The accuracy, sensitivity and specificity of rapid point-of-care testing for CD4+ T cell count enumeration and TB diagnosis.

Abstract

Objectives: The PIMA CD4+ T cell count analyser has been favourably evaluated for use in point-of-care (POC) situations in Mozambique and Zimbabwe, has also been recommended by the World Health Organisation (WHO), however, there is limited information on its use in Primary Healthcare (PHC) settings in KwaZulu Natal (KZN) South Africa. The main aim of this study therefore assessed the accuracy, sensitivity and specificity of the Alere PIMA Point of Care (POC) analyser CD4+ T cell count enumeration compared to the South African National Health Laboratory Services (SA-NHLS) methodology, which uses Beckman Coulter with Panleucogating (PLG/CD4). The potential role of using the PIMA CD4 analyser as a predictor of antiretroviral therapy (ART) eligibility was also assessed. Material and Methods: The study took place at Lancers Road clinic, a busy primary health clinic (PHC) facility under the eThekwini Health Unit. An extra two millilitres of venous blood was drawn from the same blood draw as for the routine CD4 NHLS test (Beckman Coulter) into another EDTA tube for the comparison of the enumeration of CD4+ T cells using the PIMA analyser during January – July 2013. Results: A total of 268 patients were recruited for the PIMA analyser comparison with NHLS PLG/CD4 while a sub-set of 100 blood samples were also analysed on the FACS calibur. In the 100 samples the PIMA analyser results correlated better with the FACS calibur results (mean bias of 7.52, Bland Altman limits of agreement -111 to 126 and correlation of 0.970) than with the NHLS PLG/CD4 results (mean bias of -12.78, Bland Altman limits of agreement -226.041 to 200.481 and correlation of 0.90). In the 268 samples the overall mean difference between the PIMA analyser – NHLS PLG/CD4 was 17.5 cells/μl (95% CI 6.2 to 28.8). The percentage similarity (SIM) between the two (Mean ± SD) was 106 ± 15.5; indicative of acceptable agreement between the two tests. When categorised by the following CD4+ T cell counts of: ≤350 cells/μl; 351-500 cells/μl; ≤500 cells/μl and > 500 cells/μl , the mean difference of PIMA analysers – NHLS PLG/CD4 was 33 cells/μl (95% CI 23 to 42); 22 cells/μl (95% CI -3.5 to 47); 30 cells/μl (95%CI 21 to 39); and cells/μl (95% CI -78 to 6.1) respectively. Under the current South African guidelines of ≤350 cells/μl CD4+ T cells, the sensitivity of the PIMA analyser was 83.5% and specificity 92%. At this threshold of ≤ 350 cells/μl there were 35 (13%) misclassifications, of which 27 were false negatives. This implies that 27 patients would have been falsely deemed ineligible for ART according to the PIMA analyser. The mean difference between the PIMA analyser and NHLS PLG/CD4 in this group of 27 patients was 112 cells/μl. The positive predictive value was high at 95% such that 95% of the patients eligible for treatment according to PIMA analysers would have also been deemed eligible for treatment on the NHLS PLG/CD4 test. Using future South African treatment guidelines threshold of CD4+ T cell counts ≤500 cells/μl , a high sensitivity of 94% was observed at the sacrifice of lower specificity of 78%. According to the NHLS PLG/ CD4 test result, 164/268 (61%) of patients were eligible for ART (CD4+ T cell count ≤350 cells/μl) compared to 145/268 (54%) with the PIMA analyser POC CD4+ T cell test. Of those eligible for ART according to the ART register at Lancers Road PHC, 110/164 (only 67%) of these patients were initiated on ART. Of those who did not return for their results 35/268 (13%), twenty of 35 (57%) were eligible for ART according to the NHLS PLG/CD4 laboratory CD4 test result, all of whom were not initiated on ART. Conclusion: The overall agreement between the PIMA analyser POC and NHLS PLG/ CD4+ T cell count enumeration in adult HIV positive individuals was acceptable with clinically insignificant mean bias. Together with high positive predictive value, and sensitivity and acceptable specificity the PIMA analyser POC lends itself to an excellent facilitator of improved healthcare.