGFR* ERPF* HbA1c* (%) Body weight* (kg) ACR on 24‐hour urine collection Plasma glucose Electrolytes PRA Angiotensin II Aldosterone Urinary electrolytes Urea SCr* CrCl*
Notes	Additional information Additional data requested from authors: yes; no response to request
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "patients were randomly assigned in a double‐blind fashion to receive losartan ... or placebo" Comment: insufficient information provided on how the randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding the concealment of allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "patients were randomly assigned in a double‐blind fashion to receive losartan ... or placebo" Comment: insufficient information provided on how the blinding of either the participants or the study personnel was undertaken
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding any blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition rates are not reported. Unclear from the report what the final attrition rate is. Unclear whether an ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	High risk	Quote: "This work was supported by a Merck medical school grant and an Apex Diabetes Australia Research grant. During the period of this work, CAH was supported by grants from the Austin Hospital Medical Research Foundation and the National Health and Medical Research Council" Comment: pharmaceutical industry funding
Conflicts of interest	High risk	Quote: "MEC is on an advisory panel for Merck on diabetic nephropathy, he has received honoraria for speaking engagements for Merck, and his laboratory has received funding from Merck for studies on a new drug to treat retinopathy" Comment: conflicted interests
Other bias	Low risk	Washout period for cross‐over trials: 4 weeks washout between treatment arms. Comment: no other potential sources of bias identified

Imanishi Micro 2001.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study 4 weeks Duration of treatment 1 week Duration of follow‐up No follow‐up past the end of the treatment period (approximately 4 weeks in total)
Participants	Study characteristics Setting: single centre (hospital inpatient centre) Location: Japan Inclusion criteria: type 2 DM with simple retinopathy (divided into microalbuminuria and normoalbuminuria) and normal BP Exclusion criteria: history of non‐diabetic kidney disease; history of heart disease; history of UTI; SCr ≥ 97.4 µmol/L; needing antihypertensive drugs; found to have hypertension at study entry (4) Baseline characteristics (microalbuminuria group) Microalbuminuria: 28.8 to 288 mg/24 hours Number (randomised/analysed): 12/8 Mean age ± SD: 59 ± 10 years Sex (M/F): 7/5 BMI: 24.4 ± 4.4 kg/m² Mean BP ± SD (mm Hg): SBP (136 ± 9); DBP (82 ± 6) Mean HbA1c ± SD (%): 8. 1 ± 1.8 Diabetes Type 2 DM Simple diabetic retinopathy: 100% Co‐morbidities: not reported CKD definition: microalbuminuria 20 to 200 µg/min Mean SCr ± SD: 60.4 ± 13.5 mmol/L
Interventions	Pre‐randomisation/run‐in Plasma glucose levels were brought under control during ≥ 2 weeks of hospitalisation Intervention group Microalbuminuria Low sodium diet: ≈ 80 mmol/day Duration: 7 days Control group Microalbuminuria Regular sodium diet: ≈ 200 mmol/day Duration: 7 days Washout No washout; immediately switched to other group, with no time intervening Co‐interventions or additional treatments None Follow up details No follow‐up past the end of the treatment period (approximately 4 weeks in total, including run‐in)
Outcomes	Reported outcomes (at 7 days) / outcomes relevant to this review* SBP*, DBP* CrCl* UAE* BMI* Fasting plasma glucose HbA1c* (%) SCr* ERPF* PRA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the patients were put on a diet ... in random order" Comment: insufficient information provided on how randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded study
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: unblinded study
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition rates are not reported. Unclear from the report what the final attrition rate is. Unclear whether an ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: provided but not published
Funding source	Low risk	Quote: "This study was supported by the Hohansha Foundation (Osaka, Japan) and a grant from the Osaka City General Hospital for medical research" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: no washout period applied, participants immediately switched to the other group, "with no time intervening" Comment: no other potential sources of bias identified

Imanishi Normo 2001.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study 4 weeks Duration of treatment 1 week Duration of follow‐up No follow‐up past the end of the treatment period (approximately 4 weeks in total)
Participants	Study characteristics Setting: single centre (hospital inpatient centre) Location: Japan Inclusion criteria: type 2 DM with simple retinopathy (divided into microalbuminuria and normoalbuminuria) and normal BP Exclusion criteria: history of non‐diabetic kidney disease; history of heart disease; history of UTI; SCr ≥ 97.4 µmol/L; needing antihypertensive drugs; found to have hypertension at study entry (3) Baseline characteristics (normoalbuminuria group) Normoalbuminuria: < 28.8 mg/24 hour Number (randomised/analysed): 11/8 Mean age ± SD: 61 ± 10 years Sex (M/F): 5/6 BMI: 22.0 ± 2.1 kg/m² Mean BP ± SD (mm Hg): SBP (132 ± 11); DBP (73 ± 7) Mean HbA1c ± SD (%): 8.6 ± 1.1 Diabetes Type 2 DM Simple diabetic retinopathy: 100% Co‐morbidities: not reported CKD definition: UAE < 20 µg/min Mean SCr ± SD: 53.8 ± 13.9 mmol/L
Interventions	Pre‐randomisation/run‐in Plasma glucose levels were brought under control during 2 or more weeks of hospitalisation Intervention group Normoalbuminuria Low sodium diet: ≈ 80mmol/day Duration: 7 days Control group Normoalbuminuria Regular sodium diet: ≈ 200 mmol/day Duration: 7 days Washout No washout; immediately switched to other group, with no time intervening Co‐interventions or additional treatments None Follow up details No follow‐up past the end of the treatment period (approximately 4 weeks in total, including run‐in)
Outcomes	Reported outcomes (at 7 days) / outcomes relevant to this review* SBP*, DBP* CrCl* UAE* BMI* Fasting plasma glucose HbA1c* (%) SCr* ERPF* PRA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the patients were put on a diet ... in random order" Comment: insufficient information provided about how the randomisation methods were undertaken
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded study
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: unblinded study
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition rates are not reported. Unclear from the report what the final attrition rate is. Unclear whether an ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: provided but not published
Funding source	Low risk	Quote: "this study was supported by the Hohansha Foundation (Osaka, Japan) and a grant from the Osaka City General Hospital for medical research" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: no washout period applied, participants immediately switched to the other group, "with no time intervening" Comment: no other potential sources of bias identified

