In this study we investigated the possible association between total daily iron intake during pregnancy, haemoglobin in early pregnancy and the risk of gestational diabetes. To our knowledge there are no similar studies published previously. We discovered that there was a tendency for a higher incidence of GDM in the highest fifth of total iron intake. High iron intake increased the risk of GDM especially for those women who had good haemoglobin levels in early pregnancy: we discovered a twofold OR for GDM for women in the highest fifth of iron intake with a haemoglobin level of over 120 g/l in early pregnancy.
Our study has several strengths. First unlike most of the previous studies we had detailed information of iron intake from both food and supplements. Second, we had information on baseline oral glucose tolerance test (OGTT) and could therefore exclude women with undiagnosed pre-pregnancy diabetes. Additionally, Finnish Medical Birth Registry could be utilised to cover those GDM cases that were undiagnosed in the 26–28 weeks’ OGTT. We also had information of haemoglobin measurements of the study population during pregnancy although measured by nurses. All of our data was collected prospectively. The participation rate in the original trial was high (88% in both the intervention and the usual care groups) improving the generalisability of the results to other women with risk factors for GDM.
A limitation of this kind of study is that estimating iron intake during pregnancy can be challenging because dietary habits and use of supplements can vary a lot during pregnancy due to nausea and other changes in well-being. Iron intake was assessed at a certain point in time (at 26–28 weeks’ gestation covering the previous month), thus it does not cover the intake during the whole pregnancy. In our study, we decided to use information of iron intake in mid-pregnancy to get the best estimation on iron intake during pregnancy and before the onset of GDM. We did not use information of iron intake abstracted in the beginning of pregnancy because it covered only dietary habits and supplement use during 1 month before the pregnancy when use of supplemental iron is rare. However, our information on iron intake covers only 1 month in the mid-pregnancy. Additionally we could not assess iron absorption, for example with blood measurements, which also can vary a lot depending on body iron status and the contents of meals.
There are also limitations concerning the haemoglobin measurements used in our study. In the maternity clinics haemoglobin levels are usually screened with a capillary haemoglobin measurement using finger-stick samples. This method has been demonstrated to be reliable in determining haemoglobin values but it is however susceptible to handling errors30
and therefore there can be variation in the test results. Additionally, we had a lot of missing values in haemoglobin measurements and thus we decided to use the haemoglobin measurement at 16–18 weeks’ gestation for those with missing values at weeks 8–12 (n=28). Although measurements at weeks 16–18 do not present the haemoglobin levels at early pregnancy our objective was to analyse separately women who were not anaemic and thus including women who had a haemoglobin level of over 120 g/l still at weeks 16–18 would help us not to underestimate the proportion of women who were not anaemic in the beginning of the pregnancy. However, for these reasons our results with respect to haemoglobin measurements should be considered with caution and further studies with reliable and comprehensive haemoglobin measurements are warranted to confirm our results.
In a somewhat similar study Bo et al22
discovered that women who used supplemental iron during mid-pregnancy had a twofold to threefold risk of GDM. They however had no information of dietary intake of iron or haemoglobin levels or other measurements of iron stores. Their findings are however in line with our results which suggest that dietary intake could play a relatively small part in the association whereas high supplemental iron intake could be responsible for most of the increase in GDM risk. Somewhat different results have been reported recently by Bowers et al21
and Qiu et al.23
Bowers et al21
investigated dietary and supplemental iron intake during 1 year before pregnancy utilising the material from the Nurses Health study. They observed no significant effect of total, non-heme or supplemental iron intake on the risk of GDM. It can be argued though that usually iron supplementation outside of pregnancy is rare and concerns mainly those who are anaemic. However, they did find a significant and positive association between haeme iron intake and GDM. Similarly Qiu et al23
demonstrated an association between haeme iron intake during the time before conception and in early pregnancy and the risk of GDM. To our knowledge only one placebo-controlled clinical trial has been conducted to investigate the association between iron supplementation and GDM.24
This study did not observe any association. However, it can be argued that the intake of supplemental iron was quite low in this study because of only about 50% compliance to a daily supplement of 60 mg of elemental iron.
Iron is a highly reactive component with a possibility to participate in harmful reactions.31
The human body can excrete iron with very limited mechanisms and thus iron intake is highly regulated according to body iron needs.32
Iron could interfere with glucose metabolism in several ways. For example, following mechanisms have been proposed: Iron decreases insulin extraction and metabolism in the liver, which leads to peripheral hyperinsulinaemia.33
Iron overload results in oxidative stress in pancreatic β-cells, which leads to destruction of the pancreatic islets and thus decreases insulin secretion.34
The exact mechanisms, which link iron to diabetes are still unsolved. However, the association can be argued to be biologically plausible.
Accumulating body of evidence supports the hypothesis of excess iron as a risk factor for glucose metabolism disorders such as gestational diabetes. Iron deficiency anaemia has been shown to reduce the incidence of GDM.19
Respectively high haemoglobin level (>130 g/l) in early pregnancy has been demonstrated to be an independent risk factor for GDM.18
It seems that iron can affect glucose metabolism even with no overt iron overload. Women diagnosed with GDM have been observed to have increased iron stores compared to women without GDM.15
However, many of these studies have been cross-sectional and it has been criticised that higher serum ferritin levels could in fact reflect inflammation in the body and could be rather a result than a cause for diabetes.16
In summary it seems that high iron intake might be a factor, which increases the risk of GDM especially in women with already good iron stores. Use of iron supplements during pregnancy is common even in women with good haemoglobin levels. Our results suggest that routine use of iron supplements should be reconsidered in non-anaemic women with risk factors for GDM. To confirm the hypothesis there is need for a large prospective study—ideally a randomised controlled trial—with reliable and comprehensive information on iron intake from food and supplements during pregnancy accompanied with serum measurements to determine the level of body iron stores.