Background: Discordant absolute CD4 count and CD4 percent (CD4%) results often complicate treatment decisions regarding need for OI prophylaxis and initiating antiretroviral therapy.

Methods: Using an observational, urban cohort of HIV patients (pts) from Maryland, we calculated risk of developing an AIDS-defining OI associated with a specific CD4 and CD4%. All CD4-CD4% pairs obtained after 1/96 were used. Development of an OI was assessed over 6 mos (med 70 days) after the CD4-CD4% pair was obtained. CD4 was categorized as < 50, 50–150, 151–250, and 250–350 cells/mm³. We categorized CD4% as < 7, 7–14, 15–21, and > 21%. Negative binomial regression using a generalized estimating equation to adjust for repeated measures was used to compute the incidence rate ratio (IRR) for developing an OI by CD4 and CD4% categories. We adjusted for sex, race, HIV risk factor, and use of HAART.

Results: We analyzed 15,736 CD4-CD4% pairs from 2,184 pts who had 608 OIs. The IRR for developing an OI was by CD4 was 24.2 (95% CI: 16.0, 36.5) for < 50, 6.2 (4.0, 9.5) for 50–150 and 2.7 (1.7, 4.4) for 150–250, compared to 250–350. Independently, the IRR by CD4% was 14.4 (9.3, 22.6) for < 7, 3.7 (2.4, 5.9) for 7–14, 1.9 (1.1, 3.1) for 15–21, compared to > 21. However, adjusting for absolute CD4, the CD4% was no longer associated with OI development. Adjusted IRR by CD4% was 1.41 (0.82, 2.55) for < 7, 1.21 (0.72, 2.04) for 7–14, 1.26 (0.75, 2.10) for 14–21, compared to > 21. The IRR for development of an OI is shown in the figure below for each CD4-CD4% category. The IRR shown is the relative increase in OI rate for each CD4-CD4% category, compared to CD4 250–350 with CD4% > 21. Within each CD4 stratum, there were no significant differences in the IRR based on CD4%. (Note: CD4 250–350 with CD4% < 7, and CD4 < 50 with CD4% > 21 could not be computed because of lack of data.)

Conclusions: Our results suggest that CD4 count is the most useful direct measure of the risk of development of OI. The CD4% added no further predictive information after accounting for the CD4 count, suggesting that CD4 count alone can be used in making treatment decisions.