Epidemiology of duodenal ulcer perforation: a study on hospital admissions in Norfolk, United Kingdom.
Canoy DS., Hart AR., Todd CJ.
BACKGROUND: Studies on the incidence of perforated duodenal ulcer are limited and in the United Kingdom, data are largely based on findings observed over two decades ago. To provide updated epidemiological data on duodenal ulcer perforation, the incidence of the disease in Norfolk, United Kingdom was determined. METHODOLOGY: Medical records of patients with duodenal ulcer perforation were reviewed to confirm the diagnosis and obtain information on possible risk factors, namely, Helicobacter pylori infection, smoking and intake of non-steroidal anti-inflammatory drugs. The patients were admitted between 1 January 1996 and 31 December 1998, and were residents of Norfolk, United Kingdom. RESULTS: Sixty-eight cases of duodenal ulcer perforation were identified, 36 (52.9%) were males and 32 (47.1%) were females. The age-standardised incidence rate was 3.77 per 100,000 population per year (95% confidence interval 3.72-3.83). The mean age upon admission for all cases was 72.3 years (standard deviation: 17.8). The mean age for males was 67.7 years (standard deviation: 19.4) and for females 77.6 years (standard deviation: 15.7), which differed significantly (difference in means: 9.9, 95% confidence interval 1.5-18.3). There were 29 deaths (42.7%), of which 19 were females. After adjustment for covariates, the odds ratio of mortality in women was 4.57 (95% confidence interval 1.28-16.29). There were 25 (36.8%) smokers and 22 (32.4%) patients were non-steroidal anti-inflammatory drug users. Helicobacter pylori infection was assessed in only 14 (20.6%) patients; 2 were positive, 3 were negative, and in the rest the results were unrecorded. CONCLUSIONS: The incidence rates were lower compared to previous studies in the United Kingdom conducted in the 1960's and 1980's, which could reflect either improved health care or decreasing exposure to known risk factors. Furthermore, the difference in age distribution of incident cases between males and females may explain the higher mortality in females.