Kwakernaak HTZ 2014.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study October 2009 to December 2012 Duration of treatment periods 6 weeks Duration of follow‐up No follow‐up past the end of the total treatment period (30 weeks in total, including a 6‐week run‐in)
Participants	Study characteristics Setting: multicentre (3 sites; University Medical Center Groningen, ZGT Hospital Almelo, and Medical Center Leeuwarden) Country: The Netherlands Inclusion criteria: ≥ 18 years; type 2 DM; DKD; albuminuria present at screening and after run in period > 30 mg/day; urinary albumin concentration > 20 mg/L; or urinary ACR > 2.5 mg/mmol for men and > 3.5 mg/mmol for women; CrCl ≥ 30 mL/min with < 6 mL/min decline in previous year Exclusion criteria: BP ≥ 180/110 mm Hg; overt nephrotic syndrome at baseline; second primary kidney disease in addition to DKD; type 1 diabetes; renovascular hypertension; cardiovascular or cerebrovascular events within 3 months before study; serum potassium ≥ 6.0 mmol/L; transplantation or immunosuppressive treatment; contraindication for the use of lisinopril or hydrochlorothiazide; pregnancy or lactation; poor compliance with medication Baseline characteristics (both groups) Number: 45 Mean age ± SD: 65 ± 9 years Sex (M/F): 38/7 Ethnicity (white): 100% Mean BMI ± SD: 32 ± 5 kg/m² BP (SBP/DBP): < 180/110 mm Hg Mean HbA1c ± SD (%): 7.1 ± 0.9 Diabetes: type 1 Co‐morbidities: macrovascular disease CKD definition: CrCl ≥ 30 mL/min; with < 6 mL/min decline in previous year CKD diagnosis: stages 2 to 5 Albuminuria (mg/day), geometric mean (95% CI): 711 (485 to 1043) Mean CrCl ± SD: 101 ± 47 mL/min Mean eGFR ± SD: 65 ± 25 mL/min/1.73 m²
Interventions	Pre‐randomisation/run‐in period Patients titrated to maximum dose of ACEi (lisinopril 40 mg/day) All other RAS blockers and diuretics discontinued Duration: 6 weeks Intervention group Hydrochlorothiazide: maximum dose 50 mg/day Dietary change: target 50 mmol dietary sodium/day Duration: 6 weeks Control group Hydrochlorothiazide: maximum dose 50 mg/day Dietary change: target 200 mmol dietary sodium/day Duration: 6 weeks Washout period Placebo phase of 6 weeks in between dietary change periods considered to be sufficient washout time Co‐interventions or additional treatments Hydrochlorothiazide oral tablet, maximum dose 50 mg/day, for 6 weeks Follow up details No follow‐up past the end of the treatment period (30 weeks in total)
Outcomes	Reported outcomes (at end of each 6‐week period) / outcomes relevant to this review* 24‐hour UAE* SBP*, DBP* CrCl* eGFR* SCr* Body weight* Serum sodium Serum potassium Serum urea Serum albumin Serum glucose Total cholesterol Urinary excretion of protein, urea and potassium Urinary ACR Urinary PCR
Notes	Additional information Compliance with sodium restriction assessed by 24‐hour UAE at the middle and end of each 6‐week period
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “an independent pharmacist used a computer program to randomise patients in blocks of two. No stratification was needed as the trial has a crossover design. An independent pharmacist randomised treatment sequences. Patients were sequentially enrolled according to moment of recruitment"
Allocation concealment (selection bias)	Low risk	Quote: "the randomisation code remained a secret during the entire trial"
Blinding of participants and personnel (performance bias) All outcomes	High risk	DIETARY COMPONENT Comment: the dietary component (regular sodium or sodium restriction) was not possible to blind. Open‐label Quote: “the drug intervention was double blind, whereas the dietary intervention was open label” DRUG COMPONENT Comment: the drug component (hydrochlorothiazide or placebo) was blinded within each cross‐over phase Quote: “all patients, investigators, and health‐care providers were masked, apart from the pharmacist who did the randomisation" Quote: "on completion of the trial, the principal investigator (AJK) provided the pharmacist and the medical ethics committee with a written statement that the trial was completed, after which masking ended”
Blinding of outcome assessment (detection bias) All outcomes	High risk	DIETARY COMPONENT Comment: the dietary component (regular sodium or sodium restriction) was not possible to blind. Open‐label Quote: “the drug intervention was double blind, whereas the dietary intervention was open label” DRUG COMPONENT Comment: the drug component (hydrochlorothiazide or placebo) was blinded within each cross‐over phase Quote: “all patients, investigators, and health‐care providers were masked, apart from the pharmacist who did the randomisation" Quote: "on completion of the trial, the principal investigator (AJK) provided the pharmacist and the medical ethics committee with a written statement that the trial was completed, after which masking ended”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: all participants were accounted for from start to end of the trial. Only 2 participants (attrition rate = 4%) withdrew. ITT analysis was performed for the original participant sample, including those 2 withdrawals (ITT: N = 45)
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: provided online as supplementary materials but not published prior to trial Comment: baseline BP and bodyweight are not recorded in Table 1. However, the difference is mentioned (unclear if the baseline data is before or after run‐in period)
Funding source	Low risk	Quote: "the sponsor of this trial is the University Medical Center Groningen, Groningen, The Netherlands. The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. No funding was received for this trial"
Conflicts of interest	Low risk	Quote: "we declared that we have no competing interests"
Other bias	High risk	Washout period for cross‐over trials: no washout applied, possible carryover effects Quote: "to prevent systematic errors resulting from the crossover design, the different treatment periods were done in random order. The trial protocol (appendix) did not include a washout period between treatment periods because of the randomisation procedure and the short half‐life of both interventions" Comment: no other potential sources of bias identified

