The mortality risk sits in the combination, not in either exposure

Design

National survey cohort with audiometry linked to death records, Cox proportional hazards

Population

874 US adults aged 70 or over with pure tone audiometry, median follow-up 9.3 years

Primary outcome

All-cause mortality by hearing status and comorbidity count

Effect

Hearing loss aHR 1.4 (95% CI 1.0-2.0); 3 or more comorbidities aHR 1.7 (0.9-3.1); both together aHR 2.3 (1.7-3.2), P for interaction 0.03

Eight hundred and seventy-four adults aged 70 or over with pure tone air conduction audiometry in a national US survey were linked to death records through 2020. Over a median 9.3 years, 398 (46%) died. Sixty-five per cent had audiometric hearing loss, and of those only 27% used a hearing aid.

After adjusting for sex, age, marital status and smoking, hearing loss carried an adjusted hazard ratio for death of 1.4 (95% CI 1.0-2.0), and having three or more comorbidities against fewer than three carried 1.7 (95% CI 0.9-3.1). Neither interval convincingly excludes no effect. The combination did: those with both hearing loss and three or more comorbidities had an adjusted hazard ratio of 2.3 (95% CI 1.7-3.2), and the test for interaction was significant at P = 0.03.

The mechanism the authors point to is communication. A patient managing five conditions is managing a schedule of appointments, a changing medication list and instructions delivered verbally, usually in a room with background noise, by someone wearing a mask or facing a screen. Hearing loss degrades every one of those transactions, and the more conditions there are to manage, the more transactions there are to degrade. That is a hypothesis, not something this study establishes - but it is testable, it is cheap to act on, and the 27% hearing aid use in a population where two-thirds have measurable loss says the current answer is not being acted on at all.

• Ask about hearing at review in anyone with multiple conditions, and check whether an existing aid is being worn and working
• Audiometric loss was present in 65% and aids in use in 27% of those - assume under-treatment rather than absence
• Write down instructions for a patient with hearing loss rather than relying on the verbal version
• Do not accept 'poor adherence' or 'seems confused' as a description until hearing has been considered
• Refer for audiology as a chronic disease intervention, not a quality-of-life extra

Why it matters

It reframes hearing loss from a quality-of-life issue into a variable that modifies how dangerous someone's other conditions are.

Don't overread it

Observational - it cannot show that treating hearing loss would improve survival, which the authors say explicitly.

The statistics, in plain English

The two main-effect estimates are both compatible with no effect on their own - hearing loss at 1.0-2.0 just touches 1, and the comorbidity interval of 0.9-3.1 crosses it. Ordinarily that would be the end of the analysis. What makes this interesting is that the interaction term is significant, meaning the two exposures combine to more than the sum of their separate effects, and the combined estimate (1.7-3.2) is comfortably away from 1. With 874 participants and 398 deaths the study is small for this kind of analysis, so treat the interaction as a hypothesis worth testing rather than an established effect. And this is observational: hearing loss may be marking frailty and general ageing rather than causing deaths.

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.