Low activity tracked with sharply higher diabetes odds

Physical inactivity is one of the most commonly cited modifiable risk factors for type 2 diabetes, but most of the supporting evidence comes from research cohorts with careful activity measurement and detailed laboratory work. Everyday general practice looks nothing like that: activity is usually estimated in a brief conversation, and records often lack HbA1c, fasting glucose, dietary intake, diabetes duration and full medication detail. The practical question is whether the crude activity estimate a doctor jots down in a routine consultation still tracks meaningfully with diabetes status. That determines whether asking about activity is worth doing as part of everyday cardiometabolic risk assessment.

This was a cross-sectional observational study of 863 adult patients drawn from a real-world primary care cohort. Participants were sorted into low, moderate and high physical activity groups based on self-reported habitual weekly activity as estimated during routine physician interviews, with the category thresholds chosen to align with public health recommendations. The primary outcome was documented type 2 diabetes. Secondary variables captured from the records included lipid profile, blood pressure, body mass index, waist circumference, smoking status and statin therapy. Multivariable logistic regression compared the odds of type 2 diabetes across activity categories, using the high activity group as the reference.

Type 2 diabetes prevalence showed a graded inverse pattern across the three activity categories: 29.7% in the low activity group, 16.7% in the moderate group and 8.9% in the high group. In the fully adjusted model, low activity was associated with an odds ratio of 4.32 for type 2 diabetes relative to high activity, and moderate activity with an odds ratio of 2.07, both statistically significant. Among the secondary measures, LDL cholesterol differed significantly across groups, with the lowest values in the high activity group. Most other cardiometabolic parameters were comparable between the activity categories.

The direction of the association matches the large body of evidence tying inactivity to type 2 diabetes, and the value here is showing that the signal is visible even with a pragmatic clinical estimate of activity rather than accelerometry or a validated questionnaire. What it cannot do is establish sequence. People with established diabetes and its complications may move less because of their condition, and the authors are explicit that missing HbA1c, glucose, dietary and medication data rule out causal or mechanistic conclusions. The graded pattern across three categories is suggestive but remains a snapshot of one clinic population at one point in time.

This is a reason to think a simple question about weekly activity carries real information about cardiometabolic risk, which is why the authors argue for routine activity assessment in general practice. It is not evidence that becoming more active by a specified amount would cut anyone's odds by the figures reported, because the study measured everyone at a single moment and diabetes itself can reduce activity. The step-down in prevalence between the moderate and high groups is a hint that there may be room to gain above the minimum, but that comparison shares the same limitation. Treat it as a mapping of risk, not a dose-response prescription.