Kwakernaak Placebo 2014.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study October 2009 to December 2012 Duration of treatment periods 6 weeks Duration of follow‐up No follow‐up past the end of the total treatment period (30 weeks in total, including a 6‐week run‐in)
Participants	Study characteristics Setting: multicentre (3 sites; University Medical Center Groningen, ZGT Hospital Almelo, and Medical Center Leeuwarden) Country: The Netherlands Inclusion criteria: ≥ 18 years; type 2 DM; DKD; albuminuria present at screening and after run in period > 30 mg/day; urinary albumin concentration > 20 mg/L; or urinary ACR > 2.5 mg/mmol for men and > 3.5 mg/mmol for women); CrCl ≥ 30 mL/min with < 6 mL/min decline in previous year Exclusion criteria: BP ≥ 180/110 mm Hg; overt nephrotic syndrome at baseline; second primary kidney disease in addition to DKD; type 1 diabetes; renovascular hypertension; cardiovascular or cerebrovascular events within 3 months before study; serum potassium ≥ 6.0 mmol/L; transplantation or immunosuppressive treatment; contraindication for the use of lisinopril or hydrochlorothiazide; pregnancy or lactation; poor compliance with medication Baseline characteristics (both groups) Number: 45 Mean age ± SD: 65 ± 9 years Sex (M/F): 38/7 Ethnicity (white): 100% Mean BMI ± SD: 32 ± 5 kg/m² BP (SBP/DBP): < 180/110 mm Hg Mean HbA1c ± SD (%): 7.1 ± 0.9 Diabetes: type 1 Co‐morbidities: macrovascular disease CKD definition: CrCl ≥ 30 mL/min; with < 6 mL/min decline in previous year CKD diagnosis: stages 2 to 5 Albuminuria (mg/day), geometric mean (95% CI): 711 (485 to 1043) Mean CrCl ± SD: 101 ± 47 mL/min Mean eGFR ± SD: 65 ± 25 mL/min/1.73 m²
Interventions	Pre‐randomisation/run‐in period Patients titrated to maximum dose of ACEi (lisinopril 40 mg/day) All other RAS blockers and diuretics discontinued Duration: 6 weeks Intervention group Placebo Dietary change: target 50 mmol dietary sodium/day Duration: 6 weeks Control group Placebo Dietary change: target 200 mmol dietary sodium/day Duration: 6 weeks Washout period Placebo phase of 6 weeks in between dietary change periods considered to be sufficient washout time Co‐interventions or additional treatments Placebo oral tablet, for 6 weeks Follow up details No follow‐up past the end of the treatment period (30 weeks in total)
Outcomes	Reported outcomes (at end of each 6 week period) / outcomes relevant to this review* 24‐hour UAE* SBP*, DBP* CrCl* eGFR* SCr* Body weight* Serum sodium Serum potassium Serum urea Serum albumin Serum glucose Total cholesterol Urinary excretion of protein, urea and potassium Urinary ACR Urinary PCR
Notes	Additional information Compliance with sodium restriction assessed by 24‐hour UAE at the middle and end of each 6‐week period
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: “an independent pharmacist used a computer program to randomise patients in blocks of two. No stratification was needed as the trial has a crossover design. An independent pharmacist randomised treatment sequences. Patients were sequentially enrolled according to moment of recruitment"
Allocation concealment (selection bias)	Low risk	Quote: "the randomisation code remained a secret during the entire trial"
Blinding of participants and personnel (performance bias) All outcomes	High risk	DIETARY COMPONENT Comment: the dietary component (regular sodium or sodium restriction) was not possible to blind. Open‐label Quote: “the drug intervention was double blind, whereas the dietary intervention was open label” DRUG COMPONENT Comment: the drug component (hydrochlorothiazide or placebo) was blinded within each cross‐over phase Quote: “all patients, investigators, and health‐care providers were masked, apart from the pharmacist who did the randomisation" Quote: "on completion of the trial, the principal investigator (AJK) provided the pharmacist and the medical ethics committee with a written statement that the trial was completed, after which masking ended”
Blinding of outcome assessment (detection bias) All outcomes	High risk	DIETARY COMPONENT Comment: the dietary component (regular sodium or sodium restriction) was not possible to blind. Open‐label Quote: “the drug intervention was double blind, whereas the dietary intervention was open label” DRUG COMPONENT Comment: the drug component (hydrochlorothiazide or placebo) was blinded within each cross‐over phase Quote: “all patients, investigators, and health‐care providers were masked, apart from the pharmacist who did the randomisation" Quote: "on completion of the trial, the principal investigator (AJK) provided the pharmacist and the medical ethics committee with a written statement that the trial was completed, after which masking ended”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Comment: all participants were accounted for from start to end of the trial. Only 2 participants (attrition rate = 4%) withdrew. ITT analysis was performed for the original participant sample, including those 2 withdrawals (ITT: N = 45)
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: provided online as supplementary materials but not published prior to trial Comment: baseline BP and body weight are not recorded in Table 1. However, the difference is mentioned (unclear if the baseline data is before or after run‐in period)
Funding source	Low risk	Quote: "the sponsor of this trial is the University Medical Center Groningen, Groningen, The Netherlands. The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. No funding was received for this trial"
Conflicts of interest	Low risk	Quote: "we declared that we have no competing interests"
Other bias	High risk	Washout period for cross‐over trials: no washout applied, possible carryover effects Quote: "to prevent systematic errors resulting from the crossover design, the different treatment periods were done in random order. The trial protocol (appendix) did not include a washout period between treatment periods because of the randomisation procedure and the short half‐life of both interventions" Comment: no other potential sources of bias identified

Lopes De Faria 1997.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre (diabetes outpatient clinic) Country: UK Inclusion criteria: 10 to 55 years; IDDM; normal BP (< 140/90 mm Hg); normoalbuminuria (< 20 µg/min 4 to 6 months preceding study); < 31 years old at diagnosis of diabetes; body weight within 20% of ideal range (scale not mentioned); regular isocaloric diet (25% protein, 25% fat, 50% carbohydrate) Exclusion criteria: no other medications other than insulin; evidence of endocrinologic, hepatic, cardiac, or kidney disease; oral contraceptives or oestrogens; hypertension Baseline characteristics Number: 10 Mean age ± SD (years): 30 ± 3 years Sex (M/F): 7/3 Diabetes: IDDM Co‐morbidities: none CKD definition: CKD not inclusion criteria
Interventions	Pre‐randomisation/run‐in period None reported Intervention group Patients' own regular salt diet: ≈ 100 mmol sodium/day Duration: 7 days Control group Regular salt diet: ≈ 100 mmol sodium/day Sodium chloride 0.5 g oral capsules (32 taken/day to elevate salt intake to ≈ 300 mmol/day Duration: 7 days Washout period 7 days interval between diets Co‐interventions Potassium intake fixed in both groups to approximately 79 to 80 mmol/day No participant received ACEi or ARB Follow up details No follow‐up past the end of the treatment period (2 weeks in total)
Outcomes	Reported outcomes / outcomes relevant to this review* Compliance with diets: urinary electrolyte excretion (days 5, 6, and 7) BP* (days 5, 6, and 7) Body weight* (days 5, 6, and 7) Mean 24‐hour urinary sodium excretion (daily) Plasma sodium Urinary potassium Sodium p‐amino‐hippurate (every 24 hours) Polyfructosane HCT Blood glucose HbA1c* (%) Electrolytes eGFR* Increased RPF MAP* PRA
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the order of sodium intake was randomized ... between the two diets" Comment: insufficient information provided about how the randomisation methods were undertaken
Allocation concealment (selection bias)	High risk	Comment: open‐label. All participants knew they were starting on the 'regular salt' first before changing to 'higher salt' capsules week. No placebo appears to have been provided for the first week
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: open‐label. All participants and personnel knew which week they had to change tablets from low salt to high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: open‐label. No information provided regarding the blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition are not reported. Unclear what final attrition is and no ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	Unclear risk	Comment: funding sources are not clearly stated
Conflicts of interest	Low risk	Quote: "JLdF and RF were recipients of scholarships from Conselho National de Pesquisa, Brazil. SdC was a visiting research assistant from Ospedale Casa Sollievo della Sofferenza, Italy"
Other bias	Low risk	Washout period for cross‐over trials: 7 days interval between diets Comment: no other potential sources of bias identified

