AEGiS-15IAC: Field evaluation of five commercial rapid HIV antibody test kits in Tanzania, 2003.

15th International AIDS Conference

Bangkok, Thailand - July 11-16, 2004

Field evaluation of five commercial rapid HIV antibody test kits in Tanzania, 2003.

Int Conf AIDS 2004 Jul 11-16; 15:(abstract no. TuOrB1147)

Ipuge Y, Lyamuya E, Kagoma C, Ocheng D, Martin C, Mbago T, Merlin T, Collins C, Downer M, Scott C
US Centers for Disease Control and Prevention -Tanzania, Atlanta, United States

BACKGROUND: Use of rapid HIV antibody tests in voluntary counseling and testing (VCT) sites is common, but little data describe performance of commercially available rapid HIV kits in field settings.

METHODS: We undertook a field trial to evaluate performance of five rapid HIV test kits in three VCT sites in Dar es Salaam, Tanzania. Selected sites varied by work environment and estimate of HIV prevalence. Personnel at each site tested residual whole blood specimens collected from individuals consenting to rapid HIV testing with each of five commercial rapid HIV antibody kits. Plasma and dried blood spots were prepared for reference enzyme immunoassay and Western blot testing. Rapid test kits were Capillus, Determine, Hemastrip, Oraquick, and Unigold (all previously evaluated in a laboratory environment). Reference enzyme immunoassays were Uniform II antigen/antibody test and Uniform II + O antibody test.

RESULTS: 2,587 specimens were tested by all five rapid test kits and reference EIA methods. 362/2,587 (14.03%) tested positive and 2,195/2,587 (84.84%) tested negative by all five rapid tests. The five rapid tests gave concordant results on 2,558/2,587 (98.87%) specimens. HIV prevalence and concordance rates by individual site revealed: Site A: 42.45% prevalence, 99.05% concordance; Site B: 7.93% prevalence, 98.46% concordance; Site C: 16.59% prevalence and 99.60% concordance. A chi-square test indicated that concordance rates by site were not equal (Chi-square = 6.01, p = 0.47). Site B, with the lowest prevalence, had a slightly lower concordance rate.

CONCLUSIONS: Five commercially available rapid HIV antibody tests gave highly concordant (98.87%) test results when used in VCT sites in Tanzania. These results suggest that considering characteristics other than sensitivity and specificity may be useful in developing a national testing algorithm for Tanzania.

Keywords: AEGIS, HIV Antibodies, Evaluation Studies, Prevalence, HIV Seropositivity, Blotting, Western, Immunoenzyme Techniques, Sensitivity and Specificity, Acquired Immunodeficiency Syndrome, Tanzania, immunology, methods

040711
TuOrB1147