Prediction models for sarcopenia risk in dialysis patients: a systematic review and critical appraisal.

Background: Many studies have developed or validated predictive models to estimate the risk of sarcopenia in dialysis patients, but the quality of model development and the applicability of the models remain unclear.

Objective: To systematically review and critically evaluate currently available predictive models for sarcopenia in dialysis patients.

Methods: We systematically searched five databases until March 2024. Observational studies that developed or validated predictive models or scoring systems for sarcopenia in dialysis patients were considered eligible. We included studies of adults (≥ 18 years of age) on dialysis and excluded studies that did not validate the predictive model. Data extraction was performed independently by two authors using a standardized data extraction table based on a checklist of key assessments and data extraction for systematic evaluation of predictive modeling research. The quality of the model was assessed using the Predictive Model Risk of Bias Assessment Tool.

Results: Of the 104,454 studies screened, 13 studies described 13 predictive models. The incidence of sarcopenia in dialysis patients ranged from 6.6 to 34.4%. The most commonly used predictors were age and body mass index. In the derivation set, the reported area under the curve or C-statistic is between 0.81 and 0.95. The area under the curve reported by the external validation set is between 0.78 and 0.93. All studies had a high risk of bias, mainly due to poor reporting in the outcome and the analysis domains, and three studies had a high risk of bias in terms of applicability.

Conclusion: Future research should focus on validating and improving existing predictive models or developing new models using rigorous methods.