Seasonality of hospital admissions and birth dates among inpatients with eating disorders: a nationwide population-based retrospective study.

PURPOSE: Seasonal variation exists in the psychopathology of eating disorders. However, it is still unknown whether there is seasonal variation in eating disorder symptom severity. This study investigated seasonal trends in hospital admissions and birth dates among patients with eating disorders in Taiwan (25°N). Subgroup analyses by gender and comorbid affective disorders were also of interest. METHODS: Data on all hospital admissions between 2000 and 2013 were collected from the Taiwan National Health Insurance Research Database, and 1954 patients with eating disorders were identified. Hospital admissions and birth dates were recorded by day. The four seasons and cross-seasons were defined by solstices and equinoxes. The expected distribution of births was determined using data from all patients hospitalized from 2000 to 2013 (n = 13,139,306). RESULTS: Hospital admissions among patients with eating disorders exceeded the rate of expected hospital admissions in the summer season (p < 0.001) and the autumn cross-season (p < 0.001). However, the seasonal (p = 0.421) and cross-seasonal (p = 0.24) distributions of birth dates among these patients did not differ from the expected distributions. Interestingly, hospital admissions among patients with comorbid affective disorders exceeded the rates of hospital admissions among non-affective patients during the spring (p = 0.004). Moreover, the number of non-affective patients born during autumn exceeded the birth rates of affective patients during this season (p = 0.001). Gender and comorbid affective disorders were not associated with cross-seasonal differences in either hospitalizations or dates of birth. CONCLUSIONS: Affective psychopathology in inpatients with eating disorders may substantially contribute to symptom severity that waxes and wanes with the seasons. Moreover, the seasonal distribution of birth dates was significantly different in patients without comorbid affective disorders.