ONTARGET, TRANSCEND and PROFESS – Clarifying, confusing or misleading

ONTARGET tested the hypothesis of non‐inferiority for the angiotensin receptor blocker (ARB) telmisartan vs the angiotensin‐converting enzyme inhibitor (ACEI) ramipril treatment of patients at high cardiovascular risk, which was confirmed with precision (p =ecolon;0.004), and also the hypothesis of the combination being superior to ramipril, which was rejected with about the same precision [1]. A total of 25 620 participants averaging 66 years of age, including 27% women, 74% with known stable coronary and 21% cerebrovascular disease, 37% with diabetes and 69% with known hypertension (blood pressure averaging 142/82 mmHg) were treated for more than 4 years, with only 43 persons lost to follow‐up. Numbers of participants reaching the primary cardiovascular endpoint were 1412 and 1423, respectively, on ramipril and telmisartan, and 1386 on the combination. Blood pressures were lowered by 6.0/4.6, 6.9/5.2 and 8.4/6.0 mmHg, respectively in the three groups, as expected in high‐risk people, of whom 57% were taking beta‐blockers and about 30% were taking both diuretics and calcium‐channel blockers (CCB) at the beginning of the trial. A large proportion of participants were also taking aspirin and statins. The data were consistent through all pre‐specified subgroups with no differences in any cardiovascular endpoint. Despite exclusion of patients with known ACEI intolerance from the trial, ARB treatment was better tolerated, with fewer drug discontinuations and adverse effects than ramipril treatment.

TRANSCEND included 5926 patients similar to those in ONTARGET, but with proven intolerance to ACEIs. These patients were randomized to either telmisartan or placebo on top of pre‐existing therapy [2]. Although on‐treatment mean blood pressure was 4.0/2.2 mmHg lower in the telmisartan group, these patients did not show any significant reduction in the primary cardiovascular outcome (cardiovascular death, stroke, myocardial infarction and heart failure; RR=0.92, 0.81–1.05, p= 0.23). However, when the main secondary endpoint (cardiovascular death, stroke and myocardial infarction but no heart failure as in HOPE [3]) was considered, a small beneficial effect closely approaching significance was found.

PROFESS recruited more than 20000 patients who had suffered a non‐cardioembolic ischaemic stroke within the previous 120 days, and randomized them to either telmisartan or placebo treatment [4]. There was a small reduction in the primary endpoint (recurrent stroke) in the telmisartan treatment group (RR = 0.95, 0.86–1.04) that was not statistically significant (p = 0.23), despite an average between group blood pressure difference of 3.8/2.0 mmHg.

The outcomes and clinical significance of these trial data are debated. ONTARGET showed that the ARB telmisartan and the ACEI ramipril are equally effective in preventing cardiovascular events in high‐risk patients and that the combination provides no added benefit and causes more adverse effects than either monotherapy in this patient population. This leaves most clinicians with two distinct choices, ACEI or ARB in high‐risk persons, including those with stable coronary disease, hypertension and diabetes. ONTARGET does not rule out use of the ACEI plus ARB combination in severe heart failure, where angiotensin escape mechanisms are expressed and dual blockade may be needed [5], [6] . However, in the ONTARGET high‐risk population without overt heart failure, the combination did not pay off. The lower blood pressure seen with the combination would be expected to lead to fewer endpoints, but this was not achieved. Further scrutiny of renal data in ONTARGET pointed towards problems with decreased renal function in a significant number of patients, while albumin excretion decreased [7]. This may question the niche for combined ACEI plus ARB therapy in chronic kidney disease with proteinuria. This issue needs to be tested with real outcome data. Furthermore, potential adverse effects on renal function may counterbalance the benefits of the lower blood pressure seen with dual blockade. The role of blood pressure, or possibly excessive blood pressure reduction in susceptible patients [8], is being scrutinized in analyses of ONTARGET, which was a trial of full doses of drugs with potent RAS blockade, particularly the combination of both principles, in prevention of events in high‐risk subjects, and not a hypertension trial. In this context, it should be revisited that the patients in the lowest quartile of baseline blood pressure did not benefit from ramipril in HOPE (compared to placebo), a finding that only made it to the website of New England Journal of Medicine in 1999 but never appeared in the full publication [3].

Interpretation of ONTARGET and TRANSCEND results therefore remains challenging. The patient population chosen for the trials and the concomitant treatments administered complicate interpretation of the effects of the tested therapies; these high‐risk patients were receiving multiple therapies (over 60% statins, over 80% antiplatelet agents, as well as a large number of other antihypertensive agents), thereby diminishing the differential effects of the active treatments under study. This is reflected in the finding that the incidence of cardiovascular events in the actively treated group of HOPE was similar to that of the placebo group in TRANSCEND. In addition, TRANSCEND was underpowered and should have been extended and endpoint driven.

The PROFESS findings were unexpected since stroke is closely blood pressure dependent. The blood pressure achieved in PROFESS was similar to that in PROGRESS [9] (138/82 mmHg), another trial in patients with previous stroke, which showed a highly significant 28% reduction in recurrent stroke, admittedly with a greater blood pressure difference between the active treatment and placebo arms (9.0/4.0 mmHg vs 3.8/2.0 mmHg in PROFESS). In PROGRESS, a smaller blood pressure difference (5/3 mmHg) induced by monotherapy with the ACEI perindopril did not significantly reduce recurrent stroke, and thus it cannot be excluded that prevention of recurrent stroke may require a larger blood pressure difference than observed in PROFESS. Kaplan–Meier curves of cumulative recurrent strokes appeared to diverge after 2.5 years in PROFESS, but randomized treatment lasted an average of only 2.5 years. A delayed onset of the benefit cannot be ruled out, but this potential outcome contrasts with observations from previous trials. For example, in HOPE [3] and PROGRESS [9], the cumulative stroke curves diverged significantly at 6 months. Furthermore, in PROFESS, fewer than 50% of included patients were followed for more than 2 years, thus reducing the statistical power of subsequent observations and substantiating the point that the study should have continued. Finally, all patients in PROFESS also received antiplatetet therapy in a factorial design, and these additional therapies may have obscured the benefits of a small blood pressure difference. Despite these limitations, PROFESS, mainly because of its large size, raises doubts about the effectiveness of telmisartan or ARBs in general [10] in prevention of recurrent stroke.

In conclusion, ONTARGET, TRANSCEND and PROFESS, though carried out by large and dedicated teams of investigators, and clarifying some issues, also raise important questions. ONTARGET showed that telmisartan is non‐inferior to ramipril in the high‐risk patients, mostly with established cardiovascular disease, and that the combination is not indicated and even causes harm in these patients. However, TRANSCEND and PROFESS were underpowered for the questions raised and partly undermined the ONTARGET data, calling into question the benefit of telmisartan in secondary prevention of cardiovascular events, particularly stroke, in high‐risk patients. Furthermore, the findings of ONTARGET call into question the benefits of dual RAS blockade in high‐risk patients with nephropathy.