Dengue infection and miscarriage: a prospective case control study

• Main Authors: Tan, P.C., Soe, M.Z., Si Lay, K., Wang, S.M., Sekaran, S.D., Omar, S.Z.
• Format: Article
• Published: Public Library of Science 2012
• Subjects: R Medicine
• Online Access: http://eprints.um.edu.my/3191/; http://www.ncbi.nlm.nih.gov/pubmed/22590658

BACKGROUND: Dengue is the most prevalent mosquito borne infection worldwide. Vertical transmissions after maternal dengue infection to the fetus and pregnancy losses in relation to dengue illness have been reported. The relationship of dengue to miscarriage is not known. METHOD: We aimed to establish the relationship of recent dengue infection and miscarriage. Women who presented with miscarriage (up to 22 weeks gestation) to our hospital were approached to participate in the study. For each case of miscarriage, we recruited 3 controls with viable pregnancies at a similar gestation. A brief questionnaire on recent febrile illness and prior dengue infection was answered. Blood was drawn from participants, processed and the frozen serum was stored. Stored sera were thawed and then tested in batches with dengue specific IgM capture ELISA, dengue non-structural protein 1 (NS1) antigen and dengue specific IgG ELISA tests. Controls remained in the analysis if their pregnancies continued beyond 22 weeks gestation. Tests were run on 116 case and 341 control sera. One case (a misdiagnosed viable early pregnancy) plus 45 controls (39 lost to follow up and six subsequent late miscarriages) were excluded from analysis. FINDINGS: Dengue specific IgM or dengue NS1 antigen (indicating recent dengue infection) was positive in 6/115 (5·2%) cases and 5/296 (1·7%) controls RR 3·1 (95% CI 1·0-10) P = 0·047. Maternal age, gestational age, parity and ethnicity were dissimilar between cases and controls. After adjustments for these factors, recent dengue infection remained significantly more frequently detected in cases than controls (AOR 4·2 95% CI 1·2-14 P = 0·023). INTERPRETATION: Recent dengue infections were more frequently detected in women presenting with miscarriage than in controls whose pregnancies were viable. After adjustments for confounders, the positive association remained.