Risk factors during the early postpartum period for type 2 diabetes mellitus in women with gestational diabetes

Abstract

For women with gestational diabetes mellitus (GDM), the evaluation of glucose tolerance (GT) in the early postpartum period is universally recommended. Nevertheless, few studies have evaluated the risk factors for T2DM on the basis of GT data obtained during the early postpartum period. We aimed to identify the risk factors for type 2 diabetes mellitus (T2DM) by evaluating GT in the first 12 weeks postpartum (12wPP) in women with GDM and to categorize the risk using a combination of the principal risk factors. This retrospective multicenter observational study included 399 East Asian women with GDM who underwent a 75-g oral glucose tolerance test (OGTT) within 12wPP, which was repeated annually or biennially and used to identify the postpartum development of T2DM. Forty-three women (10.8%) developed T2DM during a median follow-up period of 789 ± 477 days. The independent risk factors for T2DM were pre-pregnancy obesity (BMI ≥25 kg/m²), early postpartum impairment in glucose tolerance (IGT), and an early postpartum glycated hemoglobin (HbA1c) ≥5.7%. The odds ratios (95% confidence intervals) for T2DM were 3.2 (1.3–7.8) in women with either early postpartum IGT or pre-pregnancy obesity, 9.2 (3.0–28.3) in those with early postpartum IGT, pre-pregnancy obesity, and HbA1c <5.7%, and 51.4 (16.1–163.9) in those with early postpartum IGT, pre-pregnancy obesity, and HbA1c ≥5.7%, compared with those without obesity or IGT. T2DM risk in East Asian women with GDM should be stratified according to pre-pregnancy obesity and early postpartum IGT, and these patients should be followed up and receive appropriate care for their risk category.

GESTATIONAL DIABETES MELLITUS (GDM) is characterized by mild carbohydrate intolerance that commences or is first recognized during pregnancy [1]. GDM is associated with various adverse maternal and perinatal outcomes [2]. Although shortly after delivery, the glucose tolerance of most women diagnosed with GDM is restored to normal [3], a meta-analysis of 20 studies showed that the lifetime risk of developing T2DM in women with GDM diagnosed using the previous diagnostic criteria is 7.4-fold higher than that of women with normal glucose tolerance (NGT) during pregnancy [4]. The Hyperglycemia and Adverse Pregnancy Outcome (HAPO) follow-up study showed that women with GDM diagnosed using the International Association of Diabetes and Pregnancy Study Group (IADPSG) criteria [5] still have a 5.4-fold higher risk of developing T2DM a median 11.4 years after their pregnancy than those who did not have GDM [6].

A few studies have shown that there is a high risk of the postpartum progression of abnormalities in glucose metabolism, including T2DM, in women diagnosed with GDM using the IADPSG criteria [7-9].

A meta-analysis performed by Rayanagoudar and colleagues of the risk of developing T2DM in women with GDM found that this was higher for women who used insulin, had a high body mass index (BMI), had a family history of DM, were of non-white ethnicity, were older mothers, were diagnosed early with GDM, had a high fasting plasma glucose (FPG), 1-h glucose, or 2-h glucose concentration when diagnosed with GDM, or had a high concentration of glycated hemoglobin (HbA1c) [10]. However, importantly, all 39 studies included in the meta-analysis used non- IADPSG criteria to diagnose GDM.

In most studies of the risk of developing T2DM in women with GDM, glucose tolerance was assessed at the time of GDM diagnosis. However, the use of a 75-g oral glucose tolerance test (OGTT) [11-13], fasting plasma glucose measurement [14], or HbA1c measurement [14] is recommended worldwide for the re-evaluation of glucose tolerance within 12 weeks of delivery in women with GDM. Nevertheless, glucose tolerance has rarely been assessed in the early postpartum period when determining the risk of subsequent T2DM in published studies, and those that did report these data were performed a quarter-century ago [15, 16]. Recently, the Women’s Preventive Services Initiative recommended that studies should be performed to determine the predictors of subsequent T2DM in women after childbirth, particularly for those with initially negative or borderline screening test results [17]. Indeed, the OGTT data at the diagnosis of GDM may not always be available at the time of early postpartum reevaluation with OGTT.

In this study, we aimed to identify the risk factors for T2DM in East Asian women with GDM diagnosed using the IADPSG criteria, on the basis of glucose tolerance data obtained within 12 weeks of delivery, not on the results of OGTT at the diagnosis of GDM, and to categorize the risk using a combination of the principal risk factors.

