Seventy-four thousand nine hundred ninety-two deaths occurred within 28 days of the date of upper gastrointestinal hemorrhage, giving an overall case fatality rate of 14.5% (95% confidence interval [95% CI]: 14.4%–14.6%). Of these, 10,977 deaths (15%) occurred after discharge from hospital but within 28 days of hemorrhage. Only

312 (3%) of postdischarge deaths were coded as a subsequent hospital admission within the HES dataset. The population characteristics for nonvariceal and variceal hemorrhage are shown in Table 1. The median age for nonvariceal bleeds was 71 years (interquartile range, 50–81 years) and, for variceal bleeds, was 55 years (interquartile range, 45–66 years). Forty-six percent of those presenting DNA-PK inhibitor with nonvariceal hemorrhage had no comorbidity recorded, compared with 67% of those presenting with variceal hemorrhage after the exclusion of liver disease from the calculation of comorbidity. The population age structure and comorbidity varied over the study period (Figure 2) with a peak in the proportion of nonvariceal admissions over 80 PI3K inhibitor years old in 2002. This matched

the peak in case fatality in the same year (Table 1). There was a reduction over time in the proportion of those presenting with variceal hemorrhage who were less than 60 years old (Figure 2). The comorbidity for both groups increased over the study period. Median length of stay for nonvariceal hemorrhage was 4 days (interquartile range, 1–8 days) and for variceal hemorrhage was 7 days (interquartile range, 4–12 days). Idoxuridine The length of stay reduced over the study period for nonvariceal hemorrhage from 4 (interquartile range, 2–8 days) to 3 (interquartile range, 1–6 days) (P < .001 nonparametric test for trend), but there was no reduction for variceal hemorrhage. The overall 28-day case fatality following a nonvariceal hemorrhage admission was 14% and, following a variceal hemorrhage admission, was 23% (Table 1). From 1999 to 2007,

the unadjusted 28-day mortality following nonvariceal hemorrhage reduced from 14.7% to 13.1% (unadjusted odds ratio [OR], 0.87; 95% CI: 0.84–0.90). The unadjusted mortality following variceal hemorrhage reduced from 24.6% to 20.9% (unadjusted OR, 0.81; (95% CI: 0.69–0.95). Twenty-eight-day mortality for an acute admission with hemorrhage reduced over the study period for nonvariceal hemorrhage from 11.3% to 9.3% (unadjusted OR, 0.81; 95% CI: 0.77–0.85) and, for variceal hemorrhage, from 21.3% to 17.3% (unadjusted OR, 0.77; 95% CI: 0.62–0.95). Twenty-eight-day mortality for cases with an inpatient hemorrhage also reduced over the study period, for nonvariceal hemorrhage from 20.0% to 18.4% (unadjusted OR, 0.91; 95% CI: 0.86–0.95) and, for variceal hemorrhage, from 32% to 29% (unadjusted OR, 0.88; 95% CI: 0.67–1.14).