Early postpartum HbA1c after hyperglycemia first detected in pregnancy—Imperfect but not without value

Abstract

Background

South African women of childbearing age are disproportionally affected by obesity and at significant risk of Type 2 Diabetes Mellitus (T2DM). Unless pregnant, they do not readily undergo screening for T2DM. With a local focus on improved antenatal care, hyperglycemia is often first detected in pregnancy (HFDP). This may erroneously be attributed to Gestational Diabetes Mellitus (GDM) in all without considering T2DM. Glucose evaluation following pregnancy is essential for early detection and management of women with T2DM in whom persistent hyperglycemia is to be expected. Conventional testing with an oral glucose tolerance test (OGTT) is cumbersome, prompting investigation for alternate solutions.

Aim

To compare the diagnostic performance of HbA1c to the current gold standard OGTT in women with HFDP 4–12 weeks post-delivery.

Methods

Glucose homeostasis was assessed with OGTT and HbA1c in 167 women with HFDP, 4–12 weeks after delivery. Glucose status was based on American Diabetes Association criteria.

Results

Glucose homeostasis was assessed at 10 weeks (IQR 7–12) after delivery. Of the 167 participants, 52 (31%) had hyperglycemia, which was comprised of 34 (20%) prediabetes and 18 (11%) T2DM. Twelve women in the prediabetes subgroup had diagnostic fasting plasma glucose (FPG) and 2-hour plasma glucose (2hPG), but in two-thirds of the patients (22/34) only one time point proved diagnostic. The FPGs and the 2hPGs of six women with HbA1c-based T2DM were both within the prediabetes diagnostic range. According to the HbA1c measurements, 85% of 52 participants with gold standard OGTT defined hyperglycemia (prediabetes and T2DM) as well as 15 of 18 women with postpartum persistent T2DM were correctly classified. According to FPG, 15 women with persistent hyperglycemia would have been missed (11 with prediabetes and four with T2DM; 29%). When compared to an OGTT, a single HbA1c of 6.5% (48mmol/mol) postpartum demonstrated a sensitivity of 83% and specificity of 97% for the identification of T2DM.

Conclusion

HbA1c may improve access to postpartum testing in overburdened clinical settings where the required standards of OGTT cannot be guaranteed. HbA1c is a valuable test to detect women who will benefit most from early intervention but cannot unequivocally replace OGTT.

Introduction

Increasing prevalence of Type 2 Diabetes Mellitus (T2DM) in reproductive-aged women, in part driven by obesity, has contributed to the current high rate (1 in 6) of births with hyperglycemia [1–4]. In low-to-middle income countries (LMIC), such as South Africa, women of childbearing age are disproportionally affected by obesity, predisposing them to Gestational Diabetes Mellitus (GDM) and T2DM [5–8]. T2DM screening is limited in the local community outside of pregnancy, with two-thirds of those identified being unaware of their diagnosis [9–12]. To improve antenatal care, most international and some South African guidelines advocate universal GDM screening. Due to resource limitations, selective screening of women with the highest risk for T2DM remains the norm locally [12–15]. Hyperglycemia first detected in pregnancy (HFDP) which meets the criteria for overt T2DM, is often mistaken for GDM [15–17]. T2DM is expected to persist post-delivery, whereas GDM characteristically resolves (within 24–48 hours) postpartum [18–20]. Women, with GDM, however, remain at very high risk of future T2DM [18, 19].

The early detection and optimal management of T2DM has long-term benefit and is referred to as the so called "legacy effect" [21–23] Landmark studies support the hypothesis that progression to T2DM is preventable in women with GDM [24, 25]. In the Diabetes Prevention Program Outcomes Study, intensive lifestyle modifications and metformin compared to placebo, in women with prior GDM, reduced the incidence of T2DM by 35% and 40%, respectively. Against the backdrop of the metabolic “legacy effect” several professional organizations, including the World Health Organisation (WHO) and the International Association for Diabetes in Pregnancy Study Groups (IADPSG), have earmarked the early postpartum period as a window of opportunity to firstly assess glucose status after HFDP and then to ensure early intervention if indicated [26–28]. Early postpartum glucose evaluation is standard practice at our institution, where as many as 26% of women remain hyperglycemic following delivery [17].

The 75-gram oral glucose tolerance test (OGTT) is the gold standard for postpartum glucose evaluation. It is a cumbersome test that requires fasting, glucose ingestion and a minimum of 2-hours at a healthcare facility [29]. Glycated hemoglobin (HbA1c), the monitoring modality for longer-term glycemia, is an alternate and convenient diagnostic tool to identify non-pregnant people with T2DM and prediabetes [12, 29]. HbA1c testing requires a single blood sample with no need for prior fasting or glucose ingestion. Currently, an HbA1c of 6.5% (48mmol/mol) is the accepted threshold for the diagnosis of T2DM. Many factors interfere with HbA1c measurement and interpretation and it is recommended to repeat an HbA1c to exclude laboratory errors in asymptomatic individuals suspected to have T2DM [29]. It is also important to consider factors that affect glycation, such as iron deficiency anemia and to a lesser extent, iron deficiency as well as ethnic differences, when utilizing HbA1c for diagnosis [8, 30].

Population differences in optimal HbA1c thresholds to detect T2DM have, however, been noted and the universal applicability of this diagnostic threshold questioned [31–33]. The role of a single HbA1c measurement to accurately distinguish between women with GDM and overt T2DM in the early postpartum period, remains uncertain [34].

Few global studies have compared the diagnostic utility of HbA1c with OGTT in the early postpartum period, with the predictive value varying depending on the prevalence and degree of hyperglycemia [35–41]. The utility of a single puerperal HbA1c measurement as a diagnostic modality following HFDP has not been established in South African women. It is important to investigate simplified diagnostic methods with appropriate thresholds to streamline postpartum testing in LMICs with limited resources. This study examined the diagnostic utility of HbA1c compared with an OGTT, to identify persistent hyperglycemia 4–12 weeks after delivery, in South African women with HFDP.

