Added VALUE of an ancillary study on ambulatory blood pressure monitoring

The Valsartan Antihypertensive Long-Term Use Evaluation (VALUE) investigators undertook a double-blind, randomized clinical trial to test the hypothesis that, for the same degree of blood pressure lowering, inhibition of the renin system at the level of the angiotensin II type 1 receptor would be more effective in preventing cardiac morbidity and mortality than calcium channel blockade [1–3]. Across 943 centres and 31 countries, 15 245 high-risk patients were randomized to antihypertensive treatment based on amlodipine (5–10 mg per day) or valsartan (80–160 mg per day). Median follow-up was 4.2 years [2]. The VALUE investigators assumed that the primary endpoint, a composite of fatal and non-fatal cardiac endpoints, would run at a rate of 25 events per 1000 patient-years. Within less than 5 years, the number of cardiac endpoints exceeded the projected number of 1450 by 149. Cardiac endpoints occurred at similar rates in the two treatment groups because the incidence of myocardial infarction was 25.8% lower (P = 0.02) on amlodipine compared to valsartan, whereas, on amlodipine, the risk of heart failure increased by 12.7% (P = 0.12). The incidence of fatal and non-fatal stroke was 17.1% lower (P = 0.08) on amlodipine compared to valsartan [2]. In this issue of the journal, Lederballe-Pedersen et al. [4] report an ancillary study on ambulatory blood pressure monitoring embedded in the VALUE trial. The substudy included 47 centres (5.0%) from three countries (9.7%). The intention-to-treat analysis included 659 patients (4.3%). After 1 year, patients randomized to amlodipine and valsartan had nearly equal 24-h ambulatory blood pressures, averaging 131.5 and 132.5 mmHg systolic, and 75.2 and 74.8 mmHg diastolic, respectively. The authors of the ancillary study have to be congratulated for their courage to mount a study on ambulatory blood pressure monitoring within the framework of a large simple trial [5]. Nonetheless, the results of the VALUE substudy must be interpreted within the context of their limitations. Because it included less than 5% of the 15 245 VALUE patients, it cannot be representative for the trial as a whole. Two observations exemplify the degree of selection. First, the over-representation of patients with electrocardiographic left ventricular hypertrophy in the valsartan group of the ancillary study (15.7 versus 8.0%; P = 0.002) did not occur in the main trial (11.9 versus 12.5%). Second, the systolic/diastolic gradient on conventional blood pressure measurement at 12 months in favour of amlodipine (0.4/0.5 mmHg) was substantially smaller than in the overall trial population (2.0/1.5 mmHg). Of the VALUE patients, 92.3% were receiving antihypertensive treatment at enrolment [2]. As in most other large simple trials [6], the VALUE patients were rolled over from previous antihypertensive therapy to the study medication without a washout period to confirm the diagnosis of hypertension and to establish untreated baseline levels for blood pressure [1–3]. In the ancillary VALUE study [4], the ambulatory blood pressure was not even recorded at baseline. The usual way to study blood pressure as an intermediate outcome in parallel-group trials of antihypertensive medications is to compare baseline-adjusted changes describing the transition from no treatment to the full dose of the study medications [7]. The ancillary VALUE study [4] did not allow this state-of-the-art approach. In the main trial, especially systolic blood pressure proved hard to lower [3]. The investigators even launched the systolic blood pressure initiative to improve control rates [3]. By contrast to the predominant role of systolic blood pressure as a predictor of cardiac events in patients with a mean age of approximately 65 years [8], the VALUE researchers chose the 24-h diastolic blood pressure as the main outcome variable driving the sample size calculations to test the equivalence of the randomized first-line drugs in terms of blood pressure lowering. Lederballe-Pedersen et al. [4] noticed that, during the first 7 h after dosing, the ambulatory blood pressure was lower on valsartan compared to amlodipine, whereas amlodipine reduced blood pressure more during the last 4 h of the dosing interval. Valsartan is a highly selective antagonist of angiotensin II at the angiotensin type 1 receptor with an approximately 30 000-fold lower affinity for the angiotensin type 2 receptor [9]. After oral administration, the bioavailability of valsartan is approximately 