THE EFFECTS OF THE DASH DIET AND SODIUM REDUCTION ON BLOOD PRESSURE IN PERSONS WITH DIABETES

Abstract

Elevated blood pressure and blood pressure-related morbidity are extraordinarily common in persons with diabetes. The Dietary Approaches to Stop Hypertension dietary pattern and dietary sodium reduction are recommended as lifestyle interventions in individuals with diabetes. However, these recommendations have largely been based on studies conducted in persons without diabetes. In this review, we summarize available evidence from trials that tested the effects of these two dietary interventions on blood pressure in people with diabetes. Overall, of the 3 trials (total n=151) that tested the effects of the DASH dietary pattern in persons with diabetes, 2 trials documented that the DASH dietary pattern lowered blood pressure. While 16 trials (total n=445) tested the effects of sodium reduction in persons with diabetes, results were inconsistent, likely because of design limitations, e.g. brief duration, small sample size, and low baseline blood pressure levels, as well as differences in the mode of intervention delivery (behavioral interventions, feeding studies, and sodium supplements). In conclusion, there is a substantial need for additional research on the blood pressure lowering effects of the DASH diet and sodium reduction in people with diabetes and hypertension, given the high prevalence of hypertension and the dearth of high-quality trials in this population.

Nearly 1 in 10 adults (463 million people) have diabetes worldwide, and this number is projected to rise to 700 million by 2045 (1). Type 2 diabetes is increasing in prevalence in most countries around the world (1). Most persons with diabetes also have hypertension, which is an important risk factor for stroke (2), myocardial infarction (3) and heart failure (4). Compared to persons without diabetes, persons with diabetes experience twice the risk of stroke and myocardial infarction (2, 5). Furthermore, the longer the duration of diabetes, the higher the stroke risk (6). Cardiovascular disease, including stroke and myocardial infarction, account for 50% of deaths among persons with diabetes (7). Elevated blood pressure is also a strong determinant of the microvascular complications of diabetes, including diabetic nephropathy (8) and retinopathy (9).

A substantial body of evidence has documented that lowering blood pressure through anti-hypertensive drug therapy reduces the risk of cardiovascular events and stroke in people with diabetes (10, 11). Blood pressure lowering has also been shown to have beneficial effects on diabetic nephropathy (12, 13) and retinopathy (14). However, aside from robust trial evidence that weight loss lowers blood pressure in persons with diabetes (15), dietary approaches to reducing blood pressure among people with diabetes has not been well-studied.

In people without diabetes, there is strong evidence that the Dietary Approaches to Stop Hypertension (DASH) dietary pattern and dietary sodium reduction (16–19) lower blood pressure. The DASH eating plan emphasizes potassium-rich fruits and vegetables, and low-fat dairy products; includes whole grains, poultry, fish and nuts; and is reduced in red meat, sweets, and sugar-containing beverages (20). Most evidence for the DASH diet comes from short-term feeding studies, but several long-term studies using behavioral interventions have shown this dietary pattern to be effective for blood pressure lowering (21, 22). Likewise, there is substantial evidence that reducing sodium intake lowers blood pressure in persons without diabetes; the evidence base includes results of animal studies (23), migration studies (24), ecologic studies (25, 26), longitudinal observational studies (27), clinical trials (17), and meta-analyses of trials (28–31). The causal relationship of sodium intake with blood pressure was recently reaffirmed in the U.S. National Academies of Sciences, Engineering and Medicine’s Dietary Reference Intakes for Sodium and Potassium which concluded that there is a high strength of evidence that reducing sodium intake lowers blood pressure (32).

Based on data from the general population without diabetes, the American Diabetes Association (ADA) recommends that persons with diabetes limit sodium intake to <2,300 mg per day (33, 34). The ADA also includes the DASH diet as one of several recommended health eating plans for persons with diabetes based on evidence that the DASH diet lowers blood pressure in the general population without diabetes (34). While one can posit that dietary interventions might have similar or perhaps greater BP-lowering effects in persons with diabetes, researchers have commonly conducted trials separately by diabetes status; policymakers have likewise made separate treatment recommendations. At least for drug trials that tested the effects of different blood pressure goals, results did differ by diabetes status (35, 36).

