Cofactors and Outcomes of Infant HIV Diagnosis in PMTCT versus Hospital

Abstract

<bold>Background:</bold> Many HIV-infected infants remain undiagnosed until they present with severe illness and hospitalization. Defining cofactors and outcomes associated with late pediatric HIV diagnosis can inform new approaches to identify HIV infected children before severe illness. <bold>Methods:</bold> A cohort of HIV-infected infants was recruited from PMTCT clinics and pediatric wards in Nairobi, Kenya during the pre-randomization phase of a clinical trial. Infants were diagnosed with HIV and referred to the study; infants initiated antiretroviral therapy (ART) and were followed for 2 years. Univariate logistic regression was used to identify cofactors for diagnosis site (hospital versus PMTCT), and survival analysis, chi-squared tests, and independent T-tests were used to compare infant survival, retention, hospitalization, and ART initiation between the two sites. <bold>Results:</bold> At recruitment sites, HIV prevalence was significantly higher among infants tested in hospital than in PMTCT programs (41% vs 11%, p<0.001). Among screened infants ages 0-12 months, infants diagnosed in hospital were significantly older than those diagnosed in PMTCT (5.5 vs 2.3 months, respectively, p<0.001). Among 99 infant-caregiver pairs enrolled in the study, 37% were diagnosed in a PMTCT clinic and 63% in hospital. Infants diagnosed in hospital were 81% less likely to have received PMTCT (95%CI=0.071-0.52), and their mothers were 83% less likely to have received ART or PMTCT (95%CI=0.061-0.48) compared to infants diagnosed within a PMTCT program. Infants diagnosed in hospital were more likely to be later disease stage (WHO 3 or 4) (OR=33, 95%CI=8.3-184), and had higher median HIV viral loads (6.7 versus 6.2 log10, respectively). Infants diagnosed in hospital were more than 3 times as likely to die as infants diagnosed in PMTCT (HR=3.1, 95%CI=1.3-7.6), a relationship that persisted after controlling for CD4% at enrollment (aHR=2.7, 95%CI=1.1-6.8). The two groups did not differ in time to ART initiation, loss to follow-up, or subsequent hospitalization. <bold>Conclusions:</bold> Infant HIV diagnosis in hospital was associated with failure to access PMTCT and more advanced disease. Infants diagnosed in hospital were at a higher risk for mortality, underscoring the critical importance of earlier diagnosis.