Skip to main content

Advertisement

Table 5 Multivariable-adjusted regression models associating CSF iron biomarkers with demographic and HIV disease-related factors in CHARTER study participants

From: Cerebrospinal fluid (CSF) biomarkers of iron status are associated with CSF viral load, antiretroviral therapy, and demographic factors in HIV-infected adults

Partially adjusted (age, sex, race/ethnicity) Multivariable-adjusted β (95% CI) p-value
CSF iron
 Age 0.004 (−0.007, 0.015) n.s.
 Sex (women vs. men) 0.102 (−0.134, 0.337) n.s.
 Race/ethnicitya 0.104 (0.037, 0.171) <0.01
  W compared to B 0.329 (0.130, 0.529) <0.01
  W compared to H −0.014 (−0.318, 0.289) n.s.
  H compared to B 0.327 (−0.019, 0.674) 0.01
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model
Fully adjusted (age, sex, race/ethnicity, ART, ZDV use, plasma virus detectability, and CSF VL) Multivariable-adjusted β-estimate (95% CI) p-value
CSF iron
 Detectable plasma virus (Y vs. N) 0.003 (−0.222, 0.230) n.s.
 CSF HIV RNA concentration (VL) 0.072 (−0.079, 0.223) n.s.
 ART (on vs. off) 0.141 (−0.149, 0.431) n.s.
 Current ZDV use (Y vs. N) 0.037 (−0.206, 0.280) n.s.
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model
Partially adjusted (age, sex, race/ethnicity, ZDV use) Multivariable-adjusted β (95% CI) p-value
CSF transferrin
 Age 0.012 (0.001, 0.022) 0.03
 Sex (women vs. men) −0.225 (−0.445, 0.006) 0.04
 Race/ethnicitya 0.036 (−0.025, 0.097) n.s.
  W compared to B 0.110 (−0.078, 0.297) n.s.
  W compared to H 0.103 (−0.185, 0.390) n.s.
  H compared to B 0.070 (−0.233, 0.372) n.s.
 Current ZDV use (Y vs. N) −0.254 (−0.467, −0.040) 0.03
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model
Fully adjusted (age, sex, race/ethnicity, ART, ZDV use, plasma virus detectability and CSF VL) Multivariable-adjusted β (95% CI) p-value
CSF transferrin
 Detectable plasma virus (Y vs. N) −0.172 (−0.376, 0.032) n.s.
 CSF HIV RNA concentration (VL) 0.297 (0.145, 0.450) <0.01
 ART (on vs. off) 0.153 (−0.107, 0.413) n.s.
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model
Partially adjusted (age, sex, race/ethnicity, ZDV use) Multivariable-adjusted β (95% CI) p-value
CSF ferritin
 Age 0.017 (0.006, 0.027) <0.01
 Sex (women vs. men) −0.324 (−0.550, −0.097) <0.01
 Race/ethnicitya 0.020 (−0.042, 0.081) n.s.
  W compared to B 0.072 (−0.121, 0.264) n.s.
  W compared to H 0.017 (−0.297, 0.332) n.s.
  H compared to B 0.049 (−0.235, 0.333) n.s.
 Current ZDV use (Y/N) −0.196 (−0.413, 0.021) n.s.
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model
Fully adjusted (age, sex, race/ethnicity, ART, ZDV use, plasma virus detectability, and CSF VL) Multivariable-adjusted β (95% CI) p-value
CSF ferritin
 Detectable plasma virus (Y vs. N) −0.242 (−0.449, −0.035) 0.02
 CSF HIV RNA concentration (VL) 0.142 (−002, 0.286) 0.05
 ART (on vs. off) −0.017 (−0.285, 0.251) n.s.
  1. The natural logarithm of each biomarker variable was used in all analyses due to non-normal distributions. For each biomarker, partially and fully adjusted separate multiple linear regression models are shown, which included the covariates listed, with β-estimates and their 95% confidence intervals and p-values. p-values <0.05 are statistically significant
  2. For each biomarker, significance of covariates in partially adjusted models did not change appreciably with full adjustment
  3. W Whites, B Blacks, H Hispanics (self-reported ancestry), VL viral (HIV RNA) load, ZDV zidovudine, ART antiretroviral therapy, CI confidence interval, CSF cerebrospinal fluid
  4. aFor each dichotomous comparison by race/ethnicity (ancestry), only one ancestry term (e.g., whites compared to black individuals) was entered into the model