Skip to main content
. 2017 Jan 25;12(1):e0170857. doi: 10.1371/journal.pone.0170857

Table 4. Associations between CD4/CD8 ratio categories (relative to ratio>1.0) with DLCO and FEV1 in linear regression models.

Outcome CD4/CD8<0.4 CD4/CD8 0.4–1.0
DLCO %-predicteda
  • Unadjusted model

  • Adjusted modelb

  • Adjusted model plus CD4 and HIV viral loadc

  • Adjusted model plus CD4, HIV viral load, nadir CD4 and sCD14d


  • -11% (-17%, -4.5%)*

  • -8.7% (-14%, -3.1%)*

  • -8.7% (-14%, -2.9%)*

  • -9.0% (-17%, -1.4%)*


  • -5.5% (-12%, 0.7%)

  • -5.0% (-10%, 0.36%)

  • -5.0% (-10%, 0.47%)

  • -5.7% (-12, 0.67)

FEV1%-predicteda
  • Unadjusted model

  • Adjusted modelb

  • Adjusted model plus CD4 and HIV viral loadc


  • -5.1% (-13%, 3.1%)

  • -7.4% (-16%, 1.0%)

  • -3.4% (-14%, 6.7%)


  • -3.1% (-11%, 4.9%)

  • -5.1% (-13%, 2.9%)

  • -3.7% (-12.%, 4.7%)

aEstimate of difference in %-predicted values (95% CI) for CD4/CD8 ratios relative to ratio >1.0

bModel adjusted for age, sex, race/ethnicity, BMI, pack-years of tobacco use, chronic diseases (chronic heart disease, anemia, diabetes, hypertension) and inhalational and intravenous drug use

cAdjusted model with additional inclusion of CD4 cell count and viral load

dAdjusted model with additional inclusion of CD4 cell count, viral load, nadir CD4 and soluble CD14

*P<0.05

DLCO = diffusion capacity of carbon monoxide, FEV1 = forced expiratory volume in 1 second