University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Primary HIV-1 infection among infants in sub-Saharan Africa: HPTN 024
Global Health Sciences Literature Digest
Published October 5, 2009
Journal Article

Read JS, Mwatha A, Richardson B, et al. Primary HIV-1 infection among infants in sub-Saharan Africa: HPTN 024. 2009 Jul 1;51(3):317-22.


To assess clinical signs and diagnoses associated with primary HIV-1 infection among infants

Study Design

Secondary analysis of data from a clinical trial, HPTN 024. The primary objectives of HPTN 024 were to evaluate the efficacy of antibiotics to reduce mother-to-child transmission of HIV-1 and preterm birth.


Data for the analysis came from a clinical trial conducted at four sites in three African countries: Blantyre and Lilongwe, Malawi; Dar es Salaam, Tanzania; and Lusaka, Zambia.


The study population for this analysis comprised live-born infants (singletons or, if a multiple gestation, first-born infants), born to HIV-1-infected women enrolled in HPTN 024 who had negative HIV-1 RNA assay results at birth and at 4-6 weeks of age, and who were still breastfeeding at the 4- to 6-week visit. A total of 2,659 women enrolled in HPTN 024. Of these, 2,292 were HIV-1 infected, and 2,052 delivered live-born infants, including 2,026 singletons and 26 first-born twins. Of these live-born infants, 1,317 (64%) had negative HIV-1 RNA assay results at birth and at 4-6 weeks of age and were still breastfeeding at the 4- to 6-week visit. Of these 1,317 infants, 84 (6.4%) became HIV-1 infected after 4-6 weeks but before 12 months of age and 1,233 were uninfected.


All HIV-1-infected women participating in HPTN 024 and their infants were provided nevirapine according to the HPTN 012 regimen for prevention of mother-to-child transmission. Infant study visits were conducted within 48 hours of birth, at 4-6 weeks, and at 3, 6, 9, and 12 months. At these visits, the infant's caregiver was interviewed and the infant was examined. Although each clinical site provided counseling regarding the risks and benefits of breastfeeding, replacement feeding or other interventions related to preventing transmission by breastfeeding of HIV-1 infection were not implemented as part of the trial. Antiretroviral treatment for mothers and children was not available at any of the clinical sites at the time the trial was conducted.

Primary Outcomes

Prevalence, sensitivity, specificity, and predictive values of the clinical signs and diagnosis of primary HIV infection


Of the 84 HIV-1-infected infants, the mean age at acquisition of HIV-1 infection was 7.6 months (range: 3.0-13.6 months).The most common diagnoses were malaria and pneumonia, and the most common clinical signs were cough and diarrhea. Primary infection was associated with significantly increased odds of diarrhea (odds ratio [OR]=2.4), pneumonia (OR=3.5), otitis media (OR=3.1), and oral thrush (OR=2.9). For the clinical signs and diagnoses evaluated, sensitivity was low (1%-16.7%) and specificity was high (88.2%-99%) (see Table below). Positive predictive values (PPVs) ranged from 0.1%-1.4%, with malaria, cough, and diarrhea having the highest PPVs (all 1.4%), and negative predictive values (NPVs) ranged from 28%-51.1%, with pneumonia having the highest NPV.

Of these 84 infants who became HIV-1 infected during follow-up, infection was documented to have occurred within 3 months of the last negative visit in 36 infants. In other words, 48 infants had interval-censored (interval >3 months) HIV-1 results, making it difficult to precisely define when the infection occurred. To assess the sensitivity of the results to the lack of precision in the timing of HIV-1 infection, a secondary analysis excluding the 48 infants with interval-censored HIV-1 diagnostic test results was conducted. Of the 36 HIV-1-infected infants in this secondary analysis, the mean age at infection was 6.1 months (range: 3.0-12.3 months). Overall, there were 36 primary infection visits and 71 non-primary infection visits for these infants. The most common diagnoses at primary infection visits were malaria (19.4%) and pneumonia (8.3%), and the most common clinical signs were cough (22.2%), diarrhea (13.9%), and dermatitis (11.1%). Primary infection was associated with significantly increased odds of cough (OR=2.4; 95% confidence interval [CI]: 1.1-5.4; P=0.0258] and dermatitis (OR=3.1; 95% CI: 1.1-8.8; P=0.0326). For the clinical signs and diagnoses evaluated, sensitivity was generally higher than that observed for the entire cohort of HIV-1-infected infants (2.8%-22.2%), but specificity was similarly high (88.2%-98%). The PPVs ranged from 0.2%-1.8%, and NPVs ranged from 48.4%-51.7%.