Given that evidence on the benefits of the DASH diet and sodium reduction among people with diabetes is extrapolated from the general population, our objective was to summarize the evidence from randomized, controlled trials on the effects of the DASH dietary pattern and sodium reduction on blood pressure among adults with type 1 and type 2 diabetes. We searched PubMed to identify English-language articles related to the DASH diet and reduced sodium interventions in people with diabetes. Our search strategy included MeSH terms and keywords related to DASH, low sodium diets, and diabetes (see Supplement). Our primary outcome of interest was blood pressure.

Effect of DASH Diet on Blood Pressure in People with Diabetes

Only three randomized controlled trials (RCTs) tested the effects of behavioral interventions that promoted the DASH diet in persons with type 2 diabetes, and two of the studies documented a significant blood pressure reduction with the DASH diet compared to a control diet (37, 38) (Table 1). One was a 4-week parallel arm study (n=40) conducted in Brazil among patients with diabetes and uncontrolled blood pressure, comparing the DASH diet to a control diet (38). The DASH diet group received recommendations on following the DASH diet eating plan and to increase physical activity whereas the control group received recommendations according to the ADA guidelines and told to maintain their usual physical activity. This study showed a significantly greater reduction in 24-hour ambulatory blood pressure measurement (between-group difference in systolic blood pressure, −11.0mmHg, p<0.001; diastolic blood pressure, −5.0mmHg, p=0.018) but not office blood pressure measurement with the DASH diet compared to the control diet. The effect of the DASH diet may have been confounded by sodium reduction as there was a significant within-group reduction in urinary sodium in the DASH group but not the control group (between-group difference p=0.017).

Table 1.

Trials testing the effect of the DASH diet in people with diabetes

Study	Design	Type of intervention	Comparison	N	Age (yrs)	Males (%)	Race	Population	Study Dur (wks)	BP Δ in DASH group	BP Δ in control group	Between-group difference in BP Δ	Contrast in 24-hour urine sodium
Azadbakht, 2011 (37)	Crossover study	Behavioral	DASH vs. control‡	31	NR	42	NR	Type 2 diabetes	8	SBP: −13.6± 3.5 mmHg DBP: −9.5± 2.6 mmHg	SBP: −3.1±2.7 mmHg DBP: −0.7±3.3 mmHg	†Between-group diff in SBP −10.5mmHg, p=0.02 †Between-group diff in DBP −8.8mmHg, p=0.04	NR
Hashemi, 2019 (39)	Parallel study	Behavioral	DASH vs. control§	80	NR	37–40	NR	Type 2 diabetes, not on insulin, prehypertension, not on BP meds	12	Males: SBP −10.3±9.6 mmHg, DBP −5.2± 9.9 mmHg Females: SBP −6.4± 17.0mmHg, DBP −1.7± 12.1mmHg P=0.003 for within-group change	Males: SBP −7.3±16.7 mmHg, DBP −8.6± 14.7mmHg Females: SBP 0.6±16.2 mmHg, SBP 1.5±9.9mmHg P=0.338 for within-group change	†Between-group diff in SBP −3.0mmHg for males and −7.0 mmHg for females, p=0.686, †Between-group diff in DBP 3.4 mmHg for males and −3.2 mmHg for females, p=0.892
Paula, 2015 (38)	Parallel study	Behavioral	DASH vs. control*	40	61.8–62.5	45	White: 80–90%	Type 2 diabetes, uncontrolled BP, BMI≤40	4	All ABPM reductions were sign Office SBP: −17.6 mmHg (−44.0 to 14.0), p<0.001 Office DBP: −3.9 mmHg (−14.0 to 7.0), p=0.045	Only 24-hr diastolic and nighttime, systolic and diastolic, ABPM reductions were sign Office SBP: −6.3 mmHg (−54.0 to 24.0), p=0.246 Office DBP: −0.9 mmHg (−21.0 to 11.0), p=0.589	†Between-group diff in ABPM SBP: −11.0mmHg, p<0.001 †Between-group diff in ABPM DBP: −5.0mmHg, p=0.013 †Between-group diff in office SBP: −11.3mmHg p=0.021 †Between-group diff in office DBP: −3.0mmHg, p=0.013	Sign reduction in UNa in DASH group (p<0.001) but not control group (p=0.400) Between-group diff NR, p=0.017

