We conducted a nested case-control study of 1,377 cases of upper gastrointestinal bleeding or perforation (UGIB) and 10,000 controls to evaluate the association of individual non? steroidal antiinflammatory drugs (NSAIDs), utilization char? acteristics, and other risk factors for these conditions. Age was the strongest risk factor for UGIB. Male gender, history of complicated peptic ulcer disease, and current use of steroids were also risk factors for UGIB. The adjusted odds ratio (OR) for current NSAID use was 4.3 [95% confidence interval (CI) = 3.7-5.0]. The ORs for current NSAID use were similar for fatal cases and for the gastric, duodenal, prepyloric, and mul? tiple sites of lesion, but the OR was substantially increased for perforations (OR = 16.9; 95% CI = 9.1-31.5). Women age 80 years or older experienced the greatest effect of NSAID use. Current users of multiple NSAIDs and recent switchers showed ORs of 9.0 and 6.2, respectively. Ibuprofen showed the lowest OR and diflunisal, the highest. ORs for low, medium, and high NSAID daily dose were 2.9, 4.2, and 5.8, respectively. This trend was present among new, short-term, and long-term users. Simultaneous use of multiple NSAIDs as well as use of a single individual NSAID at high doses greatly increases the risk of complicated peptic ulcer disease. (Epidemiology 1997; 8:18-24) Keywords: nonsteroidal antiinflammatory agents, corticosteroids, drug interaction, gastrointestinal toxicity, medical record linkage, drug utilization, age. Current users of nonsteroidal antiinflammatory drugs (NSAIDs) have a three-to fivefold increase in risk of upper gastrointestinal bleeding and perforation (UGIB). Epidemiologic studies in this area have recently focused on the evaluation of effects of individual NSAIDs, pat? terns of drug use, and other risk factors for this condi? tion.1-9 We present here the results of a nested casecontrol study conducted in the general population of Saskatchewan, Canada. This study was implemented in parallel with another study1 of a general practice patient population in the United Kingdom, with the aim of addressing gastrointestinal safety of individual NSAIDs From 'Pharmacoepidemiology Research, CIBA-GEIGY S.A., ? 1996 by Epidemiology Resources Inc. 18(toxicity ranking, effect of duration of therapy, individ? ual dose effect, patterns of use) and risk among the elderly. The present study was also designed to provide detailed data on absolute rates of complicated peptic ulcer disease. Population and MethodsThe Canadian province of Saskatchewan has 1.1 million residents. Over 95% are entitled to receive benefits through the Department of Health.10 As a by-product of providing these health services, the Saskatchewan De?partment of Health has been accumulating a large vol? ume of health care information in computerized data? bases over a number of years. This study examined the relation between exposure to NSAIDs and several serious categories of outcomes in Saskatchewan during the period January 1, 1982, to December 31, 1986.5,1112 During the stu...

We studied the behavior of radiogrammetric and densitometric measurements in relation to season and body weight in a group of 30 healthy premenopausal women. Measurements were made at 6-month intervals, in summer/fall when bone density increases and in winter/spring when bone density declines. Total body bone mineral content (TBBMC) and regional bone mineral content (RBMC) were measured using dual-energy X-ray absorptiometry (DXA). Metacarpal radiogrammetry was carried out with computed radiography. Weight and body mass index increased significantly in winter (P < 0.05) and total body and RBMC decreased (P < 0.001). The opposite occurred in summer: weight and body mass index decreased significantly (P < 0.05) and total body and regional bone mineral content increased (P < 0.001). Differences in TBBMC persisted when the measurement was corrected for weight (TBBMC/W) (P < 0.001), but not for metacarpal cortical thickness corrected for weight. In the first measurement made there were significant relations between weight and both TBBMC (P < 0.001) and metacarpal cortical thickness (P < 0.005). The relation between weight and TBBMC remained significant in later measurements, but the relation between weight and metacarpal cortical thickness ceased to be significant in the second and fourth measurements. Our results show that there is an important seasonal variation in bone mass and that DXA is more sensitive than radiogrammetry in registering these changes.

No abstract