Methods

Study design and setting

The study followed a descriptive, cross-sectional design. The study population included all women with recent hyperglycemia first detected in pregnancy (HFDP) who attended the postpartum clinic 4–12 weeks after delivery from 1 November 2019 to 31 October 2022. Glucose assessment included an OGTT and HbA1c measurement. Hemoglobin was routinely measured postpartum. Women in this study received prenatal care in the multidisciplinary, diabetes clinic at Tygerberg Hospital, Cape Town, South Africa. As recommended by the World Health Organization, all pregnant women in South Africa are provided with iron and folate supplements to prevent anemia [42, 43]. At Tygerberg, this is also standard practice.

The HFDP diagnosis was based on the criteria outlined in the National Institute of Health Care Excellence (NICE) guidance (fasting plasma glucose (FPG) ≥ 5.6mmol/l ± OGTT 2h-hour plasma glucose (2hPG) ≥ 7.8mmol/l). Women who had postpartum hemorrhage, ongoing blood loss or who received a blood transfusion were excluded from study entry. An incomplete OGTT glucose set and a hemoglobin (Hb) concentration ≤ 10g/dl also served as exclusion criteria. The research complied with the World Medical Association Declaration of Helsinki ethical principles for medical research involving human subjects and was approved by the Stellenbosch University Health Research and Ethics Committee as well as by local hospital authorities (HREC ref: S21/11/255). Based on the fully anonymized data and standard practice assessment, the ethics committee waived the need for informed consent.

Data collection

Antenatal clinical information and data were retrieved from the Tygerberg Electronic Content Management System (ECM). Structured, standardized data sheets were used to record clinical and anthropometric data at the early postpartum visit. The assessment was done by dedicated nursing staff, senior medical officers, diabetologists, obstetricians and endocrinologists. Biochemical results were obtained via the National Health Laboratory System (NHLS) electronic platform. Strict confidentiality was maintained as the ECM and NHLS databases, as well as personal computers used for all data collection, are password protected. Patient information and data were entered onto a spreadsheet using Excel version 2019 Microsoft Office Professional Plus (Microsoft Corp, Redmond, WA, USA). A unique number was linked to each patient’s data set and the patient de-identified.

Demographics and clinical characteristics

A questionnaire, administered at the postpartum visit, included demographic, antenatal and early postpartum information e.g. maternal age and gestation at booking, parity, known family history of diabetes, ethnicity, marital and socio-economic status, the self-reported highest level of education, mode of feeding and personal medical history.

Anthropometry and blood pressure

Basic anthropometric measurements (weight and height) were taken. Weight was measured with minimal clothing using a manual hospital scale positioned on a flat surface, and recorded to the nearest kilogram (kg). Height was measured without shoes or headwear to the nearest centimetre (cm) with a calibrated stadiometer. Body mass index (BMI) was calculated (kg/m²) and classified according to the World Health Organisation (WHO) global recommendations [44]. Brachial blood pressure was measured with an appropriate cuff size on the left arm while seated, using a calibrated automated sphygmomanometer (Dinamap Carescape V100).

Biochemistry

The HbA1c test, in addition to a standard 75-gram OGTT (FPG and 2hPG), was performed after an overnight fast. Laboratory analysis was done at the NHLS, a South African National Accreditation System (SANAS) accredited laboratory based at Tygerberg Hospital. NHLS laboratories are accredited by SANAS for compliance with international standards (ISO 15189:2007) [45].

Blood samples were collected in sodium fluoride tubes (Becton Dickenson) for the measurement of plasma glucose using the hexokinase method on the Roche Cobas 6000 (Roche Diagnostics, Mannheim, Germany) platform which has a measuring range from 0.11–41.6 mmol/L with a reported coefficient of variation (CV) of 1.3% at a glucose level of 5.38 mmol/L and 1.1% at a level of 13.4 mmol/L respectively. Internal and external quality controls were performed according to laboratory protocol. The standard turnaround time for glucose samples’ (blood sampling to laboratory analysis) was 107 minutes for FPG and 74 minutes for 2hPG. Whole blood collected in potassium-EDTA tubes (Becton Dickinson) was used to measure HbA1c with the turbidimetric inhibition immunoassay on the Tina-quant^® HbA1c assay Generation 3 on the Roche Cobas 6000^® (Roche Diagnostics, Mannheim, Germany). The measuring range of HbA1c was 4.2%-20.1% (22 to 195mmol/mol), with a CV of 1.3% at HbA1c 5.3% (34mmol/mol) and 1.1% at HbA1c 9.9% (85mmol/mol). The HbA1c determination was done by calculation according to the approved International Federation of Clinical Chemistry(IFCC and Laboratory Medicine method, traceable to the Diabetes Control and Complications Trial. (DCCT) [46]. The HbA1c assay is standardized according to IFCC and transferable to DCCT/NGSP.

Categorization of study participants

OGTT (fasting and 2 hour post glucose-load plasma glucose).

Based on the OGTTs, women were categorized according to the American Diabetes Association (ADA) criteria into two main subgroups, namely (i) normoglycemia (FPG <5.6mmol/l and 2hPG <7.8mmol/l or (ii) hyperglycemia (FPG ≥ 5.6mmol/l ± 2hPG ≥ 7.8mmol/l) [29]. Women with hyperglycemia were then further subclassified as (a) prediabetes (impaired fasting glucose (IFG) FPG 5.6–6.9 mmol/l and 2hPG < 7.8mmol/l; impaired glucose tolerance (IGT) FPG < 5.6mmol/l and 2hPG 7.8–11.0mmol/l or both FPG 5.6–6.9mmol/l and 2hPG 7.8–11.0mmol/l) or (b) T2DM (FPG ≥ 7.0 mmol/l ± 2hPG ≥11.0mmol/l). Women with discordant glucose categorization at 0 and 120 mins were allocated according to the highest value.

HbA1c.

Based on HbA1c, women were also categorized as (i) normoglycemia: HbA1c < 5.7% (< 39 mmol/mol) or (ii) hyperglycemia. Women with hyperglycemia based on HbA1c were then further subclassified as (a) prediabetes (HbA1c ≥ 5.7–6.4% (≥39-47mmol/mol) or (b) T2DM (HbA1c ≥ 6.5% (≥ 48mmol/mol) [29].

