Chang 2006.

Study characteristics
Methods	Study design: Cluster‐randomised controlled trial Country: Taiwan Setting: Intervention conducted in a veteran's retirement home in the northern region; setting where mortality outcomes were recorded was the Department of Health Aim of study: To examine the effects of using potassium‐sparing salt on cardiovascular disease mortality and medical expenditure in elderly veterans Unit of allocation: Retirement home kitchens Start date: October 1995 End date: June 1999 Relevant study limitations as reported by study authors: NR Sample size calculation: NR
Participants	Baseline Characteristics LSSS intervention Age: in years, mean (SD): 75.21 (7.37) Gender: Male, % (n/N): 100 (768/768) Ethnicity/race: Chinese Smoking: NR Body Mass Index (BMI): n = 248, in kg/m2, mean (SD): 23.3 (3.5) Blood pressure status: Hypertensive, % (n/N): 40.2 (309/768) Antihypertensive medication used: NR Cardiovascular disease or stroke: NR Diabetes mellitus: NR Renal impairment: Serum creatinine ≥ 3.5 mg/dL (>= 309 µmol/L), % (n/N): 0 (0/768) Dietary potassium intake: NR Dietary sodium intake: NR Urinary potassium excretion: Potassium‐to‐creatinine ratio, mean (SD): 0.277 (0.143) Urinary sodium excretion: Sodium‐to‐creatinine ratio, mean (SD): 1.34 (0.92) Control Age: in years, mean (SD): 74.67 (6.7) Gender: Male, % (n/N): 100 (1213/1213) Ethnicity/race: Chinese Smoking: NR Body Mass Index (BMI): n = 391, in kg/m2, mean (SD): 23.0 (3.3) Blood pressure status: Hypertensive, % (n/N): 40.4 (490/1213) Antihypertensive medication used: NR Cardiovascular disease or stroke: NR Diabetes mellitus: NR Renal impairment: Serum creatinine ≥ 3.5 mg/dL (>= 309 µmol/L), % (n/N): 0 (0/1213) Dietary potassium intake: NR Dietary sodium intake: NR Urinary potassium excretion: Potassium‐to‐creatinine ratio, mean (SD): 0.284 (0.142) Urinary sodium excretion: Sodium‐to‐creatinine ratio, mean (SD): 1.14 (0.74) Inclusion criteria: Kitchens of a veteran's retirement home. Veterans registering into the retired home during the study period, as well as those already resident in the home Exclusion criteria: Kitchen preparing meals for a squad of bedridden residents, among others; residents with abnormal kidney function as assessed by serum creatinine concentrations of ≥ 3.5 mg/dL (>= 309 µmol/L) Pretreatment: None reported Method of recruitment of participants: Potential participants included veterans already resident in the retired home, as well as those who registered into the retired home during the study period. At baseline, resident participants from 10 squads were screened for eligibility before randomisation of the clusters (kitchens). Informed consent obtained: Yes, unclear whether written or oral Clusters: 5 clusters (Intervention group n = 395; n = 373; Control group n = 390; n = 410; n = 413). ICC calculation and outcome: NR Subgroups planned/measured: NR Subgroups reported: NR Participant flow Assessed for eligibility: n = 6 clusters; n = 1553 participants from n = 5 included clusters Excluded (number with reasons): n = 1 cluster (preparing meals for a squad of bedridden participants, among others); participants excluded through cluster exclusion NR, n = 2 excluded from other clusters (kidney function did not meet serum creatinine < 3.5 mg/dL) Randomised: n = 5 clusters with n = 1551 participants Allocated to LSSS intervention(s): Clusters n = 2; participants n = 768 (n = 635 at baseline; n = 133 during the study period) Allocated to control: Clusters n = 3; participants n = 1213 (n = 916 at baseline; n = 297 during the study period) Received allocated LSSS intervention(s): NR Did not receive allocated LSSS intervention(s): NR Lost to follow‐up (LSSS