The randomisation of 1518 outpatients from 51 centres began on 1 June 2000 and ended on 1 November 2001. A total of 760 patients were allocated to the intervention group and 758 to the control group. We stopped the trial on 1 August 2002 when 400 primary events had been reported. Another 35 events occurred before the closing date and were confirmed in the last meeting of the event committee.
The baseline characteristics of the two groups were similar (). The mean age was 65 years, 71% were men, most patients were in New York Heart Association (NYHA) class II or III, and about 80% had left ventricular systolic dysfunction. Follow-up was completed in 1511 (99.5%) randomised patients ().
Baseline characteristics. Values are numbers (percentages) unless stated otherwise
Flowchart of enrolment, randomisation, and follow-up of patients
The mean length of follow-up was 16 (range 7-27) months. The primary outcome occurred in 200 (26.3%) patients in the intervention group and in 235 (31%) patients in the control group (relative risk reduction 20%, 95% confidence interval 3% to 34%, P = 0.026) (). We found no difference between the adjusted and unadjusted effect of the intervention for the primary outcome (adjusted for NYHA class, age, baseline treatment, comorbidity, and systolic dysfunction).
Fig 2 Kaplan-Meier curves for rate of death from any cause or admission to hospital for heart failure (panel A), rate of death from any cause (panel B), and rate of admission for heart failure (panel C). HF=heart failure; RR=relative risk (with 95% confidence (more ...)
The reduction in the incidence of the primary end point was mostly due to a relative risk reduction in the incidence of admissions for heart failure of 29% (9% to 44%, P = 0.005): 128 (16.8%) in the intervention group as compared with 169 (22.3%) in the control group. The effect on all cause mortality was not significant (relative risk reduction = 5%, -27% to 23%, P = 0.69) ().
Summary of primary and secondary end points. Values are numbers (percentages) unless stated otherwise
The rate of all cause admissions was also lower in the intervention group: 261 patients were admitted at least once compared with 296 in the control group (reduction = 15%, 0.1% to 28%, P = 0.049). Significantly fewer cardiovascular admissions were recorded in the intervention group than in the control group (183 v 228 patients, reduction = 24%, 7% to 28%, P = 0.006) ().
The reduction in the primary end point with the intervention was similar in direction and magnitude in all pre-specified subgroups (for example, NYHA class, ventricular function, drug treatment) (). Although we observed no significant interaction between subgroups, the power to analyse each subgroup was limited.
Subgroup analysis: admission to hospital for heart failure
Quality of life and adherence
A total of 1159 patients completed the Minnesota living with heart failure questionnaire at the final visit (excluding 238 deaths and 121 patients with visual defects, altered cognitive functions, or cerebrovascular events). Patients in the intervention group had better quality of life than control patients at the end of the study (mean total score in intervention group 30.6 v 35.0 in control group; mean difference = 4.4, 95% confidence interval 1.8 to 6.9, P = 0.001). We also found a difference in the physical score (11.2 v 12.8, P = 0.007) and emotional score (6.7 v 7.9, P = 0.002).
At the end of the trial, significantly more patients in the intervention group than the control group were taking β blockers (450 (59.2%) v 391 (51.6%), P = 0.003), spironolactone (207 (27.2%) v 172 (22.6%), P = 0.03), digoxin (254 (33.4%) v 217 (28.6%), P = 0.04), and furosemide (588 (77.3%) v 535 (70.5%), P = 0.007). A similar trend occurred in the use of angiotensin converting enzyme inhibitors (595 (78.3) v 575 (75.8%), P = 0.24). More patients in the control group stopped taking any drugs (138 (18.2%) v 61 (8.0%), P < 0.001) and reported dietary transgressions (492 (64.9%) v 154 (20.2%), P < 0.001).