Luik 2002.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre (University Hospital Department) Country: The Netherlands Inclusion criteria: type 1 DM (age of onset < 35 years); HbA1c < 8.0% (for adequate metabolic control, average of 61 ± 18 units insulin during days before trial start); normotensive (< 140/85 mm Hg); normoalbuminuria (< 30 mg/24 hours) Exclusion criteria: not reported Baseline characteristics Number: 24 Mean age ± SD: 28.2 ± 6 years Sex (M/F): 15/9 Mean BMI ± SD (kg/m²): intervention group (23.5 ± 2.2); control group (23.7 ± 2.2_ Hypertension: normotensive Diabetes: Type 1 DM, ketosis prone (mean age of onset 15.8 ± 6.5 years) Co‐morbidities: none reported CKD definition: CKD not inclusion criteria
Interventions	Pre‐randomisation/run‐in period None reported Intervention group Dietary changes: 50 mmol sodium/day, via dietary changes, as advised by a dietician Duration: 7 days Intervention group Dietary changes: 200 mmol sodium/day, via dietary changes, as advised by a dietician Duration: 7 days Washout period 10 to 17 days Co‐interventions or additional treatments No participants received ACEi or ARB Follow up details No follow‐up past the end of the treatment period (2 weeks in total)
Outcomes	Reported outcomes (various time points) / outcomes relevant to this review* eGFR* ERPF* Serum electrolytes SCr* Serum liver enzymes Blood count Plasma glucose concentration HbA1c* SBP*, DBP* Albuminuria Urinary ACR* CrCl* BMI* Body weight*
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "the sequence of the diet was randomized by drawing an allocation number from closed envelopes"
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: open‐label. All participants and personnel knew which week they had to change their diet to low salt and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: open‐label. No information provided regarding the blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition are not reported. Unclear what final attrition is. ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	Low risk	Quote: "this work was supported by grants from the Dutch Kidney Foundation and Diabetes Fonds Nederland (Diabetes Research fund)" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Low risk	Washout period for cross‐over trials: "Crossover to treatment groups was separated by 10 ‐ 17 days to rule out any carryover effects" Comment: no other sources of bias identified

Miller 1995.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre (Clinical Investigation Unit of Toronto Hospital) Country: Canada Inclusion criteria: insulin‐dependent diabetic males Exclusion criteria: orthostatic hypertension; retinopathy; increased creatinine or evidence of microalbuminuria; signs of autonomic neuropathy Baseline characteristics Number: 9 Mean age ± SE: 25.6 ± 1.5 years Sex (M/F): 9/0 Mean body weight ± SEM (kg): intervention group (80 ± 4); control group (81 ± 4) Body fat: non‐obese Smoker: non‐smokers Mean MAP ± SE (mm Hg): intervention group (78 ± 2); control group (79 ± 1) Hypertension: normotensive Mean HbA1c ± SE (%): 8.2 ± 1.6 Diabetes IDDM, diagnosed within 5 years of the study Mean duration of diabetes: 3.5 ± 0.5 years (range 2 to 5 years) Co‐morbidities: none CKD definition: CKD not a criteria Albumin (microalbuminuria): 8.5 ± 1.3 µg/min Mean CrCl ± SE: 120 ± 12 mL/min Mean 24‐hour UNa excretion ± SE (mmol/day): intervention group (15 ± 2); control group (247 ± 22) Mean 24‐hour urinary potassium excretion SE (mmol/day): intervention group (77 ± 15); control group (68 ± 7)
Interventions	Pre‐randomisation/run‐in period None reported Intervention group Dietary changes: restriction to 10 to 20 mmol sodium/day, as advised by a dietician Duration: 7 days Control group Dietary changes: increase to 200 mmol sodium/day, as advised by a dietician Duration: 7 days Washout Not reported, appears no washout period was applied Co‐interventions or additional treatments No participants received ACEi or ARB Follow up details No follow‐up past the end of the treatment period (2 weeks in total)
Outcomes	Reported outcomes (7 days) / outcomes relevant to this review* HbA1c* 24‐hour UNa excretion* 24‐hour urine potassium excretion Central venous pressure CrCl* Microalbuminuria MAP* Heart rate (bpm) PRA Plasma norepinephrine Forearm vascular response Aldosterone HCT
Notes	Additional information Additional non‐diabetic control group not included in this review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "diets were randomly determined by the research dietician" Comment: investigator randomisation is not true randomisation, and no further details are provided about the methods used by the dietician to randomise participants
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: open‐label. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "investigators were blinded to the diets" Comment: no further details are provided about how the investigators were kept blind. It is also unclear whether the 'investigators' are the study personnel or the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition are not reported. Unclear what final attrition is. ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	Low risk	Quote: "This work was supported by Grant 92‐27 from the physicians of Ontario through the Physicians Services Inc. Foundation. JA Miller is the recipient of a Research Scholarship from the Kidney Foundation of Canada" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: not mentioned, appears no washout period was applied. Possible carryover effects Comment: no other potential sources of bias identified

