CD4 Lymphocyte Enumeration and Hemoglobin Assessment Aid for Priority Decisions: A Multisite Evaluation of the BD FACSPresto™ System

Abstract

Background: The BD FACSPresto^™ system uses capillary and venous blood to measure CD4 absolute counts (CD4), %CD4 in lymphocytes, and hemoglobin (Hb) in approximately 25 minutes. CD4 cell count is used with portable CD4 counters in resource-limited settings to manage HIV/AIDS patients. A method comparison was performed using capillary and venous samples from seven clinical laboratories in five countries. The BD FACSPresto system was assessed for variability between laboratory, instrument/operators, cartridge lots and within-run at four sites.

Methods: Samples were collected under approved voluntary consent. EDTA-anticoagulated venous samples were tested for CD4 and %CD4 T cells using the gold-standard BD FACSCalibur^™ system, and for Hb, using the Sysmex^® KX-21N^™ analyzer. Venous and capillary samples were tested on the BD FACSPresto system. Matched data was analyzed for bias (Deming linear regression and Bland-Altman methods), and for concordance around the clinical decision point. The coefficient of variation was estimated per site, instrument/operator, cartridge-lot and between-runs.

Results: For method comparison, 93% of the 720 samples were from HIV-positive and 7% from HIV-negative or normal subjects. CD4 and %CD4 T cells venous and capillary results gave slopes within 0.96-1.05 and R² ≥0.96; Hb slopes were ≥1.00 and R² ≥0.89. Variability across sites/operators gave %CV <5.8% for CD4 counts, <1.9% for %CD4 and <3.2% for Hb. The total %CV was <7.7% across instrument/cartridge lot.

Conclusion: The BD FACSPresto system provides accurate, reliable, precise CD4/%CD4/Hb results compared to gold-standard methods, irrespective of venous or capillary blood sampling. The data showed good agreement between the BD FACSPresto, BD FACSCalibur and Sysmex systems.

Keywords: CD4; Capillary; HIV-1 diversity; HIV/AIDS; Hemoglobin; Precision; Recent infections; Venous.

Fig. (1)

Clinical evaluation of the BD FACSPresto^™ system flowchart. Evaluation of the performance of the BD FACSPresto system using venous and capillary blood specimens from subjects attending a routine clinic visit.

Fig. (2)

CD4 absolute counts, %CD4, and Hb bias in venous and capillary blood. Bland-Altman plots illustrate the biases for venous (2A, 2C, and 2E) and capillary (2B, 2D, and 2F) samples with limits of agreement. Biases for CD4 cell counts are shown in 2A and 1B, for %CD4 cells in 2C and 2D, and hemoglobin in 2E and 2F. The x-axis displays the average (CD4 counts, %CD4 cells, or Hb) and the y-axis is the difference (CD4 counts, %CD4 cells, or Hb).

Fig. (3)

Deming regression analysis, Deming regression plots for CD4 cell counts, %CD4 and Hb in venous and capillary blood. BD FACSPresto vs BD FACSCalibur or Sysmex systems. Deming regression results are depicted for venous blood (3A, 3C, and 3E) and capillary blood (3B, 3D, and 3F). The CD4 count results are shown from weighted Deming regression in 3A and 3B; for %CD4 cells and Hb, the unweighted Deming regression in 3C and 3D and in 3E and 3F respectively. The x-axis displays the predicate method for CD4 cell counts, %CD4 cells, or Hb, and the y-axis corresponds to the BD FACSPresto system. Plots show the R², slope, and the intercept values for each parameter. The identity regression line is in gray and the calculated regression is in red.