Blood pressure-dependent and independent effects of amlodipine and angiotensin receptor blockers on the risk of stroke

Abstract

Calcium channel blockers (CCBs) by the use of amlodipine substantially reduce the risk of stroke, and the benefit of initial treatment with this drug versus other antihypertensive drugs could be largely, though not entirely, explained by blood pressure differences between randomized groups of trials [1]. When contrasted with a long-acting CCB, amlodipine, angiotensin receptor blockers (ARBs) are less effective in lowering blood pressure and also demonstrate inferiority for the prevention of stroke [1]. It remains uncertain, however, whether mechanisms independent of blood pressure also affect the size of the treatment benefit, and there have been few formal attempts to investigate any such classand drug-specific effects. Using data from trials of amlodipine or ARBs that were included in recent meta-analyses [2,3], we conducted analyses to determine the relative contribution of blood-pressure-dependent and independent mechanisms to the reduction in risk of stroke produced by these drugs. Instead of comprehensive electronic and manual searches, we identified 10 and 17 odds ratios (ORs) for stroke from randomized controlled trials of amlodipine (77,093 individuals) and ARBs (104,360 individuals), respectively, from the comprehensive metaanalyses [2,3] (Table 1). We used sophisticated methods designed by the Blood Pressure Lowering Treatment Trialists’ Collaboration (BPLTTC) [4]. For each trial, the mean difference in follow-up systolic blood pressure levels between the randomized groups was sought. The association between the difference in follow-up systolic blood pressure levels and the log OR for stroke was investigated using random-effects meta-regression analysis with inverse variance weighting. Initially, using a single regression model, separate regression lines were fitted for trials in which amlodipine or ARBs were the treatment. The slopes of these lines were compared, so as to test for a differential effect of blood pressure reduction on risk in trials of amlodipine compared with ARBs. If these slopes were not significantly different (pN0.05), a parallel lines regression model was fitted, which assumes equal effects of blood pressure reduction with amlodipine and ARBs. Using this model, the null hypothesis of no separation between the 2 parallel lines was tested to explore whether there was evidence of a differential effect of amlodipine and ARBs independent of blood pressure lowering. Assumptions of linear associations between differences in follow-up blood pressure levels and log ORs were tested using standard graphical methods. Analyses were carried out using the metareg routine in STATA version 8.0 (Stata Corporation, College Station, Texas). In the single regression model, the association between the magnitude of reduction in blood pressure and the size of relative risk reduction were similar for amlodipine and ARBs (p=0.326). In the parallel lines regression model, treatment with amlodipine or ARBs achieved a 4.0% (95% confidence interval [CI]; 1.2% to 7.0%; p=0.007) reduction in the risk for each 1 mm Hg reduction in blood pressure (Fig. 1). At zero blood pressure reduction, the estimated relative risk reduction with amlodipine was 11.9% (95% CI, 4.4% to 18.7%; p=0.004). However, ARBs did not confer any additional protection beyond that conferred by blood pressure reduction alone. There was a significant difference between amlodipine and ARBs in terms of their likelihood of providing protection independent of blood pressure lowering (p=0.010), suggesting greater protection with amlodipine than ARBs. The size of blood pressure reduction achieved with amlodipine or ARBs was directly associated with the size of the reductions in the risk of stroke. In addition, although treatment with amlodipine provided an additional 11.9% relative risk reduction beyond that explained by the observed blood pressure differences, no such effect was observed for ARBs. This cerebroprotective effect of amlodipine was significantly greater than any such effect produced by ARBs. These findings extend those previously reported by Wang et al. [1], who plotted only 5 ORs for amlodipine trials and 5 ORs for ARB trials and performed meta-regression analyses separately for trials involving amlodipine and ARB groups. The present analyses provide much more reliable information about the blood pressure-dependent and independent effects of amlodipine and ARBs. The analyses included 10 ORs for amlodipine trials and 17 ORs for ARB trials and involved more sophisticated statistical methods (parallel lines regression model) specifically aimed at the identification of blood pressure-dependent and independent components of the treatment effects. For the risk of stroke, metaregression analyses by the BPLTTC [4] showed similar blood pressure-dependent effects of angiotensin-converting enzyme inhibitors and ARBs but no blood pressure-independent effects for both drug classes. Maximization of the benefit for stroke may be achieved with a regimen that includes amlodipine together with other drugs in an effort to optimize the size of the blood pressure reduction achieved. The authors of this manuscript have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology (Shewan and Coats 2010;144:1–2).