Sex differences in creation of do-not-resuscitate orders for critically ill elderly patients following emergency surgery

Background: Emergency surgery patients are older, often critically ill, and at high risk of morbidity and mortality. We studied factors associated with issuance of a do-not-resuscitate (DNR) order and impact on morbidity and mortality in emergency surgery patients. Methods: Prospective study of all patients since January 1, 2000 who underwent emergency surgery before admission to the intensive care unit. Data collected were: age, sex, admission diagnosis (case-mix), raw- (AIII) and age-adjusted (aAIII) APACHE III scores, issuance of a DNR order, and morbidity and mortality. Primary outcomes were DNR status and morbidity and mortality. Groups were stratified by age (patients > 75 years versus younger) and sex. Statistics were calculated by chi(2) test, analysis of variance, and logistic regression. Results: In all, 723 emergency surgery patients (gastrointestinal: 35%; traumatic: 20%; neurologic: 17%) met study criteria and had morbidity and mortality of 17.8% (AIII-predicted: 31%). Ninety-two patients (12%) were made DNR, of whom 82.6% died. Women received DNR status more often (16.4% versus 9.5%, p < 0.01) but morbidity and mortality was the same for women and men (18.2% versus 17.5%, p = 0.85). By logistic regression, sex most predicted new DNR status (odds ratio [OR] 2.512, p = 0.005) compared with Multiple Organ Dysfunction score (OR 1.410, p < 0.0001), Age (OR 1.054, p < 0.0001) and aAIII (OR 0.995, 0.355), with goodness of fit of 3.876 (p = 0.868) and Nagelkerke R-2 of 0.511. Percent correct was 88.9, implying good discrimination. Conclusions: Female sex and, to a lesser extent, age were associated with issuance of DNR in series of patients who received emergency surgery. The association of DNR with female sex is an unexpected finding and may indicate clinician bias and necessitate the performance of further analysis.