25%, and decreases further by approximately 40% in the presence of food [9,10]. The elimination half-life varies from 5 to 10 h [9,10]. The usual daily doses range from 80 to 320 mg [9,10]. A mean daily dose of 131.3 mg, as administered in the VALUE ancillary study [4], is not likely to cover the whole day. Amlodipine is a dihydropyridine calcium blocker, which has a bioavailability of 60–65% after oral administration and does not undergo a presystemic or first-pass metabolism [11,12]. After intake, the plasma concentration rises gradually, peaking 6–8 h after administration with a terminal elimination half-life of 40–50 h [11,12]. Patients on chronic treatment with amlodipine at an average daily dose of 7.5 mg, who skipped one dose, maintained blood pressure control for 48 h [13]. The average daily dose of amlodipine in the VALUE substudy was 7.5 mg [4]. The reported differences in the diurnal blood pressure profiles therefore reflect the well-known pharmacokinetic profiles of the study medications in VALUE [9–13]. The VALUE investigators noticed a slightly, but significantly, higher 24-h pulse rate on amlodipine compared to valsartan (72.3 versus 70.5 beats/min), which they attributed to sympathetic stimulation in response to blood pressure lowering with amlodipine. The latter explanation goes against most of the published evidence [14–17]. Indeed, under much better standardized conditions than those used in the ancillary VALUE study [4], the diurnal patterns of plasma noradrenaline and plasma renin activity remained unchanged in patients with essential hypertension 1, 4 and 7 days after the start of amlodipine [14]. Similarly, in hypertensive patients, chronic treatment with amlodipine was associated with lowered [15] or unchanged [16] levels of noradrenaline in plasma [15,16] or urine [15], decreased plasma adrenaline [16], or a reduced low-to-high frequency ratio of heart rate variability in the frequency domain [15]. In patients with heart failure [17], amlodipine at a daily dose of 10 mg did not affect heart rate, plasma renin and aldosterone, plasma noradrenaline, muscle sympathetic activity, or the arterial baroreceptor reflex. Intermittent measurement of pulse rate at the brachial artery is only a crude proxy of sympathetic activity. Because of the generally weaker blood pressure-lowering activity of valsartan compared to amlodipine [2,3], patients randomized to the angiotensin receptor blocker moved faster and more frequently to combination therapy. In the substudy, 30.1% of patients in the valsartan group were on β-blockers, compared to 28.4% in the amlodipine group [4]. During the course of the ancillary VALUE study [4], only 80 composite cardiovascular endpoints occurred, which provided little power to compare the predictive value of various blood pressure measurements. Although not substantiated by any formal statistical testing, the authors proposed that the ambulatory blood pressure and systolic blood pressure conferred a better prediction than the conventional blood pressure and diastolic blood pressure, respectively. The prior probability from other studies in hypertensive cohorts [18–21] in support of the aforementioned conclusions is large. The VALUE investigators also suggested that the night-time blood pressure was more predictive than the daytime level. The multivariate-adjusted hazard ratios for 10-mmHg blood pressure increments were 1.336 versus 1.332 systolic and 1.403 versus 1.218 diastolic, respectively. In the Dublin Outcome Study [18], and in the placebo arm of the Systolic Hypertension in Europe Trial [21], patients were untreated at the time of the ambulatory recording and the night-time blood pressure was a significantly better predictor of outcome than the daytime level. As highlighted by the different diurnal blood pressure profiles on valsartan and amlodipine [4], the VALUE study emphasizes the need to further investigate the extent by which antihypertensive drugs might influence the predictive significance of the night-time versus the daytime level. Antihypertensive drugs are usually taken during daytime and the blood pressure-lowering activity might wear off at night. Controlling blood pressure during the day, but not at night, might contribute to the worse prognosis in non-dippers compared to dippers, or enhance the prediction based on the night-time blood pressure. In conclusion, and as is often the case, subsidiary studies contribute to a better understanding of the results of the main trial but also raise new issues that need to be addressed in future research.