OP.4A.07] THREE DIMENSIONAL META-ANALYSIS OF HYPERTENSION TRIALS WITH HEART FAILURE AS AN OUTCOME

Abstract

Objective: Odds ratios (OR) are usually reported in meta-analyses summarizing data from multiple trials. However, absolute risk reduction (ARR) and drop-out rates (DOR) are important in the application of clinical trials in different clinical settings. We report a novel random effects meta-analysis method where OR, ARR and DOR are simultaneously evaluated. Design and method: The effects of different classes of antihypertensive medications in preventing the occurrence of heart failure (HF) was examined using OR and ARR data from 54 randomized clinical trials (331,728 patients, 12,969 HF occurrences), of which 16 trials (87,187 patients, 2,793 HF events) included DOR data. A three dimensional meta-analysis was performed using a novel algorithm based on the R statistical program. Results: The expected relationship between OR and ARR for a given baseline risk was observed. Overall, higher efficacy in preventing HF (low OR 0.82, 95% CI 0.75 to 0.90) was associated with higher ARR (4.52%, 95% CI 2.46% to 6.58%), but also with higher DOR (OR 1.32%, 95% CI 1.60% to 1.94%). Diuretics were more effective in decreasing the OR of the occurrence of HF at follow-up (OR 0.52, 95% CI 0.38 to 0.70) and were associated with higher ARR (14.64%, 95% CI 7.30% to 21.98%). However, diuretics were associated with higher DOR (2.42%, 95% CI 1.51% to 3.90%). Verapamil was associated an increased occurrence of HF during follow-up (OR 1.17, 95% CI 0.99 to 1.37) as well as with an increased rate of occurrence of HF during follow-up (by 2.34%, 95% CI 0.07% to 4.75%). The values for the other drug classes were between those for diuretics and verapamil. Angiotensin receptor blockers and intensive control of blood pressure regardless of antihypertensive class were associated with the lowest DOR (1.15%, 95% CI 0.91% to 1.44% and 1.14%, 95% CI 0.84% to 1.54% respectively). Conclusions: This three dimensional meta-analysis may be useful in guiding clinicians in choosing the appropriate antihypertensive medication class in clinical care. Odds ratios may be used to summarize clinical trials, while absolute risk reduction and drop out rate are more appropriate in answering specific clinical questions.