Compared with usual care, stepping down antihypertensive therapy in nursing home patients aged 80 years and older did not yield any reduction in major adverse events or correlate with any other clinical benefit in the multicenter, randomized RETREAT-FRAIL trial.

These findings conflict with prior observational studies predicting fewer adverse events and perhaps even a mortality benefit from de-intensifying blood pressure-lowering therapies in frail older adults, according to Athanase Benetos, MD, PhD, the trial’s principal investigator and a professor of geriatric medicine at the University Hospital of Nancy-Lorraine in France. 

The hypothesis of the study anticipated a 25% reduction in all-cause mortality in the experimental arm. Instead, the rate of death over a median of 38.4 months of follow-up was numerically, although not significantly, higher in the experimental arm, Benetos reported.

The late-breaking trial was presented at the 2025 annual congress of the European Society of Cardiology, and the results were published simultaneously in The New England Journal of Medicine.

No Clear Benefit

In the study, 1048 nursing home residents aged 80 years and older with blood pressure below 130 mm Hg who were being treated with at least two antihypertensive therapies were randomly assigned to a protocol-driven strategy of step-down therapy or usual care. The primary endpoint was all-cause death. 

At the end of follow-up, 61.7% of the step-down group and 60.2% of the usual care group had died (P = .78). 

During follow-up, the rate of a composite of major adverse cardiovascular events in the step-down group was slightly higher (19.3% vs. 17.3%), as was the rate of noncardiovascular death (53.8% vs. 53.5%). Neither difference approached significance. Fifty percent of patients in each group experienced a fall over the course of follow-up, although the fracture rate was nonsignificantly lower in the step-down group (7.8% vs. 9.2%).

The median number of antihypertensive therapies fell from 2.6 to 1.5 in the experimental arm over the course of the trial and from 2.5 to 2 in the usual care arm. Systolic and diastolic blood pressure increased by 4.1 and 1.8 mm Hg, respectively, among those in the step-down arm relative to the usual care arm.

There were no significant between-group differences in an array of adverse events monitored over the study, including relative changes on standardized tests measuring cognition, balance, activities of daily living, or quality of life, Benetos reported.

A large body of evidence suggests the cardiovascular benefit of antihypertensive therapy does not decline with age in the absence of frailty, Benetos said, but he cited several studies that have questioned the safety and efficacy of combination antihypertensive drugs in frail patients. The 2015 PARTAGE study, which Benetos led, is one example. 

In the nonrandomized PARTAGE study, nursing home patients older than 80 years with controlled systolic blood pressure (defined as ≤ 130 mm Hg) taking two or more antihypertensive medications were compared with those taking one or none. 

Over 2 years of follow-up, two or more blood pressure-lowering medications compared with one or fewer was associated with a nearly twofold increase in mortality (P < .001). The difference persisted over all three of the sensitivity analyses conducted. 

In more recent randomized trials, such as OPTIMISE, which compared reduction to maintenance of multiple antihypertensive therapies in older adults, and DANTON, which compared elimination of antihypertensive therapies in the older adults with dementia, the value of step-down therapy has been more difficult to interpret. Both found that step-down therapy could be performed without loss of blood pressure control, but neither associated the reduction in blood pressure medications with improved outcome.

On the basis of observational data suggesting potential harm from intensive antihypertensive regimens in frail older adults, the 2024 ESC hypertension guidelines suggest individualization of therapy in this population. Specifically, a single-drug antihypertensive regimen was called “reasonable” due to a “less certain” benefit-to-risk ratio of more intensive therapy in this population. 

Guidelines: Drug Personalization in Frailty

RETREAT-FRAIL has now generated level 1 evidence with which to judge both the efficacy and safety of antihypertensive drugs in patients age 80 years or older who are frail as defined by the need for nursing home care, Benetos said. Although he acknowledged the reduction in antihypertensive medications in patients who received usual care was unanticipated, possibly diluting differences between the arms, he said the results are generalizable given the trial’s hard primary endpoint of all-cause death adjudicated by independent observers unaware of group assignment.

What RETREAT-FRAIL shows is that just lowering the number of antihypertensive drugs in the frail older adults has no specific outcome advantages, as suggested by prior observational studies, according to Eugene Yang, MD, a professor of medicine at the University of Washington in Seattle.

“There was hope that step-down therapy might actually reduce mortality,” said Yang, the immediate-past chair of the American College of Cardiology Prevention of Cardiovascular Disease Council. RETREAT-FRAIL did not show such an effect, but the trial is noteworthy because its findings fill a knowledge gap, he said.

“My takeaway is that reducing blood pressure medication in old and frail patients has no significant impact on outcomes, whether in terms of safety or morbidity and mortality,” Yang said. However, he noted it also did not cause additional harm, suggesting individualized therapy, including a more streamlined number of therapies, might still be a reasonable choice in patients preferring a simplified regimen.

“This aligns well with current ESC and European Society of Hypertension recommendations,” he said.

Benetos reports no relevant financial relationships. Yang reports financial relationships with Chroma, Genentech, Idorsia, Measure Labs, Microsoft, Mineralys, Qure, Sky Labs and TenPoint7.