Miller 1997.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre (Clinical Investigation Unit of Toronto Hospital) Country: Canada Inclusion criteria: early IDDM; males Exclusion criteria: orthostatic hypertension; retinopathy; decreased creatinine or evidence of microalbuminuria; signs of autonomic neuropathy Baseline characteristics Number: 12 Age ± SE: 23 ± 2 years Sex (M/F): 12/0 Mean body weight ± SE (kg): intervention group (79 ± 3); control group (80 ± 2) Mean MAP ± SE (mm Hg): intervention group (79 ± 2); control group (80 ± 1) Hypertension: normotensive (< 140/90 mm Hg) HbA1c: < 10% Diabetes Early IDDM Duration of diabetes: 2.8 ± 0.4 years Co‐morbidities: none CKD definition: CKD not a criteria Mean Na+ excretion ± SE (mmol/day): intervention group (19 ± 2); control group (201 ± 10) Mean UNa excretion ± SE (µmol/min): intervention group (124 ± 33); control group (254 ± 34)
Interventions	Pre‐randomisation/run‐in period None reported Intervention group Dietary changes: 20 mmol sodium/day, as advised by a dietician Duration: 7 days Intervention group Dietary changes: 200 mmol sodium/day, as advised by a dietician Duration: 7 days Washout 2 weeks of normal diet between each treatment period Co‐interventions or additional treatments No participants received ACEi or ARBs Follow up details No follow‐up past the end of the treatment period (2 weeks in total)
Outcomes	Reported outcomes (at 7 days) / outcomes relevant to this review* Body weight* 24‐hour urine sodium 24‐hour UNa excretion* 24‐hour urine potassium excretion 24‐hour urea excretion MAP* (mm Hg) Heart rate (bpm) PRA Plasma norepinephrine eGFR* Aldosterone HCT
Notes	Additional information Additional non‐diabetic control group not included in this review
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "the sequence was randomly determined" Comment: appears to be investigator randomisation which is no true randomisation, and no further details are provided about the methods used by the investigators to properly randomise
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: open‐label. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: open‐label. No information provided regarding the blinding of the outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition are not reported. Unclear what final attrition is. ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	Low risk	Quote: "this work was supported by Grant 92‐27 from the physicians of Ontario through the Physicians Services Inc. Foundation. JA Miller is the recipient of a Research Scholarship from the Kidney Foundation of Canada" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Low risk	Washout period for cross‐over trials: 2 weeks of normal diet between each treatment period Comment: no other potential sources of bias identified

Mulhauser 1996.

Study characteristics
Methods	Study design Parallel RCT Duration of study Not reported Duration of treatment 1 month Duration of follow‐up No follow‐up past the end of the treatment period (4 weeks in total)
Participants	Study characteristics Setting: multicentre (3 sites) Country: Germany Inclusion criteria: 18 to 60 years; IDDM on intensified insulin therapy; diabetes duration > 5 years; increased proteinuria (> 60 mg/24 hours); stable renopathy Exclusion criteria: UTI; other drugs or oral contraception (except insulin); pregnancy; untreated SBP ≥ 140 < 160 mm Hg and/or DBP ≥ 85 < 100 mm Hg Baseline characteristics Number: intervention group (8); control group (8) Mean age ± SD (years): intervention group (35 ± 11); control group (27 ± 9) Sex (M/F): intervention group (5/3); control group (7/1) Mean BMI ± SD (kg/m²): intervention group (27 ± 9); control group (25.2 ± 3.1) Hypertension: 'hypertensive' Mean HbA1c ± SD (%): intervention group (8 ± 1); control group (7.5 ± 0.7) Diabetes IDDM Mean duration of diabetes ± SD (years): intervention group (23 ± 5); control group (22 ± 10) Co‐morbidities: none reported CKD definition: not inclusion criteria for this trial
Interventions	Pre‐randomisation/run‐in period 4 weeks usual diet 2 weeks of dietary training to reduce sodium intake to about 90 mmol/day Intervention group Placebo (oral): lactose 1 g, 6 times/day Diet changes: keep at a maximum 90 mmol/day Duration: 4 weeks Control group Sodium supplement (oral): 100 mmol/day Diet changes: keep at a maximum 90 mmol/day Duration: 4 weeks Co‐interventions or additional treatments None, other than keeping diet at a maximum 90 mmol/day No participants received ACEi or ARBs Follow up details Not reported
Outcomes	Reported outcomes (at 4 weeks) / outcomes relevant to this review* Body weight* Energy intake (calories/kg/day) Dietary sodium intake (mmol/day) UNa excretion (mmol/day)* Urinary potassium excretion (mmol/day) Protein intake (g/kg) HbA1c* Insulin dosage (IU/kg/day) Serum cholesterol (mmol/L) HDL‐cholesterol (mmol/L)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "eight patients were randomised to receive ..." Comment: insufficient details provided about how the randomisation methods were carried out, and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "all relevant laboratory results of the blinded study period, including excretion of urinary electrolytes, were kept blinded to patients and investigators until the last patient had finished the study"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "all relevant laboratory results of the blinded study period, including excretion of urinary electrolytes, were kept blinded to patients and investigators until the last patient had finished the study" Comment: implies that the outcome assessors were probably also kept blind
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition are not reported. Methods state that 2 participants withdrew during the run‐in period (before receiving treatment), but the results and analysis includes 8 participants per group. Unclear what the final attrition is ITT analysis was not performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori published protocol or trial registration details Trial registration: not reported A priori published protocol: not reported
Funding source	High risk	Quote: "the coated tablets were kindly provided by Cassella‐Riedel (Frankfurt, Germany)". "The study has been supported by Cassella Riedel, Frankfurt, Germany, and by the E Klockner Stiftung, Duisburg, Germany" Comment: pharmaceutical industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	Low risk	Washout period for cross‐over trials: not applicable for parallel design Comment: no other potential sources of bias identified

