Single-platform, volumetric, CD45-assisted pan-leucogating flow cytometry for CD4 T lymphocytes monitoring of HIV infection according to the WHO recommendations for resource-constrained settings

BMC Research Notes

Table 2 CD4 T cell counting in absolute count and percentage in 258 HIV-1-infected individuals, by the Apogee Auto40 flow cytometer at the hôpital Militaire D’Instruction, N’Djamena, Chad, and by the FACSCalibur at the hôpital de la Liberté, N’Djamena, at various CD4 T cell count ranges according to the FACSCalibur results

Categories		< 200 cells/μl	200 – 350 cells/μl	> 350 cells/μl
Number		N = 27	N = 43	N = 188
Absolute CD4 T cells (cells/ μl)*
Apogee Auto40	Mean ± SD	118 ± 161	278 ± 129	2990 ± 3745
FACSCalibur	Mean ± SD	100 ± 135	215 ± 103	3017 ± 3756
Absolute bias (limits of agreement)^a	Mean (±1.96 SD)	−4.0 (−33.8;+25.8)	10.0 (−70.0;+90.0)	12.0 (−160.7;+184.7)
Relative (%) bias (limits of agreement)^b	Mean (±1.96 SD)	62.0 (−134.0;+258.0)	7.0 (−27.2;+41.2)	−3.0 (−12.4;+6.4)
Percent CD4 T cells (%CD4)**
Apogee Auto40	Mean ± SD	10.0 ± 12.7	19.5 ± 6.3	55.1 ± 44.5
FACSCalibur	Mean ± SD	9.0 ± 11.3	18.0 ± 4.2	53.3 ± 41.7
Absolute bias (limits of agreement)^a	Mean (±1.96 SD)	1.0 (−1.8;+3.8)	2.0 (−2.3;+6.3)	2.0 (−3.4;+7.4)
Relative (%) bias (limits of agreement)^b	Mean (±1.96 SD)	6.0 (−10.3;+22.3)	7.0 (−13.2;+27.2)	2.0 (−4.6;+8.6)

* Results in absolute CD4 T cell counts by Auto40 and FACSCalibur were correlated using the non-parametric Passing–Bablok regression analysis (< 200 cells/μl: r = 0.86, slope = 1.11, intercept = −12.0; 200 – 350 cells/μl: r = 0.68, slope = 1.31, intercept = −96.5; > 350 cells/μl: r = 0.98, slope = 1.02, intercept = −24.8);
** Results in percentage CD4 T cell counts by Auto40 and FACSCalibur were correlated using the non-parametric Passing–Bablok regression analysis (< 200 cells/μl: r = 0.89, slope = 1.00, intercept = 0.0; 200 – 350 cells/μl: r = 0.89, slope = 1.00, intercept = 0.0; > 350 cells/μl: r = 0.95, slope = 1.00, intercept = 0.0);
^a The Bland Altman analysis was carried out to calculate the absolute bias and limits of agreement which are the 95% confidence intervals (±1.96 × SD) of the mean bias of all paired measurements in a given category [27];
^b The Pollock analysis was carried out to calculate the relative bias and limits of agreement which are the 95% confidence intervals (±1.96 × SD) of the mean bias of all paired measurements in a given category [28].
SD: Standard deviation.