Despite the modest differences in sensitivity for certain clinical signs, the results of this secondary analysis were generally comparable to the main study results, with the exception of pneumonia, noninfectious dermatitis, and oral thrush.

Table. Sensitivity, Specificity, and Predictive Values
Clinical Signs and DiagnosesSensitivity (102 Visits), %Specificity (5,650 Visits), %PPV, %*NPV, %*
Otitis media5.998.00.449.2
Oral thrush5.997.90.549.2
Neonatal sepsis1.
Noninfectious (or unspecified) dermatitis5.996.00.548.7

*HIV-1 transmission rate estimated from cohort is 7%.

NPV = negative predictive value, PPV = positive predictive value.


The authors conclude that certain clinical signs and diagnoses, although more common during primary HIV-1 infection, had low sensitivity and high specificity. Efforts to expand access to laboratory assays for the diagnosis of primary HIV-1 infection among infants of HIV-1-infected women should be emphasized.

Quality Rating

There is no widely accepted rating system for secondary analysis studies such as this, making it difficult to assess its quality. Also, it is important to note that the HPTN 024 trial was not designed to evaluate primary HIV-1 infection in infants; the schedule of study visits resulted in relatively wide intervals between visits; and diagnoses were made by clinicians at the trial sites, without a priori diagnostic criteria imposed by the trial. Therefore, clinical diagnoses were based on clinical findings, with or without laboratory or radiographic studies, and were not necessarily made uniformly across sites.

In Context

To date, there have been only two studies of clinical correlates of primary HIV-1 infection among breastfeeding infants of HIV-1-infected women. Rouet et al (1) conducted a case-control study in West Africa, comparing 22 HIV-1-infected children with postnatal acquisition of HIV-1 to controls (breastfed, HIV-1-uninfected children). In this study, primary HIV-1 infection was associated with generalized lymphadenopathy, dermatitis, and a mononucleosis-like illness. Richardson et al (2) evaluated 56 Kenyan infants with 125 primary infection visits and 3,491 nonprimary infection visits. Overall, primary HIV-1 infection was associated with lymphadenopathy, failure to thrive, and rash. Among infants less than aged 2 months, primary HIV-1 infection was associated with lymphadenopathy. Among older infants, primary HIV-1 infection was associated with dehydration, pneumonia, and hospitalization. As in this analysis, clinical manifestations of primary HIV-1 infection were had high specificity but low sensitivity.

Programmatic Implications

Recognition of the timing of HIV-1 acquisition remains important because early initiation of antiretroviral therapy may provide clinical benefits, such as decreasing the risk of neurodevelopmental delay and of life-threatening infections like Pneumocystis pneumonia. The World Health Organization (WHO) has developed clinical case definitions and clinical staging systems for HIV-1 infection, and the WHO and the United Nations Children's Fund developed the "integrated management of childhood illness" strategy to provide guidelines for the diagnosis and management of ill children at the primary care level. Evaluations of these clinical staging systems, however, especially among young infants, have had low sensitivity and are consistent with the results of this study. The WHO has released revised case definitions of HIV-1 infection to be used for surveillance and clinical staging classifications for HIV-1-related disease among adults and children.(3) Included in these guidelines are clinical criteria for the presumptive diagnosis of severe HIV-1 disease among HIV-1-seropositive, HIV-1-exposed children younger than 18 months where virologic testing is not available so that early initiation of antiretroviral therapy can occur; the results of formal evaluations of these criteria are eagerly anticipated.


  1. Rouet F, Elenga N, Msellati P, et al. Primary HIV-1 infection in African children infected through breastfeeding. AIDS. 2002;16:2303-9.
  2. Richardson BA, Nduati R, Mbori-Ngacha D, et al. Acute HIV infection among Kenyan infants. Clin Infect Dis 2008;46:289-95.
  3. World Health Organization. WHO Case Definitions of HIV for Surveillance and Revised Clinical Staging and Immunological Classification of HIV-Related Disease in Adults and Children. Geneva, Switzerland: World Health Organization; 2006. Accessed September 8, 2008.