Clinical predictors of HIV-1 infection among preschool children in Dar es Salaam, Tanzania.

Abstract

Seroprevalence of HIV-1 infection was determined in children aged between eighteen months and five years, attending maternal and child health (MCH) clinics in Dar es Salaam, Tanzania. A total of 889 children were eligible for the study, however seven children could not be enrolled because their mothers/guardians absconded and blood could not be drawn from 21 children due to refusal of mothers/guardians and from another 12 children due to technical reasons. Therefore, the participation rate was 95.5%. Of the 849 children screened, 14 (1.65%) were found to have IgG anti HIV-1 antibodies in their sera. The main clinical features found in children with symptomatic HIV-1 disease were weight loss, generalized lymphadenopathy, recurrent fevers, and prolonged diarrhoea. The utility of clinical features suggestive of HIV-1 infection (according to CDC classification) in identifying HIV-1 infection in children was evaluated and found to have high sensitivity (100%), specificity (96.9%) and negative predictive value (100%), but a low positive predictive value (35%). Marked variations in progression to symptomatic phase were noted, whereby some relatively young children had progressed to symptomatic phase (CDC class P-2A), while some older children were still in the asymptomatic stage (CDC class P-1 C). None of the symptomatic HIV-1 infected children presented with neurological disease, severe opportunistic infections, or malignancies. Although reduced mid-upper arm circumference and weight-for-age were associated with HIV seropositivity, these clinical parameters had low positive predictive values compared to the CDC classification. | Serologic testing of 849 children 18 months-5 years of age attending the Lugalo and Mwananyamala maternal-child health clinics in Dar es Salaam, Tanzania, during May-August 1994 identified 14 cases (1.65%) of HIV infection. The sample represented 95.5% of children making mandatory monthly clinic visits during the 3-month study period. The main clinical symptoms in HIV-infected children were weight loss, generalized lymphadenopathy, recurrent fever, and prolonged diarrhea, all included in the Centers for Disease Control and Prevention (CDC) classification scheme. Children with symptomatic HIV infection were younger than those with asymptomatic infection. None of the children with symptomatic HIV infection presented with neurologic impairment, severe opportunistic infections, or cancers. The clinical features included in the CDC classification for HIV had a 100% sensitivity, a 96.9% specificity, and a 100% negative predictive value, but the positive predictive value was only 35%. Although the model with the best fit included mid-upper arm circumference less than 14 cm and a reduced weight-for-age (odds ratios, 3.8 and 1.9, respectively), the positive predictive values for these two factors were only 4.3% and 4.1%, respectively. The 1.65% HIV seroprevalence rate recorded in this community-based study is lower than the 2.4% rate estimated among newborns in Dar es Salaam, presumably because of high infant mortality and hospitalization among HIV-infected newborns. Use of a simplified version of the CDC classification of HIV infection in children is recommended for routine clinical use in Tanzania.