Statistical analyses

The sample size calculation was based on our previous studies with a T2DM prevalence of 25% postpartum and indicated a minimum of 111 subjects. Statistical analysis was performed using GraphPad Prism software (Version 9.5.0). The Shapiro-Wilk and Kolmogorov-Smirnov normality testing was performed to assess data distribution. Normally distributed data are presented as mean ± standard deviations (SD) and non-parametric data as median and interquartile range (IQR). Descriptive statistics were used for socio-demographic parameters. The Wilcoxon matched pairs signed rank test was used to assess differences in parameters between time points (booking vs postpartum). A receiver operator characteristic (ROC) curve was plotted to evaluate the diagnostic performance of HbA1c against the mean glucose derived from the OGTT. Threshold level for significance was accepted at p<0.05. Sensitivity, specificity, negative predictor value (NPV) and positive predictor values (PPV) were calculated as follow: sensitivity = TP/(TP+FN); specificity = TN/(TN+FP); NPV = TN/ (FN+TN); PPV = TP/(TP+FP). TP = true positive; TN = true negative; FP = false positive; FN = false negative.

Results

In total, 181 women with HFDP attended the early postnatal visit, but fourteen were excluded because glucose datasets were missing in ten cases, Hb- values were below 10g/dL in three cases, and the Hb result was unavailable in one case (see flow sheet; Fig 1). We thus report on the 167 study participants with HFDP who had an OGTT glucose dataset with HbA1c 4–12 weeks postpartum. Participant characteristics are stratified based on postpartum glucose categories in Table 1.

[Figure omitted. See PDF.]

Hb = Haemoglobin, HFDP = hyperglycemia first detected in pregnancy, OGTT = oral glucose tolerance test.

[Figure omitted. See PDF.]

Antenatal characteristics

Participant characteristics are noted in Table 1. The median age of the total cohort was 34 years (IQR 29–37), and near half of women (73/167; 44%) were of advanced maternal age (≥ 35 years). Most women were multigravidas (89%), either black African (57%) or of mixed ancestry (31%) and educated up to secondary level (78%). Sixty-four of the multigravida women (64/148; 43%) reported at least one previous fetal loss, of which five (5/148; 3%) were in the third trimester.

The median BMI at booking was 38 kg/m² (IQR 33–45). Individual BMI’s, based on WHO categories (overweight BMI 25.0–29.9 kg/m²; obese BMI ≥ 30kg/m²⁾ were overweight or obese in all but five study participants [overweight in 20 (12%), obese in 142 (86%)].

Study participants booked early in the index pregnancy at a median gestation of 12 weeks (IQR 8–17). Only eight women (8/167;5%) had their initial evaluation after 24 weeks. HFDP was diagnosed at a median gestation of 26 weeks (IQR 24–31). At diagnosis, fasting glucose was available in 166/167, 2-hour glucose in 156/167 and HbA1c in 160/167 women, respectively. Forty-six women (46/167; 28%) met the diagnostic criteria for IADPSG overt diabetes. In 41 of these women, a diagnosis of overt diabetes was based on the presence of elevated plasma glucose at least one time point during the OGTT. In 22 women both an elevated fasting (≥ 7mmol/l) and an elevated 2-hour value of ≥11mmol/l were present, 11 had high fasting glucose only, and eight only an elevated 2-hour value. At diagnosis a HbA1c ≥ 6.5% was documented in 34 study participants. Twenty nine of these women had a concomitant diagnostic OGTT criterium for overt diabetes, whereas in five patients the diagnosis was based on a HbA1c only.

Hyperglycemia was managed and glycemic targets were achieved with dietary modifications in 41 (25%); 115 women required additional metformin (69%), and in only a minority of the cohort, both metformin and insulin were needed (11/167; 7%). Third trimester HbA1c was available in more than half (95/167; 56%) of the cohort with a median value of 6% (IQR 5.6–6.5). By using paired analysis to compare the median 3^rd trimester HbA1c to the median postpartum HbA1c [5.6% (IQR 5.4–6.2)], the 3^rd trimester values were higher by 0.4%.

Delivery information was available for the 164 live births. The median gestation at delivery was 38 weeks (IQR 37–38), while preterm birth (< 37 weeks) occurred in 29 (29/164, 18%). Labour was induced in more than half of the cohort (95/164;58%). This largely reflected the local policy of delivery at 38 weeks’ gestation for patients with HFDP on glucose-lowering pharmacotherapy. Normal vaginal delivery (NVD) predominated (100/164; 61%), preceded by induction of labour in most (66/100; 66%). Of the 64 caesarean sections performed, 46 (72%) were emergency procedures for fetal indications Fifteen neonates (15/164; 9%) had macrosomia (>4000g) at birth. Three intra-uterine deaths (IUDs) occurred at 28, 30, and 31 weeks. Two of these occurred during maternal admission with severe diabetic ketoacidosis (DKA) (glucose and pH values of 21.5mmol/l and 7.1; 27mmol/l and 7.11, respectively). Preeclampsia and a urinary tract infection precipitated these DKA’s.

Postpartum assessment

After delivery, the median number of weeks to postpartum OGTT was 10 (IQR 7–12). Pair-wise analysis indicated a significant decrease (p<0.001) in both weight and BMI (kg/m²) at the time of the postpartum assessment compared to booking, with the largest weight change observed in the postpartum normoglycemic group (Table 1). Two-thirds of women (112/167; 67%) reported exclusive breastfeeding.

