Introduction
Gestational diabetes mellitus (GDM) is one of the most common pregnancy disorders prone to serious outcomes such as ketoacidosis, preeclampsia, macrosomia, growth restriction, fetal distress, neonatal hypoglycemia, and later cardiovascular disease [
1,
2]. The prevalence of GDM varies globally, with rates ranging from 1 to 30% depending on screening methods, diagnostic criteria, geography, and ethnicity [
1,
3]. To minimize the negative impact of GDM on pregnancy outcomes, strict glycemic control is crucial. Current antenatal care and management rely on self-managing blood glucose measurements, as recommended by the American Diabetes Association (ADA) and the International Association of the Diabetes and Pregnancy Study Groups (IADPSG), in evaluating glucose control in diabetic pregnant women [
4]. However, compliance with blood glucose monitoring is not always consistent, thus alternative methods to assess glycemic control are needed.
HbA1c levels are common clinical indicators for glycemic control in non-pregnant individuals. It reflects an average glycemia for the past 8–12 weeks, can be measured without fasting making it easier for patients and the results are relatively stable and repeatable [
5,
6]. However, HbA1c during normal pregnancy is investigated with in one trimester or two trimesters [
7,
8], and a few in early pregnancy. Pregnant women with high HbA1c values above normal (5.7 − 6.4%, or around 38 to 47 mmol/mol) are prone to preterm delivery and preeclampsia [
8‐
11]. Early diagnosis of abnormal HbA1c levels allows for preventive measures to be taken.
Studies have established reference intervals for HbA1c levels during pregnancy in non-diabetic women across gestation periods in other ethnic populations [
8,
12]. However, there is limited information available on HbA1c levels in early pregnancy in China. A significant association has been found between maternal glucose levels at 10–14 weeks of pregnancy and birth weight at term [
13]. Factors such as genetic background and insufficient sample size may contribute to varying HbA1c values in different ethnic populations. Therefore, the HbA1c levels in the early trimester remain undefined and require further research.
The objective of this study was to establish the reference intervals for HbA1c in non-diabetic pregnancy in China. A large, hospital-based sample of pregnant women was recruited. HbA1c in each trimester was measured using a high-performance liquid chromatography (HPLC) assay standardized to the Diabetes Control and Complications Trial (DCCT) values.
Discussion
The study reported the HbA1c reference intervals (2.5th and 97.5th percentile) in 4, 134 non-diabetic pregnant Chinese women as: 4.0 − 5.5%, 3.9 − 5.3%, and 4.1 − 5.7% in the first, second, and third trimester respectively. Results showed that HbA1c decreased in the second trimester compared to the first trimester, then increased physiologically in the late trimester. We highlighted that preexisting diabetes or GDM treatment goals (for example, HbA1c targets < 6% recommended by ADA [
19]) should consider specific times in the gestational period to avoid hypoglycemic episodes in clinical practice. These findings demonstrate the importance of assessing blood glucose fluctuation using trimester-specific HbA1c measurement during pregnancy.
The HbA1c reference intervals in non-diabetic Chinese pregnant women in the 97.5th percentile for the T1, T2, and T3 groups were ≤ 5.5%, 5.3%, and 5.7%, respectively. The results were per previous studies that assessed HbA1c in normal pregnancies. For example, a study of 250 healthy pregnant women in the Netherlands indicated that HbA1c: 4.2–5.4% in early pregnancy, 3.9–5.5% in mid-pregnancy, and 4.1–5.8% in late pregnancy [
20]. Also, O’Shea et al. determined that the comparable reference intervals for HbA1c of Irish non-diabetic pregnant women were 4.3–5.4%, 4.4–5.4%, and 4.7–5.7% in the early, middle, and late pregnancy, respectively [
21]. A study of 725 normoglycemia pregnant subjects in Mexico showed that the range of HbA1c was 4.5–5.6% in the first trimester, 4.4–5.5% in the second trimester, and 4.5–5.6% in the third trimester [
12]. Hence, the 97.5th centiles for HbA1c in our study were similar to previous studies that reported a range of 4.9 to 5.7% in the first trimester [
20,
22] and 5.5–5.9% in the third trimester [
8,
23]. However, HbA1c levels ranging from 5.4 to 5.7% in the second trimester were lower than in previous observations [
21‐
23]. This may be attributed to potential variabilities in the study populations, such as ethnicity, maternal BMI et al., and sample size differences in the first trimester among those studies. Notably, the sample size of the two studies was 88 [
22], and 84 [
12], while our study had 706 participants. Furthermore, the mean (SD) of HbA1c was 4.7 (0.4) %, consistent with data from the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study of more than 20,000 non-diabetic pregnancies that reported maternal mean (SD) values of HbA1c was 4.7 (0.4) % [
24]. The HAPO study explored the relationship between glycemia status at 28 weeks to 32 weeks of gestation and perinatal outcomes, providing the strongest data on this topic worldwide.
