Use of age and CD4 cell count as criteria for identification of recent HIV infection in resource‐limited countries

Abstract

BackgroundIdentification of recent HIV infection is crucial for estimating HIV incidence and transmitted drug resistance (TDR) prevalence. Due to limited availability of diagnostic assays, WHO TDR surveys use age <25 yrs and/or CD4 >500 cells/mm3 at HIV diagnosis as epidemiological criteria to maximize inclusion of recently infected (within 3 yrs) and ARV‐naïve individuals. Accuracy of these criteria and variation by geographical region is unknown.MethodsA literature review of studies on HIV seroconverters (SC) published through March 2012 was performed. Age at SC and CD4 decline in absence of treatment were abstracted. Accuracy of alternative TDR survey criteria was explored.Results11 studies provided age at SC: 7 in Africa, 2 in Latin America, 2 in Asia. Median age at SC ranged between 24 and 33 years in studies in Kenya and Zambia, respectively and was 29 [interquantile range (IQR) 24, 34] in a large cohort study from Africa. Median age at SC was 29 years in studies on MSM in Brazil and China. 7 studies reported CD4 count decline: 5 in Africa, 1 in Latin America and 1 in Asia. Studies used ordinary least square regression or mixed models. None described median CD4 count 3 yrs after SC. The estimated mean CD4 count 3 yrs after SC ranged from 350–420 cells/mm3 in Africa and was 237 and 282 cells/mm3 in Asia and Latin America, respectively.ConclusionHIV SC occurs at all ages (median 29 yrs) in the assessed geographical regions. Enhancing feasibility of TDR survey implementation by including individuals >25 yrs decreases specificity, particularly in low HIV prevalence settings (Table).Use of age <25 yrs can maximized specificity to detect recent infection, but misses almost 75% of recent infections thus limiting feasibility of TDR survey implementation, particularly in low HIV prevalence settings. Lower mean CD4 count 3 yrs after SC was observed in Asia and Latin America compared to Africa. Regional differences may be explained by heterogeneity in eligibility criteria and statistical methods as well as contextual factors such as HIV subtypes or co‐morbidities. Confirmed CD4 >500 cells/mm3 would maximize specificity but greatly reduce the number of individuals included in TDR surveys. This review highlights the limitations of using age and CD4 count at HIV diagnosis as criteria to identify recent infections. These criteria should be revised when results of well‐designed studies to calibrate misclassification errors become available. Transmitted Drug Resistance Survery Criteria Evidence of Recent Infection Feasibility in Low‐Prevalence Settings Feasibility in High‐Prevalence Settings Age < 22 years STRONG POOR FAIR Age < 25 years INTERMEDIATE FAIR FAIR Age < 30 years WEAK GOOD GOOD Confirmed* CD4 > 500 cells/mm3 STRONG VERY POOR VERY POOR Confirmed* CD4 > 350 cells/mm3 INTERMEDIATE POOR POOR CD4 > 500 cells/mm3 INTERMEDIATE FAIR FAIR CD4 > 350 cells/mm3 WEAK GOOD GOOD Confirmed = At least two consecutive CD4 counts