Trevisan Micro 1998.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 6 days Duration of follow‐up No follow‐up past the end of the treatment period (12 days in total)
Participants	Study characteristics Setting: single centre (diabetes hospital clinic) Location: Italy Inclusion criteria: IDDM; known history of persistent microalbuminuria (urinary AER: 20 to 200 µg/min); non‐obese (BMI < 27 kg/m²); untreated seated BP < 140/90 mm Hg (normotensive) Exclusion criteria: smoker; endocrine liver or non‐DKD; oral contraceptives Baseline characteristics (microalbuminuria group) Number: 7 Mean age ± SE: 38 ± 5 years Sex (M/F): 6/1 Ethnicity (Caucasian): 100% Mean BMI SE: 23 ± 1 kg/m² Smoker: non‐smokers Mean BP ± SE (mm Hg): SBP (125 ± 6); DBP (78 ± 5) Hypertension: normotensive Mean HbA1c ± SE (%): 9.2 ± 0.6 Diabetes IDDM Mean duration of diabetes ± SE: 24 ± 4 years Co‐morbidities: none CKD definition: none Urinary AER (µg/min): 77.3 (39 to 198)
Interventions	Pre‐randomisation/run‐in None reported Intervention group Dietary changes: aimed at 25 mmol NaCl/day Duration: 6 days Intervention group Dietary changes: aimed at 250 mmol NaCl/day, given as 500 mg NaCl tablets Duration: 6 days Washout period None, consecutive diets Co‐interventions or additional treatments Day 5 IV insulin to achieve plasma glucose concentrations 5 to 7 mmol/L No participants received ACEi or ARBs Follow up details No follow‐up past the end of the treatment period (12 days in total)
Outcomes	Reported outcomes / outcomes relevant to this review* UNa excretion* (mmol/day) Urinary potassium (mmol/day) PRA (ng/mL/hour) Aldosterone (pmol/L) Atrial natriuretic peptide (pmol/L) eGFR* ERPF* FF AER (µg/min) BMI* Body weight*
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "All subjects underwent two … diets … in random order" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what final attrition is. Appears ITT analysis was performed, but totals numbers are unclear
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Low risk	Quote: "This work was supported by Consiglio Nazionale delle Ricerche Grants 9603475CT04 and 9504361CT04" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: no washout applied, "All subjects underwent two consecutive 6‐day diets" Comment: no other potential sources of bias identified

Trevisan Normo 1998.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 6 days Duration of follow‐up No follow‐up past the end of the treatment period (12 days in total)
Participants	Study characteristics Setting: single centre (diabetes hospital clinic) Location: Italy Inclusion criteria: IDDM; known history of persistent microalbuminuria (urinary AER: 20 to 200 µg/min); non‐obese (BMI < 27 kg/m²); untreated seated BP < 140/90 mm Hg (normotensive) Exclusion criteria: smoker; endocrine liver or non‐DKD; oral contraceptives Baseline characteristics (normoalbuminuria group) Number: 9 Mean age ± SE: 42 ± 4 years Sex (M/F): 6/3 Ethnicity (Caucasian): 100% Mean BMI ± SE: 22 ± 1 kg/m² Smoker: non‐smokers Mean BP ± SE (mm Hg): SBP (120 ± 7); DBP (72 ± 2) Hypertension: normotensive Mean HbA1c ± SE (%): 8.9 ± 0.9 Diabetes IDDM Mean duration of diabetes ± SE: 22 ± 4 years Co‐morbidities: none CKD definition: none Urinary AER (µg/min): 7.5 (2 to 12)
Interventions	Pre‐randomisation/run‐in None reported Intervention group Dietary changes: aimed at 25 mmol NaCl/day Duration: 6 days Intervention group Dietary changes: aimed at 250 mmol NaCl/day, given as 500 mg NaCl tablets Duration: 6 days Washout period None, consecutive diets Co‐interventions or additional treatments Day 5 IV insulin to achieve plasma glucose concentrations of 5 to 7 mmol/L No participants received ACEi or ARBs Follow up details No follow‐up past the end of the treatment period (12 days in total)
Outcomes	Reported outcomes / outcomes relevant to this review* UNa excretion* (mmol/day) Urinary potassium (mmol/day) PRA (ng/mL/hour) Aldosterone (pmol/L) Atrial natriuretic peptide (pmol/L) eGFR* ERPF* FF AER (µg/min) BMI* Body weight*
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "all subjects underwent two … diets … in random order" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of the allocation to treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what final attrition is. Appears ITT analysis was performed, but totals numbers are unclear
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Low risk	Quote: "this work was supported by Consiglio Nazionale delle Ricerche Grants 9603475CT04 and 9504361CT04" Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: no washout applied, "All subjects underwent two consecutive 6‐day diets" Comment: no other potential sources of bias identified

Vedovato Micro 2004.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: multicentre (Diabetes Clinics at Padova University Hospital and Bergamo Hospital) Country: Italy Inclusion criteria: type 2 DM; BP < 140/90 mm Hg (in the absence of antihypertensives); normal or slightly increased UAE Exclusion criteria: not reported Baseline characteristics (microalbuminuria group) Number: 20 Mean age ± SEM: 57 ± 1 years Sex (M/F): 15/5 Mean BMI SEM: 29 ± 2 kg/m² Mean BP ± SEM (mm Hg): SBP (130 ± 2); DBP (80 ± 2) Hypertension: normotensive Mean HbA1c ± SEM (%): 8.2 ± 0.6 Diabetes Type 2 DM Mean duration of diabetes ± SEM: 9 ± 2 years Co‐morbidities: none reported CKD definition: none Urinary AER (µg/min): 91 (43 to 180)
Interventions	Pre‐randomisation/run‐in None reported Intervention group Dietary changes: aimed at 25 mmol NaCl/day Duration: 7 days Control group Dietary changes: aimed at 250 mmol NaCl/day, as 500 mg sodium chloride oral tablets Duration: 7 days Washout period None, consecutive diets Co‐interventions or additional treatments None No antihypertensive agents given so presumed not to receive ACEi or ARB Follow up details None, only follow‐up at the end of treatment period (2 weeks in total)
Outcomes	Reported outcomes / outcomes relevant to this review* SBP*, DBP*, MAP* Insulin sensitivity Mean 24‐hour UNa excretion* Mean 24‐hour urinary potassium excretion Aldosterone PRA Body weight* Urinary AER* (µg/min) eGFR* ERPF*
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "all subjects underwent, in random order, two consecutive 7‐day diet periods" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of allocation to randomised treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors. Unblinded trial so unlikely to be done
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what final attrition is. Appears ITT analysis was performed, but unclear
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Low risk	Quote: "this study was supported by a research grant from the University of Padua, Italy". Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trial: no washout applied, "Two consecutive 7‐day diet periods" Comment: no other sources of bias identified

