Depression severity is associated with increased risk behaviors and decreased CD4 cell counts

Abstract

Depression is a common comorbidity among HIV-infected individuals. We studied the relationship between depressive symptoms, risk behaviors (risky-sexual behavior, tobacco, alcohol, and illicit drug use) and HIV outcomes. This cross-sectional study conducted in 2009 at the Washington University HIV Clinic included screening for depression with patient health questionnaire, survey of sexual behavior, illicit drug, alcohol, and tobacco use within 30 days. Sociodemographics, plasma HIV RNA levels, CD4 cell counts, and sexually transmitted disease test results were obtained from medical records. Multivariate logistic and linear regression models were used to assess the association between depressive symptoms severity and risk behaviors, HIV outcomes and combination antiretroviral therapy (cART) adherence. A total of 624 persons completed the assessment of whom 432 (69%) were male and 426 (68%) African-American. The median CD4 cell count was 410 cells/mm(3) and 479 persons (77%) were on cART of whom 112 (23%) had HIV RNA level > 400 copies/mL. Overall, 96 (15%) had symptoms of major depressive disorder. Depressive symptom severity was associated with increased likelihood of high-risk drinking (odds ratio [OR], 2.4; 95% confidence interval [CI], 1.1-5.1), current tobacco use (OR, 1.8; 95% CI, 1.1-2.9), illicit drug use (OR, 1.7; 95% CI, 1.0-2.8), and risky-sexual behavior (OR, 1.5; 95% CI, 0.8-2.7). Suboptimal cART adherence (visual analog scale < 95%) was also associated with depressive symptoms severity (p < 0.05). After adjustment for age, sex, race, receipt of cART, and cART adherence, depressive symptoms severity was independently associated with lower CD4 cell count (p < 0.05) but not with higher HIV RNA level (p = 0.39). Depression adversely affects HIV-infected individuals, requiring greater effort at utilizing multidisciplinary interventions.

Keywords: HIV; adherence; antiretroviral therapy; depression; risk behaviors.

Figure 1

Probability of each risk behaviors estimated by PHQ-9 scores. Fitted logistic regression estimating the probability of engaging in each risk behaviors by PHQ-9 scores. Jittered dots represent the PHQ-9 score data. These multivariate models are adjusted by age, sex, race, current CD4 count, Log HIV viral load, receipt of cART and cART adherence.

Figure 2

Adjusted odds ratios for depression (PHQ-9 ≥ 10) predicting each risk behavior. Adjusted by age, sex, race, HIV RNA level, current CD4 cell count, receipt of cART and cART adherence. 95% CI, reference for the odds ratios is no depression (PHQ-9 score of 0).

Figure 3

Probability of cART adherence < 95% in the multivariate model. Fitted logistic regression estimating the probability of cART adherence < 95%. Jittered dots represent PHQ-9 score data. This multivariate model includes age, race, tobacco use and receipt of PI-based vs. NNRTI-based cART.

Figure 4

Linear regression models demonstrating the association between depression severity and HIV RNA level (top) or CD4 cell count (bottom) in unadjusted (left) or adjusted (right) analyses. These multivariate models are adjusted by age, sex, race, current CD4 count, Log HIV viral load, receipt of cART and cART adherence. Solid line shows the fitted regression model and light lines indicating uncertainty in the fitted regression with posterior simulations.