Abbreviations: ABPM, Ambulatory blood pressure measurement; BMI, Body Mass Index; DASH, Dietary Approaches to Stop Hypertension; DBP, diastolic blood pressure; NR, not reported; sign, significant; SBP, systolic blood pressure; UNa, urine sodium

^*Control group received dietary recommendations according to American Diabetes Association

^†Calculated between-group difference

^‡Control diet included macronutrient composition of 50–60% carbohydrates, 15–20% protein, <30% total fat, and <5% caloric intake from simple sugars

^§Control diet included diabetic dietary pattern according to the American Diabetes Association recommendations

Another trial was an 8-week crossover study (n=31) conducted in Iran among patients with diabetes comparing the DASH diet to a control diet (37). Patients were randomly assigned to the DASH diet or control diet and advised on the dietary intervention by a dietician. This study demonstrated a significantly larger blood pressure reduction in the DASH vs. control diet (between-group difference in systolic blood pressure, −10.5mmHg, p=0.02; diastolic blood pressure, −8.8mmHg, p=0.04). These two studies were behavioral intervention trials, in which partial adherence prevents an assessment of the maximum potential effect of the DASH diet, which is best assessed in the setting of a controlled feeding study.

The third study (n=80) showing non-significant findings was a behavioral intervention conducted in Iran among patients with diabetes and prehypertension (39). Patients were randomly assigned to the DASH or control diet for 12 weeks. While the study found a significant reduction in systolic blood pressure within the DASH group, the between-group difference in blood pressure was not significant.

Effect of Sodium Reduction on Blood Pressure in People with Diabetes

Sixteen RCTs (total n=445) have studied the effects of sodium reduction in persons with diabetes using behavioral interventions (n=5), feeding studies (n=7) and sodium supplements (n=4). Six of the studies were conducted among people with Type 1 diabetes while ten studies were conducted among people with Type 2 diabetes. Among the five studies evaluating behavioral interventions (Table 2), two showed a significant between-group blood pressure reduction with a low sodium diet compared to control (i.e. usual diet) or high sodium diet (40, 41). One was a 12-week parallel arm study from the United Kingdom of patients with diabetes and hypertension (n=34) which compared moderate sodium reduction (not specified) to a control diet (41). At follow-up, there was a significant between-group systolic blood pressure reduction (p<0.03 for both supine and erect) but not for diastolic blood pressure (41). Correspondingly, there was a significant reduction in urine sodium in the moderate sodium restriction group compared to the control group (between-group difference, p<0.05). A 24-week crossover study (n=45) in Netherlands compared a low sodium (1200mg) to regular sodium diet (4800mg) combined with hydrochlorothiazide or placebo (40). This study showed a significant between-group difference in systolic and diastolic blood pressure with a greater reduction in blood pressure during the low sodium diet compared to regular diet in both the placebo and hydrochlorothiazide groups (40). Correspondingly, urine sodium excretion decreased going from the regular sodium diet to low sodium diet (between-group difference, p<0.0001).

Table 2.

