Mp. Meyer et al., SYMPTOMATIC HIV-INFECTION IN INFANCY - CLINICAL AND LABORATORY MARKERS OF INFECTION, South African medical journal, 87(2), 1997, pp. 158-162
Objective. To investigate the usefulness oi immunological tests in the
diagnosis of HIV infection in young symptomatic children (< 15 months
of age), Design. Tests were evaluated in HIV-infected (HIV antibody-
and PCR-positive) patients and non-infected individuals, Setting, Hosp
italised patients in a referral centre (Red Cross War Memorial Childre
n's Hospital, Cape Town), Patients, All admissions under 15 months of
age who had HIV antibody requested were eligible, provided there was s
ufficient serum (150 mu l) for further study, Overall, there were 201
symptomatic cases and 49 healthy controls, Twenty of the symptomatic c
ases were HIV antibody-positive and 19 of these were HIV-infected on t
he basis of a positive PCR for HIV viral product. Results. Of the test
s we evaluated (total IgG, IgM, IgA and rheumatoid factors of the same
classes), raised total IgG level (cut-off 18 g/l or above) was the mo
st useful, We used a commercial radial immunodiffusion plate which was
found to have excellent reproducibility (inter-assay coefficient of v
ariation 3.2%), The test detected 16 of 19 infected infants (sensitivi
ty 84%, negative predictive value 98%), With the exception of the find
ing of oral thrush (odds ratio 7; P < 0.001), the clinical signs at pr
esentation did not distinguish those who were HIV antibody-positive fr
om those who were negative, Conclusions, In our study of hospital admi
ssions, the finding of elevated IgG and HIV antibody was diagnostic of
HIV infection, (The positive predictive value of the combination was
100%.) Likewise, the presence of raised IgG levels and oral candidosis
had a high specificity for HIV infection (98%) but the sensitivity wa
s low (37%).