Non-virologic Methods to Diagnose Treatment Eligibility in HIV-exposed Infants

Overview

This study is designed to develop and evaluate a set of non-virologic diagnostic algorithms to monitor HIV-exposed children of unknown infection status for treatment eligibility during the first year of life. The results of this cross sectional study are expected to be used in development of a series of non-virologic algorithms to determining treatment eligibility among HIV-exposed children in settings where polymerase chain reaction (PCR) testing is not available and to guide the judicious use of PCR testing among HIV-exposed children in settings where PCR is available. These results will directly inform program implementation in Zambia.

Despite recent scientific advances in prevention of mother-to-child transmission (PMTCT), substantial donor investment, and national commitments to PMTCT program roll-out, pediatric HIV/AIDS remains a largely uncontrolled epidemic. Unless some intervention is undertaken, more than half of children who become HIV-infected will die by their second birthday. Yet in sub-Saharan Africa, the majority of infected children die without ever having confirmation of their HIV status, having never had the chance to begin life-saving antiretroviral therapy (ART). Recognizing the need to provide children with greater access to care, the World Health Organization (WHO) has advised countries that at least 10% of patients receiving ART should be children. However, most programs fall well short of this target. Unlike diagnosis in adults, definitive diagnosis of HIV in children less than 18 months of age requires virologic testing. Maternal IgG antibodies cross the placenta into the fetal circulation through pregnancy and are also secreted in breast milk. As infants may retain these maternal antibodies for up to 18 months after birth, conventional HIV antibody tests are of limited use in determining positive infection status. Instead it is necessary to directly identify HIV infection using a virologic method, typically polymerase chain reaction (PCR) testing to detect viral DNA or RNA. PCR testing is expensive and, in most developing countries, available only through specialty laboratories. Because PCR testing is more complex than other common HIV laboratory assays (e.g. CD4+ cell counting), it requires technicians with considerably more training and technical skill, involves longer sample preparation time, and can be considerably more costly than other diagnostic tests. Thus, most countries with large populations of HIV-infected children have extremely limited capacity to carry out PCR testing. Even where PCR is available, pediatric HIV diagnosis may remain difficult in breastfeeding populations because infants are at continued risk of infection until breastfeeding is completely stopped. A negative PCR test during breastfeeding does not guarantee that the infant will remain uninfected. The alternative - serial testing for HIV infection - would result in an exponential increase in cost. Indeed, a convincing argument can be made that rather than monitoring HIV-exposed breastfeeding children for HIV infection, it may be more appropriate to simply monitor for treatment eligibility, followed by a definitive antibody test at 15-18 months or a few months after breastfeeding has stopped. Those children who meet criteria for treatment could be commenced on ART; those who do not would continue to undergo periodic screening until they were weaned and old enough to have a final (definitive) antibody test. For HIV-exposed infants without access to HIV PCR testing - the vast majority in Africa - development of such an algorithm would have the obvious benefit of providing reliable evaluation for treatment eligibility in the first 18 months of life. Even in settings where HIV PCR testing is available - but may be limited - the introduction of such algorithms could reduce the volume of HIV PCR tests required at a programmatic level, and thus lead to substantial cost-savings. The overall goal of this protocol is to develop and evaluate a set of non-virologic diagnostic algorithms that can be used to monitor HIV-exposed children of unknown infection status for treatment eligibility during the first 12 weeks of life, with validation of our model among HIV-exposed children up to 12 months of age. We will evaluate the performance of different algorithms in two separate but common scenarios in Africa: (1) where infant PCR testing is not available, and (2) where infant PCR may be available, but poorly accessible due to cost and resource considerations. Although DNA PCR testing has become available in Zambia, only three sites currently provide this type of testing and capacity may be limited for country-wide screening. Our study was designed to determine HIV infection and antiretroviral therapy (ART) eligibility using more commonly used "non-virologic" tests, such as CD4, CD8, HIV antibody, total lymphocyte count, hemoglobin, and clinical staging. Previous studies have demonstrated that these indicators by themselves do not perform well in the absence of DNA PCR testing. In this exploratory study, we hypothesize that combinations of these tests - in formalized algorithms - can yield high performance and be used in rural settings where reliable DNA PCR services are not yet available or remain limited.