Bidirectional association between eating disorder and temporomandibular joint disorder: A retrospective longitudinal nationwide population-based cohort study.

Background/purpose: An increasing body of evidence indicates correlations between the symptoms of temporomandibular disorder and those of eating disorder (ED). However, further investigation is required to elucidate the temporal and causal relationships between the aforementioned disorders. Materials and methods: This retrospective cohort study was conducted using data from the Taiwan National Health Insurance Research Database. Temporomandibular joint disorder (TMJD) was analyzed both as the cause and consequence of ED. We collected the data (from January 1, 1998, to December 31, 2011) of patients with antecedent TMJD (N = 15,059) or ED (N = 1219) and their respective controls (1:10), matched by age, sex, income level, residential location, and comorbidities. This study included patients who had received a new diagnosis of an ED or a TMJD between January 1, 1998, and December 31, 2013. Cox regression models were used to assess the risk of ED or TMJD development in patients with antecedent TMJD or ED. Results: TMJD patients had an approximately 3.70-fold (95 % confidence interval [CI]: 1.93-7.10) risk of ED development. Similarly, patients with ED had an approximately 4.78-fold (95 % CI: 2.52-9.09) risk of TMJD development. Subgroup analyses based on ED subtypes indicated antecedent TMJD and bulimia nervosa as the predictors of increased bulimia nervosa and TMJD risks (hazard ratios: 6.41 [95 % CI: 2.91 to 14.11] and 5.84 [95 % CI: 2.75 to 12.41]), respectively. Conclusion: Previous TMJD and ED are associated with increased risks of subsequent ED and TMJD; these findings suggest the presence of a bidirectional temporal association between TMJD and ED.