Response to letter regarding article “Association of Blood Pressure Measurements with Peripheral Arterial Disease Events: A Reanalysis of the ALLHAT Data”

We appreciate the interest of Drs. Messerli and Bangalore in our analysis of blood pressure and peripheral arterial disease (PAD) events using data from ALLHAT¹. They cite concerns that treatment with beta-blockers could exacerbate PAD. Current clinical practice guidelines do not recommend the preferential use (or avoidance) of any particular class of antihypertensive medication to lower blood pressure in adults with hypertension and PAD^{2, 3}, including beta-blockers, but do note the effectiveness of angiotensin converting enzyme inhibitors/angiotensin receptor blockers to reduce cardiovascular events in this population.

There are few clinical trials comparing beta-blockers to other antihypertensive medication classes to examine PAD events specifically. A 2013 meta-analysis of six randomized clinical trials found no evidence that beta-blockers worsened symptoms of PAD such as claudication or maximal walking distance, although the included studies were generally determined to be of low quality⁴. To shed more light on this topic, Drs. Messerli and Bangalore suggest an analysis examining atenolol use with PAD events using data from ALLHAT. In ALLHAT, if participants were not able to achieve the target blood pressure with the assigned randomization drug, then they could be given one or more open-label medications at the treating physician’s discretion. Atenolol, along with clonidine and reserpine, were the second-line open-label medications provided by the study, and 34.4% of the cohort in our analysis had documented use of atenolol during follow-up. The study protocol also allowed the use of other open-label antihypertensive medications, including other beta-blockers aside from atenolol. Only beginning in 1996 (two years after study enrollment began) did the revised study protocol begin to request specification of which classes of open-label medications were used (if any), but this variable is almost entirely missing from the publically available ALLHAT dataset. Given that over 50% of the study cohort had atherosclerotic cardiovascular disease at baseline and that several thousand more participants experienced a coronary heart disease event during follow-up, we posit a high likelihood that many participants may have had undocumented use of beta-blockers other than atenolol. With these concerns about incomplete data, as well as lack of detailed information on specific timing of medication initiation and discontinuation during trial follow-up, we chose not to include medication use during follow-up in our analysis.

Drs. Messerli and Banglaore speculate that future trials will show that beta-blockers should be avoided to prevent PAD events, and they may be correct. But until such trials are conducted, we favor the principle of “life over limb” in patients with PAD, who are at high risk for cardiovascular events, and we therefore support the continued use of beta-blockers when indicated in these patients for the present time.