The Telmisartan Randomized Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease (TRANSCEND) assessed whether the administration of the AT1 receptor blocker telmisartan, in patients who were intolerant to angiotensin-converting enzyme (ACE) inhibitors, could reduce cardiovascular outcomes or new onset diabetes in high-risk patients . The patients had to be aged 55 years or older, be intolerant of ACE inhibitors, and have established coronary artery disease, peripheral vascular disease or cerebrovascular disease, or diabetes with end organ damage. In the original trial, telmisartan did not reduce the composite primary outcome (cardiovascular death, myocardial infarction, stroke, or hospitalization for heart failure: hazard ratio 0.92: P = 0.2) but did reduce the three-fold outcome, omitting heart failure (hazard ratio 0.87: P = 0.048).
In this issue of the Journal, Foulquier et al. present a set of post-hoc analyses, designed to test the hypothesis that telmisartan might have a differential effect, on both the primary outcome and a number of secondary outcomes in hypertensive versus nonhypertensive patients . Hypertension was defined as patients with a medical history of hypertension or with sitting SBP at least 140 mmHg or a DBP at least 90 mmHg. One major difficulty with the analyses presented, is that by this definition 86% were hypertensive so that the two groups are quite unbalanced: this is made worse by the fact that many of the 14% so-called nonhypertensive patients were actually well controlled hypertensive patients . The small number of nonhypertensive patients (828 versus 5098) bedevils the analyses as that sample size is plainly too small to demonstrate significant effects for the majority of outcomes tested.
The new analyses show that SBP in the hypertensive subgroup was reduced from 143 mmHg at baseline, by 7.4 mmHg in the telmisartan group and 3.5 mmHg in the placebo group. On the other hand, in the nonhypertensive group, SBP was reduced by 0.8 mmHg in the telmisartan group compared with an increase of 5.4 mmHg in the placebo group . As far as outcomes are concerned, it may be best to quote the authors who state in the results section of the abstract: ‘There was no evidence for a significantly differential treatment effect of telmisartan in hypertensive and nonhypertensive patients for any end-points’. And in the conclusions of the abstract they state; ‘The effect of telmisartan in hypertensive and nonhypertensive patients at high cardiovascular risk was not different’. Although this is disappointing, and while it could well reflect the true situation, it is also possible that it distorts the truth and reflects the inherent weakness arising from the very small numbers in the nonhypertensive group, leading to large confidence intervals, so that when the authors analyze for ‘interactions’ or for ‘heterogeneity’ between results in hypertensive and nonhypertensive groups, these results are invariably ‘nonsignificant’.
But, as the authors report, there are some indications that certain benefits may be greater in the hypertensive group, for at least they are able to demonstrate a significant reduction for some outcomes in this subgroup. This applies to the three-fold composite outcome of cardiovascular death, myocardial infarction, and stroke, as well as for myocardial infarction alone, and for new onset albuminuria, all of which appear to be significantly reduced in the hypertensive group but not in the nonhypertensive patients .
The authors very correctly point out the limitations of these analyses, including the fact that they are posthoc, observational, and underpowered in the nonhypertensive group. They conclude that the possibility of differential benefits of telmisartan therapy in the hypertensive subset should be seen as hypothesis generating and tested through further research. This highlights a more general limitation that applies to so many of the best and largest randomized clinical trials, no matter how well conceived and conducted. That limitation is the very real and necessary limitation on trial size and patient numbers imposed by the realities of obtaining funding for even the most necessary and important trials addressing questions of fundamental importance. The limitations of funding are such that trial size and sample size are guided by the need to achieve sufficient power to address the primary question posed and to detect significant differences in the primary outcomes specified. This limitation immediately restricts the capacity of most trials to address the very many related questions that beg to be answered. This includes those related to secondary outcomes or to critically important subgroup analyses, whether prespecified or not. Unfortunately, these limitations are unlikely to be alleviated, and may even grow more severe, thus limiting the potential of even the best conceived trials to provide critically important information.
J.C. has received research grants and honoraria from Servier International for the PROGRESS and ADVANCE trials, and grants from the National Health & Medical Research Council of Australia for several large-scale trials of blood pressure lowering in acute stroke.
Conflicts of interest
There are no conflicts of interest.
1. Yusuf S, Teo K, Anderson C, Pogue J, Dyal L, Copland H. TRANSCEND InvestigatorsEffects of the angiotensin-receptor blocker telmisartan on cardiovascular events, in high-risk patients intolerant to angiotensin-converting enzyme inhibitors: a randomised controlled trial. Lancet
2. Foulquier S, Böhm M, Schmieder R, Sleight P, Teo K, Yusuf S, et al. Impact of telmisartan on cardiovascular outcome in hypertensive patients at high risk: a Telmisartan Randomised AssessmeNt Study in ACE iNtolerant subjects with cardiovascular Disease subanalysis. J Hypertens