Trials testing the effect of sodium reduction using behavioral interventions among people with diabetes

Study	Design	Comparison	N	Age (yrs)	Males (%)	Race	Population	Study Dur (wks)	BP Δ in lower sodium group	BP Δ in higher sodium group	Between-group difference in BP Δ (lower sodium-higher sodium)	Contrast in 24-hour urine sodium
Dodson, 1989 (41)	Parallel study*	Moderate sodium restriction† vs. control	34	61	71/65	NR	Type 2 diabetes, hypertension, no proteinuria, no past or current use of insulin	12	Erect: SBP −12.0%, p<0.001 DBP −3.1%, p<0.05 Supine: SBP −11.0%, p<0.01 DBP −3.1%, NS	Erect: SBP −4.2%, NS DBP −5.1%, NS Supine: SBP −4.0%, NS DBP −1.0%, NS	Between-group diff in supine SBP −12.9 (−4.3 to −21.0) mmHg, p<0.03 Between-group diff in erect SBP −11.9 (−3.9 to −19.0 mmHg, p<0.03 ‡Between-group diff in supine DBP −1.8mmHg, NS ‡Between-group diff in erect DBP 2.3mmHg, NS	UNa excretion greater in sodium restriction group (p<0.001) vs. control group which had NS reduction in UNa; between-group diff NR, p<0.05
Houlihan, 2002 (42)	Crossover study	Placebo vs. Losartan 50mg→ crossover of dietary period: Low Sodium (50–70mmol or 1150–1610mg)§ vs. normal Sodium (>100mmol or 2300mg)§	20	60.6–63.1	95	NR	Type 2 diabetes, with hypertension, elevated urinary albumin excretion rates	2	Placebo: change in 24-h SBP −0.7±2.7 mmHg Change in 24-h DBP 2.3±1.9 mmHg Losartan: change in 24-h SBP −9.7±2.7 mmHg Change in 24-h DBP −5.1±1.4 mmHg	Placebo: change in 24-h SBP −1.1±2.1 mmHg Change in 24-h DBP −1±1.1 mmHg Losartan: change in 24-h SBP 0±2.1 mmHg Change in 24-h DBP 0.4±0.9 mmHg	Placebo group: NS between-group diff in 24-h SBP, 1.8 mmHg (−5.9 to 9.5); DBP 3.3 mmHg (−1.7 to 8.3) Losartan group: greater diff in low vs. normal sodium phase; between-group diff in 24-h SBP, −9.7mmHg (−17.2 to −2.2), p=0.002; DBP diff, −5.5 mmHg (−8.4 to −2.6), p=0.002	UNa excretion decreased from baseline during low sodium diet in placebo vs. losartan groups, 80±22 mmol/day (p=0.01) and 85±14 mmol/day (p=0.01) respectively UNa excretion did not change from baseline during normal sodium diet for placebo and losartan groups Between-group diff NR
Kwakernaak, 2014 (40)	Crossover study‖	Low sodium (50mmol or 1200mg)§ vs. regular sodium (200mmol or 4800mg)§; Plus HCTZ vs placebo added to ACE inhibition	45	65	84	White: 100%	Type 2 diabetes, albuminuria	24	NR	NR	‡Placebo group: between-group diff in SBP: −6.0 mmHg, p=0.0080; DBP −3.0 mmHg, p=0.0067 ‡HCTZ group: between-group diff in SBP −6.0 mmHg, p=0.0009; DBP −4.0 mmHg, p=0.0055	UNa excretion decreased from 224±73 mmol/day during regular sodium diet to 148±65 mmol/day during low sodium diet when combined with placebo and 164±73 mmol/day with HCTZ (both p<0.0001 vs. regular sodium intake)
Luik, 2002 (43)	Crossover study	Low sodium (50mmol or 1150mg)§ vs. high sodium (200mmol or 4600mg)§	24	29	63	NR	Type 1 diabetes, normotensive, normoalbuminuric	1	NR	NR	‡Between-group diff in MAP, −3.0mmHg, NS	UNa excretion lower on low sodium vs. high sodium diet, 38±13.1 mmol/day and 249±70.7 mmol/day respectively, between-group diff NR, p<0.001
Parvanova, 2018 (44)	Crossover study‖	Low sodium (100 mEq or 2400mg)§ vs. high sodium (200 mEq or 4800mg)§ Plus Paricalcitol vs. placebo	115	64	89	NR	Type 2 diabetes, macroalbuminuria, RAS inhibitor therapy	12	‡Within-group diff in SBP −4.5mmHg and DBP −1.8mmHg p<0.01	‡Within-group diff in SBP 0.2mmHg and DBP 1.2mmHg, NS	‡Between-group difference, −4.7mmHg, p-value NR	UNa excretion greater in high vs. low sodium group, between-group diff: 35.3 ± 58.6 mmol/day, p=0.0002

