Session number: 1008
Abstract title: Subjects in Heart Failure Trials Are Not the Same as Your Patients
Author(s):
M Ansari - San Francisco VAMC, University of California, San Francisco
JR Teerlink - San Francisco VAMC, University of California, San Francisco
M Alexander - Division of Research, Kaiser Permanente Medical Care Program, Northern California Region
A Tutar - University of Louisville, Louisville, Kentucky
BM Massie - San Francisco VAMC, University of California, San Francisco
Objectives: Clinical practice guidelines for the management of patients with chronic heart failure (CHF) are intended to help translate results of randomized controlled trials into specific treatment recommendations. This approach assumes selected patients enrolled in clinical trials are representative of the general CHF population. Therefore, we sought to compare CHF patients in a large regional HMO as well as a local VA hospital with those enrolled in major CHF clinical trials.
Methods: Outpatients with a CHF diagnosis from 7/96-12/96 (HMO sample) or 10/98-10/99 (VA sample) were evaluated, of whom a total of 778 patients were analyzed after exclusions for prior CHF hospitalizations, absence of Framingham CHF criteria, lack of continuous follow-up, and presence of terminal comorbidities. The cohorts were compared with the participants of 5 landmark trials that have defined CHF practices.
Results: Compared to study participants, our patients were older, included nearly twice as many women (HMO sample only), and had a higher mean EF. The proportion of our study sample that would have met eligibility for these trials was <50% for all the trials and <20% for 4 of the 5 trials. Exclusions overall were mainly for EF>45%, however, even when only low EF (<45%) patients were analyzed, the percentage of our sample patients meeting entry criteria remained poor (<35%) except for the DIG trial, which was the most inclusive. Other trial specific exclusions were age, COPD, significant valve disease, low HR, absence of a required background medication or use of excluded ones.
Conclusions: Clinical trials are poorly representative of CHF outpatients. The greatest difference is exclusion of patients with preserved EF from trials, but other factors are the younger age of patients, low proportion of women, absence of valvular heart disease, and exclusion of commonly used medications.
Impact statement: The use of inclusion and exclusion criteria may not affect the internal validity of trials, however, they may markedly reduce the external validity of the results. Care must be taken in extrapolating results from highly selected study participants to the general population.