Sarcopenic obesity outranks obesity alone for diabetes risk

Design

prospective cohort study with Cox models and a landmark analysis of phenotype transitions

Population

479,607 diabetes-free UK Biobank participants; landmark analysis in 53,107

Primary outcome

incident type 2 diabetes over a median 14.2 years

Effect

32,948 cases; sarcopenic obesity HR 3.54 (95% CI 3.34-3.74); transition to it HR 2.90 (2.07-4.07); persistent HR 3.07 (1.63-5.79)

Body mass index says nothing about what the weight is made of. This analysis of 479,607 diabetes-free UK Biobank participants classified body composition phenotypes using handgrip strength, skeletal muscle mass as a proportion of weight, and fat mass percentage, then followed them for a median of 14.2 years, during which 32,948 developed type 2 diabetes.

Sarcopenic obesity carried the highest risk, with a hazard ratio of 3.54 (95% CI 3.34-3.74) — higher than obesity alone and higher than sarcopenia alone. A landmark analysis in 53,107 participants added the part that matters clinically: transitioning into sarcopenic obesity during follow-up carried a hazard ratio of 2.90 (2.07-4.07), and remaining in it 3.07 (1.63-5.79). The phenotype is not simply a marker of who was always going to develop diabetes.

This is observational, and UK Biobank volunteers are healthier and less diverse than the population, which usually attenuates rather than inflates associations. Handgrip strength is the practical takeaway: it is the cheapest of the three measures, needs a dynamometer and thirty seconds, and identifies the person whose risk is not visible on the scales. In Indian practice, where sarcopenic obesity is common at lower body mass indices and central adiposity dominates, a normal BMI is even less reassuring than it is in this cohort.

• Measure grip strength alongside weight and waist in anyone being assessed for diabetes risk
• A normal BMI with low muscle mass is not a low-risk phenotype, particularly in South Asian patients
• Ask about resistance exercise, not only aerobic activity, when advising on prevention
• Watch for muscle loss during weight loss, especially on incretin therapy in older patients
• Ensure adequate protein intake in older adults advised to lose weight

The statistics, in plain English

A hazard ratio of 3.54 with an interval from 3.34 to 3.74 is precisely estimated because the cohort is very large — but precision is not the same as freedom from confounding, and people with sarcopenic obesity differ in age, activity, illness and smoking in ways adjustment handles imperfectly. The landmark analysis is the stronger part of the design: by requiring the phenotype change to precede the diagnosis, it reduces the chance that early undiagnosed diabetes caused the muscle loss rather than the reverse. Its intervals are much wider (2.07-4.07, 1.63-5.79) because far fewer people had repeat measurements.

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.