The Hyperechoic Appearance of the Deltoid Muscle on Shoulder Ultrasound Imaging as a Predictor of Diabetes and Prediabetes

Document Type


Publication Date


Publication Title

J ultrasound med


OBJECTIVES: To evaluate whether the ultrasound appearance of the deltoid muscle in diabetic patients differs from that in obese nondiabetic patients.

METHODS: Ultrasound images of the deltoid muscle from 137 type 2 diabetic patients (including 13 prediabetic patients) and 49 obese nondiabetic patients were blindly reviewed by 2 musculoskeletal radiologists, and by a third when arbitration was needed, to determine whether the appearance was "normal," "suspected diabetes," or "definite diabetes." Age, sex, race, body mass index (BMI), insulin use, and hemoglobin A1c were analyzed. This retrospective study included patients presenting between October 2005 and November 2017. Statistical analyses included a 2-sided sample t test or Wilcoxon rank sum test and a chi(2) or Fisher exact test. Statistical significance was defined as P < .05.

RESULTS: The type 2 diabetic patients included 98 women and 39 men aged 29 to 92 years, and the nondiabetic patients included 19 women and 30 men aged 18 to 75 years. A consensus diagnosis of definite diabetes by the musculoskeletal radiologists based on a hyperechoic deltoid was a powerful predictor of diabetes, with a positive predictive value of 89%. A hyperechoic deltoid was also a powerful predictor of prediabetes. Of the 13 prediabetic patients, all had the same hyperechoic appearance of the diabetic deltoid, regardless of BMI. Although obese diabetic patients more often had a diagnosis of definite diabetes, the BMI alone could not explain the increased echogenicity, as obese nondiabetic patients' deltoid muscles did not appear as hyperechoic and were correctly categorized as not having definite diabetes with 82% specificity.

CONCLUSIONS: The characteristic hyperechoic deltoid appearance is a strong predictor of both diabetes and prediabetes and differs from that of obese nondiabetic patients.

PubMed ID



ePub ahead of print