Hyperthyroidism is a risk factor for developing adhesive capsulitis of the shoulder: a nationwide longitudinal population-based study.

Conceived and designed the experiments: S.W.H., J.W.L. Performed the experiments: S.W.H., H.W.L. Analyzed the data: H.W.L., W.T.W. Prepare Tables and Figure: T.H.L., C.W.W. Wrote the paper: S.W.H. All authors reviewed the manuscript.

We analyzed records that were obtained from the Longitudinal Health Insurance Database 2005 (LHID2005), which is maintained by the Taiwan National Health Insurance (NHI) Institutes. The LHID contains data of all health insurance claims, including demographic, inpatient care, ambulatory care, and prescription drug data, as well as International Classification of Diseases, Ninth Revision, Clinical Modification diagnostic codes. Overall, the records of 1 000 000 beneficiaries, who were enrolled in 2005, were randomly sampled from the records of the 25.56 million people contained in this database (the NHI program insures approximately 99% of the population in Taiwan). To protect the privacy of personal data, the records obtained from the database were de-identified. In addition, informed consent was waived in this secondary data analysis study. The study cohort contained all patients who had been diagnosed with a hyperthyroid disease (ICD-9-CM codes 242.9) between January 1, 2004 and December 31, 2007 according to data on ambulatory medical care visits. The hyperthyroid group consisted of patients who had received a principal diagnosis of hyperthyroidism (ICD-9-CM codes 242.9) during an ambulatory medical care visit between January 1, 2004 and December 31, 2007. To improve the diagnosis accuracy, only patients who had at least 2 consecutive ambulatory visits in which the principal diagnosis was hyperthyroidism were recruited in this study (N = 4779). Patients who were diagnosed with adhesive capsulitis (ICD-9-CM codes 726.0) prior to a thyroid disorder diagnosis (N = 57, hyperthyroidism) or multiple diagnosed with other thyroid disease such as thyroid cancer or hypothyroidism (N = 85) or whose records were missing variables, such as date of birth, sex, and age < 18 years (N = 165) were excluded. A total of 4472 hyperthyroid patients met the inclusion and exclusion criteria and were enrolled in this study. Patients in the control group were matched with those in the study cohort (5 control patients per case patient) according to age (70 y) and sex by using the remaining patient records obtained from the LHID. Patients were excluded if they had been diagnosed with a thyroid disorder between 2004 and 2010, or had been diagnosed with adhesive capsulitis before 2004. We obtained baseline variables, including age, sex, urbanization level (stratified into 3 levels: urban, suburban and rural), and confounding factors, such as diabetes mellitus (DM; ICD-9-CM code 250), hypertension (ICD-9-CM codes 401–405), hyperlipidemia (ICD-9-CM codes 272.0–272.4), autoimmune diseases (rheumatoid arthritis: ICD-9-CM code 714.0; SLE ICD-9-CM code 710.0), chronic liver disease (ICD-9-CM code 571), cancer (ICD-9-CM codes 140-208), and coronary heart disease (ICD-9-CM codes 410–414), for all patients. We used the initial diagnosis of adhesive capsulitis (ICD-9-CM codes 726.0) as the study endpoint. All participants were followed from the index date until the occurrence of the endpoint or until December 31, 2010, whichever was first, and final-date observations were censored observations. The Pearson chi-squared test or the Fisher exact test was applied to compare demographic characteristics and comorbidities. The Cox proportional hazard model was used to evaluate the hazard rates of adhesive capsulitis between the study and control cohorts, after adjusting for potential confounding factors, constituting patient age (as continuous), sex, diabetes mellitus, coronary heart disease, hypertension, hyperlipidemia, urbanization level, autoimmune disease, chronic liver disease, and cancer. To fulfill the proportional-hazard assumption, exploratory diagnostic log-log survival plots were applied to verify the proportionality of each dichotomous variable in the model and meet the proportional-hazard assumption. We plotted the stroke hazard curves based on the Cox model of the patient and control cohorts after adjusting for potential confounding factors. The SAS statistical package (SAS System for Windows, Version 9.1.3, SAS Institute Inc., Cary, NC, USA) and SPSS version 20 were used for analysis. A P value < .05 was considered statistically significant.