Table 2 provides information on the antenatal and postpartum biochemical parameters of glucose homeostasis stratified according to postpartum OGTT defined glucose subcategories. Median (IQR) fasting plasma glucose (FPG), 2-hour post-OGTT plasma glucose (2hPG) and plasma HbA1c are tabulated. Concordance with HbA1c categorization is also demonstrated. The postpartum OGTT indicated hyperglycemia in 52/167 (31%) participants. Thirty-four women (20%) had prediabetes and 18 (11%) had T2DM. In the prediabetes category, both FPG and 2hPG values were diagnostic in 12 patients, whereas a third of women (11/34; 32%) presented with an isolated abnormal FPG or an abnormal 2hPG respectively. Of the 18 women with T2DM, both FPG and 2hPG values were diagnostic in six patients, the diagnosis was based on FPG only in eight patients, and in four patients only an abnormal 2hPG was present. Based on fasting glucose only, 15 women with persistent hyperglycemia (11 with prediabetes and four with T2DM) in our cohort would not have been identified (15/52; 29%). An HbA1c correctly classified 85% of the 52 participants with gold standard OGTT defined hyperglycemia and 15 of 18 women with postpartum persistent T2DM.

[Figure omitted. See PDF.]

The sensitivity, specificity, positive and negative predictive values to correctly identify normoglycemia and specific subcategories of hyperglycemia based on OGTT criteria, are detailed in S1 Fig. At the early postpartum visit, our OGTT classified 115 of our total cohort as normoglycemic. In the OGTT normoglycemic category, HbA1c indicated hyperglycemia in 38/115 (33%), 35 with prediabetes and three with diabetes. The sensitivity and specificity of a single HbA1c postpartum to correctly identify the degree of hyperglycemia compared to the OGTT were best for T2DM (sensitivity 83%; specificity 97%).

The receiver operating characteristic (ROC) curves for the diagnostic performance of HbA1c and FPG against OGTT are presented in Fig 2. HbA1c had a higher sensitivity to detect hyperglycemia than FPG at conventional diagnostic thresholds in all three glucose categories (Fig 2a–2f). By lowering the HbA1c threshold to ≥5.7% (39mmol/mol) for T2DM, the sensitivity improved to 94% [AUC 0.92 (95% CI 0.82–1.0)], but significantly worsened specificity to 67% (Fig 2h).

[Figure omitted. See PDF.]

Discussion

Although not ideal, a single postpartum HbA1c at conventional diagnostic thresholds is clinically useful in detecting hyperglycemia in women with prior HFDP. The ability of a single HbA1c to differentiate between the two hyperglycemia classes, i.e., prediabetes and T2DM, was limited, with the test more accurate in predicting T2DM. It is noteworthy that 33% of patients with OGTT normoglycemia had an HbA1c measurement of ≥ 5.7% (39mmol/mol) and were therefore over diagnosed. ROC curves based on the gold standard postpartum OGTT indicated that HbA1c was comparable as a diagnostic test but did not outperform the ability of fasting blood glucose to detect hyperglycemia. The sensitivity of HbA1c improved to an ideal 94% when the conventional threshold for T2DM was adjusted to the prediabetes cut-off [HbA1c ≥ 5.7% (39mmol/mol)]. This, however, came at the expense of a significant decline in specificity from 97% (conventional threshold) to 63% (lowered threshold).

In pregnancies complicated by HFDP, pragmatic solutions to optimally categorize and diagnose glucose homeostasis early postpartum are especially valuable in resource-limited environments. Identifying persistent hyperglycemia postpartum is important, as timeous pharmacological intervention in those with T2DM has significant long-term health benefits. HbA1c represents an attractive diagnostic alternative to a postpartum OGTT. The test can be performed without any patient-related restrictions, in contrast to obligatory fasting with an FPG or when performing an OGTT and is thus ideal in the out-patient setting.

Non-communicable diseases associated with obesity are among the top causes of death in South Africa [47]. Globally, the rising prevalence of T2DM parallels that of obesity and South African women of childbearing age have a disproportionately high burden [6]. Recent estimations noted an escalation in the prevalence of overweight and obesity of 60% and 35.2%, respectively, from 1998 and 2017 [6]. A concerning 97% of women in our study cohort were overweight or obese. T2DM, prediabetes and obesity lead to several maternal and child health problems extending far beyond pregnancy. The risk of developing T2DM in patients with prediabetes varies and is influenced by the degree of hyperglycemia, as well as the individual background genetics and pancreatic β-cell reserve. Individuals with both IFG and IGT, develop T2DM sooner than those with isolated defects [48]. In our study, postpartum HbA1c overestimated the number of women with T2DM, incorrectly classifying 12 women with OGTT prediabetes and three with OGTT normoglycemia as having T2DM, based on OGTT criteria. Most (7/12) of the women diagnosed with OGTT prediabetes had both IFG and IGT on OGTT. It can thus be argued that HbA1c tends to overestimate the presence of T2DM, mostly in a subset of women with prediabetes at highest risk for future T2DM.

The landmark randomised controlled trial “The Diabetes Prevention Program Outcomes Study” indicated that intensive lifestyle modification and metformin are highly effective in delaying the progression to T2DM in women with a history of GDM [24, 25]. Women with OGTT prediabetes and a HbA1c ≥ 6.5% (48mmol/mol) represent a subset at higher risk within the prediabetes range and may benefit from earlier pharmacological intervention with metformin.

Fasting plasma glucose, an OGTT and HbA1c are well-established screening tools for T2DM. FPG and HbA1c are convenient, but the diagnostic accuracy varies according to cut-offs, the background diabetes prevalence in the population and ethnicity. In this study, an early postpartum HbA1c ≥ 6.5% (48mmol/mol) had a sensitivity of 83% to detect T2DM based on OGTT glucose criteria, identifying 15/18 women. In comparison a FPG of ≥ 7mmol/L alone had a sensitivity of 72% to detect T2DM and missed four of eighteen women with T2DM. Our small study thus indicated that the sensitivity of a HbA1c was ±20% better compared to a diagnostic FPG in T2DM. Chivese and colleagues assessed the diagnostic accuracy of HbA1c vs FPG and OGTT in non-pregnant individuals from seven African countries [49]. The ethnic backgrounds resembled the ethnic representation in our cohort and included predominantly Black African women and women of Mixed Ancestry. In their study, the sensitivity of a HbA1c ≥ 6.5% (48mmol/mol) to diagnose T2DM was 58% vs OGTT, a value significantly lower than the 83% documented in our cohort. As expected, sensitivity improved by lowering the cut off [HbA1c ≥ 6% (42mmol/mol)] in their study, but this came at the expense of decreased specificity. Another recent study suggested that an HbA1c cut-off of 6.5% (48mmol/mol) may miss up to 50% of non-pregnant persons with T2DM, perhaps attributable to ethnic and population differences [50]. This study suggests an HbA1c ≥6.1% may be more appropriate in the mixed ancestry population.

