The value of industry-sponsored studies of initial antihypertensive therapies

Despite serious concern,1^,2 James Wright continues to dismiss a body of evidence from actual practice that newer classes of antihypertensive drugs may improve adherence to therapy and therefore blood pressure control.3^,4^,5 We believe that it is he who has “misse[d] the point.”6 The issue here is not that observational data should replace data from clinical trials, but simply that results from real-world studies are also worth considering in the initial choice of antihypertensive therapy.

Wright's summary dismissal of this evidence is based on the contention that the results of observational studies are contradictory and that those showing worse compliance with diuretics reflect the vested interest of the sponsoring companies. The clincher, he avows, is that such studies are irremediably biased anyhow whereas blinded randomized trials are not. His first and third points are clearly incorrect. All studies of adherence in patients with newly diagnosed hypertension have produced similar results, demonstrating greater adherence to angiotensin-converting-enzyme inhibitors.7 The observational studies that have not shown these results8^,9 are about patients with chronic hypertension, that is, patients who already have established prescription therapy; the findings of the latter studies are thus irrelevant in this debate.10 Although observational studies are clearly more prone to confounding, randomized trials are by no means immune to bias. More importantly, randomized trials severely aggravate the Hawthorne effect, and blinding of patients precludes proper study of the question of the relation of compliance to drug characteristics: the method Wright advocates will not answer this question.

Funding by drug manufacturers should always be taken into account when interpreting results, but it does not justify summary dismissal of the findings. Apart from disclosing the funding sources for our study, we reported our methodology in detail: neither our data source (Saskatchewan Health) nor our straightforward methods were in any way influenced by the company that partly funded the study (the company was interested in irbesartan, a drug that was not included in our report).11

Wright's own position on these matters would have been more widely communicated if he had addressed the studies of actual practice, not just presented arguments against their validity. We remain convinced that proper assessment of first-line antihypertensive therapy means looking at all available evidence, including the observational studies. Wright's failure to acknowledge these data is a serious and misleading omission and discourages a deeper understanding of the reasons for our poor performance in managing hypertension.