We identified 4472 hyperthyroid patients in the age- and sex-matched control cohorts. Compared with the control cohort ( Table 1 ), hyperthyroid patients are more likely to have comorbid autoimmune diseases ( P < .001), diabetes mellitus ( P < .001), hypertension ( P < .001), hyperlipidemia ( P < .001), coronary heart disease ( P < .001), and chronic liver disease ( P < .001). Figure 1 presents the hazard curves of hyperthyroid patients diagnosed with adhesive capsulitis and the control cohort during the 7-year follow-up period, after adjusting for patient age, sex, diabetes mellitus, coronary heart disease, hypertension, hyperlipidemia, autoimmune diseases, chronic liver disease, cancer, and urbanization level (log-rank test, P < .001). Table 2 shows the incidence and hazard ratios (HRs) of adhesives capsulitis in hyperthyroid patients and the control cohort. Of the 4472 hyperthyroid patients, 162 patients (671/100 000 person-years) exhibited adhesive capsulitis during the 24 122 person-year follow-up period. The crude HR of stroke was 1.26 (95% confidence interval [CI], 1.06 to 1.49), which was larger than that obtained for the control group. Table 3 presents the adjusted HRs of adhesive capsulitis in both the hyperthyroid patients and control cohort. The adjusted HR obtained for hyperthyroid patients was 1.22 (95% CI, 1.03 to 1.45), which was statistically significant. In addition, other confounding factors of adhesive capsulitis were analyzed. The results showed that age (adjusted HR: 1.05, 95% CI, 1.04 to 1.05), the male sex (adjusted HR: 069, 95% CI, 0.58 to 0.82), hyperlipidemia (adjusted HR: 1.49, 95% CI, 1.25 to 1.77), and chronic liver disease (adjusted HR: 1.37, 95% CI, 1.14 to 1.66) were all statistically significant regarding the development of frozen shoulder during the 7-year follow-up period.

