Heat risk depends on how many conditions a patient has, not which ones

Design

matched case-control study using a linked claims and health examination database

Population

21,724 Japanese employees aged 18-74 with heatstroke, matched 1:3 to 65,172 controls

Primary outcome

odds of heatstroke by wet-bulb globe temperature and comorbidity count

Effect

at 33C, OR 2.04 (95% CI 1.84-2.25) with no conditions vs 2.66 (2.40-2.94) with four

A Japanese claims database linked to health examinations matched 21,724 working-age adults with heatstroke to 65,172 controls by age and sex. Exposure was the daily maximum wet-bulb globe temperature; the comorbidity measure was a plain count of eleven behavioural, cardiometabolic and chronic conditions, with no weighting.

The interaction between temperature and condition count was real on both scales. Against a reference of 25C with no conditions, the odds ratio at 33C was 2.04 (95% CI 1.84-2.25) in someone with no conditions and 2.66 (2.40-2.94) in someone with four. On the additive scale the relative excess risk due to interaction was 0.65 (0.45-0.85). Individual conditions contributed modestly: after adjusting for heat-sensitive drugs, smoking, liver dysfunction and obesity stayed positively associated, while the cardiovascular and cerebrovascular associations attenuated.

The finding is the arithmetic, not the list. This is the structure of multimorbidity that general medicine deals with every day and that single-organ guidelines have no way to express - four ordinary conditions in one person behaving differently from four conditions in four people. Heat health warnings are issued by temperature alone, and in an Indian summer the patients whose risk they understate are sitting in every general clinic.

• Count conditions when judging heat risk rather than looking for one qualifying diagnosis.
• Four conditions raised the odds at 33C from about twofold to about 2.7-fold.
• Smoking, liver dysfunction and obesity held their association after adjusting for medication.
• Give heat advice by patient, not by weather threshold - warning systems use temperature alone.
• Ask outdoor workers with two or more chronic conditions about water access and shade at work.

Why it matters

Heat warnings are issued by temperature; this says the same temperature is a different event for the patient carrying four diagnoses.

Don't overread it

Case-control data from claims records - it shows that risk accumulates with comorbidity, not that treating any condition lowers heat risk.

The statistics, in plain English

The interaction term of 1.009 per degree and per extra condition looks negligible until you multiply it out across eight degrees and four conditions, which is how it becomes the difference between a doubling and a near-tripling of odds. This is a case-control design using claims data, so it establishes that risk clusters this way, not that any one exposure caused any one event. Recorded conditions in claims data also undercount the undiagnosed.

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.