Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Antivir Ther. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
Antivir Ther. 2010; 15(4): 571–577.
doi: 10.3851/IMP1557

Table 2

Factors Associated with Hemoglobin A1c Values in HIV-Infected and HIV-Uninfected Diabetic Women from Bivariate and Multivariate Linear Regression of Log Hemoglobin A1c

VariableBivariate modelaMultivariate modelb

Percent Changec95% Confidence IntervalP- valuePercent Changec95% Confidence IntervalP- value

HIV-infected-1.32-2.56, -0.060.040-0.28-1.36, 0.810.617

Age (per 10 years)0.22-0.39, 0.830.483

 White-2.89-4.16, -1.60< 0.001-1.80-2.99, -0.590.004
 Other-2.04-3.42, -0.700.003-1.31-2.37, -0.240.017

Family history of diabetes mellitus1.08-0.06, 2.240.063

Diabetic medication6.104.83, 7.38< 0.0015.714.50, 6.94< 0.001

BMI category (kg/m2)
< 25referent
≥ 25 – 29-0.50-2.03, 1.050.523
≥ 301.700.29, 3.120.018

Hemoglobin (g/dL)-0.37-0.70, -0.030.033

Mean corpuscular volume (per 10 fL)-1.86-2.30, -1.42< 0.001-1.61-2.01, -1.20< 0.001

Albumin (g/dL)-0.89-1.97, 0.200.109

HCV RNA positive-0.71-1.90, 0.500.249
aAdjusted for log plasma glucose concentration
bFinal model from stepwise multiple regression model of all variables in this table (including log fasting plasma glucose and HIV serostatus, which were forced into the model)
cReflects the percentage change on a multiplicative scale relative to the referent group, obtained as 100(expβ-1) where β is the regression coefficient (or corresponding lower or upper confidence limit for this coefficient) using log-transformed hemoglobin A1c as the outcome