USPHSHCSG 1977.

Methods	Randomized, double‐blind, placebo‐controlled trial conducted in ambulatory young people in USA.
Participants	389 participants. Mean age 44.3 years, range 21‐55 years. 28% African‐Americans. 80% men. Baseline mean SBP/DBP 146.9/99 mmHg and pulse pressure 48 mmHg. Inclusion criterion: DBP 90‐115 mmHg. Target: none (medication was not titrated). Exclusion criteria: diabetes mellitus, renal insufficiency, or hypercholesterolaemia; abnormal ECG including single or double Master test; radiographic cardiomegaly; grade III or IV retinopathy; clinical history or findings of previous arterial thrombosis or vascular insufficiency, whether coronary, cerebral or peripheral; CHF; angina pectoris; valvular heart disease; or secondary or correctable hypertension; and known sensitivities to the intervention agents. Follow‐up: 10 years.
Interventions	Treatment: chlorothiazide 500 mg twice daily + rauwolfia serpentina 100 mg twice daily. Control: placebo.
Outcomes	All‐cause mortality, CHD, stroke, CHF, SBP and DBP.
Notes	Study carried out in middle‐aged population with mild hypertension which is itself a low‐risk factor for studying mortality and morbidity data in such a population. Study terminated if participant had a stroke.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quotes: "Subjects were randomly assigned to treatment or placebo and then matched by race and sex for two broad age groups (under 46 and 46‐55). The randomization was carried out within each of the six participating clinics" (page 100). "The distribution of all pre‐treatment characteristics into the active drug and placebo groups was uniform." Comment: randomization carried out, but method used for generation of random numbers not specified.
Allocation concealment (selection bias)	Low risk	Quote: "At the conclusion of the trial period, subjects were randomly assigned either active or placebo treatment, and this medication was substituted for the identical placebo of the trial period and administered in double‐blind fashion. Active therapy consisted of chlorothiazide, 500 mg, plus rauwolfia serpentina, 100 mg, in one tablet taken twice daily. There was no intervention on diet or smoking or other behavioral factors." Comment: allocation carried out randomly but methodology used to conceal allocation not mentioned; however, baseline characteristics well matched.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quotes: "At the conclusion of the trial period, subjects were randomly assigned either active or placebo treatment, and this medication was substituted for the identical placebo of the trial period and administered in double‐blind fashion." "The complications were also classified in terms of those events considered likely to be the consequence of elevated pressure per se and those which are predominantly associated with vascular sclerosis (Table 2). All such events were reviewed by two consultants otherwise unassociated with the trial, who were provided with all pertinent information except knowledge of the treatment regimen." Comment: participants and treating physicians were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quotes: "administered in double‐blind fashion." "Follow‐up continued for another 4 months, the last 2 weeks of which included home blood pressures again. At that point the annual examination procedures were repeated. Thereafter, the regimens were unblinded and investigators were at liberty to treat as clinically indicated." "Of the complications observed, only stroke required termination from the regimen to which they were randomized. For myocardial infarction it was elective, depending on the clinical circumstances. Thus, by design, most subjects continued on the same double‐blind follow‐up after their first morbid event and were at risk for additional subsequent events. Others were followed on known medication." Comment: blinding of outcome assessor not mentioned.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quotes: "Of importance in considering the morbidity and side effects data in this report is the fact that there was no differential dropout rate between the treatment and control groups (33.2% vs. 34.7%). This applies to those who simply failed to return ("Lost to Follow‐up") as well as those who voluntarily "withdrew" from assigned therapy but remained under follow‐up. The number for whom vital status is unknown is also similar in the two groups (14 vs.12)." "During that time, 206 (52.9%) were terminated from their assigned regimen (Table 5). Dropouts accounted for 132 (33.9%), of whom 75 have been lost to regular follow‐up. The vital status of 26 of the dropouts is unknown. Drug intolerance necessitated terminations in 23 cases and major morbid events in 27, four of which were deaths. The remainder of those terminated have continued under regular follow‐up on known medications, including 24 who were terminated as treatment failures on the basis of progressive elevation of blood pressure to above a predetermined level." "The dropout rate of 33.9% overall is within that allowed for in the calculation of sample size (5% per year of follow‐up). At the beginning of closeout, one‐half remained on their assigned coded medication." Comment: attrition rate high. Attrition rates per year not mentioned as cut‐off was kept at 5% lost to follow‐up, and if > 5% the outcome measures will be affected.
Selective reporting (reporting bias)	Unclear risk	Comment: secondary and tertiary outcomes as mentioned in the objectives not mentioned in results section.
Other bias	Unclear risk	Comment: conflict of interest not reported. Source of funding not stated.