Contribution of NS1 and IgM based-rapid tests to dengue diagnosis in Colombia before zika emergence

Abstract

Objective: Dengue diagnosis is considered to be mainly clinical; however, rapid tests that detect IgM or NS1/IgM are being used in health services. This study assessed the contribution of rapid tests to dengue diagnosis in an endemic area before the emergence of zika virus in Colombia. Methods: Cross-sectional study of diagnostic tests based on a secondary analysis of a previous study in 14 health care institutions in Valle del Cauca department. Results of dengue rapid test, clinical diagnosis, and reference tests ELISA NS1, ELISA IgM, and RT-PCR were obtained for 632 participants. The sensitivity, specificity, predictive values and likelihood ratios of the use alone, serial and parallel combinations of NS1, IgM, NS1/IgM of the rapid test and clinical diagnosis were compared using Cochran's Q and MacNemar tests for paired data. Results: The sensitivity of clinical diagnosis (61.4% 95%IC 56-66.7) was higher than the sensitivity of rapid tests (37% 95% IC 29.6-44.7) (P<0.001). The serial used of NS1/IgM rapid test when clinical diagnosis was negative increased the sensitivity to 79.5% and, the serial use when clinical diagnosis was positive increased the specificity (from 66.3% to 98.7%). However, the latter decreased the sensitivity to 32.2%. While all negative likelihood ratios (LR-) were close to 1, the serial use of rapid tests when clinical diagnosis was positive had LR+ higher than 10. Conclusion: The clinical diagnosis is more sensitive than rapid tests, but by itself does not confirm or rule out dengue. A positive result in rapid tests is useful to confirm dengue but a negative result does not rule it out.