Vedovato Normo 2004.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: multicentre (Diabetes Clinics at Padova University Hospital and Bergamo Hospital) Country: Italy Inclusion criteria: type 2 DM; BP < 140/90 mm Hg (in the absence of antihypertensives); normal or slightly increased UAE Exclusion criteria: not reported Baseline characteristics (normoalbuminuria group) Number: 21 Mean age ± SEM: 60 ± 2 years Sex (M/F): 16/5 Mean BMI ± SEM: 29 ± 2 kg/m² Mean BP ± SEM (mm Hg): SBP (130 ± 2); DBP (80 ± 2) Hypertension: normotensive Mean HbA1c ± SEM (%): 8.2 ± 0.6 Diabetes Type 2 DM Mean duration of diabetes ± SEM: 9 ± 2 years Co‐morbidities: none reported CKD definition: none, not inclusion criteria for this review, but must be normoalbuminuria Urinary AER (µg/min): 10 (3 to 18)
Interventions	Pre‐randomisation/run‐in None reported Intervention group Dietary changes: aimed at 25 mmol NaCl/day Duration: 7 days Control group Dietary changes: aimed at 250 mmol NaCl/day, as 500 mg sodium chloride oral tablets Duration: 7 days Washout period None, consecutive diets Co‐interventions or additional treatments None No antihypertensive agents given so presumed not to receive ACEi or ARB Follow up details None, only follow‐up at the end of treatment period (2 weeks in total)
Outcomes	Reported outcomes / outcomes relevant to this review* SBP*, DBP*, MAP* Insulin sensitivity Mean 24‐hour UNa excretion* Mean 24‐hour urinary potassium excretion Aldosterone PRA Body weight* Urinary AER* (µg/min) eGFR* ERPF*
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "All subjects underwent, in random order, two consecutive 7‐day diet periods" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of allocation to randomised treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors. Unblinded trial so unlikely to be done
Incomplete outcome data (attrition bias) All outcomes	High risk	Comment: withdrawals and study attrition not reported. Unclear what final attrition is. Appears ITT analysis was performed, but unclear
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Low risk	Quote: "This study was supported by a research grant from the University of Padua, Italy". Comment: no industry funding
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trial: no washout applied, "Two consecutive 7‐day diet periods" Comment: no other sources of bias identified

Yoshioka Adva Alb 1998.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre Country: Japan Inclusion criteria: type 2 DM; advanced albuminuria (UAE > 144 mg/24 hours) Exclusion criteria: history of heart disease; non‐diabetic CKD; UTI; SCr > 1.0 mg/dL; treated currently or previously with antihypertensive agents or with BP > 160/95 mm Hg Baseline characteristics (advanced albuminuria group) Number: 4 Mean age ± SE: 61 ± 7 years Sex (M/F): 11/4 (study total) Mean BMI ± SE: 23.3 ± 1.9 kg/m² Mean BP ± SE (mm Hg): SBP (131 ± 4); DBP (75 ± 4) Hypertension: must be < 160/95 mm Hg Mean HbA1c ± SE (%): 8.1 ± 0.5 Diabetes Type 2 DM Mean duration of diabetes ± SE: 12 ± 3 years Co‐morbidities: none reported CKD definition: none Mean UAE ± SE (range): 479 ± 91 mg/24 hours (309 to 635)
Interventions	Pre‐randomisation/run‐in Standard diet with ordinary salt (170 mEq sodium/day) Duration: for at least 2 weeks until plasma glucose had been controlled Intervention group Dietary changes: low salt diet 85 mEq sodium/day Duration: 7 days Control group Dietary changes: low salt diet 225 mEq sodium/day Duration: 7 days Washout period None, no time intervening Co‐interventions or additional treatments None Participants on antihypertensive agents or previously receiving these agents were excluded Follow up details Not reported
Outcomes	Reported outcomes / outcomes relevant to this review* eGFR* MAP* UNa excretion* (mEq/24 hours) PRA (ng/mL/hours)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the order of these diets was random" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of allocation to randomised treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: all participants are accounted for from start to end of trial. No participants dropped out, 0% attrition. Unclear whether ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Unclear risk	Comment: not reported
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: none applied, "No time intervening" Comment: no other potential sources of bias identified

Yoshioka Micro 1998.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre Country: Japan Inclusion criteria: type 2 DM; microalbuminuria (UAE 28.8 to 144 mg/24 hours) Exclusion criteria: history of heart disease; non‐diabetic CKD; UTI; SCr > 1.0 mg/dL; treated currently or previously with antihypertensive agents or with BP > 160/95 mm Hg Baseline characteristics (microalbuminuria group) Number: 7 Mean age ± SE: 55 ± 4 years Sex (M/F): 11/4 (study total) Mean BMI ± SE: 23.8 ± 1.8 kg/m² Mean BP ± SE (mm Hg): SBP (133 ± 6); DBP (75 ± 4) Hypertension: must be < 160/95 mm Hg Mean HbA1c ± SE (%): 8.0 ± 0.5 Diabetes Type 2 DM Mean duration of diabetes ± SE: 12 ± 3 years Co‐morbidities: none reported CKD definition: none Mean UAE ± SE (range): 65 ± 13 mg/24 hours (31 to 122)
Interventions	Pre‐randomisation/run‐in Standard diet with ordinary salt (170 mEq sodium/day) Duration: for at least 2 weeks until plasma glucose had been controlled Intervention group Dietary changes: low salt diet 85 mEq sodium/day Duration: 7 days Control group Dietary changes: low salt diet 225 mEq sodium/day Duration: 7 days Washout period None, no time intervening Co‐interventions or additional treatments None Participants on antihypertensive agents or previously receiving these agents were excluded. Follow up details Not reported
Outcomes	Reported outcomes / outcomes relevant to this review* eGFR* MAP* UNa excretion* (mEq/24 hours) PRA (ng/mL/hour)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the order of these diets was random" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of allocation to randomised treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: all participants are accounted for from start to end of trial. No participants dropped out, 0% attrition. Unclear whether ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Unclear risk	Comment: not reported
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: none applied, "no time intervening" Comment: no other potential sources of bias identified

Yoshioka Normo 1998.

