Presence of Diabetes Linked to Greater Likelihood of Frozen Shoulder

New findings from a systematic review and meta-analysis suggest people with both type 1 and type 2 diabetes have a greater likelihood of developing frozen shoulder than those without diabetes.

The six case-control studies in the meta-analysis reported a pooled odds ratio (OR) of 3.69, with investigators noting the value of the pooled estimate was robust to the omission of any individual studies. Two identified cohort studies confirmed the evidence from the case-control studies.

“Given the existing evidence that has been summarized in this review, clinicans should consider checking whether patients with diabetes are experiencing shoulder pain at their routine follow-up appointments,” wrote corresponding author Brett Paul Dyer, PhD, MSc, Primary Care Centre Versus Arthritis, School of Medicine, Keele University.

Prior knowledge indicates the most common comorbidity in people with frozen shoulder is either type 1 or type 2 diabetes, with a much higher prevalence than the general population. It is suggested that diabetes may be a cause of frozen shoulder through glycation processes and/or inflammatory processes leading to scapular fibrosis and subsequent contracture.

A systematic search of MEDLINE, Embase, AMED, PsycINFO, Web of Science Core Collection, CINAHL, Epistemonikos, PEDro, OpenGrey, and the Grey Literature Report was conducted in January 2019 and updated in June 2021. For inclusion in the current analysis, studies were required to have a longitudinal, prospective, or retrospective, observational study design.

Cohort studies were required to have a study population consisting of people without frozen shoulder at inclusion and have established whether diabetes was present at baseline. Meanwhile, case-control studies were required to have a study population consisting of people with frozen shoulder, with diabetes defined as the exposure of interest. Within those providing sufficient data, random-effects meta-analysis was used to derive summary estimates of the association between diabetes and the onset of frozen shoulder.

Eight studies consisting of 346,278 people fulfilled the identified inclusion criteria. Of these studies, six had case-control designs and two had cohort designs. Among the case-control studies, data show the percentage of female cases ranged from 52% to 75% and the mean age for cases ranged from 52.8 years to 57.2 years.

The Quality in Prognosis Studies tool recorded risk of bias scores for each study and indicated a 75% agreement between reviews across the individual bias domains. The reviewers agreed on four of the eight overall risk of bias scores. One of the cohort studies scored as at a moderate risk of bias for the overall study rating, while the other seven studies were rated as being at a high risk of bias overall.

The six case-control studies with 5388 people were pooled in a meta-analysis and the odds of developing frozen shoulder for people with diabetes was 3.69 (95% CI, 2.99 to 4.56) times the odds for those without diabetes.

Meanwhile, the two identified cohort studies used Cox proportional hazards models and recorded results showed that individuals with diabetes had greater risk of developing frozen shoulder.

A cohort study (n = 315,308) with 3-year follow-up reported an age-adjusted, sex-adjusted and dyslipidemia-adjusted HR of 1.32 (95% CI, 1.22 to 1.42). A second cohort study (n = 25,582) with 8-year follow-up adjusted for age, income, stroke, hypertension, hyperlipidemia, obesity, and chronic obstructive pulmonary disease reported an HR of 1.67 (95% CI, 1.46 to 1.91).

Investigators noted that the exisiing research is limited by the high risk of unmeasured confounding.

“To better understand the nature of the relationship between diabetes and the onset of frozen shoulder, it is necessary to have high-quality cohort studies that use causal inference methods that are appropriate for aetiological modeling,” Dyer added.

The study, “Diabetes as a risk factor for the onset of frozen shoulder: a systematic review and meta-analysis,” was published in BMJ Open.