Materials and Methods

Study population

In this retrospective multi-center observational study, we collected GDM registration data from three perinatal centers in Japan: the National Center for Child Health and Development in Tokyo (Tokyo), Osaka Women’s and Children’s Hospital in Izumi City (Osaka), and the National Hospital Organization Nagasaki Medical Center in Omura City (Nagasaki). Women with a singleton pregnancy and GDM who delivered between 1995 and 2011 were included, and most participants were Japanese. Women registered at the National Center for Child Health and Development gave birth between March 2002 and September 2011 and were followed up until October 2014, those registered at Osaka Women’s and Children’s Hospital gave birth between April 1995 and June 2010 and were followed up until February 2011, and those registered at the National Hospital Organization Nagasaki Medical Center gave birth between January 2003 and February 2014 and were followed up until December 2014.

GDM was diagnosed using a two-step approach. First, women with a positive 50-g glucose challenge test or who were at a high risk of GDM underwent a diagnostic 75-g OGTT. Second, using the results from the 75-g OGTT, and until June 2010, a diagnosis of GDM was made using the Japan Society of Obstetrics and Gynecology criteria [18] (two or more abnormal plasma glucose values during the 75-g OGTT, with abnormal values being defined as ≥5.6 mmol/L (100 mg/dL) when fasted, ≥10.0 mmol/L (180 mg/dL) at 1 h, and ≥8.3 mmol/L (150 mg/dL) at 2 h. In all three facilities, women with one or more abnormal blood glucose values during the 75-g OGTT were managed as though they had GDM during pregnancy, and underwent further 75 g-OGTTs postpartum and annually or biennially thereafter. After June 2010, the diagnosis of GDM was made using the IADPSG criteria (one or more abnormal plasma glucose value during the 75-g OGTT, with abnormal values being defined as ≥92 mg/dL fasted, ≥180 mg/dL at 1 h, and ≥153 mg/dL at 2 h) [5]. The IADPSG criteria were adopted by the Japan Society of Obstetrics and Gynecology at this time and used at all three study facilities thereafter.

Those women diagnosed with GDM who were taking corticosteroids or those with any values missing from the 75-g OGTTs were excluded. Women diagnosed with GDM using the Japan Society of Obstetrics and Gynecology criteria before May 2010 were re-evaluated using the IADPSG criteria, and those with overt diabetes in pregnancy, according to IADPSG criteria (fasting plasma glucose [FPG] ≥126 mg/dL or HbA1c ≥6.5%), were excluded.

Therefore, the final analysis was conducted on data from those women diagnosed with GDM using the IADPSG criteria who underwent a 75-g OGTT within 12 weeks of delivery, who had not been diagnosed with T2DM, and who had undergone postpartum follow-up using a 75-g OGTT on more than one occasion. We defined “regular” postpartum follow-up as the performance of a 75-g OGTT every 1–2 years. If there were more than 3 years between OGTTs, we regarded the last 75-g OGTT before this period as the final procedure. In women who were followed up for more than 2,000 days after delivery, we used the last 75-g OGTT prior to this 2,000-day time point as the final procedure. Women who were not able to discontinue insulin treatment after childbirth were excluded, because this implied a postpartum diagnosis of T2DM.

Flowchart of study participation

A total of 974 women with GDM and a singleton pregnancy (Tokyo 321, Osaka 336, and Nagasaki 317) were eligible for the study. Of these women, 77 were excluded as follows: one was diagnosed GDM while taking corticosteroids, six had missing OGTT data during pregnancy, and 70 were diagnosed with overt diabetes in pregnancy using the IADPSG criteria. Of the remaining 897 women diagnosed with GDM using the IADPSG criteria, three were diagnosed with clinical postpartum diabetes because insulin treatment could not be stopped after childbirth, 607 underwent a 75-g OGTT within 12 weeks of childbirth, 109 underwent a 75-g OGTT beyond 12 weeks postpartum, and 178 were not followed up postpartum. Of the 607 women who underwent a 75-g OGTT within 12 weeks of giving birth, 25 were diagnosed with diabetes, and 183 did not undergo regular follow-up testing after the first postpartum test. Therefore, data from 399 women were included in the analysis (Fig. 1). 67.6% of women with GDM were evaluated within 12 weeks of delivery using a 75-g OGTT and 65.7% of these were followed up regularly after the early postpartum 75-g OGTT.

Fig. 1

Flowchart showing the number of participants

Glycemic measurements

A 75-g OGTT was performed within 12 weeks of childbirth, after a 10–12 h overnight fast. Blood samples were obtained for the determination of glucose and insulin at 0, 30, 60, and 120 min, and HbA1c was also measured. NGT was defined by a fasting glucose <100 mg/dL and a 2-h glucose <140 mg/dL, impaired glucose tolerance (IGT) was defined by a 2-h glucose of 140–199 mg/dL, and impaired fasting glucose (IFG) was defined by a fasting glucose of 100–125 mg/dL [19].

