Reversion and conversion of <it>Mycobacterium tuberculosis </it>IFN-γ ELISpot results during anti-tuberculous treatment in HIV-infected children

Abstract

Abstract Background Recent interest has focused on the potential use of serial interferon gamma (IFN-γ) release assay (IGRA) measurements to assess the response to anti-tuberculous (TB) treatment. The kinetics of IFN-γ responses to Mycobacterium tuberculosis (MTB) antigens in HIV-infected children during treatment have not however been previously investigated. Methods IFN-γ responses to the MTB antigens, ESAT-6, CFP-10 and PPD were measured by an enzyme-linked immunospot assay (IFN-γ ELISpot) at presentation and at one, two and six months after starting anti-tuberculous treatment in HIV-infected children with definite or probable TB. Responses at different time points were compared using a Mann-Whitney U test with paired data analysed using the Wilcoxon signed rank test. A Fisher's exact or Chi-squared test was used to compare proportions when test results were analysed as dichotomous outcomes. Results Of 102 children with suspected TB, 22 (21%) had definite TB and 24 (23%) probable TB. At least one follow up IFN-γ ELISpot assay result was available for 31 (67%) of the 46 children. In children with definite or probable TB in whom the IFN-γ ELISpot assay result was positive at presentation, anti-tuberculous treatment was accompanied by a significant decrease in both the magnitude of the IFN-γ response to individual or combined MTB-specific antigens (ESAT-6 median 110 SFCs/106 PBMC (IQR 65-305) at presentation vs. 15 (10-115) at six months, p = 0.04; CFP-10 177 (48-508) vs. 20 (5-165), p = 0.004, ESAT-6 or CFP-10 median 250 SFCs/106 PBMC (IQR 94-508) vs. 25 (10-165), p = 0.004) and in the proportion of children with a positive IFN-γ ELISpot assay (Fisher's exact test: ESAT-6 15/0 vs 5/11, p = 0.0002, CFP-10 22/0 vs 8/17, p = 0.0001, ESAT-6 or CFP-10 22/0 vs. 9/17, p= 0.002). However almost half of the children had a positive IFN-γ ELISpot assay after six months of anti-tuberculous treatment. In addition, there was conversion of the IFN-γ ELISpot assay result during anti-tuberculous therapy in six of 12 children in whom the initial IFN-γ ELISpot assay was negative. Conclusions In HIV-infected children with definite or probable TB, anti-tuberculosis treatment is accompanied by a reduction in the magnitude of the IFN-γ ELISpot response to MTB-antigens. However, serial IFN-γ ELISpot measurements appear to have limited clinical utility in assessing a successful response to anti-tuberculous treatment in HIV infected children.