Abbreviations: ACE, angiotensin-converting enzyme; DBP, diastolic blood pressure; HCTZ, hydrochlorothiazide; MAP, mean arterial pressure; NR, not reported; NS, non-significant; SBP, systolic blood pressure; UNa, urine sodium

^*Report only parallel study findings but there was a subsequent crossover study in which participants of sodium restriction group were randomized to salt pill vs. placebo

^†Dietary advice consisted of no salt to be added at table or in cooking and avoidance of heavily salted foods.

^‡Calculated within or between-group difference

^§Sodium content converted from mmol or mEq of sodium or sodium chloride to mg of sodium based on guidance from CDC: https://www.cdc.gov/salt/sodium_toolkit.htm

^‖Crossover study refers to medications. Sodium diet assignments remained consistent throughout study.

The other three behavioral intervention studies did not show a significant reduction in blood pressure for the lower versus higher sodium intervention within the placebo group (42–44). However, one study from Australia which randomized patients to placebo vs. losartan then to a crossover study of low sodium (1150–1610mg) vs. high sodium (>2300mg) diet found a greater reduction in blood pressure during the low sodium vs. high sodium phase among patients on losartan (42). Notably, one of these null studies was conducted among young participants with Type 1 diabetes without elevated blood pressure (43).

Among the seven feeding studies evaluating sodium reduction (Table 3), all were of one-week duration (45–51), and three of the studies showed significant blood pressure reduction for the lower sodium compared to higher sodium diet (45, 46, 48). One crossover study (n=32) from Japan compared a low sodium (1840mg) to normal sodium (4600mg) diet (46). At follow-up, systolic blood pressure was lower in the low sodium group vs. normal sodium group (p<0.001) among patients stratified by normoalbuminuria/normotension and microalbuminuria/normotension. However, urine sodium excretion was not measured in this study so adherence to the diet is unknown. Two studies in Italy showed a higher blood pressure when switching from a low to high sodium diet (570mg vs. 5700mg sodium) but only among those with microalbuminuria; no changes were seen with participants with normoalbuminuria (45, 48). In these two studies, there was no significant contrast in urine sodium excretion despite the large contrast in sodium content between diets.

Table 3.

Trials testing the effect of sodium reduction in feeding studies among people with diabetes

