“A French team of investigators studied the question of whether a protocol-driven strategy of progressive reduction of blood pressure lowering drugs would alter a primary outcome of death rates in elderly frail patients who were residing in a nursing home. The authors also measured many relevant secondary endpoints.
Just over 1000 patients underwent randomization to either the so-called step-down approach or usual care. The hypothesis was that reducing BP meds, and allowing BP to rise a bit would improve mortality.
Inclusion in the trial required patients to be older than 80 years, be on more than one BP-lowering drug and have a systolic blood pressure of ≤ 130 mmHg. [..]
While inclusion criteria mandated age > 80 years, the actual age of patients was 90 years and 80% were female. The average systolic blood pressure was 113 mm Hg and the average number of medications for blood pressure was 2.5-2.6. For an eye-opener, the average number of concomitant meds was 6.6. That’s 9 medications for 90-year-old nursing home patients.
More than 10,000 nursing home patients were screened to enroll 1048 patients in the trial. That’s important, because, obviously, this is a highly selected group.
Using a frailty scale, almost 40% of patients were described as having severe or very severe frailty; only 10% were labeled as “fit, well or managing well.”
Over the follow-up period of just over 3 years, the primary endpoint of death was similar in both groups–61.7% in the step-down group and 60.2% in the usual care arm. (Hazard ratio, 1.02; 95% CI, 0.86 to 1.21; P =0.78).
Death from cardiovascular causes were also similar (53.8% vs 53.3%). All other outcomes were similar, including heart failure events, a composite of major adverse cardiac events and measures of cognition or strength.
Between baseline and the last trial visit, the mean number of BP drugs being used decreased from 2.6 to 1.5 in the step-down group and from 2.5 to 2.0 in the usual-care group. Actual systolic BP changed by only 4 mmHg in the two groups. [..]
The authors showed that the step-down algorithm resulted in fewer medication, only a mildly higher systolic blood pressure (4 mm Hg over 3 years) and no change in any endpoint—including cardiac and neurologic outcomes.
This data supports my often-repeated advice to older patients that the sole purpose of preventive medications, such as BP-lowering drugs, is that we give them to 40- and 50-year-olds so that they live to 80 or 90. Once a person makes it to old age, they can incur more harm than benefit from the drugs.
This trial doesn’t show that patients live longer with less medication, but it does show that less medication has no ill effect on any serious outcome. That’s a positive because when you are that age eating food is far better than eating pills.
Normally a trial that screens 10,000 to enroll 1000 should be cautiously translated to clinic because it is a highly select group of patients. But I think translation here can be generalized to a wide swath of older frail patients. Why? Because… a) trials are hard, and we are unlikely to see many trials like RETREAT-FRAIL. And b) because common sense still applies to translation of medical evidence.
Common sense suggests that the benefit-harm ratio of preventive medications changes in older patients—as this group has less life-years to gain, more competing risks of morbidity and mortality, lower drug clearance and less physiologic reserve.
Consider also that using preventive meds in 90-year-olds in the first place is outside of evidence-based practice as I know of few if any “positive” trials that enroll patients this old or frail.
Wisdom also holds that we don’t want older patients to have stroke from severe hypertension, but, in most cases, we can combine the results RETREAT-FRAIL with common sense to reduce pill burden in these patients.
In my opinion, reducing the work of being a patient is one the most valuable things a clinician can do for older patients with frailty.”
Full post, J Mandrola, Sensible Medicine Substack, 2025.9.8