Validation of shock index for predicting mortality in older patients with dengue fever

Background Older adults have a higher mortality for dengue fever (DF). However, the best method for predicting mortality is still unclear. Aims We conducted this study to evaluate the shock index (SI) for this issue. Methods A retrospective case-control study was conducted by recruiting older patients (>= 65 years old) with DF who visited the study hospital in southern Taiwan during the 2015 DF outbreak. Demographic data, vital signs, past histories, decision groups, complications, and mortality were included in the analyses. We evaluated the accuracy of SI >= 1 for predicting 30-day mortality in this population. Results A total of 626 patients with a mean age of 74.1 years and nearly equal sex distribution were recruited. The mean of SI (+/- standard deviation [SD]) was 0.6 (+/- 0.2) and patients with a SI >= 1 accounted for 3.5% of the total patients. Logistic regression showed that patients with SI >= 1 had a higher mortality than those with SI < 1 (odds ratio: 8.49; 95% confidence interval: 1.76-17.92). The area under the receiver-operating characteristic was 0.76, and the Hosmer-Lemeshow goodness of fit test was 0.48. The SI >= 1 had a sensitivity, specificity, positive predictive value, and negative predictive value of 14.8%, 97.0%, 18.2%, and 96.2% for predicting mortality. Conclusions The SI >= 1 is an easy tool that can be potentially used to predict 30-day mortality in older DF patients, especially in DF outbreak. It has a high specificity and negative predictive value for excluding patients with high-risk mortality.