Statins and a falling frailty index, with nothing to show for it on the walking test

Design

Observational longitudinal clinic cohort with linear mixed models, statin users matched to non-users for follow-up duration

Population

3,444 people with HIV, 69% men, median age 44.7 years, with at least three frailty index assessments

Primary outcome

Frailty index trajectory after statin initiation, with walking speed, physical performance battery and grip strength as secondary outcomes

Effect

Statin-by-time interaction coefficient -2.562 (95% CI -3.176 to -1.987); no association with any secondary physical performance outcome

Three thousand four hundred and forty-four people with HIV attending a metabolic clinic in Italy, none on a statin at baseline and all with at least three frailty index assessments, were followed after statin initiation. Sixty-nine per cent were men and median age was 44.7 years; 23.1% started a statin for primary prevention and 2.0% for secondary. Frailty was measured on a 34-item index from which lipid variables were deliberately excluded, to avoid the index moving simply because the statin changed the lipids.

In adjusted linear mixed models, statin use was associated with a falling frailty index over time - a statin-by-time interaction coefficient of minus 2.562 (95% CI minus 3.176 to minus 1.987), with estimates scaled by 100 for readability. Statin users had a worse cardiometabolic profile at the point of starting.

And the secondary outcomes showed nothing. Walking speed, short physical performance battery and hand grip strength were all unchanged. That divergence is the part to weigh. A frailty index counts accumulated deficits, many of them diagnoses and abnormal results, and a drug that prevents or improves several of those will move the index without the patient being any less frail in the way a clinician would recognise across the room. The objective physical measures are the ones that would have shown that, and they did not move.

• Prescribe statins in people with HIV on cardiovascular grounds, which is where the evidence already sits
• Do not offer frailty prevention as an additional reason to start one on this evidence
• Note that the physical performance measures - gait speed, performance battery, grip - showed no difference
• Median age 44.7 means this is a middle-aged cohort, not an older one
• A falling frailty index without a change in physical function is a measurement question, not a clinical result

Why it matters

It tests whether a frailty index that moves means a patient who has actually become less frail, and the physical measures say not necessarily.

Don't overread it

Observational, in a middle-aged cohort, with no change in any objective physical performance measure.

The statistics, in plain English

The interaction coefficient describes how the slope of the frailty index over time differs between statin users and non-users, and the confidence interval excludes zero comfortably. But this is observational, and statin users were not like non-users - they had worse cardiometabolic profiles at initiation, which is why they were started. Adjustment handles measured differences and not unmeasured ones, and the direction of the residual confounding here is genuinely hard to predict: sicker at baseline argues one way, while being well enough engaged in care to start preventive medication argues the other. The null results on gait speed and grip strength are the most interpretable finding, because those measures are not built out of diagnostic codes.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.