Tuberculosis Case Finding in HIV-Infected Pregnant Women in Kenya Reveals Poor Performance of Symptom Screening and Rapid Diagnostic Tests

• Authors: Sylvia M. LaCourse, University of Washington; Lisa Cranmer, Emory University; Daniel Matemo, Kenyatta Natl Hosp; John Kinuthia, Kenyatta Natl Hosp; Barbra A. Richardson, University of Washington; Grace John-Stewart, University of Washington; David J. Horne, University of Washington
• Journal Title: Journal of Acquired Immune Deficiency Syndromes
• Volume: Volume 71, Number 2
• Publisher: Lippincott, Williams & Wilkins | 2016-02-01, Pages 219-227
• Type of Work: Article | Post-print: After Peer Review
• Publisher DOI: http://doi.org/10.1097/QAI.0000000000000826
• Subjects: Health Sciences, Epidemiology
• Keywords: Science & Technology; Life Sciences & Biomedicine; Immunology; Infectious Diseases; tuberculosis; HIV; pregnancy; Xpert; urine TB LAM; screening; SOUTH-AFRICA; ACTIVE TUBERCULOSIS; PULMONARY TUBERCULOSIS; ANTIRETROVIRAL THERAPY; POSTPARTUM WOMEN; CHILD HEALTH; PREVENTION; RISK; INDIVIDUALS; ACCURACY

Abstract

Background: Tuberculosis (TB) during pregnancy in HIV-infected women is associated with poor maternal and infant outcomes. There are limited data on TB prevalence, optimal TB screening, and performance of rapid diagnostics in pregnant HIV-infected women. Methods: We conducted a cross-sectional study among HIV-infected pregnant women seeking antenatal care in western Kenya. After a standardized questionnaire, sputum smear microscopy for acid-fast bacilli, mycobacterial liquid culture, GeneXpert MTB/RIF (Xpert), urine lipoarabinomannan, and tuberculin skin testing were performed. We determined prevalence and correlates of culture-confirmed pulmonary TB, and compared diagnostic performance of World Health Organization (WHO) symptom screening and rapid diagnostic tests to sputum culture. Results: Between July 2013 and July 2014, we enrolled 306 women. Among 288 women with a valid sputum culture result, 54% were on antiretroviral treatment, median CD4 cell count was 437 cell per cubic millimeter (IQR 342-565), and prevalence of culture-confirmed pulmonary TB was 2.4% (confidence interval: 1.0% to 4.9%). Cough >2 weeks (P 0.04) and positive tuberculin skin testing (≥5 mm, P 0.03) were associated with pulmonary TB. Women with TB were 23-fold (95% confidence interval: 4.4 to 116.6) more likely to report a household member with TB symptoms (P 0.002). WHO symptom screen (43%), acid-fast bacilli smear (0%), Xpert (43%), and lipoarabinomannan (0%) had low sensitivity but high specificity (81%, 99%, 99%, and 95%, respectively) for pulmonary TB. Conclusions: HIV-infected pregnant women had appreciable prevalence of pulmonary TB despite modest immunosuppression. Current TB screening and diagnostic tools perform poorly in pregnant HIV-infected women. Adapted TB screening tools that include household member TB symptoms may be useful in this population.