Minoxidil versus placebo in the treatment of arterial wall hypertrophy in children with Williams Beuren Syndrome: a randomized controlled trial

Patients with proven diagnosis of Williams Beuren syndrome by genetic testing, with normal or high blood pressure, treated with antihypertensive agent(s) or not, male or female, aged over 6 and under 18-year, childbearing potential female negative for pregnancy test, or accepting an effective birth control for sexually active female were eligible. Participants with pulmonary hypertension secondary to mitral stenosis, history of myocardial infarction within one-month prior randomization, or with known allergies to minoxidil or any of its components, history of asthma, renal failure (creatinine clearance <40ml/min/1.73m², Schwartz formula), intolerant to lactose, receiving vasodilator anti-hypertensive agents, were excluded. The study protocol was approved by the ethics committee “Comité de Protection des Personnes dans la Recherche Biomédicale Sud Est II” (file number 2008-005-AM7). All subjects and their parents provided written informed consent before enrolment.

Participants were randomized to receive hard capsules containing 2.5, 5, 7.5 or 10 mg of minoxidil or a matched placebo. Hard capsules (size 4; volume 0.21 mL) were filled manually after compounding minoxidil and lactose as excipient. Uniformity of drug content and mass was confirmed in 20 individual hard capsules (batch: 100 hard capsules) before release. Placebo hard capsules (size 4; volume 0.21 mL) were filled with lactose.

Adherence to treatment was calculated as the proportion of the capsules taken among the capsules necessary for the trial. The number of capsules taken was calculated from the number of capsules delivered to participants after deducting the number of capsules returned after twelve-month follow-up. Minoxidil was prescribed following the summary of product characteristics for treating high blood pressure in children. Arterial hypertension was defined following the André curve (Andre J, Deschamps J, Gueguen R. La tension artérielle chez l'enfant et l'adolescent. Valeurs rapportées à l'âge et à la taille chez … Arch Fr Pediatr. 1980). Normotensive children, 12 years old and younger, received the minimum usual dose of 0.2 mg/kg per day. Hypertensive children, 12 years old and younger received the minimum usual dose of 0.2mg/kg per day, then doses were increased by 0.1 mg/kg/day every three days when necessary to reach the maximum dose of 1 mg/kg per day. Normotensive children over 12 years old received 5mg per day. Hypertensive children over 12 years old received the minimum usual dose of 0.2mg/kg per day, and then doses were increased to reach the maximum of 40mg per day. The dosing was decreased in case of adverse events. Dosing of previous antihypertensive treatments were adapted following the blood pressure level. Other calcium blockers were not allowed during the study.

The primary outcome was the difference in IMT from the beginning to the end of 12-month treatment period on the right common carotid artery (CCA). IMT measurements were obtained with Sequoia™ 512 ultrasound (ACUSON, Siemens) with a 7.5 to 8 MHz probe, 10 mm before carotid bifurcation, 1 cm below the carotid bulb with the patient lying in the supine position and with the neck rotated to the opposite side of examination. Acquisitions were done on B-mode followed by Time Motion (TM) mode twice among three different possible angle views, i.e., transversal, anterolateral and posterolateral position of the probe. On B-mode, a longitudinal scanning view in the longest extension was performed simultaneous to a one derivation electrocardiogram (ECG) in order to synchronize images on diastole. TM mode should provide a clear image of the artery layers with a view of the CCA on five cardiac cycles with a minimum depth of 4 cm and a constant zoom of 10 to 20 mm allowing measurement of end-systolic and end-diastolic diameters. All images were recorded on DMO and CD Rom and read by two radiologists unaware of the allocated treatment group, with a specific software. Each member of the committee measured all IMTs separately. When the discrepancy between the two assessors was more than 10% for the same position of the probe, a common measurement of IMT was organized. A third assessor adjudicated the IMT when committee members failed to reach consensus.

Secondary outcomes were to assess the efficacy of minoxidil on:

Harms related information were collected at each visit using the case report form. Site monitoring was performed to make sure that all adverse events have been reported appropriately during the trial. MedDRA coding was used to report harms. Hypertrichosis, oedema, sinus tachycardia, and pericarditis were among expected adverse events.

Based on the study of Aggoun et al. [46] comparing the IMT in 21 children with WBS (0.6 ± 0.07 mm) to 21 controls matched on age (0.5 ± 0.03 mm), 23 patients should be enrolled in each group to show an improvement of 0.1 mm in the IMT after twelve month treatment with minoxidil with a bilateral alpha risk of 0.05 and a power of 90%. Randomization sequence were generated by the department of biostatistics of Hospices Civils de Lyon and concealed through a web-based platform. Centres were stratified into two groups i) Bordeaux and Lyon, ii) Angers, Lille and Poitiers.

