Results
A total of 206 infants, consisting of 114 males (55.3%), with the mean age of 8.2 months (SD 2.0, range 6-12 months), were enrolled randomly after informed consent. Interestingly, we found 97 individuals (47%) with some form of thal minor, 39 with α-thal (18.9% of the total population), 45 with β-globin mutation mainly Hb E (21.8%), and 13 with combined α and β globin abnormalities (6.3%). None of these individuals had the genotype found in thal diseases such as HbH disease or Hb E/β-thal; therefore, they were all classified as thal minor and were subsequently used as a group for further analysis. Details of all comprehensive genotype data are shown in Table 1 . Infants with thal minor had no history of blood transfusion and no hepatosplenomegaly.
We found no significant difference in all clinical characteristics: age, gender, growth and nutrition parameters, and the iron markers such as serum ferritin, TS, and hepcidin levels. The exception was having a family history of anemia or thal being more common in infants with thal minor (43.3%) than those without thal (12.8%), as shown inTable 2. In our logistic regression analysis, having a history of anemia or thal showed an increased risk of thal minor in infants: odd ratio 5.18 (95% CI: 2.60-10.33), P =0.001. Interestingly, the number of infants with thal minor with IDA at first diagnosis was higher than those with normal globin genotypes (32.0% vs. 20.2%); moreover, the number of infants with normal iron status and ID were significantly different among infants with and without thal minor (P =0.037) (Table 2). To determine the effects of iron status (normal iron, ID, or IDA) on hematological parameters, we performed subgrouping analysis within the groups of infants with and without thal minor (Table 3 ). There were significant differences in Hb, hematocrit (Hct), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and red blood cell distribution width (RDW) within both groups suggesting the significant role of iron in determining Hb, Hct and red blood cell indices. With the exception of increased RBC counts and RDW, other RBC parameters were lowest in IDA, followed by ID and normal iron. Of note, infants with thal minor who had IDA (N=31) displayed the statistically lowest values in MCV, MCH, and highest RDW (Table 2), suggesting co-inheritance of globin mutations can have epistatic effects on hematology on top of the primary effects of iron status.
In this regard, we further compared the group of infants with normal iron and IDA to see the effects of thal minor (see Supplementary Table 1 ). Coinheritance of thal had significant effects on Hb, Hct, RBC, MCV, MCH, RDW but not MCHC (suppl. Table 1) only in those with normal iron status. However, we found only RBC and MCV to be significantly different in infants who already had IDA, suggesting this epistatic effect was operative on those only two parameters. In addition, we found an effect of iron on significantly decreased levels of Hb A2 in infants without thal minor but not in those with thal. On the other hand, the basal Hb F in infants with thal minor was generally higher than those without thal minor, suggesting a delay in globin switching, one of the consequences of globin abnormality.39,40 Finally, the iron status was not significantly associated with the levels of persistent Hb F expression within the groups of both infants with and without thal minor (Table 3 ).
To determine the effects of thal minor on hepcidin expression, we compared serum hepcidin, serum ferritin, and TS among healthy infants (no thal and normal iron status) with those having different types of thal minor as shown in Figure 1 . Details of measurements in each group are shown in Supplementary Table 2A (with iron-replete) and 2B (with iron deplete). The levels of hepcidin tended to be slightly lower in those with thal minor and lowest in those with combined α and β globin mutations (Figure 1C).This finding was consistent with a slightly increasing trend of serum ferritin (Figure 1A) and TS (Figure 1B) in those with combined thal minors, although there were no statistically significant differences. We also found no differences in these parameters as a group (normal vs . thal minor) and by different genders (male vs . female) (see Supplementary Figure 1 ).
The primary diagnosis of infants with IDA using Hb levels, SF and TS values in our study was further confirmed in the majority of those by the therapeutic response to iron therapy. Thirty-four out of 53 IDA infants with (n=31) and without thal minor (n=22) who can be reached by telephone appointments have received iron therapy (4-6 mg/kg/day) for 8-12 weeks, and all showed therapeutic response. Their hematology during fellow up visits revealed a significant increase in all RBC parameters compared to the baseline within their groups (P <0.05). Interestingly, IDA infants without thal minor (n=18) had slightly greater increment for Hb and MCV after iron therapy than IDA infants with a thal minor (n=16); however, there were no statistically significant differences between the groups (P =0.099 and 0.278, respectively). The mean Hb increment after iron therapy was 1.7 g/dL (SD 1.1, range 0.1-4.1) vs 1.1 (SD 0.9, range 0.2-3.1 g/dL) and mean MCV increment 4.9 fL (SD 3.7, range 0.3-11.8 fL) vs 3.5 (SD 3.4, range 0-9.5) fL, in IDA infants without and with thal minor respectively. This result has confirmed our diagnosis of IDA in both groups at the baseline. (Supplementary Table 3 )