Hemorrhagic peptic ulcer is one of the most frequently encountered emergency conditions in daily practice1. The overall incidence of hemorrhagic peptic ulcers has decreased worldwide, although the overall incidence in elderly patients remains high. Loperfido et al reported that the mean age of patients with acute upper gastrointestinal bleeding had significantly increased from 1983e1985 to 2002e2004 in the United States. In a retrospective study in Norway, Bakkevold3 reported that the incidence of hemorrhagic peptic ulcers was significantly higher in elderly patients than in younger patients. Because of the high incidence of hemorrhagic peptic ulcers in elderly patients and the rapid acceleration in the growth of the aging population, it is expected that many gastroenterologists or endoscopists will encounter more elderly patients with hemorrhagic peptic ulcers. Hospital volume is a major factor that influences the outcomes of various endoscopic treatments such as endoscopic retrograde cholangiopancreatography or endoscopic sub-mucosal dissection.
Little information is available on the relationship between case volume and the outcomes of elderly patients with bleeding peptic ulcers. This study investigated the effect of case volume on the outcomes of elderly patients with bleeding peptic ulcers, based on a national administrative database.
In total, 14,569 elderly patients (i.e., _ 80 years old) treated by endoscopic hemostasis for bleeding peptic ulcers were referred to 1073 hospitals from 2010 to 2012 in Japan. We compared inhospital mortality (30-day and overall), length of hospital stay (LOS), and medical costs in relation to case volume. A hospital was categorized as a low-volume hospital (i.e., < 5 cases/y), a medium-volume hospital (i.e., 5e9 cases/y), or a high-volume hospital (HVH) (i.e., > 9 cases/y).
Multiple logistic regression revealed that HVHs did not have lower in-hospital mortality rates, compared to low- or medium-volume hospitals [for 30-day mortality: the odds ratio (OR) was 0.97 with a 95% confidence interval (CI) of 0.76e1.24; p ¼ 0.831; for overall mortality: OR, 0.86; 95% CI, 0.70e1.07; p ¼ 0.197]. However, multiple linear regressions showed that HVHs had significantly shorter LOS and lower medical costs. The coefficient for LOS was _5.02 days (95% CI, from _6.04 days to _4.01 days; p < 0.001), whereas the coefficient for LOS for medical costs was _1393.00 United States dollar (US$; 95% CI, from _1793.30 to _992.70 US$; p < 0.001).
This study demonstrated that a higher case volume was significantly associated with shorter LOS and lower medical costs among elderly patients with bleeding peptic ulcers. However, there was no significant effect of case volume on mortality in Japan.