Each investigator should call the coordination centre or connect to the web site in order to randomize patients after checking for eligibility criteria.

Frequency of harmful events was reported in each arm.

Minoxidil and placebo had the same aspects and taste. All participants, investigators and the coordination centre were blind to the allocation group. Assessors of the principal outcome were also blind to the allocated group. The anti-poison centre of Hospices Civils de Lyon hold the allocation list in order to verify the justification of non-blinding inquiries and to guide investigators. The success of blinding was not assessed.

An independent data monitoring committee was set up to supervise the conduct of the study and ensure participant’s safety. No interim analysis of efficacy was planned.

IMT of the right primitive carotid artery was measured at inclusion and after 12 and 18 months by two radiologists at two probe positions. The Bland and Altman method was used to assess the agreement between the two radiologists. Before the analysis, we verified that the measures were performed on the same probe positions between visits. All the measurements, three visits, 2 probe positions by visit, 2 to 4 readings by probe position, allowed us to use modeling techniques to gain power and precision. Thus, we used a linear mixed model with a random intercept to take into consideration the intra-patient correlation between repeated measurements for estimating the effect size. IMT was the dependent variable; the visits, the centers’ strata, the treatment, and the interactions between the treatment and the visits were the independent variables. The estimation method was the restricted maximum likelihood (REML). The underlying hypotheses of the mixed model were checked with graphical methods (using the mixed procedure of SAS). The Wald test was used to test whether the coefficient of the interaction between the treatment and the follow-up visit at 12 months was statistically significant. The result was checked with the likelihood ratio (after re-estimation of nested models by the maximum likelihood method). In a sensitivity analysis, the position of the 3 possible position of the probe and the quality of the measurement (bad versus good) were introduced separately as fixed effects into the model to measure their influence on the estimated effect of minoxidil on IMT. The influence of age and systolic blood pressure as fixed effects for continuous variable was also explored for adjustment. The efficacy analysis was based on intention to treat principle.

IMT of the right humeral artery was measured once by visit (only one probe position). For all other measurements on arteries, we measured and analyzed results as described for the IMT. The maintenance of the treatment effect after 6 months was tested between groups using the Wald statistics on the coefficient of the interaction between the treatment and the visit at 18 months.

All confidence intervals are bilateral, at 95%. Continuous variable were summarized by mean and standard deviation (SD), and Wilcoxon test statistics was used to compare them between groups.

From 10 March 2009 to 18 February 2014, from a total of 64 eligible patients, 21 were finally randomized, twelve in the placebo and nine in the minoxidil group. One patient stopped the treatment prematurely 16 days after randomization because of the occurrence of an adverse event four days after starting the treatment and stopped participating in the study after the first visit at month three. On the fourth of April 2014, the independent data monitoring committee advised to stop the trial, since the probability for the study to enroll 25 more participants in an acceptable time seemed unachievable. Finally, twenty participants attended all follow-up visits.

The mean age was 12.3 years (SD = 4.4 yrs) in the minoxidil and 10.8 years (SD = 3.8 yrs) in the placebo group. The mean BMI was 17.9 Kg/m² (SD = 3.9 Kg/m²) in the minoxidil, and 17.5 Kg/m² (SD = 3.3 Kg/m²) in the placebo group. Ten participants, four in the minoxidil and 6 in the placebo group had a previous history of cardiovascular diseases, and four were treated for hypertension (2 received ACE inhibitors in the minoxidil and 2 received Beta blockers in the minoxidil group). Ten patients had cardiovascular diseases, 4 in the minoxidil and 6 in the placebo group. In the minoxidil group 2 patients had supravalvular aortic stenosis, 1 pulmonary valvular stenosis, and 1 hypertrophic cardiomyopathy. In the placebo groups three had supravalvular stenosis, one patient had aortic coarctation, 1 coarctation de l’aorte, 1 mitral leak, and 1 stenosis of pulmonary artery.

Participants’ characteristics at baseline are presented in Table 1. Mean age difference between minoxidil and placebo groups (1.6 (95% CI -2.1, 5.2), p = 0.38) was not statistically different. Cardiac ultrasound showed SVAS in three participants in the minoxidil group and one in the placebo group, and three ventricular hypertrophies in the minoxidil group and two in the placebo group, which could be related to a cardiopathy independently of high blood pressure. Renal Doppler detected two stenosis in the placebo group and one in the minoxidil group, and no clinically significant deviations were observed on biological parameters.

