Construction and Validation of a Risk Prediction Model for Statin-Associated Muscle Symptoms (SAMS) in Atorvastatin Users

Abstract

Background

Statin-associated muscle symptoms (SAMS) are a common cause of statin intolerance. This study aimed to develop and validate a risk prediction model for SAMS in atorvastatin users.

Methods

A retrospective study was conducted on patients who received atorvastatin treatment at a single hospital from January 2017 to December 2024. Based on the treatment initiation period and outcome, patients were assigned to a training set (2017–2021) or a test set (2022–2024). Demographic characteristics, medical history, concurrent medications, and laboratory data (creatine kinase [CK], lipid profiles) measured prior to atorvastatin initiation were collected. The primary outcome was the occurrence of SAMS within 1 year, diagnosed according to established criteria. A nomogram prediction model was constructed using independent risk factors identified from multivariate analysis.

Results

Among 576 patients, 170 (29.5%) developed SAMS. In the training set (n = 400), multivariate analysis identified male sex as protective (OR: 0.577), while hypothyroidism (OR: 2.390), renal disease (OR: 1.868), polypharmacy (OR: 2.264), fibrate use (OR: 3.742), and higher CK levels (OR: 1.015) were independent risk factors. The nomogram incorporating these factors showed excellent discrimination, with an area under the curve (AUC) of 0.863. Clinical validation in the test set (n = 176) confirmed the model’s robustness, yielding an AUC of 0.801.

Conclusion

The developed nomogram provides a reliable tool for individualized prediction of SAMS risk in patients initiating atorvastatin therapy.