Study characteristics
Methods	Study design Cross‐over RCT Duration of study Not reported Duration of treatment 7 days Duration of follow‐up No follow‐up past the end of the treatment period (2 weeks in total)
Participants	Study characteristics Setting: single centre Country: Japan Inclusion criteria: type 2 DM; normoalbuminuria (UAE < 28.8 mg/24 hours) Exclusion criteria: history of heart disease; non‐diabetic CKD; UTI; SCr > 1.0 mg/dL; treated currently or previously with antihypertensive agents or with BP > 160/95 mm Hg Baseline characteristics (normoalbuminuria group) Number: 8 Mean age ± SE: 65 ± 3 years Sex (M/F): 11/4 (study total) Mean BMI ± SE: 22.1 ± 0.8 kg/m² Mean BP ± SE (mm Hg): SBP (134 ± 5); DBP (72 ± 1) Hypertension: must be < 160/95 mm Hg Mean HbA1c ± SE (%): 8.8 ± 0.4 Diabetes Type 2 DM Mean duration of diabetes ± SE: 12 ± 3 years Co‐morbidities: none reported CKD definition: none Mean UAE ± SE (range): 16 ± 7 mg/24 hours (8 to 27)
Interventions	Pre‐randomisation/run‐in Standard diet with ordinary salt (170 mEq sodium/day) Duration: for at least 2 weeks until plasma glucose had been controlled Intervention group Dietary changes: low salt diet 85 mEq sodium/day Duration: 7 days Control group Dietary changes: low salt diet 225 mEq sodium/day Duration: 7 days Washout period None, no time intervening Co‐interventions or additional treatments None Participants on antihypertensive agents or previously receiving these agents were excluded. Follow up details Not reported
Outcomes	Reported outcomes / outcomes relevant to this review* eGFR* MAP* UNa excretion* (mEq/24 hours) PRA (ng/mL/hour)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "the order of these diets was random" Comment: insufficient details provided about how the randomisation methods were carried out and by whom
Allocation concealment (selection bias)	Unclear risk	Comment: no information provided regarding concealment of allocation to randomised treatment groups
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: unblinded trial. All participants knew which week they had to change their diet to low salt, and then one week on high salt
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: no information provided regarding the blinding of outcome assessors
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: all participants are accounted for from start to end of trial. No participants dropped out, 0% attrition. Unclear whether ITT analysis was performed
Selective reporting (reporting bias)	High risk	Comment: all outcomes planned in the methods were reported in the results. However, no access to an a priori trial registration or published protocol Trial registration: not reported Protocol registered/published: not reported
Funding source	Unclear risk	Comment: not reported
Conflicts of interest	Unclear risk	Comment: not reported
Other bias	High risk	Washout period for cross‐over trials: none applied, "No time intervening" Comment: no other potential sources of bias identified

ACEi: angiotensin‐converting enzyme inhibitors; ACR: albumin:creatinine ratio; AER: albumin excretion ratio; ARB: angiotensin receptor blocker; BMI: body mass index; BP: blood pressure; CKD: chronic kidney disease; CrCl: creatinine clearance; DBP: diastolic blood pressure; DKD: diabetic kidney disease; DM: diabetes mellitus; ERPF: effective renal plasma flow; FF: filtration fraction; (e)GFR: (estimated) glomerular filtration rate; GTT: glucose tolerance test; HbA1c: glycated haemoglobin; HCT: haematocrit; HDL ‐ high‐density lipoprotein, IDDM: insulin‐dependent diabetes mellitus; ITT: intention‐to‐treat analysis; IV: intravenous; M/F: male/female; MAP: mean arterial pressure; NIDDM: non‐insulin diabetes mellitus; NSAIDs: nonsteroidal anti‐inflammatory drugs; PCR: protein:creatinine ratio; PRA: plasma renin activity; RAS: renin‐angiotensin system; RPF: renal plasma flow; SBP: systolic blood pressure; SCr: serum creatinine; UAE: urinary albumin excretion; UNa: urinary sodium; UTI: urinary tract infection 

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
DNETT Japan 2010	Intervention: efficacy of multifactorial intensive therapy, not comparing high and low salt intake
Dodson 1984	Intervention: multiple randomised interventions
Ekinci 2009	Intervention: multiple interventions ‐ additional sodium supplement used Other: baseline differences between participants
Gilleran 1996	Intervention: multiple randomised interventions. Effects of dietary sodium substitution with potassium and magnesium in hypertensive type II diabetics
Helou 2016	Intervention: not comparing a difference in salt intake, multidisciplinary care versus usual care Other: publication type is a study protocol
HHK 2018	Intervention: not comparing a difference in salt intake, multidisciplinary care versus tech versus both versus education Other: publication type is a study protocol
Imanishi 1997	Intervention: study to evaluate effects of ACEi Outcomes: sodium excretion not measured at the baseline or end of intervention, as required by inclusion criteria Other: baseline differences between participants
LowSALT CKD 2012	Participants: the study population includes some non‐diabetic patients
Petrie 1998	Intervention: frusemide was given to high salt intake group and would have increased sodium excretion in that group compared with low salt group Previously included study
PROCEED 2018	Intervention: 24‐hour urinary sodium excretion difference between intervention and control group was < 34 mmol so did not satisfy inclusion criteria
Suckling 2016	Participants: not exclusive to diabetics, 43% of the participant sample have impaired glucose intolerance, which does not meet the inclusion criteria of type 1 or 2 diabetes
Ushigome 2019	Intervention: study in diabetic patients but intervention did not require measurements of urinary sodium
ViRTUE‐CKD 2016	Participants: the study includes non‐diabetic patients

ACEi: angiotensin‐converting enzyme inhibitor

Differences between protocol and review

Data on weight were included for analysis in addition to BMI.

This review was updated to inform the latest KDIGO blood pressure guidelines. Additional primary and secondary outcomes were added to cover all clinically relevant and critical outcomes (KDIGO 2020). These are: death (any cause), doubling of SCr, stroke, heart failure, myocardial infarction, and adverse events.

Contributions of authors

• Writing of protocol and review: EH, TC, RJS, FJH, GAM

• Screening of titles and abstracts: EH, TC, RJS, FJH, GAM

• Assessment for inclusion: EH, TC, RJS, FJH, GAM

• Quality assessment: EH, TC, RJS, FJH, GAM

• Data extraction: EH, TC, RJS, FJH, GAM

• Data entry into RevMan: EH, TC, RJS, FJH, GAM

• Data analysis: EH, TC, RJS, FJH, GAM

• Disagreement resolution: TC, RJS, FJH, GAM

Sources of support

Internal sources

• No sources of support provided

External sources

• BEAT‐CKD grant 1092957, Australia

  TC is employed by funding from this grant.

Declarations of interest

• Elisabeth Hodson: no relevant interests were disclosed

• Tess Cooper: no relevant interests were disclosed

New search for studies and content updated (no change to conclusions)