The adherence to treatment was 99.6% (SD = 5.7 %) in the minoxidil and 98.7% (SD = 11.0%) in the placebo group. The principal outcome was available for 9 patients in the placebo group and 8 patients in the minoxidil group (Fig. 1). Data for three patients were not analysed, because two had uninterpretable scan views of IMTs and one was lost from follow-up. A total number of 384 measurements of IMT were available for the analysis at 12 and 18 months.

After 12-month treatment, the IMT in the minoxidil group increased in CCA by 0.03 mm (95 % CI -0.002, 0.06) compared with 0.01 mm (95 % CI -0.02, 0.04 mm) in the placebo group (p = 0.4), difference between the groups was 0.02 mm (95 % CI -0.02, 0.06 mm) (Fig. 2).

After 18 months, the IMT increased in CCA by 0.06 mm (95% CI, 0.02 , 0.10mm) more in the minoxidil group compared with the placebo group (p = 0.008). The IMT of the right humeral artery dropped at 12 months in both groups, difference between the minoxidil and the placebo group was 0.03 mm (95 % CI -0.04, 0.09 mm) at 12 months (p = 0.4), and 0.07 mm (95 % CI 0.01, 0.14 mm) at 18 months (p = 0.04).

The position of the probe, the quality of the measurement, the age at inclusion and the systolic blood pressure did not have any statistically significant effect on the IMT variation at 12 and 18 months. The estimated effect of minoxidil on IMT variation remained unchanged after adjustment (Additional file 1: Table S2).

The lumenal diameter of the CCA adjusted for the time of the cardiac cycle increased more in the minoxidil group (difference 0.36 mm, 95% CI, 0.16, 0.56 mm; p = 0.0006). This effect persisted (0.26 mm (95 % CI, 0.05, 0.46 mm), p= 0.013) 6 months after the end of the treatment. The distensibility of the CCA increased in both group, difference between groups was -1.9 % (95 % CI -6.4, 2.7 %, p=0.4) at 12 months. At 18-month visit the distensibility increased more in the placebo group, difference between groups was -6.1 % (95 % CI -10.6, -1.6 %, p=0.008).

The diameter of ascending aorta increased by 1.89 in the minoxidil and 3.42 in the placebo group (p = 0.81). The diameter of the humeral artery adjusted for the time of the cardiac cycle increased by 0.25 mm (95% CI, 0.02, 0.47 mm) more (p = 0.03) in the minoxidil group compared with the placebo group. This effect persisted (0.32 mm (95 % CI, 0.08, 0.55 mm), p= 0.008) after 18 months. The distensibility of the humeral artery increased in the placebo group, and decreased in the minoxidil group, the difference between groups was -2.9% (95 % CI -12.8, 7.0 %, p=0.6). At 18-month visit the distensibility decreased in both group, the difference between groups was 3.5 % (95 % CI -6.8, 13.8 %, p=0.5).

Pulse wave velocity measurement variation, available for 11 patients after 12 months, were -0.8 m/s (SD = 2.0 m/s) in the minoxidil group and -1.5 m/s (SD = 3 m/s) in the placebo group (p = 0.7).

After 12 months, the 24-H mean SBP increased by 3.6 mmHg in the minoxidil group and by 0.8 mmHg in the placebo group (p = 0.5), and the 24-H mean DBP increased by 1.6 mmHg in the minoxidil and by 0.9 mm Hg in the placebo group (p = 0.6). The variations of CCA, humeral artery, PWV and blood pressure at 12-month visit are presented in Table 2. The variations of CCA and humeral artery at 18-month visit are presented in Table 3.

Finally, 2 patients in the minoxidil group and 1 in the placebo group still presented a SVAS at 12 months. One patient with SVAS stopped the trial early and was lost from follow-up. No renal artery stenosis was detected at the end of trial for two patients, one in the minoxidil and on in the placebo group who completed the trial.

A total of 47 adverse events were categorized with the use of the Medical Dictionary for Regulatory Activities classification, 3 severe, 12 moderate, and 32 mild occurred in the minoxidil group and 68, 1 severe, 15 moderate, and 52 mild in the placebo group. As it was expected, hypertrichosis occurred only in participants of the minoxidil group, and was reversible after the end of the treatment. The details on adverse events are reported in Table 4 and in the Additional file 1: Table S1.