Utility of total lymphocyte count as a surrogate marker for CD4+ T cell count in the initiation of haart at Kenyatta National Hospital
Gitura, Bernard M
MetadataShow full item record
Objectives: To evaluate Total Lymphocyte Count as a surrogate marker for CD4 + T cell count in ARV treatment initiation in a Kenyan population of HIV sero-positive patients at Kenyatta National Hospital. STUDY DESIGN AND SETTING Cross-sectional descriptive study, carried out at a national hospital HIV treatment and follow-up outpatient facility: Comprehensive Care Centre, Kenyatta National Hospital. Methods: A total of 225 HIV Elisa positive, ARV naive patients visiting the Comprehensive Care Centre from January 2006 to March 2006, stratified by WHO clinical stages were consecutively recruited; and evaluated by clinical assessment, CD4 count, and TLC. The WHO prescribed cutoff of 1200 cells/mm3 was evaluated, utilizing standard diagnostic test utility testing criteria ( sensitivity,specificity,positive and negative predictive value (PPV & NPV),against a CD4 count cut-off of 200 cells/mm3 in its ability to appropriately classify patients as either above or below this cut-off. A Receiver Operator Characteristic (ROC) curve was thus generated to obtain optimal TLC cut-off for this study. RESULTS A significant linear correlation was found between TLC and CD4 cell count for the whole group with a spearman rank correlation of 0.761. (p < 0.01); and was also independently observed in the four WHO clinical stages. The classification utility of TLC 1200 cells/mrrr' cut-off was suboptimal; sensitivity of 37 %; specificity of 99 % and the NPV was 56 %. The ROC curve generated an optimal TLC cut-off of 1900 cells/mm3 to be of greatest utility with a sensitivity of 81.1%, specificity of90.3%, PPVof90.8% and NPV was 80.2%.This implies that a TLC cutoff of 1900cells/mm3 correctly identifies eight of ten HIV positive patients as having a CD4 < 200 cells/mm3 and only misclassify two such patients. Serial CD4 testing can then be performed on the minority of patients who despite a TLC >1 1900 cells/mm3 are, on basis of clinical data, suspect of more advanced disease warranting ARV therapy. This would reduce the number of patients tested for and focus the application of CD4 testing and thus reduce attendant cost in care provision in CD4 resource poor settings. CONCLUSION Our data showed a good positive correlation between TLC and CD4 cell count, however the WHO recommended TLC cutoff of 1200/mm3 was found to be of low sensitivity in classifying patients as having a CD4 counts < 200 cells/mm3• This would result in underestimation of advanced stage of disease and to withholding ARVs treatment to persons who need treatment. We recommend a TLC cut-off of 1900 cells/mnr' for our population to classify patients as either above or below the CD4 count cut off of 200 cells/mrrr' as an indicator of when to start antiretroviral therapy.