Skip to main content
. 2017 Jun 28;2017(1):109–119. doi: 10.1093/emph/eox010

Table 2.

Logistic regression models evaluating associations between iron status and prevalent infectious disease (reference: iron replete)

Model 1:
 Model 2:
 Model 3:
Crudea
 Adjusted for identified confounding variablesb
 Adjusted for all individual characteristicsc
OR 95% CI P-value aOR 95% CI P-value aOR 95% CI P-value
Outcome: physician’s diagnosis of any infectious disease
    Iron-deficient erythropoesis 0.69 0.32, 1.51 0.359 0.84 0.34, 2.06 0.698 0.79 0.31, 2.01 0.614
    Iron deficiency anemia 1.42 0.62, 3.28 0.407 2.42 0.87, 6.79 0.091 2.90 1.02, 8.30 0.047
    Non-iron deficiency anemia 1.56 0.67, 3.62 0.300 1.92 0.75, 4.94 0.174 2.10 0.80, 5.52 0.131
Outcome: physician’s diagnosis of malaria
    Iron-deficient erythropoesis 1.25 0.42, 3.70 0.691 1.37 0.40, 4.74 0.620 1.44 0.41, 5.08 0.568
    Iron deficiency anemia 1.80 0.54, 6.02 0.337 3.75 0.95, 14.76 0.059 4.45 1.10, 18.00 0.036
    Non-iron deficiency anemia 1.11 0.27, 4.51 0.884 1.41 0.34, 5.92 0.637 1.37 0.32, 5.84 0.666
Outcome: physician’s diagnosis of respiratory infectious disease
    Iron-deficient erythropoesis 0.62 0.25, 1.54 0.302 0.63 0.21, 1.92 0.413 0.51 0.15, 1.70 0.272
    Iron deficiency anemia 0.95 0.34, 2.64 0.928 1.58 0.46, 5.43 0.465 1.75 0.50, 6.07 0.379
    Non-iron deficiency anemia 1.03 0.37, 2.87 0.950 1.69 0.57, 4.99 0.345 1.87 0.62, 5.67 0.268
Outcome: elevated biomarkers of inflammationd
    Iron-deficient erythropoesis 1.00 0.55, 1.79 0.991 0.76 0.39, 1.46 0.408 0.75 0.38, 1.48 0.415
    Iron deficiency anemia 2.63 1.19, 5.80 0.017 2.25 0.87, 5.82 0.095 2.71 0.99, 7.41 0.053
    Non-iron deficiency anemia 1.57 0.74, 3.31 0.239 1.53 0.67, 3.51 0.310 1.61 0.69, 3.73 0.271
a

N = 283 for models of physician’s diagnoses; N = 265 for model of elevated biomarkers of inflammation.

b

Controlling for age, sex and triceps skin-fold thickness; N = 239 for models of physician’s diagnoses; N = 224 for model of elevated biomarker of inflammation.

c

Controlling for age, sex and triceps skin-fold thickness, early weaning, home construction materials, and household cattle ownership; N = 233 for models of physician’s diagnoses; N = 219 for model of elevated biomarker of inflammation.

d

Elevated C-reactive protein (CRP) or α1-acid glycoprotein (AGP) [28].

aOR: adjusted odds ratio; OR, odds ratio; CI, confidence interval.