Performance evaluation of BD FACSPresto™ point of care CD4 analyzer to enumerate CD4 counts for monitoring HIV infected individuals in Nigeria

Background: Despite the upsurge in support and intervention of donor agencies in HIV care and treatment programing in Sub-Sahara African, antiretroviral (ART) programs are still confronted with access and coverage challenges which influence enrolment of new patients. This study investigated the validity of point of care BD FACSPresto™ CD4 analyzer for CD4+ cell count, overall agreement, correlation, sensitivity, and specificity in comparison to a reference standard flow cytometry method. We also assessed the feasibility of use among nonlaboratorians. Methods Blood samples from 300 HIV infected individuals were analyzed for CD4+ T cell and CD4%, using finger prick capillary sample from 150 PMTCT clients and 150 ART clients at Murtala Mohammed Specialist Hospital, Kano, Nigeria. Their venous samples were compared on a flow cytometry reference method using BD FACSCount CD4+ count system. The accuracy of the BD FACSPresto machine in comparison to BD FACSCount was evaluated. Statistical analysis was carried out using STATA (version 12). Bland-Altman method and correlation analysis were used to analyze agreement between both measurements. In addition, sensitivity and specificity of both measurements were determined. Statistical significance was set at p-value <0.05. Results The mean bias and limit of agreement for CD4+ count between BD FACSPresto and BD FACS count machine were 7.49 (95% CI: 2.44 to 12.54) and -8.14 to 96.39 respectively. Further analysis revealed close agreement between BD FACSPresto and BD FACSCount with no significant difference between the two methods (p = .0.95). Using a threshold of 500 cells/μL, sensitivity and specificity of BD FACSPresto were 95.1% and 97.1% respectively, compared to BD FACSCount. There was no statistically significant difference in the misclassification between BD FACSPresto and BD FACSCount results (p = 0.23). Furthermore, sensitivity and specificity were similar when BD FACSPresto machine was operated by a nurse or laboratory scientist, there was no substantial difference in testing variability observed between laboratory and non-laboratory operators using the BD FACSPresto analyzer. Conclusions Overall, BD FACSPresto Point of Care CD4+ count finger stick capillary blood results is a reliable method in comparison to venous sample cytometry method and no significant difference variability observed between laboratory personnel and non-laboratory operators. The BD FACSPresto is simple, more robust and easy to use equipment without significant variability in reliability by non-laboratory health care workers hence will be a valuable instrument in increasing access and coverage of CD4 estimations in developing countries. The introduction of the BD FACSPresto POC analyzer has a high potential in reducing patients waiting time and improving the overall quality of ART service and clients' satisfaction especially in rural settings. The study was conducted at Murtala Mohammed Specialist Hospital (MMSH), Kano, one of the secondary hospitals located in Kano state, North-western Nigeria. The site has been supported by FHI 360 (Family Health International) to provide comprehensive ART program since June 2006 till date with funding from President’s Emergency Plan for AIDS Relief (PEPFAR) through United States Agency for International Development (USAID). The site has an average of over 6,000 out-patients recorded daily. It is the largest hospital owned by the Kano State Government (State Ministry of Health) and has the highest client volume. Of the 18 ART facilities that met the inclusion criteria, MMSH was purposively selected. In this facility, patients who present for the first time are offered HIV testing and counseling. Those that test HIV positive are enrolled into the ART program. The PMTCT services is integrated into antenatal clinic (ANC) and maternal child health (MCH) services. Women testing HIV positive are enrolled into the PMTCT program. The laboratory supports HIV screening and disease monitoring tests such as CD4+ enumeration, toxicity and opportunistic Infections (OI) assays. The laboratory has BD FACSCount analyzer (Beckton Dickinson, USA) which was used as the reference method in this study. The instrument is registered for the external proficiency assessment scheme of National Health Laboratories (NHLS) South Africa. The BD FACSCount analyzer at MMSH uses CD4 absolute and CD4% count floppy diskette software. Prior to this study, two laboratory scientists and a nurse from the PMTCT clinic were trained for 2 days on CD4 count estimation using BD FACSPresto analyzer, finger prick and venous sample collection and biosafety. The BD FACSPresto™ analyzer and reagents were provided by BD Bioscience USA. Study participants were enrolled between 30th September 2015 to March 2016. Informed, consenting patients 18 years and above who tested positive for HIV infection, attending MMSH ART clinic and pregnant women 18 years and above, currently attending PMTCT clinic were considered eligible for the study. A total of 300 HIV infected patients were recruited and relevant bio data information collected. Patients who did not sign consent form were disallowed from participating in the study but not denied any relevant service. Similarly, HIV-positive