Phenotyping the shoulder patient based on ultrasound-detected pathologies: a cross-sectional study in general practice

Abstract

Background: Shoulder complaints arise from a single pathology or a combination of different underlying pathologies that are hard to differentiate in general practice. Subgroups of pathologies have been identified on the basis of ultrasound imaging that might affect treatment outcomes.

Objective: Our aim was to validate the existence of different subgroups of patients with shoulder complaints, based on ultrasound-detected pathology, and compare clinical features among them. Profiling shoulder patients into distinct shoulder pathology phenotypes could help designing tailored treatment trials.

Methods: This was a cross-sectional study in general practice. Data were extracted from 840 first visit patient records at a single diagnostic centre in the Netherlands. Exclusion criteria were age <18 years and previous shoulder surgery. Latent class analysis was used to uncover cross-combinations of ultrasound detected pathologies, yielding subgroups of shoulder patients. The uncovered subgroups were compared for demographic and clinical characteristics.

Results: We uncovered four distinct subgroups of patients with shoulder complaints: (i) Frozen shoulder group (11%), (ii) Limited pathology group (44%), (iii) Degenerative pathology group (31%) and (iv) Calcifying tendinopathy group (15%). Group comparisons showed significant differences in demographic and clinical characteristics among subgroups, consistent with the literature.

Conclusion: In a general practice population, we uncovered four different phenotypes of shoulder patients on the basis of ultrasound detected pathology. These phenotypes can be used designing tailored treatment trials in patients with shoulder complaints.

Keywords: cluster analysis; diagnostic imaging; latent class analysis; phenotype; shoulder pain; ultrasonography.