Osteoarthritis: four studies, four answers

One found that rising pressure hurt. One found nothing. A third found the opposite. The fourth found a link and dismissed it as irrelevant itself.

Four solid studies. Four different answers. None of them got the method wrong.

McAlindon and colleagues followed 200 patients with knee osteoarthritis and found a clear link: when air pressure rose, it hurt more (coefficient 1.14, p = 0.02), and when the outside temperature fell, it also hurt more (coefficient −0.01, p = 0.004). Consistent, statistically solid — and the kind of link you could find again and again in the same patients.

Then another study found nothing

Nine years later, Ferreira and colleagues set out to test exactly the same question with a different method: a case-crossover design, in which each patient's worsened pain was weighed against that same patient's own weather days. They found nothing. No visible link between temperature, humidity, air pressure or rainfall and the risk of a knee osteoarthritis patient having a bad day.

The authors' own conclusion was blunt: despite patients' accounts, weather change does not appear to affect the risk of a pain flare.

A systematic review and meta-analysis from 2023 gathered a long list of such studies and found: air pressure and relative humidity were positively correlated with pain intensity in osteoarthritis, while temperature was negatively correlated. A third answer — and one that fits better with McAlindon than with Ferreira.

The fourth study, which barely counts

The fourth study is probably the most honest of them all. Dorleijn and colleagues followed 222 patients with hip osteoarthritis for two years, with measurements every three months. They found a link: humidity was associated with patients' pain scores, air pressure with their function scores. Both were statistically solid.

And together, both explained at most 1% of the variation in symptoms — both within the individual patient over time and between patients.

The results confirmed patients' experience — and explained under one percent of it.

Four answers, the same question

McAlindon found an effect. Ferreira found none. The meta-analysis found a signal across studies. Dorleijn found an effect so small it doesn't count clinically. None of them got their method wrong. They just asked the question four different ways, of four different groups of patients, and got four different answers.

That's not an argument that weather and osteoarthritis aren't linked. It's an argument that the average of a thousand patients can never answer whether it applies to you.