Absolute lymphocyte count as a predictor of CD4 count

To determine whether the absolute lymphocyte count (ALC) (white blood count x lymphocyte percentage) can be used to predict a low CD4 count.

We conducted a retrospective data analysis of consecutive CD4 count analyses performed between January 1, 1995, through December 1, 1995, at an urban university teaching hospital. Results of consecutive CD4 counts and simultaneously measured ALCs were analyzed from samples obtained in inpatient, clinic, and emergency department settings. The ability of ALC to predict a CD4 count less than 200 cells/mm3 was analyzed by calculating sensitivities, specificities, predictive values, and likelihood ratios for a range of ALC values.

Among the 807 samples, 322 results (40%) had a CD4 count less than 200 cells/mm3. The ALC and CD4 count were correlated (r=.69, P<.0001). An ALC less than 1,000 cells/mm3 predicted CD4 counts less than 200 cells/mm3 with a sensitivity of .67 (95% confidence interval .62 to .72), specificity of .96 (.94 to .98), positive predictive value of .91 (.87 to .95), and a negative predictive value of .81 (.78 to .84). An ALC less than 2,000 cells/mm3 predicted CD4 counts less than 200 cells/mm3 with a sensitivity of .97 (.95 to .99), specificity of .41 (.37 to .45), positive predictive value of .52 (.48 to .56), and negative predictive value of .95 (.92 to .98).

A reliable relationship exists between ALC and CD4 count. In a similar population, an ALC less than 1,000 cells/mm3 is predictive of a CD4 count less than 200 cells/mm3, and an ALC greater than or equal to 2,000 cells/mm3 is predictive of a CD4 count greater than or equal to 200 cells/mm3. Physicians may find these criteria useful in identifying patients with increased risk of opportunistic infection.