Methodologic problems in clinical trials in heart failure.

Trials designed to establish the effectiveness of pharmacologic therapy in chronic heart failure illustrate methodologic problems associated with intervention trials in chronic disease. Early improvement in hemodynamic status or exercise capacity does not predict long-term changes, and hemodynamic variables do not correlate with exercise capacity. Mortality may not be a feasible outcome, suggesting that an adequate measure of quality of life is necessary to establish patient benefit. Care must be taken to ensure that subjects recruited are representative of the underlying population with respect to drug response. Need for cointervention can be dealt with by comparing only periods of comparable treatment, treating need for cointervention as a treatment failure, or using changes in non-study drugs as a measure of outcome. While cross-over designs can improve feasibility, the investigator runs the risk of inadequate followup and a time-treatment interaction. These issues must be considered in planning intervention trials in chronic disease.