Effects of antihypertensive drugs on coronary artery disease risk: a meta-analysis

1. The effects of antihypertensive drugs on lipids may also influence their effect on coronary artery disease (CAD). However, the clinical significance of these effects and the extent to which they persist during long-term therapy is uncertain. 2. We performed a meta-analysis on 23 randomized trials published between 1988 and 1994 that compared the effects of atenolol, celiprolol (a beta-blocker with beta 2-adrenoceptor intrinsic sympathomimetic activity), enalapril, nifedipine and doxazosin on plasma cholesterol, low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), triglycerides and blood pressure (BP). 3. Predicted changes in CAD risk were calculated by incorporating the results for these parameters into the Framingham equations. 4. While there were no differences in antihypertensive efficacy between the drugs, atenolol significantly (P < 0.05) reduced HDL-C and increased total cholesterol, LDL-C and triglycerides compared with celiprolol, enalapril, doxazosin and nifedipine. 5. The magnitude of the effects on lipids was not significantly influenced by the duration of therapy (up to 3 years for atenolol and doxazosin and up to 2 years for celiprolol). 6. The improvement in Framingham equation point scores (systolic BP formula) was significantly (P < 0.05) less for atenolol (-0.54; confidence intervals (CI) -0.29-(-0.78)) than for celiprolol (-1.69; CI -0.68-2.70), doxazosin (-1.67; CI -1.11-(-2.23)), enalapril (-1.43; CI 0.23-(-3.07)) and nifedipine (-1.91; CI -1.22-(-2.59)). Similar results were obtained for the Framingham diastolic BP formula. 7. These results suggest that the adverse effects of atenolol ]on plasma lipids do not improve with prolonged therapy and are theoretically great enough to reduce its efficacy in reducing CAD by approximately two thirds compared with antihypertensive drugs that do not adversely affect plasma lipids. However it must be emphasized that these are theoretical effects. In order to determine the actual differences between these drugs on CAD end points, studies using these end points are required.