The conventional HbA1c threshold may thus not be globally applicable and may be too high in certain populations. This might also apply to postpartum assessment following pregnancies complicated by HFDP. The reported diagnostic accuracy of HbA1c for T2DM postpartum is impacted by the antenatal diagnostic criteria used, the timing of the early postpartum OGTT and the background T2DM prevalence. Katreddy evaluated 203 predominantly Caucasian women ± six weeks after delivery and reported a sensitivity of 71% and 99% specificity for T2DM with the conventional HbA1c threshold of 6.5% (48mmol/mol). By using an HbA1c of 6.0% (44mmol/mol), sensitivity improved to 100% [area under the curve: 0.98 (95%CI: 0.96–1.00)] [39]. Kim et al. evaluated 699 Korean women with GDM, 6–12 weeks postpartum. The prevalence of T2DM and prediabetes was 5.2% and 49.1% respectively [40]. They found that a HbA1c was not sensitive enough for early postpartum testing with poor sensitivity to detect both T2DM and prediabetes in their cohort of women (19.4% for T2DM and 41.2% for prediabetes). The authors speculated that the usefulness and diagnostic sensitivity of HbA1c might improve with the addition of a FPG [41]. In a study by Picón et al. that included 231 Spanish women with prior GDM, the ability of a HbA1c or a FPG alone or a HbA1c combined with a FPG to identify women with persistent hyperglycemia based on OGTT criteria was evaluated. The OGTT identified 39% women with prediabetes and 6% with diabetes. A HbA1c of ≥ 6.5% (48mmol/mol) identified 19% with prediabetes and only one woman with T2DM. The combination of HbA1c and FPG did not improve the sensitivity to detect hyperglycemia compared to a FPG only. Claesson evaluated 140 women 1–5 years after delivery and combined FPG with HbA1c, but also found that the combination did not improve the performance of FPG per se [35].

In the study cohort, the yield to identify T2DM increased significantly (sensitivity 94%) when the HbA1c diagnostic threshold was adjusted to ≥ 5.7% (39mmol/mol). The sensitivity did not improve with the addition of a FPG ≥ 5.6mmol/l, but the specificity of the FPG per se was better (100% vs 67%). In the prediabetes category, a HbA1c ≥ 5.7% (39mmol/mol) outperformed FPG ≥ 5.6mmol/l as a stand-alone test, with a sensitivity of 79% for HbA1c compared to 67% for FPG.

There are several plausible reasons, for the variation in the optimal diagnostic set points of HbA1c and glucose. The concentration of plasma glucose ex-vivo decreases due to glycolysis [51]. To prevent glycolysis, fluoride tubes are used in clinical practice and were utilized in this study. Complete inhibition of glycolysis can take as long as four hours, with a resultant potential decline in measured glucose up to 0.6mmol/mol. Alternatively, glycolysis can be inhibited by separating plasma from cells within 30 minutes after venesection and transporting samples in ice slurries [51]. This is mainly used in research settings, as it is impractical in clinical practice. In our study, the reported turnaround times (74–107 minutes) from sampling to analysis were relatively short, and it is thus unlikely that glycolysis significantly impacted glucose measurements. Breastfeeding has short and long-term benefits in women with GDM, including delaying the onset of T2DM by impacting insulin sensitivity. Breastfeeding should thus be encouraged [52–54]. Most (two-thirds) women in the study were breastfeeding at the time of the assessment, raising the question of whether this may have influenced or modified glucose responses to an OGTT. A randomized crossover trial investigated the impact of breastfeeding on OGTT in women with recent GDM (±12 weeks postpartum) and concluded that breastfeeding did not impact postpartum glucose categorization [55]. HbA1c as a monitoring tool reflects the mean plasma glucose over ±120 days (the lifespan of the red blood cells) in non-pregnancy. This relationship may be affected by conditions that alter red blood cell turnover, including iron-deficient states and pregnancy per se [30, 40, 51, 56]. In iron-deficient states, HbA1c is falsely elevated, but declines by 0.4% after one month’s supplementation [30]. Consequently, one must consider that HbA1c levels in the study could have been altered, the direction dependent on the timing of iron supplementation. Iron supplementation is initiated routinely at the first antenatal visit in our setting. Most women booked early and before 24 weeks gestation, making a variation in the measured HbA1c postpartum due to iron supplementation unlikely. HbA1c, in contrast to the impact of iron deficiency, may be lowered due to physiological changes of pregnancy per se. The increased red cell turnover in pregnancy reduces glycation of hemoglobin and can lower HbA1c by as much as 0.5% in the third trimester [57]. Moreover, evidence indicates a delay between the development of hyperglycemia and increasing HbA1c [58]. As a result, postpartum HbA1c may still reflect a proportion of third-trimester glycemia. In our study, the postpartum assessment occurred ten weeks (IQR 7–12) after delivery. The 3^rd trimester HbA1c was higher than postpartum and antenatal representation are thus excluded. At present, HbA1c results are interpreted based on the assumption that erythrocyte life span is constant at 120 days when, in fact, enormous heterogeneity occurs due to biological variation [59]. Possibly this was also a contributing factor in our study, explaining the lack of correlation between 3^rd trimester and postpartum HbA1c. In view of the fact that our analysis did not include individual indices of red blood cell survival, we did not take into account such heterogeneity. It may be beneficial for future research to assess red cell survival, particularly in early postpartum assessments.

