W. Withold et al., EFFICACY OF TRANSFERRIN DETERMINATION IN HUMAN SERA IN THE DIAGNOSIS OF IRON-DEFICIENCY, European journal of clinical chemistry and clinical biochemistry, 32(1), 1994, pp. 19-25
Apparently healthy persons (n = 425) as well as 264 patients character
ized by an iron concentration in serum < 7.2 mu mol/l were examined. A
latent iron deficiency was defined as a concentration of ferritin < 2
0 mu g/l (males) and < 15 mu g/l (females), without anaemia; manifest
iron deficiency defined by an additional presence of hypochromic micro
cytic anaemia. Fifty-nine of 425 (= 14%) apparently healthy persons sh
owed a latent iron deficiency. In the remaining 366 we established the
following reference intervals for the concentration of transferrin in
serum [mu mol/l]: 25.2-45.3 (males), 29.1-54.5 (females, less than or
equal to 25 years of age) and 25.3-48.6 (females, > 25 years of age).
Eight of 59 (= 14%) apparently healthy persons with latent iron defic
iency had a transferrin concentration above the reference interval. Si
xty-one of 264 (= 23%) patients with an iron concentration < 7.2 mu mo
l/l showed a ferritin concentration < 20 mu g/l (males) and < 15 mu g/
l (females). Thirty-eight of these 61 patients (= 62%) had a manifest
iron deficiency. In 18 of these 38 patients (= 47%) the transferrin co
ncentration was increased. For our 264 patients we determined the diag
nostic validity of an increased transferrin concentration for diagnosi
s of iron deficiency, assuming an iron deficiency if the concentration
of ferritin remained below the discrimination values mentioned above:
The diagnostic sensitivity was 36%, the diagnostic specificity 97%, t
he predictive value of the positive test result 79% and the predictive
value of the negative test result 83%. It can be concluded that the d
iagnostic validity of transferrin determination is inferior to that of
ferritin measurement with respect to the diagnosis of iron deficiency
. This especially applies to diagnostic situations in which the preval
ence of iron deficiency is far lower than in the present study: e.g.,
given a prevalence of iron deficiency anaemia of 0.81% as it can be ob
served in the general population the predictive value of the positive
test result only amounts to 11.3%. Thus, determination of transferrin
concentration does not yield further information, so that this test sh
ould no longer be used in the diagnosis of iron deficiency.