This study showed that hyperthyroid patients have 1.22 times the risk of developing adhesive capsulitis compared to the general population. Until now, no relevant large-scale longitudinal population-based study has been conducted on the risk of adhesive capsulitis in hyperthyroid patients. All previously conducted studies regarding the risk of adhesive capsulitis in thyroid-disorder patients have been small-sample cross-sectional studies 12 . Limited information has been obtained regarding the temporal relationship between hyperthyroidism and adhesive capsulitis in previous studies. We adjusted for factors, such as diabetes and dyslipidemia, and identified hyperthyroid patients who were subsequently diagnosed with adhesive capsulitis. The results show that hyperthyroidism patients have a high comorbidity risk of adhesive capsulitis. The definite pathogenesis of adhesive capsulitis remains under investigation. Shoulder pain onset that indicates adhesive capsulitis is considered to be mediated through nerve stimulation. Adhesive-capsulitis shoulder pain can be caused by alpha-adrenoreceptor hyperresponsiveness, in which both nociceptive and proprioceptive fibers are stimulated, resulting in pain 13 . The inflammatory process is considered another pathogenesis of adhesive capsulitis. Rodeo et al found elevated levels of pro-inflammatory cytokines in adhesive capsulitis patients 14 . They proposed that the stimulation of inflammation caused by cytokines can engender shoulder synovitis, and this can result in a fibrotic cascade that is associated with growth factors such as TGF-beta 14 . The chronic fibrosis process of capsulitis is confirmed by observing the histologic presentation of fibroblast cell proliferation. Matrix metalloproteinase and fibrogenic growth factors are also increased in adhesive capsulitis patients 15 . Similar to the inflammatory and fibrosis process pathogenesis of adhesive capsulitis patients, hyperthyroid patients also present an inflammatory cytokine release and fibrosis phenomenon. Regarding thyroid disorder-associated ophthalmopathy patients, cell-mediated (Th1) and humoral-mediated immune responses infiltrate the orbital area. High levels of IL-2, IFN-γ, and TNF-α cytokines are secreted by Th1 cells in the retro-orbital area of patients who are diagnosed with Graves disease 16 17 . However, another study found Th2 cell-secreted cytokines, such as IFN-γ, IL-4, and IL-10, in patients 18 . Moreover, a previous study described the cytokine profiles of patients diagnosed with Graves disease and thyroid disorder-associated ophthalmopathy 17 . Cytokines have been proven as capable of inducing several proteins in orbital fibroblasts, and these cytokines have the ability to stimulate orbital fibroblast proliferation 19 . We propose that cytokine and fibroblast proliferation contributes to not only the process of thyroid ophthalmopathy but also to adhesive capsulitis. This can explain why hyperthyroid patients are vulnerable to adhesive capsulitis. SLE and RA patients were analyzed in this study, but were not found to have a high risk of adhesive capsulitis. Regarding SLE patients, cytokines have been proposed to have a pathogenic role in autoantibody production and immune complex deposition. These cytokines are interleukin-6, interleukin-17, interleukin-18, type I interferons, and TNF-alpha 20 . Cytokines have also been found in the pathogenesis of rheumatoid arthritis. One review article provided a comprehensive list of related cytokines, such as TNF-alpha, IL 1, 6, 15, 17, and 18, GM-CSF, VEGF, and TGF-beta 21 . However, this article stated that how these cytokines are organized within a hierarchical regulatory network remains unclear. In addition, the article stated that TNF-alpha plays a key role because TNF-alpha-blockage agents can are involved in successfully treating RA 21 . We propose that the pro-inflammatory cytokine and fibrosis process modulation is different in these autoimmune diseases and adhesive capsulitis. We controlled other possible hazard factors for developing adhesive capsulitis. In addition to hyperthyroidism, the results showed that hyperlipidemia patients have a high risk of adhesive capsulitis. These results support those of a previous case-controlled study that analyzed the lipid profiles of patients diagnosed with frozen shoulders and found that fasting serum triglyceride and cholesterol levels were increased in these patients compared with the levels of participants without frozen shoulders 22 . However, the pathogenesis of adhesive capsulitis in hyperlipidemia patients remains unclear. In contrast to a previous population-based longitudinal study, our results showed that diabetes mellitus patients do not have an increased risk of adhesive capsulitis 23 . This could be because the controlled group selection was different from that of the previous study either because of differing circumstances or participant matching to the hyperthyroid group. Our patients were predominantly women and were younger than those in the previous study on diabetes mellitus and frozen shoulders. A different control group selection can result in the influence of confounding factors. Long longitudinal follow-up data collection is the strength of our study. In addition, we attempted to control confounding factors, such as diabetes, hyperlipidemia, and autoimmune diseases such as SLE and RA. The long follow-up period and large number of potential confounding factors that were considered in this study enabled producing reliable results. The LHID2005 data released by the Taiwan NHI Institutes were used in this study. However, this study has several possible limitations. First, the diagnosis of adhesive capsulitis and hyperthyroidism was determined using the ICD codes listed in the NHI claim database; however, the diagnostic accuracy of the results obtained from the database were not confirmed. To increase the accuracy of diagnoses, the NHI Bureau has formed various audit committees that randomly sample claim data regularly to verify diagnostic validity and care quality. In addition, we used only consecutively coded cases to avoid inaccurate codes in the database records. These methods might improve the accuracy of registering rheumatic diseases. Second, the NHIRD records do not contain the laboratory data of hyperthyroid patients, and information regarding disease stratification severity is also limited. Third, the influences of thyroid medications, I131 radioiodine therapy, and thyroid surgery were not analyzed. Finally, our results were based on a retrospective cohort study. Information regarding lifestyles, obesity, cigarette smoking, and alcohol consumption cannot be obtained from the administrative database.

In conclusion, our 7-year longitudinal population-based case-controlled cohort study results showed that hyperthyroid patients have a 1.22-fold higher risk of adhesive capsulitis than that of patients without this condition. In addition to hyperthyroidism, the results showed that hyperlipidemia is another risk factor of adhesive capsulitis. Additional studies regarding the influence of hyperthyroidism treatment on the risk of adhesive capsulitis are recommended.