Double-blind, placebo-controlled crossover rotation of the five principal classes of antihypertensive drugs

Outcome studies in hypertension show that achieved BP is more important than choice of initial treatment. Meta-analyses report risk ratios of 1.0 for either calcium blockers (CCB) or ACE inhibitors (ACEi) vs. βblockers (βB) or diuretics (D), but 0.8 for more vs. less intensive reduction in BP by any regimen. We previously reported an open-label rotation through four of the main classes, which demonstrated an approximate doubling of patients controlled on monotherapy after rotation vs. their first drug. We have now completed a more formal study, whose objectives were to demonstrate whether there is true inter-individual response to the classes, and to determine any predictors of resistance to monotherapy. We studied 25 men and 9 women, aged 28-55 (median 47), with previously untreated hypertension (BP 160±12/101±5 mmHg after 3 readings over 3 months). They received in random order amlodipine 5 mg (CCB), lisinopril 10 mg (ACEi), bisoprolol 5 mg (βB), bendrofluazide 2.5 mg (D), doxazosin (4 mg) (α-blocker) and placebo once daily for 6 weeks, with the best drug then repeated. BP and arterial stiffness (augmentation index, AI) were measured at each visit. The principal analysis was a McNemar's χ2 test of the number of patients at target (140/90 mmHg) at stages between first and last treatment. These numbers were, respectively, 8, 15, 2, 9 at {first + later}, {later only}, {first only} and {no} phases of rotation (χ2=8.4, p=0.003). Thus rotation increased by half the proportion of patients at target. The resistant patients had the same baseline BP, but stiffer arteries (AI, 28±7% vs. 21±9%). Mean BP fell from 159±10/99±6 (placebo run-in) to 150±14/93±9 at 6 weeks with similar means thereafter. By contrast, the lowest BP of each subject was 144±15/90±9 (after both first and repeat administration) (p<0.0001). The results demonstrate the heterogeneity of pharmacological phenotype in a complex disorder where many drugs with distinct mechanisms are available. A randomly selected drug is unlikely to be the best.