intervention group): n = 0 clusters; n = 265 (died n = 189; institutionalised n = 30; moved out n = 46) Discontinued intervention (LSSS intervention group): n = 0 clusters; n = 0 participants Analysed (LSSS intervention group): n = 2 clusters with n = 692 participants Excluded from analysis (LSSS intervention group): n = 0 clusters; n = 16 participants excluded from overall group (n = 1 invalid date of birth, n = 15 dates missing on death certificates; trial arm distribution not reported) Received allocated control: NR Did not receive allocated control: NR Lost to follow‐up (control group): n = 0 clusters; n = 435 (died n = 307; institutionalised n = 50; moved out n = 78) Discontinued intervention (control group): n = 0 clusters; n = 0 participants Analysed (control group): n = 3 clusters with n = 1085 participants Excluded from analysis (control group): n = 0 clusters; n = 16 participants excluded from overall group (n = 1 invalid date of birth, n = 15 dates missing on death certificates; trial arm distribution not reported)
Interventions	Intervention Characteristics LSSS intervention Theoretical basis: The long‐term effects of potassium‐enriched salt substitute on cardiovascular mortality needs to be determined. Description: Potassium‐enriched salt (49% NaCl; 49% KCl; 2% other additives) LSSS category: ≥ 30%KCl Contains fortificant: NR Delivery: discretionary use Duration of run‐in period: none Duration of active intervention: mean of 2.6 years; up to 3.7 years Duration of follow‐up (as reported): none Total duration of study (as reported): mean of 2.6 years; up to 3.7 years Timing: Gradually introduced by cooks in the intervention kitchens; replaced 'regular' salt within a 4 week‐period. Used throughout for the daily preparation of all meals Implementation: LSSS salt used by cooks in intervention kitchens Providers: Taiwan Salt Work manufactured the LSSS, which was weighed and delivered to kitchens by research staff. Co‐interventions: NR Resource requirements: NR Integrity of delivery: Amount of LSSS salt used per month recorded in each kitchen. Regular condiments and spices e.g. soy sauce and MSG, not limited. Mean LSSS salt use: 1.41 (0.22) kg/day per kitchen serving approx. 400 participants (approx. 3.8 g per person per day). Mean consumption of monosodium glutamate, soy sauce, vinegar, hot sauce, ketchup, and pickled vegetables during the first 3 mo of the trial in a subsample of 248 participants were 560 mg, 389 mg, 91 mg, 84 mg, 10 mg, and 225 mg per person per day. Control Theoretical basis: To enable comparison of LSSS intervention with standard practice Description: Regular salt (99.6% NaCl; 0.4% other additives) Contains fortificant: NR Delivery: discretionary use Duration of run‐in period: none Duration of active intervention: mean of 2.6 years; up to 3.7 years Duration of follow‐up (as reported): none Total duration of study (as reported): mean of 2.6 years; up to 3.7 years Timing: Used throughout for the daily preparation of all meals Implementation: Regular salt used by cooks in the control kitchens Providers: Taiwan Salt Work manufactured the regular salt, which was weighed and delivered to kitchens by research staff. Co‐interventions: NR Resource requirements: NR Integrity of delivery: Amount of regular salt used per month recorded in each kitchen. Regular condiments and spices e.g. soy sauce and MSG, not limited. Mean regular salt use: NR per kitchen (approx. 5.8 g per person per day reported). Mean consumption of monosodium glutamate, soy sauce, vinegar, hot sauce, ketchup, and pickled vegetables during the first 3 mo of the trial in a subsample of 391 participants: 536 mg, 394 mg, 88mg, 55 mg, 10 mg, and 224 mg per person per day.