This is the first study in South-African women with recent HFDP where the diagnostic accuracy of a single HbA1c is compared with the gold standard OGTT early postpartum. The study had limitations. It is not a large study, antenatal data were gathered retrospectively and there was no control group. Population-specific thresholds associated with adverse outcomes in T2DM and prediabetes and the ability of postpartum HbA1c to predict future T2DM warrants further investigation as it may inform required long-term follow-up intervals. The various pathophysiological contributors to abnormal glucose homeostasis and their relation to postpartum HbA1c are also worthy of further interrogation as are tests evaluating glycemia not influenced by red blood cell kinetics such as glycated albumin.

In summary, good sensitivity of an early postpartum HbA1c to detect women with established T2DM and those at the highest risk for future T2DM (prediabetes with combined FPG and IGT) is demonstrated in this study. These women represent at risk individuals who will benefit from early intervention. In clinical practice, HbA1c is valuable when obligatory fasting and other pre-analytical OGTT testing requirements are not feasible but cannot unequivocally replace OGTT based on this study. Further evaluation in a larger cohort is required.

Supporting information

S1 Fig. The performance discordance of Hba1c categorisation compared to oral glucose tolerance testing.

https://doi.org/10.1371/journal.pone.0282446.s001

(TIF)

Citation: Coetzee A, Hall DR, van de Vyver M, Conradie M (2023) Early postpartum HbA1c after hyperglycemia first detected in pregnancy—Imperfect but not without value. PLoS ONE 18(6): e0282446. https://doi.org/10.1371/journal.pone.0282446

About the Authors:

Ankia Coetzee

Roles: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing

E-mail: [email protected]

Affiliation: Department of Medicine, Division of Endocrinology Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa

ORICD: https://orcid.org/0000-0001-9993-6439

David R. Hall

Roles: Methodology, Supervision, Writing – original draft, Writing – review & editing

Affiliation: Department of Obstetrics & Gynecology, Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa

ORICD: https://orcid.org/0000-0002-2344-3969

Mari van de Vyver

Roles: Formal analysis, Methodology, Software, Validation, Writing – original draft, Writing – review & editing

Affiliation: Department of Medicine, Division of Clinical Pharmacology, Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa

Magda Conradie

Roles: Conceptualization, Data curation, Formal analysis, Supervision, Writing – original draft, Writing – review & editing

Affiliation: Department of Obstetrics & Gynecology, Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa

ORICD: https://orcid.org/0000-0003-3092-4098

References

1. Yuen L, Saeedi P, Riaz M, Karuranga S, Divakar H, Levitt N, et al. Projections of the prevalence of hyperglycaemia in pregnancy in 2019 and beyond: Results from the International Diabetes Federation Diabetes Atlas, 9th edition. Diabetes Res Clin Pract. 2019 Nov;157:107841. pmid:31518656

2. Wang H, Li N, Chivese T, Werfalli M, Sun H, Yuen L, et al. IDF Diabetes Atlas: Estimation of Global and Regional Gestational Diabetes Mellitus Prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group’s Criteria. Diabetes Res Clin Pract. 2022 Jan;183:109050. pmid:34883186

3. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet. 2014 Aug;384(9945):766–81. pmid:24880830

4. Khan MAB, Hashim MJ, King JK, Govender RD, Mustafa H, Al Kaabi J. Epidemiology of Type 2 Diabetes—Global Burden of Disease and Forecasted Trends: J Epidemiol Glob Health. 2019;10(1):107.

5. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham CL, Anis AH. The incidence of co-morbidities related to obesity and overweight: A systematic review and meta-analysis. BMC Public Health. 2009 Mar 25;9(1):88. pmid:19320986

6. Nglazi MD, Ataguba JEO. Overweight and obesity in non-pregnant women of childbearing age in South Africa: subgroup regression analyses of survey data from 1998 to 2017. BMC Public Health. 2022 Dec;22(1):395. pmid:35216565

7. Peer N, Steyn K, Lombard C, Lambert EV, Vythilingum B, Levitt NS. Rising Diabetes Prevalence among Urban-Dwelling Black South Africans. Herder C, editor. PLoS ONE. 2012 Sep 4;7(9):e43336. pmid:22962583

8. Hird TR, Pirie FJ, Esterhuizen TM, O’Leary B, McCarthy MI, Young EH, et al. Burden of Diabetes and First Evidence for the Utility of HbA1c for Diagnosis and Detection of Diabetes in Urban Black South Africans: The Durban Diabetes Study. Staiger H, editor. PLOS ONE. 2016 Aug 25;11(8):e0161966. pmid:27560687

9. Mbanya JCN, Motala AA, Sobngwi E, Assah FK, Enoru ST. Diabetes in sub-Saharan Africa. The Lancet. 2010 Jun;375(9733):2254–66. pmid:20609971

10. Goedecke JH, Keswell D, Weinreich C, Fan J, Hauksson J, Victor H, et al. Ethnic differences in hepatic and systemic insulin sensitivity and their associated determinants in obese black and white South African women. Diabetologia. 2015 Nov;58(11):2647–52. pmid:26232099

11. Goedecke JH, Mtintsilana A, Dlamini SN, Kengne AP. Type 2 diabetes mellitus in African women. Diabetes Res Clin Pract. 2017 Jan;123:87–96. pmid:28006698

12. Webb D. 2017 SEMDSA diabetes management guidelines. South Afr J Diabetes Vasc Dis. 2018 Jul;15(1):37–40.

13. ACOG Practice Bulletin No. 190: Gestational Diabetes Mellitus. Obstet Gynecol. 2018 Feb;131(2):e49–64. pmid:29370047

14. Goyal A, Gupta Y, Singla R, Kalra S, Tandon N. American Diabetes Association “Standards of Medical Care—2020 for Gestational Diabetes Mellitus”: A Critical Appraisal. Diabetes Ther. 2020 Aug;11(8):1639–44. pmid:32564336

15. Adam S, Rheeder P. Screening for gestational diabetes mellitus in a South African population: Prevalence, comparison of diagnostic criteria and the role of risk factors. S Afr Med J. 2017 May 24;107(6):523. pmid:28604326

16. Goyal A, Gupta Y, Tandon N. Overt Diabetes in Pregnancy. Diabetes Ther. 2022 Apr;13(4):589–600. pmid:35107789

17. Coetzee A, Mason D, Hall DR, Conradie M. Prevalence and predictive factors of early postpartum diabetes among women with gestational diabetes in a single-center cohort. Int J Gynecol Obstet. 2018 Jul;142(1):54–60.