Study	Design	Comparison	N	Age (yrs)	Males (%)	Race	Population	Study Dur (wks)	BP Δ in lower sodium group	BP Δ in higher sodium group	Between-group Difference in BP Δ (lower sodium-higher sodium)	Contrast in 24-hour urine sodium
De’Oliveira, 1997 (51)	Parallel study	Low sodium (10mmol NaCl or 228mg sodium)* vs high sodium (200mmol NaCl or 4563mg sodium)*	19	56	84	White 90%, African American, 10%	Type 2 diabetes	1	NR	NR	†MAP higher in high vs. low sodium group, between-group diff −9.0mmHg, p-value NR	UNa excretion similar in low vs. high sodium group, 14±2 mmol/day and 195±7 mmol/day respectively, between-group diff NR, NS
Imanishi, 2001 (46)	Crossover study	Low sodium (80mmol or 1840mg)* vs. normal sodium (200 mmol or 4600mg)*; stratified by albuminuria: normal, micro or macro	32	59–62	59	NR	Type 2 diabetes, inpatients, simple diabetic retinopathy, excluded those on BP meds	1	NR	NR	Normoalbuminuria and normotension: SBP lower in low vs. normal sodium diet; †between-group diff −7.0mmHg, p<0.001 No between-group difference in DBP Microalbuminuria and normotension: †between-group diff in SBP −11.0 mmHg, p<0.001; DBP −6.0mmHg, p=0.0021	NR
Miller, 1995 (50)	Crossover study	Low sodium (10–20mmol or 230–460mg) vs. high sodium (250mmol or 5750mg)*; control (no diabetes) vs. diabetes	9	26	100	NR	Type 1 diabetes, within 5 years of diagnosis, normotensive, nonobese, no microalbuminuria	1	NR	NR	†MAP was similar in both groups, between-group diff −1.0mmHg, p-value NR	UNa excretion higher in high vs. low sodium group, 247±22 mmol/day and 15±2 mmol/day respectively, between-group diff NR, p≤0.05
Miller, 1997 (49)	Crossover study	Low sodium (20mmol or 460mg)* vs. high sodium (200mmol or 4600mg)*; control (no diabetes) vs. diabetes	12	23	100	NR	Type 1 diabetes, within 5 years of diagnosis, normotensive, no proteinuria	1	NR	NR	†MAP was similar in both groups, between-group diff −1.0mmHg, p-value NR	UNa excretion higher in high vs. low sodium group, 201±10 mmol/day and 19±2mmol/day respectively, between-group diff NR, p≤0.05
Trevisan, 1998 (48)	Crossover study	Low sodium (25mmol sodium chloride or 570mg sodium)* vs. high sodium (250mmol sodium chloride or 5700mg sodium)*; normal vs. micro-albuminuria	16	38–42	75	NR	Type 1 diabetes, BMI<27kg/m2; with microalbuminuria and untreated HTN vs. control without HTN and normal albumin excretion	1	NR	NR	With microalbuminuria: †Between-group diff in MAP −9.0mmHg, p<0.001 Without microalbuminuria: †Between-group diff in MAP 3.0mmHg, NS	UNa excretion higher in high vs. low sodium diet within groups (p<0.0001) but not between groups
Vedovato, 2004 (45)	Crossover study	Low (25mmol sodium chloride or 570mg sodium)* vs. High (250mmol sodium chloride or 5700mg sodium)*; control (diabetes w/o micro- albuminuria) vs. diabetes w/ micro-albuminuria	41	57–60	76	NR	Type 2 diabetes, with and without microalbuminuria, normotensive	1	No change	NR	With microalbuminuria: †Between-group diff in MAP −8.0mmHg p<0.0001 Without microalbuminuria: †Between-group diff in MAP, −1.0 mmHg, NS	No between-group difference in mean UNa excretion for those with or without microalbuminuria
Yoshioka, 1998 (47)	Crossover study	Low sodium (85mEq or 1960mg)* vs. high sodium (225mEq or 5192 mg)*; Stratified by albuminuria: normal vs micro vs. advanced albuminuria	19	60	58	NR	Type 2 diabetes, excluded those on prior or current BP treatment or elevated BP >160/95 mmHg	1	NR	NR	MAP similar between low and high salt diet when stratified by albuminuria status Microalbuminuria: †Between-group diff 3.0mmHg, p-value NR Normoalbuminuria: †Between-group diff 2.0mmHg, p-value NR	UNa excretion higher in high vs. low sodium diet among 3 groups stratified by albuminuria status, p<0.01

Abbreviations: DBP, diastolic blood pressure; MAP, mean arterial pressure; NR, not reported; NS, non-significant; SBP, systolic blood pressure; UNa, urine sodium

^*Sodium content converted from mmol or mEq of sodium or sodium chloride to mg of sodium based on guidance from CDC: https://www.cdc.gov/salt/sodium_toolkit.htm

