Seasonal pattern of peptic ulcer hospitalizations: analysis of the hospital discharge data of the Emilia-Romagna region of Italy

Background: Previous studies have reported seasonal variation in peptic ulcer disease (PUD), but few large-scale, population-based studies have been conducted. Methods: To verify whether a seasonal variation in cases of PUD (either compicated or not complicated) requiring acute hospitalization exists, we assessed the database of hospital admissions of the region Emilia Romagna (RER), Italy, obtained from the Center for Health Statistics, between January 1998 and December 2005. Admissions were categorized by sex, age (<65, 65-74, ≥ 75 yrs), site of PUD lesion (stomach or duodenum), main complication (hemorrhage or perforation), and final outcome (intended as fatal outcome: in-hospital death; nonfatal outcome: patient discharged alive). Temporal patterns in PUD admissions were assessed in two ways, considering a) total counts per single month and season, and b) prevalence proportion, such as the monthly prevalence of PUD admissions divided by the monthly prevalence of total hospital admissions, to assess if the temporal patterns in the raw data might be the consequence of seasonal and annual variations in hospital admissions per se in the region. For statistical analysis, the χ2 test for goodness of fit and inferential chronobiologic method (Cosinor and partial Fourier series) were used. Results: Of the total sample of PUD patients (26,848 [16,795 males, age 65 ± 16 yrs; 10,053 females, age 72 ± 15 yrs, p < 0.001)], 7,151 were <65 yrs of age, 8,849 between 65 and 74 yrs of age, and 10,848 ≥ 75 yrs of age. There were more cases of duodenal (DU). (89.8%) than gastric ulcer (GU) (3.6%), and there were 1,290 (4.8%) fatal events. Data by season showed a statistically difference with the lowest proportion of PUD hospital admissions in summer (23.3%) (p < 0.001), for total cases and rather all subgroups. Chronobiological analysis identified three major peaks of PUD hospitalizations (September-October, January-February, and April-May) for the whole sample (p = 0.035), and several subgroups, with nadir in July. Finally, analysis of the monthly prevalence proportions yielded a significant (p = 0.025) biphasic pattern with a main peak in August-September-October, and a secondary one in January-February. Conclusions: A seasonal variation in PUD hospitalization, characterized by three peaks of higher incidence (Autumn, Winter, and Spring) is observed. When data corrected by monthly admission proportions are analyzed, late summer-autumn and winter are confirmed as higher risk periods. The underlying pathophysiologic mechanisms are unknown, and need further studies. In subjects at higher risk, certain periods of the year could deserve an appropriate pharmacological protection to reduce the risk of PUD hospitalization.