18. Carson MP, Ananth CV, Gyamfi-Bannerman C, Smulian J, Wapner RJ. Postpartum Testing to Detect Persistent Dysglycemia in Women With Gestational Diabetes Mellitus. Obstet Gynecol. 2018 Jul;132(1):193–8. pmid:29889746

19. Rehder PM, Borovac-Pinheiro A, de Araujo ROMB, Diniz JAPM, Ferreira NLC, Branco ACR, et al. Gestational Diabetes Mellitus and Obesity are Related to Persistent Hyperglycemia in the Postpartum Period. Rev Bras Ginecol E Obstetrícia RBGO Gynecol Obstet. 2021 Feb;43(02):107–12.

20. Weinert LS, Mastella LS, Oppermann MLR, Silveiro SP, Guimarães LSP, Reichelt AJ. Postpartum glucose tolerance status 6 to 12 weeks after gestational diabetes mellitus: a Brazilian cohort. Arq Bras Endocrinol Metabol. 2014 Mar;58(2):197–204. pmid:24830597

21. Edelson PK, James KE, Leong A, Arenas J, Cayford M, Callahan MJ, et al. Longitudinal Changes in the Relationship Between Hemoglobin A1c and Glucose Tolerance Across Pregnancy and Postpartum. J Clin Endocrinol Metab. 2020 May 1;105(5):e1999–2007. pmid:32010954

22. Chalmers J, Cooper ME. UKPDS and the Legacy Effect. N Engl J Med. 2008 Oct 9;359(15):1618–20. pmid:18843126

23. Khunti K, Kosiborod M, Ray KK. Legacy benefits of blood glucose, blood pressure and lipid control in individuals with diabetes and cardiovascular disease: Time to overcome multifactorial therapeutic inertia? Diabetes Obes Metab. 2018 Jun;20(6):1337–41. pmid:29405543

24. Aroda VR, Christophi CA, Edelstein SL, Zhang P, Herman WH, Barrett-Connor E, et al. The Effect of Lifestyle Intervention and Metformin on Preventing or Delaying Diabetes Among Women With and Without Gestational Diabetes: The Diabetes Prevention Program Outcomes Study 10-Year Follow-Up. J Clin Endocrinol Metab. 2015 Apr;100(4):1646–53. pmid:25706240

25. Ratner RE, Christophi CA, Metzger BE, Dabelea D, Bennett PH, Pi-Sunyer X, et al. Prevention of Diabetes in Women with a History of Gestational Diabetes: Effects of Metformin and Lifestyle Interventions. J Clin Endocrinol Metab. 2008 Dec 1;93(12):4774–9. pmid:18826999

26. International Association of Diabetes and Pregnancy Study Groups Consensus Panel. International Association of Diabetes and Pregnancy Study Groups Recommendations on the Diagnosis and Classification of Hyperglycemia in Pregnancy. Diabetes Care. 2010 Mar 1;33(3):676–82. pmid:20190296

27. Walker JD. NICE guidance on diabetes in pregnancy: management of diabetes and its complications from preconception to the postnatal period. NICE clinical guideline 63. London, March 2008: Commentary. Diabet Med. 2008 Aug 29;25(9):1025–7.

28. Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy: A World Health Organization Guideline. Diabetes Res Clin Pract. 2014 Mar;103(3):341–63. pmid:24847517

29. American Diabetes Association Professional Practice Committee. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes—2022. Diabetes Care. 2022 Jan 1;45(Supplement_1):S17–38.

30. Rafat D, Rabbani TK, Ahmad J, Ansari MA. Influence of iron metabolism indices on HbA1c in non-diabetic pregnant women with and without iron-deficiency anemia: Effect of Iron supplementation. Diabetes Metab Syndr Clin Res Rev. 2012 Apr 1;6(2):102–5. pmid:23153978

31. Staimez LR, Kipling LM, Nina Ham J, Legvold BT, Jackson SL, Wilson PWF, et al. Potential misclassification of diabetes and prediabetes in the U.S.: Mismatched HbA1c and glucose in NHANES 2005–2016. Diabetes Res Clin Pract. 2022 Jul;189:109935. pmid:35662612

32. Cavagnolli G, Pimentel AL, Freitas PAC, Gross JL, Camargo JL. Effect of ethnicity on HbA1c levels in individuals without diabetes: Systematic review and meta-analysis. Ali R, editor. PLOS ONE. 2017 Feb 13;12(2):e0171315. pmid:28192447

33. Hare MJL, Magliano DJ, Zimmet PZ, Söderberg S, Joonas N, Pauvaday V, et al. Glucose-Independent Ethnic Differences in HbA1c in People Without Known Diabetes. Diabetes Care. 2013 Jun 1;36(6):1534–40. pmid:23275368

34. Siricharoenthai P, Phupong V. Diagnostic accuracy of HbA1c in detecting gestational diabetes mellitus. J Matern Fetal Neonatal Med. 2020 Oct 17;33(20):3497–500. pmid:30691324

35. Claesson R, Ekelund M, Ignell C, Berntorp K. Role of HbA1c in post-partum screening of women with gestational diabetes mellitus. J Clin Transl Endocrinol. 2015 Mar;2(1):21–5. pmid:29159105

36. Duke A, Yap C, Bradbury R, Hng TM, Kim C, Wansbrough A, et al. The discordance between HbA1c and glucose tolerance testing for the postpartum exclusion of diabetes following gestational diabetes. Diabetes Res Clin Pract. 2015 Apr;108(1):72–7. pmid:25661662

37. Megia A, Näf S, Herranz L, Serrat N, Yañez R, Simón I, et al. The usefulness of HbA1c in postpartum reclassification of gestational diabetes: HbA1c in postpartum reclassification. BJOG Int J Obstet Gynaecol. 2012 Jun;119(7):891–4.