^†Calculated between-group difference

Four of the sodium reduction studies used sodium supplements (52–55) (Table 4) with two of the studies demonstrating significant findings that those assigned to placebo pill had a lower blood pressure than those taking the sodium supplement (53, 55). One 4-week parallel arm study (n=14) from Germany compared 100mmol (2300mg) sodium supplement with placebo while on 90mmol (2070mg) sodium diet in people with insulin-dependent diabetes, showing a significant between-group clinic diastolic blood pressure reduction with placebo compared to the supplement but not with systolic blood pressure (53). Another 6-week crossover study (n=26) from United Kingdom compared a 90mmol (2053mg) sodium supplement with placebo, while both groups were on a reduced salt diet of 90mmol/day among patients with type 2 diabetes and/or impaired glucose tolerance and demonstrated a significant between-group clinic systolic blood pressure reduction going from high to low sodium diet (p<0.01) (55). The authors of this study noted that the difference in salt intake between groups was small due to the high prevalence of salt in many foods, so maintaining patient compliance with the dietary regimen was a challenge (55). In both studies, 24-hour urine sodium was lower in the low sodium group vs. high sodium group.

Table 4.

RCTs studying the effect of sodium reduction using sodium supplements among people with diabetes

Study	Design	Comparison	N	Age (yrs)	Males (%)	Race	Population	Study Dur (wks)	BP Δ in lower sodium group	BP Δ in higher sodium group	Between-group Difference in BP Δ (lower sodium-higher sodium)	Contrast in 24-hour urine sodium
Lopes de Faria, 1997 (52)	Crossover study	Regular sodium diet (100mmol or 2300mg)* vs. high sodium supplement (300mmol or 6900mg)*	10	30	70	NR	Type 1 diabetes, normotensive, normoalbuminuria	1	NR	NR	MAP was similar in both diets, †between-group 2.0mmHg, p=0.25	UNa excretion lower in regular vs. high sodium diet, 99±27mmol/day and 293±80mmol/day respectively; between-group diff NR, p<0.0001
Muhlhauser, 1996 (53)	Parallel study	Sodium supplement (100 mmol or 2300mg sodium)* vs. placebo, while on 90mmol or 2070mg sodium diet	14	31	75	NR	Type 1 diabetes, increased proteinuria, high normal or mildly hypertensive BP	4	Clinic SBP change, 3.2 mmHg (−2.6 to 9) Clinic DBP change, 2.2 mmHg (−1.8 to 6.2)	Clinic SBP change, −1.7 mmHg (−8.6 to 5.3) Clinic DBP change, −3.1 mmHg (−5.9 to −0.4)	Clinic SBP: 4.9 mmHg (−3.3 to 13.1), p=0.22 Clinic DBP: 5.3 mmHg (1 to 9.7), p=0.02	UNa excretion greater in supplement vs placebo group, 199±52mmol/day and 92±33mmol/day respectively, between-group diff NR, p=0.0002
Petrie, 1998 (54)	Crossover study	Sodium supplement (120 mmol or 3680mg sodium)* vs. placebo, while on 40 mmol or 920mg sodium diet*	9	57	89	White: 100%	Type 2 diabetes, diet controlled, without retinopathy or microalbuminuria	5 days	NR	NR	†Between-group diff in SBP 2.0mmHg, p=0.66 †Between-group diff in DBP 5.0mmHg, p=0.15	UNa excretion greater reduction in high vs. low sodium group, 197±76.0 mmol/day and 67±19.5 mmol/day respectively, between-group diff NR, p=0.03
Suckling, 2016 (55)	Crossover study	Salt tablet (90mmol salt or 2053mg sodium)* vs. placebo, while on reduced salt diet (90mmol salt or 2053mg sodium)*	26	56	54	White: 65% Black: 23% Asian: 12%	Type 2 diabetes, diet-controlled, untreated normal or high normal BP	6	NR	NR	BP lower in low vs. high sodium diet, between-group diff in SBP −4.2 mmHg (p<0.01) and DBP of −1.7 mmHg (p=0.055)	UNa excretion lower in low vs. high sodium group, between-group difference −48.6±9.3 mmol (p<0.001)

Abbreviations: DBP, diastolic blood pressure; MAP, mean arterial pressure; NR, not reported; NS, non-significant; SBP, systolic blood pressure; UNa, urine sodium

