Skip Navigation LinksHome > Published Ahead-of-Print > A mathematical model evaluating the timing of early diagnost...
JAIDS Journal of Acquired Immune Deficiency Syndromes:
doi: 10.1097/QAI.0000000000000307
Original Article: PDF Only

A mathematical model evaluating the timing of early diagnostic testing in HIV-exposed infants in South Africa.

Lilian, Rivka R. MSc (Med); Johnson, Leigh F. PhD; Moolla, Haroon BBusSc; Sherman, Gayle G. MD, PhD

Published Ahead-of-Print
Collapse Box

Abstract

Background: Antiretroviral therapy is often initiated too late to impact early HIV-related infant mortality. Earlier treatment requires an earlier diagnosis and the currently recommended 6-week HIV polymerase chain reaction (PCR) test needs reconsideration. This study aims to identify 1) optimal testing intervals to maximize the number of perinatal HIV-infections diagnosed and 2) programmatic issues that impact diagnosis.

Methods: A mathematical model was developed to simulate antiretroviral prophylaxis uptake and health outcomes in 240,000 HIV-exposed South African infants. The model considered routine, early testing with one PCR (at birth, 6, 10 or 14 weeks of age) and with two PCR tests (at birth and at 6, 10 or 14 weeks of age).

Results: A single 6-week test would diagnose the same number of perinatal HIV-infections as birth testing (p=0.92) but fewer infections than a 10-week test (p<0.01). Ten-week testing identifies the highest number of perinatally infected infants (p<0.01 compared to a single test at all other ages) but does not save additional life years compared to birth testing (p=0.27). Performing 2 PCR tests (at birth and 10 weeks) would identify the highest number of perinatal infections (p<0.01 versus a second 6- or 14-week test). However, 25% of perinatal HIV-infections would remain undiagnosed, largely due to the failure to return PCR test results to caregivers.

Conclusion: Six weeks may no longer be the optimal age to diagnose perinatal HIV-infections. Two early PCR tests (at birth and 10-weeks) would likely be the ideal diagnostic algorithm, but must be coupled with improved program coverage.

(C) 2014 by Lippincott Williams & Wilkins

Login

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.