Background: Visit-to-visit blood pressure (BP) variability limits the ability to detect therapeutic effects in hypertension trials.
Methods: To enable future renal denervation trials to detect smaller effect sizes and reliably identify technical improvements, we examined within-patient visit-to-visit BP variability, quantified as SD of change from baseline to final BP (SDΔ), in renal denervation (RDN) trials, trials of BP-lowering tablets, and the VOLTAGE study including 4151 patients.
Results: The control arms of RDN trials had more visit-to-visit BP variability than tablet trials (SDΔ 23.6 versus 13.5 mmHg; P < 0.001). This might be explained by more prescribed antihypertensive patients in the RDN trials (5.19 ± 0.13 versus 0.11 ± 0.11; P < 0.001). In the VOLTAGE study, as the number of medications prescribed rose from 0 to 4, SDΔ rose: 11.9, 11.2, 12.9, 14.4 and 18.0 mmHg (P < 0.001 for trend). Neither baseline BP, nor demographics, nor diabetes independently affected variability. The sample size required for a trial rises proportionally to the square of the number of medications prescribed (rather than just linearly). The relationship between the number of background medications prescribed in a cohort and the excess test-retest variance closely fitted this quadratic formula (R = 0.98, P = 0.001).
Conclusion: Visit-to-visit variability in BP is dramatically larger in patients with more background medications prescribed. If this is due to variable adherence, then future RDN trials, needing to detect smaller effect sizes, would benefit from measures to guarantee adherence. Conceivable measures include enrolling patients on no background medication, preceding each BP measurement with a period off medication, or directly supervising medication intake.