^*Sodium content converted from mmol or mEq of sodium or sodium chloride to mg of sodium based on guidance from CDC: https://www.cdc.gov/salt/sodium_toolkit.htm

^†Calculated between-group difference

Discussion

Overall, evidence on the blood pressure lowering effects of the DASH dietary pattern and sodium reduction in people with diabetes is limited. Specifically, we found only three RCTs of the DASH diet in people with diabetes with two trials reporting significant reductions in blood pressure with the DASH diet compared to control diet. We found 16 trials that tested the effects of sodium reduction on blood pressure in people with diabetes, and these studies had heterogeneous results. Methodologic limitations in these studies were common and limit strong inferences on the effects of the DASH diet and sodium reduction on blood pressure in persons with diabetes.

Previously published meta-analyses and reviews of sodium reduction studies have shown that sodium reduction lowers blood pressure in the general population (28–31). A 2010 Cochrane review of randomized controlled trials reported that sodium intake reduction lowers blood pressure in those with diabetes (56). Furthermore, they found that patients with diabetes who were normotensive had a larger reduction in blood pressure with sodium lowering compared to hypertensive patients, which they attributed to normotensive patients achieving a greater salt intake reduction compared to hypertensive patients (56). These trials (n=13) were included in our review, and we identified 3 new trials (40, 44, 55).

There are several design issues to consider when interpreting the available evidence, including several limitations - small sample size, brief duration of intervention, reliance on behavioral counseling to achieve an experimental contrast, inclusion of persons without elevated blood pressure, and non-diverse populations. We only identified three studies examining the effect of the DASH diet on blood pressure reduction in people with diabetes, which were small, behavioral intervention trials (37, 38). Adherence to diet is difficult to achieve in behavioral intervention studies (21) compared to feeding studies (57). Among the feeding studies that examined sodium reduction (45–51), all were one week in duration so the maximal and long-term effects of these diets on blood pressure are unknown. In general, studies that were longer in duration (>4 weeks) showed a significant reduction in blood pressure in the lower sodium group compared to higher sodium group (40, 41, 55). Many studies for both dietary interventions did not include participants with elevated blood pressure which makes it challenging to detect intervention effects unless sample sizes are large. Lastly, some trials might have enrolled diverse populations of persons with and without diabetes, but did not report results separately.

In terms of generalizability, study populations from the sodium reduction studies were predominantly white, male and older adults (age >60 years) limiting generalizability of study results to diverse populations. Understandably, most studies also excluded participants on antihypertensive therapy which can confound study results. However, the exclusion of persons on drug therapy also limits generalizability given that most patients with diabetes and hypertension are on blood pressure lowering medications.

The ADA recommendation for persons with diabetes to consume a diet with a sodium intake of less than 2300 mg/day is extrapolated from guidelines for the general population (33, 34). The ADA has recommended that persons with diabetes not consume less than 1500 mg of sodium per day given evidence from two observational studies demonstrating increased mortality associated with very low sodium intake (58, 59); however, the updated Dietary Reference Intakes for Sodium and Potassium from the National Academy of Medicine concluded that these observational studies had substantial methodological flaws and that better evidence was needed about the health effects of sodium intakes below 2300 mg/day (32). Several of the studies including in our review included low sodium interventions of less than 1500mg of sodium per day but showed mixed results on blood pressure reduction (40, 42, 43, 45, 48, 49, 54). Therefore, the effects on blood pressure of lowering sodium intake below 2300mg/day are uncertain in persons with diabetes.

In conclusion, a very limited body evidence, just three trials, suggests that the DASH dietary pattern lowers blood pressure in people with diabetes. While a much larger number of trials tested the effects of sodium reduction in persons with diabetes, results were inconsistent, likely because of methodologic limitations. Given the substantial burden of elevated blood pressure and its associated macrovascular and microvascular complications in persons with diabetes, additional research on the effects of the DASH diet and sodium reduction in this population is clearly needed.