38. Apostolakis M, Paschou SA, Zapanti E, Sarantopoulou V, Vasileiou V, Anastasiou E. HbA1c presents low sensitivity as a post-pregnancy screening test for both diabetes and prediabetes in Greek women with history of gestational diabetes mellitus. Hormones. 2018 Jun;17(2):255–9. pmid:29949125

39. Katreddy MV. Hemoglobin A1c in early postpartum screening of women with gestational diabetes. World J Diabetes. 2013;4(3):76. pmid:23772276

40. Kim KS, Kim SK, Cho YW, Park SW. Diagnostic value of haemoglobin A1c in post-partum screening of women with gestational diabetes mellitus. Diabet Med. 2016 Dec;33(12):1668–72.

41. Picón MJ, Murri M, Muñoz A, Fernández-García JC, Gomez-Huelgas R, Tinahones FJ. Hemoglobin A1c Versus Oral Glucose Tolerance Test in Postpartum Diabetes Screening. Diabetes Care. 2012 Aug 1;35(8):1648–53. pmid:22688550

42. Mbhenyane X, Cherane M. Compliance with the consumption of iron and folate supplements by pregnant women in Mafikeng local municipality, North West province, South Africa. Afr Health Sci. 2017 Sep;17(3):657–70. pmid:29085393

43. Policy—Roadmap for Nutrition in South Africa | Global database on the Implementation of Nutrition Action (GINA) [Internet]. [cited 2023 Apr 3]. https://extranet.who.int/nutrition/gina/en/node/7994

44. WHO Consultation on Obesity (1999: Geneva S, Organization WH. Obesity: preventing and managing the global epidemic: report of a WHO consultation [Internet]. World Health Organization; 2000 [cited 2023 Feb 1]. https://apps.who.int/iris/handle/10665/42330

45. nhlsadmin. QA Division [Internet]. National Health Laboratory Service. [cited 2023 Feb 1]. https://www.nhls.ac.za/quality-assurance/qa-division/

46. Little RR. Glycated Hemoglobin Standardization—National Glycohemoglobin Standardization Program (NGSP) Perspective. 2003 Sep 16;41(9):1191–8.

47. Mortality and causes of death in South Africa:Findings from death notification 2017 [Internet]. STATISTICS SOUTH AFRICA; 2020 Mar. https://www.statssa.gov.za/publications/P03093/P030932017.pdf

48. Tura A, Grespan E, Göbl CS, Koivula RW, Franks PW, Pearson ER, et al. Profiles of Glucose Metabolism in Different Prediabetes Phenotypes, Classified by Fasting Glycemia, 2-Hour OGTT, Glycated Hemoglobin, and 1-Hour OGTT: An IMI DIRECT Study. Diabetes. 2021 Jul 7;70(9):2092–106. pmid:34233929

49. Chivese T, Hirst J, Matizanadzo JT, Custodio M, Farmer A, Norris S, et al. The diagnostic accuracy of HbA1c, compared to the oral glucose tolerance test, for screening for type 2 diabetes mellitus in Africa—A systematic review and meta-analysis. Diabet Med. 2022;39(4):e14754. pmid:34854127

50. Zemlin AE, Matsha TE, Hassan MS, Erasmus RT. HbA1c of 6.5% to Diagnose Diabetes Mellitus—Does It Work for Us?—The Bellville South Africa Study. PLOS ONE. 2011 Aug 12;6(8):e22558.

51. Sacks DB, Arnold M, Bakris GL, Bruns DE, Horvath AR, Kirkman MS, et al. Position Statement Executive Summary: Guidelines and Recommendations for Laboratory Analysis in the Diagnosis and Management of Diabetes Mellitus. Diabetes Care. 2011 Jun 1;34(6):1419–23. pmid:21617111

52. Gunderson EP, Lewis CE, Lin Y, Sorel M, Gross M, Sidney S, et al. Lactation Duration and Progression to Diabetes in Women Across the Childbearing Years: The 30-Year CARDIA Study. JAMA Intern Med. 2018 Mar 1;178(3):328. pmid:29340577

53. Gunderson EP, Hurston SR, Ning X, Lo JC, Crites Y, Walton D, et al. Lactation and Progression to Type 2 Diabetes Mellitus After Gestational Diabetes Mellitus: A Prospective Cohort Study. Ann Intern Med. 2015 Dec 15;163(12):889–98. pmid:26595611

54. Ley SH, Chavarro JE, Li M, Bao W, Hinkle SN, Wander PL, et al. Lactation Duration and Long-term Risk for Incident Type 2 Diabetes in Women With a History of Gestational Diabetes Mellitus. Diabetes Care. 2020 Feb 10;43(4):793–8. pmid:32041900

55. Monroy G, Fernández C, Caballé T, Altimira L, Corcoy R. Breastfeeding effect on glucose tolerance assessment in women with previous gestational diabetes mellitus: A randomized controlled trial. Diabet Med. 2022;39(11):e14954. pmid:36053958

56. Kilpatrick ES. Haemoglobin A1c in the diagnosis and monitoring of diabetes mellitus. J Clin Pathol. 2008 Sep 1;61(9):977–82. pmid:18755721

57. O’Connor C, O’Shea PM, Owens LA, Carmody L, Avalos G, Nestor L, et al. Trimester-specific reference intervals for haemoglobin A1c (HbA1c) in pregnancy. Clin Chem Lab Med. 2012 May 1;50(5):905–9.

58. Stratinaki M, Milaki K, Stavrakis S, Pitarokoilis M, Charitakis E, Stavratis S, et al. Is HBA1C reliable in diagnosing diabetes in the acute phase of acute coronary syndromes? Eur Heart J Acute Cardiovasc Care. 2021 Apr 26;10(Supplement_1):zuab020.104.

59. Cohen RM, Franco RS, Khera PK, Smith EP, Lindsell CJ, Ciraolo PJ, et al. Red cell life span heterogeneity in hematologically normal people is sufficient to alter HbA1c. Blood. 2008 Nov 15;112(10):4284–91. pmid:18694998