OBJECTIVE: Treatments such as ACE inhibitors and beta-blockers reduce mortality in CHF. Demonstration of a parallel improvement of HRQL would appear of good added-value. METHOD: A critical appraisal of published controlled trials used the guidance document and the checklist developed by ERIQA. RESULTS: Although a reduction of mortality and NYHA score was demonstrated in a few large studies, none was able to show a clear and unbiased HRQL improvement vs. placebo. Several major issues that would increase the confidence in the results are not presented or missing: hypotheses of changes in HRQL scores and power estimation; justification of the choice of questionnaires (e.g. 6 questionnaires were used in a trial and presented as HRQL, some of them being a dyspnoea score or a global evaluation of change, with no explanation whether they were covering different concepts, and without selecting the most important questionnaire for the primary statistical analysis); validation data (e.g. a trial used 14 subscales taken from 3 different validated questionnaires. Is the resultant questionnaire still valid ?); statistical analysis plan; description and imputation of missing data; statement that the analysis is in intent to treat. In the majority of trials, small differences were observed for a few subscales among many, or only at some assessments over time. When the difference reaches statistical significance, it is only because of the large sample, but it has no clinical signification. Even for the physical subscale which may be of importance in CHF, small non significant effect sizes are observed. Some explanations have been proposed for these results: non severe patients included, questionnaire not reliable or responsive. CONCLUSION: There is concern to conclude that despite attempts of publications and general reviews to enhance the modest results, there is no definite demonstration that HRQL is improved by treatment which reduces mortality in CHF.