individuals attending ART clinic, pregnant adolescent girls attending PMTCT clinic below 18 years of age, and those who decline finger prick collection for the BD FACSPresto study were not included in the study. This study investigated validity of POC BD FACSPresto™ CD4 analyzer for CD4+ cell count using finger prick capillary samples in comparison to a reference standard flow cytometry method, BD FACSCount using same participants venous sample. The study was designed within ART and PMTCT clinics using paired measurements of same patient sample on both BD FACSPresto and BD FACSCount system. At the ART clinic, 150 consenting patients were sent to the laboratory where venous blood sample was collected as part of routine laboratory monitoring for HIV positive patients by two trained laboratory staff. Patients who were finger pricked for BD FACSPresto testing were also bled for routine CD4 testing on BD FACSCount analyzer. The venous EDTA blood sample collection is part of routine laboratory monitoring for all HIV positive patients. Samples were drawn for routine monitoring of patient who declined participation in the BD FACSPresto study while treatment decisions for all patients were made based on the CD4 values obtained from the BD FACSCount analyzer. The BD FACSPresto™ analyzer result was used for research purposes only and the results were not handed over to the client nor recorded in the laboratory order and request form from the clinic. At the PMTCT clinic, 150 finger pricked capillary samples were tested on BD FACSPresto analyzer by a trained Nurse while matching samples was drawn by the Nurse using vacutainer EDTA tubes and sent to the hospital laboratory for comparison with the reference flow cytometer BD FACSCount. The health care workers (HCWs); two laboratory scientists and a PMTCT nurse, who provided CD4 testing related services to patients in the study were interviewed to assess provider acceptability. Finger-prick blood sample (capillary) was collected aseptically from fingertip using 1.8 mm depth lancet finger stick (Becton Dickinson Bioscience) and transferred to the BD FACSPresto labelled cartridge. The cartridge cap was closed and then placed on the BD FACSPresto work station outside the instrument for 18 minutes to allow incubation at room temperature. BD FACSPresto test cartridge contains in built control features to check the analyzer and reagent functionality daily. After incubation, the test-strip was removed and the cartridge was inserted into the analyser to read the result which takes about 4 minutes. The absolute CD4 count, %CD4 results are displayed on the screen and printed automatically. The print out result was stored at the respective point of testing (laboratory and PMTCT clinic) and not used for patients’ clinical management. Testing on BD FACSPresto analyzer was performed based on manufacturer instructions. In the laboratory, venous blood samples collected were tested on BD FACSCount analyzer. Briefly, BD FACSCount CD4 reagents tube was brought to ambient temperature and vortexed upright for 10 seconds before it was opened for use. Fifty (50) μl of whole blood was added to the CD4 reagent tube containing CD3/CD4 PE monoclonal antibody (Becton Dickinson, USA). The tube was incubated in the dark for 30 minutes at room temperature and 50μl of fixative (5% formaldehyde in PBS) was added and vortexed before reading on Becton Dickson FACS machine according to manufacturer’s instruction using CD4 Absolute and CD4% count software [19]. To ensure reproducibility and precision of CD4 count testing, daily quality controls were run for the BD FACSCount and BD FACSPresto. The BD FACS count machine quality control pack (BD Biosciences, San Jose, CA) was analyzed by running low, medium and high bead count following manufacturer’s procedures. The outcome reading of “passed control” indicated the testing process was under control such as reagents, equipment, personnel and standard operation procedures was followed before clients’ sample were tested. For BD FACSPresto, daily quality control was ensured by analyzing CD Chex Plus BC low and normal control (Streck, Omaha, NE). Data was entered into Microsoft Excel and exported into STATA version 12 for analysis. The variable of interest was absolute CD4 count. Absolute CD4 count was reported as median values with accompanying interquartile range. Pearson correlation was calculated for each pair of results generated by BD FACSPresto and BD FACSCount. The accuracy of the BD FACSpresto machine in comparison to the BD FACSCount was evaluated using the Bland-Altman method, and assessed statistically with Pitman’s test of difference in variance. Sensitivity, specificity and misclassification rates of BD FACSPresto were calculated at CD4 threshold of 500 cells/μl (the 2013 eligibility threshold by WHO before the current UNAIDS Test and Start guideline) compared with results of BD FACSCount. McNemar test was performed to determine any significant differences in the misclassification between BD FACSPresto and BD FACSCount. Statistical significance was set at p-value<0.05. The study protocol was reviewed and approved by FHI 360 Office of International Research Ethics (OIRE), North Carolina, IRB No 614211–1. Ethical approval was also received from Kano State Hospital Management Board, Ministry of Health with reference: HMB/GEN/488/VOL1.