Background:  Access to CD4⁺ T cell counts is limited in most resource-constrained settings. There is therefore need to evaluate less expensive laboratory methods for detecting level of immunosuppression.

Methods:  Paired CD4⁺ T cell count and total lymphocyte count (TLC) data were collected from hospital records of treatment-naïve HIV-infected children. The sensitivity, specificity, and positive and negative predictive value of TLC were determined using the World Health Organization (WHO) CD4⁺ T cell cut-off values as the standard.

Results:  A total of 487 children with a median age of 36 months (1 to 144 months) were enrolled. The correlation between TLC and CD4⁺ was r = 0.66 (p <0.001). The highest correlation was seen in children with severe immunosuppression (r = 0.72, p <0.001) and age category >59 months (r = 0.68, p <0.001).The WHO-recommended TLC values for severe immunosuppression of 4000, 3000, 2500, and 2000 cells/mm³ for age categories <12, 12 to 35, 36 to 59, and >59 months, respectively, had sensitivity of 25%, 23%, 33%, and 62%, respectively; specificity of 100%, 98%, 92%, and 83%, respectively; and positive predictive values of 100%, 93%, 68%, and 74%, respectively. Raising the TLC cut-offs to 7000, 6000, 4500, and 3000 cells/mm³ for each of the stated age categories resulted in sensitivity of 71%, 64%, 56%, and 86%, respectively; and specificity of 73%, 62%, 54%, and 68%, respectively.

Conclusions:  TLC has a better correlation with CD4⁺ count in the age groups >59 months and in the subset of children with severe immunosuppression. The WHO recommended TLC cut-off for severe immunosuppression has low sensitivity, which may be improved by raising the cut-off level.