Multiple Variable Logistic Regression Analysis Research Articles

Serum bone alkaline phosphatase and calcaneus bone density predict fractures: a prospective study.

The aim of this study was to assess the ability of serum bone-specific alkaline phosphatase (bone ALP), creatinine-corrected urinary collagen crosslinks (CTx) and calcaneus bone mineral density (BMD) to identify postmenopausal women who have an increased risk of osteoporotic fractures. Calcaneus BMD and biochemical markers of bone turnover (serum bone ALP and urinary CTx) were measured in 512 community-dwelling postmenopausal women (mean age at baseline 69 years) participating in the Hawaii Osteoporosis Study. New spine and nonspine fractures subsequent to the BMD and biochemical bone markers measurements were recorded over an average of 2.7 years. Lateral spinal radiographs were used to identify spine fractures. Nonspine fractures were identified by self-report at the time of each examination. During the 2.7-year follow-up, at least one osteoporotic fracture occurred in 55 (10.7%) of the 512 women. Mean baseline serum bone ALP and urinary CTx were significantly higher among women who experienced an osteoporotic fracture compared with those women who did not fracture. In separate age-adjusted logistic regression models, serum bone ALP, urinary CTx and calcaneus BMD were each significantly associated with new fractures (odds ratios of 1.53, 1.54 and 1.61 per SD, respectively). Multiple variable logistic regression analysis identified BMD and serum bone ALP as significant predictors of fracture (p = 0.002 and 0.017, respectively). The results from this investigation indicate that increased bone turnover is significantly associated with an increased risk of osteoporotic fracture in postmenopausal women. This association is similar in magnitude and independent of that observed for BMD.

Surgical and smoking history in inflammatory bowel disease: A case—control study

: An inverse relation between appendectomy and ulcerative colitis and smoking and ulcerative colitis has been proposed. Our study examined the frequency of common surgical interventions and of smoking in 500 patients with inflammatory bowel disease. They comprised 177 patients with ulcerative colitis, 134 patients with Crohn's disease, and 189 controls matched for age, sex, and socioeconomic group. Subjects were questioned on all previous surgery and on smoking history. The appendectomy rate among controls was 17.5% (33 of 189), which was significantly greater than that of patients with ulcerative colitis: 8.5% (15 of 177; p < 0.05). However, after using multiple variable logistic regression analysis, this was no longer statistically significant. There was no significant difference in appendectomy rate between patients with Crohn's disease and controls. The three groups had comparable rates of tonsillectomy and cholecystectomy. Of patients with ulcerative colitis, 84.2% were nonsmokers at the time of diagnosis in contrast to 50.4% among the patients with Crohn's disease (p < 0.01). In this case-control study, by using multivariate logistic regression analysis, appendectomy offers no significant protection from developing ulcerative colitis or Crohn's disease. Smoking appears to be deleterious in Crohn's disease, whereas a cessation of smoking precedes the onset of ulcerative colitis in a significant number of cases.