CHARM‐Preserved.

Methods	Study design: parallel RCT Centres: 618, 26 countries (Australia, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Hungary, Iceland, Italy, Luxembourg, Malaysia, Netherlands, Norway, Poland, Portugal, Russia, Singapore, South Africa, Spain, Sweden, Switzerland, UK/Ireland, USA) Start of enrolment: March 1999 End of enrolment: July 2000 Median follow‐up: 36.6 months Run‐in period: no
Participants	Inclusion criteria: "Eligible patients were aged 18 years or older, had New York Heart Association functional class II–IV of at least 4 weeks’ duration, had a history of hospital admission for a cardiac reason, and had LVEF higher than 40%." Exclusion criteria: Important exclusion criteria for any of the studies include current serum‐creatinine > 265mmol/L (> 3 mg/dL); current serum‐potassium > 5.5 mmol/L (> 5.5 mEq/L) or a history of marked ACE inhibitor‐induced hyperkalemia resulting in either a serum potassium greater than or equal to 6.0 mmol/L (>6.0 mEq/L) or a life‐threatening adverse event; known bilateral renal artery stenosis; current symptomatic hypotension; persistent systolic or diastolic hypertension; stroke, acute myocardial infarction, or open heart surgery within the last 4 weeks; previous heart transplant or heart transplant expected to be performed within the next 6 months; presence of any noncardiac disease (eg, cancer) that is likely to significantly shorten life expectancy to less than 2 years. Randomised (N): 3023 (1514 intervention, 1509 control) Withdrawn (N): for reasons other than death (270 intervention, 204 control) Lost to follow‐up (N): (2 intervention, 1 control) Analysed (N): 3020 (1512 intervention, 1508 control) Age (years, mean, SD): intervention: 67.2, 11.1; control: 67.1, 11.1 Sex (% men): intervention: 60.8; control: 59.0 Ethnicity (%): intervention: European 90.8 , control: European 92.3 Systolic blood pressure (mmHg, mean, SD): intervention: 136.0, 18.6; control: 136.3, 18.3 Heart rate (beats/min, mean, SD): intervention: 71.2, 12.4; control: 71.4, 12.5 BMI (mean, SD): intervention: 29.3, 5.9; control: 29.0, 5.6 Serum creatinine (mg/dL): not reported B‐type natriuretic peptide (pg/mL): not reported NT pro B‐type natriuretic peptide (pg/mL): not reported LVEF (%, mean, SD): intervention: 54.0, 9.4; control: 54.1, 9.4 NYHA class I (%): 0 NYHA class II (%): intervention: 61.5; control: 60.0 NYHA class III (%): intervention: 36.7; control: 38.7 NYHA class IV (%): intervention: 1.8; control: 1.3 Hypertension (%): intervention: 65.0; control: 63.6 Diabetes (%): intervention: 28.7; control: 28.0 Atrial fibrillation (%): intervention: 29.0; control: 29.3 Hospitalisation for heart failure (%): intervention: 69.6; control: 68.8 Coronary heart disease (%): intervention: 45.0; control: 43.7 Stroke (%): intervention: 9.2; control: 8.5 Diuretic (%); intervention: diuretic: 75.2, spironolactone 11.3; control: diuretic 74.3, spironolactone 12.0 Digoxin (%): intervention: 28.5; control: 27.2 Beta‐blocker (%): intervention: 55.9; control: 55.5 ACEI (%): intervention: 19.6; control: 18.6 ARB (%): study drug MRA (%): intervention: 11.3; control: 12.0
Interventions	Intervention: candesartan. "which could be started at 4 or 8 mg once daily, the assignment code being held by an independent centre and the data safety monitoring board. The treatment dose was doubled every 2 weeks, as tolerated, according to a forced titration protocol, with recommended monitoring of blood pressure, serum creatinine, and potassium. The target dose was 32 mg once daily from 6 weeks onwards." Comparator: "matching placebo" Concomitant medication: "physicians were free to prescribe all treatments other than angiotensin‐receptor blockers." "Initially, angiotensin converting‐enzyme inhibitors were not allowed as concomitant treatment, but after publication of the Heart Outcomes Prevention Evaluation trial results, (The Heart Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin‐converting‐enzyme inhibitor, ramipril, on cardiovascular events in high‐risk patients. N Engl J Med 2000; 342: 145–53.) their use was optional in appropriate patients." "By the end of the study, 298 (20%) in the candesartan and 340 (23%) in the placebo group were receiving angiotensin‐converting‐enzyme inhibitors, 712 (47%) and 748 (50%) were receiving blockers, and 136 (9%) and 201 (13%) were receiving spironolactone. Non‐study angiotensin‐receptor blockers were used in 3% of patients in each of the two groups."
Outcomes	Planned: Planned: "The primary outcome was cardiovascular death or unplanned admission to hospital for the management of worsening CHF. Prespecified secondary outcomes were: cardiovascular death, admission to hospital for CHF, or non‐fatal myocardial infarction; cardiovascular death, admission to hospital for CHF, non‐fatal myocardial infarction, or non‐fatal stroke; cardiovascular death, admission to hospital for CHF, non‐fatal myocardial infarction, non‐fatal stroke, or coronary revascularisation; death (any cause) or admission to hospital for CHF; and development of new diabetes." Reported: as planned
Notes	The CHARM program consisted of 3 strands, one of which was CHARM‐Preserved. Contacted investigators for end scores of MLHF QoL by treatment arms. No response.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"We randomly assigned patient"
Allocation concealment (selection bias)	Low risk	"the assignment code being held by an independent centre and the data safety monitoring board"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	"We randomly assigned patients, in a double‐blind way", "matching placebo"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	"A committee unaware of treatment assignment and which component of the CHARM programme was being undertaken adjudicated the cause of death, first myocardial infarctions, and first hospital admissions for heart failure." "All final data analyses were done by the sponsor and verified independently by the statistical centre at the London school of Hygiene and Tropical Medicine, London, UK"
Incomplete outcome data (attrition bias) All outcomes	Low risk	"two patients who mistakenly received randomisation numbers but had no other data recorded and never received study medication" "Two candesartan patients and one placebo patient were lost to follow‐up" 207/204 withdrew due to AE "Anaysis was done by intention to treat."
Selective reporting (reporting bias)	Low risk	outcomes reported as planned
Other bias	Low risk	"MA Pfeffer, K Swedberg, CB Granger, JJV McMurray, and S Yusuf have served as consultants to or received research grants from AstraZeneca and other major cardiovascular pharmaceutical companies. J Östergren has served as a consultant and received research grants from AstraZeneca. P Held, E L Michelson, and B Olofsson are employees of AstraZeneca." "This study was supported by AstraZeneca R&D, Mölndal, Sweden"