Article Text

Download PDFPDF
OP165 Identifying sarcopenia from combinations of long-term conditions in UK biobank
  1. Susan J Hillman1,2,
  2. Richard M Dodds1,2,
  3. Antoneta Granic1,2,
  4. Miles D Witham1,2,
  5. Avan A Sayer1,2,
  6. Rachel Cooper1,2
  1. 1AGE Research Group, Translational and Clinical Research Institute, Newcastle University, Newcastle upon Tyne, UK
  2. 2NIHR Biomedical Research Centre, Newcastle University, Newcastle upon Tyne NHS Foundation Trust and Cumbria, Northumberland, Tyne and Wear NHS Foundation Trust, UK


Background Sarcopenia is a skeletal muscle disorder involving accelerated loss of muscle mass and function. Both the presence of two or more long-term conditions, known as multiple long-term conditions (MLTC) or multimorbidity, and specific long-term conditions (LTC), including diabetes and osteoarthritis, are associated with increased risk of sarcopenia. However, it is unclear whether some combinations of LTC are more strongly associated with sarcopenia than others. We aimed to explore this in UK Biobank using decision tree classification techniques.

Methods Our analysis included 140,001 UK Biobank participants (55% women) with two or more self-reported LTC at baseline, taken from a list of 53 conditions identified as those which should always or usually be included in MLTC studies. We used grip strength to classify participants as at ‘risk of sarcopenia’ by applying T-scores of <2 standard deviations below the mean for healthy young adults [<32 kg for men; <19 kg for women]. Decision tree classifiers were then trained on 1000 bootstrap samples of the sex-stratified analytic dataset. We examined the outputs of the decision tree ensembles to identify combinations of conditions most consistently identified with risk of sarcopenia. Odds ratios (OR) of risk of sarcopenia for the highest risk combinations were estimated using logistic regression models, and interactions between LTC explored.

Results The decision tree ensembles identified the six highest risk combinations for sarcopenia risk for men as connective tissue disease with any other condition (OR:2.55, 95% CI:2.33–2.80), osteoarthritis with stroke (OR:2.88, CI:2.34–3.54), connective tissue disease with diabetes (OR:3.41, CI:2.66–4.37), diabetes with stroke (OR:2.68, CI:2.30–3.12), diabetes with uncorrectable visual problems (OR:2.33, CI:2.06–2.64), and drug/alcohol misuse with osteoarthritis (OR:4.50, CI:2.73–7.43). For women the combinations identified were multiple sclerosis (MS) with osteoporosis (OR:3.31, CI:1.89–5.78), connective tissue disease with any other condition (OR:2.86, CI:2.70–3.03), chronic obstructive pulmonary disease (COPD) with stroke (OR:3.35, CI:2.50–4.50), coronary artery disease with osteoarthritis (OR:2.43, CI:2.19–2.71), hypertension with MS (OR:2.32, CI:1.80–2.99), and osteoarthritis with osteoporosis (OR:2.42, CI:2.15–2.71). Amongst these combinations, there was some evidence of multiplicative interactions between: osteoarthritis and stroke, diabetes and stroke, diabetes and uncorrectable visual problems, and drug/alcohol misuse and osteoarthritis in men; and COPD and stroke, and hypertension and MS in women.

Conclusion We found specific combinations of LTC associated with increased risk of sarcopenia, some of which showed evidence of synergistic effects. Knowledge of these combinations could help to identify individuals for targeted interventions, recruit participants to sarcopenia studies, and contribute to understanding the aetiology of sarcopenia.

  • sarcopenia
  • multimorbidity
  • ageing

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.