Genetic Overlap Emerges Between Frozen Shoulder and Type 2 Diabetes
Adhesive capsulitis and type 2 diabetes have long traveled together at the bedside, but this analysis moved the association into the genome. Using GWAS summary statistics from phenotype-specific FinnGen data, broader Pan-UK Biobank data, and cross-trait analytic tools, investigators found a positive genetic correlation between adhesive capsulitis and diabetes in both resources. The signal was statistically robust in each dataset and pointed to shared inherited susceptibility rather than a chance clinical pairing.
The analysis centered on FinnGen R12 M13_ADHCAPSULITIS, defined by ICD-10 M75.0, with 7,372 cases and 357,549 controls, alongside FinnGen T2D and Pan-UK Biobank T2D. Cross-trait LD score regression produced a positive correlation between adhesive capsulitis and diabetes in Pan-UK Biobank, with rg = 0.334, SE = 0.049, and P = 1.4 × 10⁻¹¹. FinnGen showed the same direction of effect, with rg = 0.189, SE = 0.042, and P = 7.0 × 10⁻⁶. In plain clinical terms, the inherited architecture of frozen shoulder appears to overlap with the inherited architecture of diabetes more than would be expected if the two diseases were unrelated.
Partial Sharing, Not a Molecular Shortcut to Causality
The strength of the correlation does not convert into causality. The investigators ran five bivariate MiXeR fits, and those models were compatible with partial overlap, but the shared component varied, and the Akaike information criterion and Bayesian information criterion diverged in the minimum-overlap comparison. In genomics, that kind of disagreement carries weight: it limits the precision of any claim about how much of the biology is truly shared. The study therefore supports overlap, not a fixed estimate of shared variants, shared pathways, or a simple one-to-one mechanistic map.
This distinction matters because adhesive capsulitis is not merely a nuisance diagnosis. In daily practice, it is often the shoulder syndrome that arrives after glycemic disease has already done its systemic work, whether through connective tissue glycation, low-grade inflammation, microvascular injury, or altered fibroblast behavior. A genetic correlation between the two conditions is biologically plausible in that context, but the present data stop short of naming the precise mechanism. The signal tells clinicians and researchers that the connection may begin earlier than symptoms, inside inherited architecture that predisposes both metabolic dysregulation and capsular fibrosis.
Genome-Wide Prioritization Without Gene-Level Certainty
Conditional and conjunctional FDR analyses prioritized regions associated with adhesive capsulitis and regions jointly associated with the two traits. Those statistical methods can help narrow the field, yet they do not establish shared mechanisms, causal variants, or effector genes. The result is a list of genomic neighborhoods that deserve further scrutiny, not a final molecular explanation. Fine-mapping and colocalization remain necessary before any locus can be treated as a credible bridge between shoulder fibrosis and glucose dysregulation.
That restraint is important for clinicians who may be tempted to overread the findings. A genome-wide association signal is not a bedside biomarker, and this paper does not create an individual-level risk calculator for frozen shoulder in people with diabetes. Nor does it justify changing screening, referral, or treatment thresholds on its own. The current value lies in biologic triangulation: the disorder pair is not just clinically associated; it is also genetically connected at the population level.
Why the Diabetes-Shoulder Link Still Makes Biological Sense
Frozen shoulder has always suggested a connective-tissue phenotype, and diabetes offers a plausible systemic driver of that phenotype. Persistent hyperglycemia can alter collagen turnover and promote tissue stiffening, while inflammatory signaling and vascular compromise may further reduce normal capsular remodeling. If inherited variation shapes those pathways, then the overlap detected here could reflect more than coincidence. It could represent a shared predisposition to aberrant tissue response, with diabetes acting as one expression of a broader metabolic-fibrotic susceptibility.
The study also compared broad Pan-UK Biobank M75 with phenotype-specific FinnGen M75.0. Genome-wide correlations were similar, but polygenicity and locus-level results differed. The investigators explicitly caution that the resource, cohort, ascertainment method, GWAS construction, case count, and sample-overlap structure changed at the same time, so the differences cannot be pinned on phenotype breadth alone. In other words, the contrast is informative, but not clean enough to isolate a single driver.
What the Numbers Support, and What They Do Not
The strongest conclusion here is consistency. Across two resources, adhesive capsulitis and type 2 diabetes shared a positive genetic correlation, and the direction of effect held across analytic frameworks. The weaker conclusion is precision. MiXeR could not deliver a stable minimum-overlap estimate, and the downstream FDR work could not prove shared causal alleles or effector genes. The study therefore sits at the level of inherited architecture, not molecular dissection.
That boundary should be preserved. Population-level genetic sharing can explain why the two conditions cluster, but it cannot tell a clinician which patient with shoulder stiffness will later develop diabetes, or which patient with diabetes will progress to adhesive capsulitis. It also cannot distinguish whether the overlap reflects direct pleiotropy, related biological pathways, or hidden influences from measurement and cohort design. Those questions require fine-mapping, functional experiments, and independent validation in other datasets.
A Useful Signal for Future Mechanistic Work
For musculoskeletal and metabolic researchers, the paper creates a sharper starting point. Frozen shoulder has often been treated as a clinical companion of diabetes; this analysis argues that the relationship reaches deeper into inherited biology. The combination of LD score regression, MiXeR, and pleioFDR points toward a shared genetic substrate, but the current evidence is still descriptive rather than explanatory. The next step is not more speculation about association. It is colocalization, gene-level mapping, and functional testing in tissue or cellular systems that can determine whether the overlap converges on fibrosis, glucose handling, or another pathway entirely.
Funding came from the Natural Science Foundation of Jilin Province, and the work used publicly available, de-identified GWAS summary statistics with no competing interests declared. Those features make the analysis straightforward in its ethics and ambitious in its scope. Its real contribution is disciplined restraint: it confirms a genetic connection between adhesive capsulitis and type 2 diabetes while refusing to overclaim what that connection means at the molecular or bedside level.