- Design
- systematic review and meta-analysis of seven cross-sectional studies, random-effects with Hartung-Knapp intervals; GRADE certainty very low
- Population
- 1,359 Indian adults with type 2 diabetes assessed by Asian Working Group for Sarcopenia criteria
- Primary outcome
- pooled prevalence of sarcopenia phenotypes
- Effect
- confirmed sarcopenia 27.4% (95% CI 18.9-37.9), I² 91.7%, prediction interval 6.9-65.6%; severe 11.0%, sarcopenic obesity 25.2%
Seven Indian cross-sectional studies covering 1,359 adults with type 2 diabetes were pooled to estimate how common sarcopenia is, using Asian Working Group for Sarcopenia criteria — which matter here, because European cut-offs misclassify Asian populations.
Pooled confirmed sarcopenia was 27.4% (95% Hartung-Knapp CI 18.9-37.9). Severe sarcopenia was 11.0% and sarcopenic obesity 25.2%. Method changed the answer: 20.2% by dual-energy X-ray absorptiometry against 31.7% by bioelectrical impedance, a difference large enough to matter when comparing figures between clinics. In the two studies with adjusted estimates, hypertension and diabetes duration of ten years or more were positively but imprecisely associated.
The imprecision is the substance of the finding, not a footnote to it. Heterogeneity was 91.7% and the 95% prediction interval — what the prevalence would plausibly be in a new cohort — runs from 6.9% to 65.6%. Certainty was very low by GRADE for every prevalence outcome, and the authors state directly that this is not a sufficient basis for a screening recommendation.
So the change is in clinical attention rather than in a protocol. An older Indian adult with long-standing type 2 diabetes, particularly with hypertension, belongs to a group in which sarcopenia is plausibly common, and the assessment costs almost nothing: grip strength with a dynamometer, a five-times sit-to-stand, and a gait speed over four metres. Where those are abnormal, resistance exercise and adequate protein intake are the response — both of which are useful anyway and neither of which requires the diagnosis to be certain.
What should not happen is a screening programme built on a prevalence estimate this imprecise, or a bioimpedance figure from one clinic being compared with an absorptiometry figure from another.
- Measure grip strength, five-times sit-to-stand and gait speed in older adults with long-standing type 2 diabetes
- Use Asian Working Group cut-offs rather than European ones in Indian patients
- Do not compare a bioimpedance prevalence with an absorptiometry one; they differed by more than 11 percentage points here
- Respond to abnormal results with resistance exercise and adequate protein intake, which are worthwhile regardless
- Do not institute a screening programme on these estimates — certainty was very low and the prediction interval spans 7% to 66%
Why it matters
Muscle function is not assessed at all in most diabetes reviews, and the assessment takes three minutes and no equipment beyond a dynamometer.
Don't overread it
Very low certainty and a prediction interval from 7% to 66% mean the true prevalence is genuinely unknown — the authors themselves rule out a screening recommendation.
The statistics, in plain English
The confidence interval (18.9% to 37.9%) describes uncertainty about the average across these seven studies; the prediction interval (6.9% to 65.6%) describes what you would expect in the next cohort, and it is the number that matters clinically. That it spans almost the whole possible range means the studies disagree so much that no single prevalence figure is usable. I-squared of 91.7% says the same thing. GRADE certainty of very low reflects cross-sectional designs, small samples and inconsistent measurement methods — so the 27.4% headline should be treated as a prompt, not an estimate.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for geriatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free