Outcomes	Primary outcomes:  Diastolic Blood Pressure (DBP): NR Systolic Blood Pressure (SBP): NR Hypertension (as reported, or SBP > 140 mmHg or DBP > 85 mmHg): NR Blood pressure control (achieving blood pressure threshold or 'control' as prespecified): NR Cardiovascular events‐stroke: NR Cardiovascular events‐myocardial infarction: NR Cardiovascular events‐other: NR Cardiovascular mortality: Outcome measurement: death registry of Department of Health, using the International Classification of Diseases 9th revision (ICD‐9) code for cause of CVD‐related death. Time points: duration of study period Blood potassium: NR Hyperkalaemia: NR Hypokalaemia: NR Secondary outcomes:  All‐cause mortality: Outcome measurement: death registry of Department of Health, using the International Classification of Diseases 9th revision (ICD‐9) code for cause of death. Time points: duration of study period Adverse events: NR Antihypertensive medication use: NR Body Mass index (BMI): NR Serum creatinine: NR Albuminuria: NR Urinary albumin‐to‐creatinine ratio (uACR): NR Fasting blood glucose: NR Blood triglycerides: NR Total blood cholesterol: NR 24‐h urinary sodium excretion: NR 24‐h urinary potassium excretion: NR
Notes	Funding source: Taiwan Salt Work Company; Academia Sinica Authors name: Hsing‐Yi Chang; Wen‐Harn Pan Institution: Institute of Biomedical Sciences, Academia Sinica, Taipei Email: pan@ibms.sinica.edu.tw Possible conflicts of interest (for study authors): "None of the authors had any conflict of interest or any financial relations with the funding agent." Sources used for data extraction: Journal article with results of the trial Trial registration details: NR
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was done by drawing lots.
Allocation concealment (selection bias)	Unclear risk	Insufficient information available
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study authors reported that only participants were blinded. Thus, there may be a possibility for performance bias for study personnel e.g. cooks, research personnel.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Official coders of the Department of Health used the International Classification of Diseases 9th revision to assign cause of death. Good agreement between these judgements and those of four physicians. Unlikely that the coders would be aware of treatment allocation; the outcomes were also not likely to be influenced by unblinded assessment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Overall attrition was 76/768 (9.9%) in the intervention and 128/1213 (10.6%) in the control group. No difference in reasons for dropout
Selective reporting (reporting bias)	Unclear risk	No study protocol or prospective trial registry entry available
Other bias	Unclear risk	There was a considerable risk of contamination in the intervention group: "Other condiments and spices such as soy sauce and monosodium glutamate were not limited, because reasonably priced low‐sodium soy sauce and monosodium glutamate were not available at the time of the trial." These condiments accounted for approx. 30% of total sodium intake in a subsample of participants during the first three months of the study. New participants joined the study as they moved into the home; it was not clear whether these persons were assessed for renal function.
Recruitment bias (cluster‐RCTs)	High risk	A significant number of participants were recruited after the study clusters (kitchens) were randomised (LSSS intervention group n = 133; Control group n = 295).
Comparability with individually randomised trials (cluster‐RCTs)	Low risk	"Recently, Cook et al (31) presented the long‐term effects of sodium reduction on CVD based on the data from the Trials of Hypertension Prevention."
Loss of clusters (cluster‐RCTs)	Low risk	No loss of clusters reported
Baseline imbalance (cluster‐RCTs)	Unclear risk	Baseline characteristics were reported as not significantly different, however, height, weight, BMI, blood pressure, hypertension as well as sodium‐ and potassium‐to‐creatinine ratios were determined in a subsample of participants; 248/768 (32.3%) in the intervention and 391/1213 (32.2%) in the control group. Presence of comorbidities not reported. No information was provided on whether this subsample was randomly selected.
Incorrect analysis (cluster‐RCTs)	Low risk	Cluster effects were accounted for in the survival curves.
Overall risk of bias	High risk	Low risk of bias for incomplete outcome data and loss of clusters, unclear risk of bias for baseline imbalance, high risk of recruitment bias