Increase in CD4 count among new enrollees in HIV care in the modern antiretroviral therapy era

Abstract

Background: Earlier HIV diagnosis and engagement in care improve outcomes and is cost effective, as a result, in 2006, the Centers for Disease Control and Prevention (CDC) revised the HIV-screening guidelines. We sought to determine whether the CD4 count (CD4) at presentation, a surrogate for time to presentation, increased during the study period. Our a priori hypothesis was that the CD4 at presentation increased during the study period, particularly after the CDC guideline revision.

Methods: We performed a retrospective cohort study and analyzed data from the HIV Research Network, a consortium of 18 US clinics caring for HIV-infected patients. HIV-infected adults (≥18 years old) newly presenting for care between 2003 and 2011 were included in this study. Multivariable linear regression examined associations with CD4 at enrollment. Calendar year was modeled as a linear spline with a change in slope at 2008, allowing determination of the mean change in CD4 per year during 2003-2007 and 2008-2011.

Results: Over 13,543 newly presenting subjects enrolled from 2003 to 2011. Median CD4 at enrollment rose from 285 to 317 cells per cubic millimeter between 2003-2007 and 2008-2011 (P < 0.001). After adjusting for age, race/ethnicity, gender, HIV risk factor, and clinic site, the mean increase in the CD4 count at presentation per year was 13.3 cells per cubic millimeter per year (95% confidence interval 6.4 to 20.1 cells per cubic millimeter per year) greater during 2008-2011 than during 2003-2007.

Conclusions: We demonstrate a small, but statistically significant, increase in CD4 at presentation after the CDC guideline revision. More efforts are needed to decrease time to presentation to HIV care.

Figure 1

Mean and Predicted Mean CD4 Count at Enrollment Among New Presenters from 2003–2011 with Corresponding 95% Confidence Intervals.* *Solid lines denote mean CD4 count at enrollment predicted by the multivariable linear regression model (Table 3A) for all new presenters. Dashed lines denote actual mean CD4 count at enrollment for all new presenters. Enrollment CD4 was predicted for each subject using the adjusted MLR (Table 3A) and the mean of the predicted enrollment CD4 was plotted for each enrollment year. As the inputs to the model (the demographic and behavioral risk composition of the underlying cohort) vary from year to year, the mean predicted CD4 may not be perfectly linear when plotted against calendar year.

Figure 2

Mean and Predicted Mean CD4 Count at Enrollment Among New Presenters from 2003–2011 Stratified by Race/Ethnicity (Panel A) and Age Category (Panel B)* * Solid lines denote mean CD4 count at enrollment predicted by the multivariable linear regression model (Table 3B) for all new presenters, stratified as indicated. Dashed lines denote actual mean CD4 count at enrollment for all new presenters, stratified as indicated. Other race is not shown. Enrollment CD4 was predicted for each subject using the adjusted MLR (Table 3B) and the mean of the predicted enrollment CD4 was plotted for each enrollment year. As the inputs to the model (the demographic and behavioral risk composition of the underlying cohort) vary from year to year, the mean predicted CD4 may not be perfectly linear when plotted against calendar year.