There are varying outcomes in the HbA1c change pattern throughout pregnancy. Research shows that biphasic changes in HbA1c measurements, initially declined to a nadir level at 24 weeks of gestation, then increase gradually to summit near term [
25]. Other studies have concluded that HbA1c levels are unchanged in normal pregnancy [
23] or decrease [
7] with advancing gestation. However, our study showed that HbA1c levels were lower in the second trimester compared with the first trimester similar to the studies by [
20,
22]. The sample size in each trimester in our study is larger, and we use the HbA1c measurement method with the traceability to the DCCT method. As a concurrence, Sánchez-González et al. [
12] documented similar conclusions and indicated the increment of HbA1c levels from the second to the third trimester for healthy pregnancies. Moreover, we report a fluctuation of the mean of HbA1c from 4.7% in the early pregnancy to 4.5% in the second pregnancy to 4.8% in the late pregnancy, suggesting a decline of HbA1c in the mid-pregnancy. This is clinically significant when defining the goal for HbA1c in diabetic individuals. During the second trimester, HbA1c levels above the normal range (5.1 − 6.4%, or 32–45 mmol/mol) were associated with an increased risk of adverse pregnancy complications, such as large for gestation age, macrosomia, preterm birth, and preeclampsia [
26,
27]. Furthermore, recent research examining continuous glucose monitoring in pregnant women who do not have GDM according to the IADPSG criteria has revealed irregular glycemic fluctuations in these pregnancies as well [
28]. These underscore the necessity for more precise diagnostic criteria to effectively identify hyperglycemia during pregnancy. Thus, mid-pregnancy women who exhibit elevated values in oral glucose tolerance tests despite remaining within the normal limits need close observation. Adopting HbA1c reference intervals based on gestational week could improve obstetrics outcomes.
HbA1c concentration should be accepted as an integrated measure of blood glucose levels in both pregnant and non-pregnant diabetic individuals over the past three-month period. Red cell turnover during gestation can impact HbA1c levels regardless of ambient glucose levels. A reduction in HbA1c concentrations before 20 weeks of gestation could be explained by an increase in new red blood cell production and a recline in fasting glucose levels [
6]. Moreover, HbA1c integrates maternal glucose levels with fetal development in the previous weeks. Maternal glucose determinations have been linked to birth weight [
3], and HbA1c levels found to have a linear relationship with infant size [
9]. In addition, maximum fetal growth occurs in the second trimester [
13,
29,
30]. These findings suggest that the fetal-placental unit consumes more glucose in the second trimester compared to the increase in maternal glucose production, leading to a decrease in HbA1c levels. Besides, during pregnancy, healthy women experience a 30% increase in basal endogenous glucose production by the end of gestation and a decrease in peripheral tissue insulin sensitivity of about 50% by late gestation [
31]. A previous study showed that a decrease in hemoglobin levels corresponds to an increase in HbA1c [
32]. We reported mean maternal hemoglobin levels of 123.8 ± 10.3 g/L in the early trimester and 112.9 ± 11.9 g/L in the late trimester, with the HbA1c levels increasing in the late trimester. Unmanaged iron deficiency can result in an HbA1c rise of 0.1–0.2% (1–2 mmol/mol) [
6,
33]. These observations may explain the increase in HbA1c values in the third trimester.
There were some strengths in this study. First, our study included the largest sample size in the early trimester. Second, we assessed the reference intervals of HbA1c measurements in a non-diabetic pregnant woman in a Chinese population.
The potential limitations should be mentioned. First, the observational design of the study raised the possibility of residual confounding and selection bias influencing the results. Second, HbA1c measurement did not adequately capture the short-term fluctuations which is also important in the blood glucose management, making continuous glucose monitoring a valuable tool for future research. Third, the interpretation of our study warrants cautious consideration, primarily due to the large sample size employed. It is important to acknowledge that the sheer size of our study cohort enhances the probability of even minor discrepancies attaining statistical significance, irrespective of their genuine clinical significance. Although the different reference intervals of HbA1c levels in three trimesters might be of interest, they should be considered their clinical relevance. To confirm the real-world significance of our findings, an outcome-derived approach is necessary in future research.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.