Diabetes was diagnosed when a participant’s blood glucose values exceeded one or more of the following thresholds during the postpartum 75-g OGTT: FPG ≥126 mg/dL, 2-h plasma glucose ≥200 mg/dL, or HbA1c ≥6.5% [19].

As a surrogate measure of insulin resistance, homeostatic model assessment of insulin resistance (HOMA-IR) was calculated, using fasting glucose and insulin values. As a surrogate measure of insulin response, the insulinogenic index (IGI) was calculated as (30 min plasma insulin–fasting plasma insulin) [μU/mL]/(30 min blood glucose–fasting blood glucose) [mmol/L].

Anthropometric and laboratory measurements

Maternal antepartum and postpartum data were obtained from medical records. At the time of GDM diagnosis, maternal age, parity, pre-pregnancy BMI, history of diabetes in first-degree relatives (father, mother, or sibling), and gestational age were recorded. The use of insulin (>20 units/day during pregnancy [20]) and birth weight (macrosomia was defined by a birth weight ≥4,000 g) were also recorded.

This study was conducted with the approval of the Institutional Review Boards of the National Center for Child Health and Development, Osaka Women’s and Children’s Hospital, and the Nagasaki Medical Center for the collection of clinical data (approval number: 894, May 2015). Informed consent was obtained in the form of an opt-out procedure.

Statistical analysis

Data are presented as n (%) for categorical variables and mean ± standard deviation for continuous variables. A Pearson chi-square test or Student’s t-test was used to compare data between the groups two-sided Mann-Whitney U test was used to compare data from women who were or were not followed up. Differences with p < 0.05 were regarded as statistically significant.

To identify independent predictive factors for T2DM, the following variables were assessed: perinatal facility, maternal age at childbirth, pre-pregnancy BMI, family history of diabetes, GDM diagnosis before 20 weeks of gestation, use of insulin during pregnancy, fetal macrosomia, plasma glucose concentrations (fasting, 1 h, and 2 h) during a 75-g OGTT in the early postpartum period (within 12 weeks of delivery), and HbA1c, IGI, and HOMA-IR in the early postpartum period.

Forward stepwise multivariate logistic regression analysis was performed that included all the clinically important parameters and all the parameters identified as statistically significant in univariate analyses. These variables were dichotomized according to cut-off points determined using receiver operating characteristic (ROC) curves. The results for each predictive factor are presented as odds ratios (ORs) and 95% confidence intervals (CIs). The time to the development of T2DM was described using a Kaplan-Meier curve. Finally, the independent predictive factors identified by our analysis were used to categorize the participants according to their risk of developing T2DM.

All analyses were performed using SPSS version 25 for Windows (IBM Inc., Armonk, NY, USA).

Results

Characteristics of the study participants

Table 1 shows the characteristics of the 399 study participants and those who were excluded from the study who fulfilled the IADPSG criteria and had not developed diabetes by the time they underwent a 75-g OGTT within 12 weeks of giving birth. Those who were excluded from the study were less likely to be using >20 IU/day insulin during pregnancy, and to have lower FPG and higher 1-h glucose concentrations during the early postpartum 75-g OGTT than the 399 study subjects. The values for the other parameters were similar between the two groups. However, the proportion of the study subjects recruited at the Nagasaki facility was higher than at the other two facilities.

Table 1 Characteristics of the 399 study participants and those who were excluded from the study who fulfilled the IADPSG criteria and had not developed diabetes by the time they underwent a 75-g OGTT within 12 weeks of giving birth

Characteristic	The study participants (n = 399)	The participants who were excluded from the study (n = 470)	p value
Facility			<0.001
Tokyo	105 (34.2)	202 (65.8)
Osaka	74 (26.7)	203 (73.3)
Nagasaki	220 (77.2)	65 (22.8)
Age at childbirth, years	34.1 ± 4.8	34.3 ± 4.8	0.61
Pre-pregnancy BMI, kg/m2	23.4 ± 5.0	23.8 ± 5.2	0.35
Primiparous, n (%)	157 (48.0)	136 (48.9)	0.79
Family history of diabetes (%)	98 (48.8)	140 (45.2)	0.43
Antenatal glucose metabolism
GDM diagnosis before 20 weeks of gestation, n (%)	60 (15.0)	62 (13.9)	0.64
apOGTT Fasting plasma glucose, mg/dL	86.6 ± 10.2	88.5 ± 10.4	0.009
apOGTT 1-h plasma glucose, mg/dL	185.9 ± 24.9	181.0 ± 30.0	0.010
apOGTT 2-h plasma glucose, mg/dL	160.6 ± 25.0	158.7 ± 29.1	0.31
Use of >20 units/day insulin during pregnancy, n (%)	87 (21.8)	37 (7.8)	<0.001
Delivery
body weight at birth, g	2,963.4 ± 487.8	3,022.2 ± 524.7	0.09
Macrosomia, n (%)	5 (1.3)	11 (2.3)	0.24

Data are mean (SD), or n (%).

Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus; OGTT, oral glucose tolerance test; ap, antepartum

Forty-three women (10.8%) were found to have T2DM during the median follow-up period of 737 days (range 127–1,994 days).

Table 2 shows the characteristics of the study participants who developed diabetes and those who had not by the time of the final follow-up appointment. The participants who developed diabetes had higher pre-pregnancy BMIs, were more likely to have a family history of diabetes, and had higher fasting, 1-h, and 2-h plasma glucose concentrations, fasting and 2-h plasma insulin concentrations, and HOMA-IR during the OGTTs, and higher HbA1c values in the early postpartum period. There were no differences in 1-h plasma insulin concentration or IGI. There were significant differences in the prevalence of DM at the final follow-up appointment among the three facilities: 18.9% in Osaka, 10.5% in Nagasaki, and 5.7% in Tokyo, respectively (p = 0.02).

Table 2 Characteristics of the study participants who had developed diabetes and those who had not at the end of follow-up

Characteristics	T2DM (n = 43)	NGT/IFG and IGT (n = 356)	p value
Facility			0.019
Tokyo	6 (5.7)	99 (94.3)
Osaka	14 (18.9)	60 (81.1)
Nagasaki	23 (10.5)	197 (89.5)
Final follow up date, postpartum day	657.2 ± 488.8	924.3 ± 636.5	0.008
Age at childbirth, years	34.3 ± 4.5	34.0 ± 4.9	0.70
Pre-pregnancy BMI, kg/m2	27.0 ± 5.6	23.0 ± 4.8	<0.001
Primiparous, n (%)	17 (58.6)	139 (46.8)	0.22
Family history of diabetes, n (%)	17 (81.0)	81 (45.0)	0.016
Use of >20 units/day insulin during pregnancy, n (%)	14 (32.6)	73 (20.5)	0.071
body weight at birth, g	2,969.3 ± 498.4	2,962.7 ± 487.2	0.93
Macrosomia, n (%)	1 (2.3)	4 (1.1)	0.50
ppOGTT
ppOGTT date, postpartum day	51.6 ± 12.2	50.1 ± 11.9	0.44
ppOGTT Fasting plasma glucose, mg/dL	94.2 ± 11.7	88.7 ± 8.6	0.002
ppOGTT 1-h plasma glucose, mg/dL	172.7 ± 32.8	148.0 ± 32.8	<0.001
ppOGTT 2-h plasma glucose, mg/dL	142.7 ± 21.2	120.2 ± 26.1	<0.001
ppOGTT Fasting plasma insulin, ng/mL	7.3 ± 5.2	4.9 ± 3.6	0.001
ppOGTT 1-h plasma insulin, ng/mL	48.6 ± 34.6	42.1 ± 25.8	0.22
ppOGTT 2-h plasma insulin, ng/mL	51.6 ± 40.3	34.5 ± 22.8	0.005
HbA1c, %	5.7 ± 0.397	5.4 ± 0.382	0.005
IGI	0.46 ± 0.29	0.62 ± 0.68	0.14
HOMA-R	1.8 ± 1.3	1.1 ± 0.84	0.002

Data are mean (SD), or n (%).

Abbreviations: T2DM, type 2 diabetes mellitus; NGT, normal glucose tolerance; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; BMI, body mass index; OGTT, oral glucose tolerance test, IGI, insulinogenic index, pp, postpartum.

Risk of T2DM developing

Table 3a shows the results of the univariate logistic regression analysis. Of the known pre-pregnancy and pregnancy risk factors for the subsequent development of T2DM, high pre-pregnancy BMI, a family history of diabetes, and a diagnosis of GDM made before 20 weeks of pregnancy were significant. The risk factors for subsequent T2DM at the early postpartum evaluation were high fasting, 1-h, and 2-h plasma glucose concentrations, high HbA1c, high HOMA-IR, and a diagnosis of IFG or IGT made by 12 weeks after childbirth.

Table 3

Odds Ratios for subsequent T2DM in women with GDM

(a) Univariate logistic regression analysis
Variable	Crude Odds Ratio (95% confidence interval)	p value
Age at childbirth ≥35	0.65 (0.34–1.2)	0.20
Pre-pregnancy BMI ≥25 kg/m2	6.4 (3.2–12.7)	<0.001
Family history of diabetes	2.2 (1.1–4.3)	0.018
GDM diagnosis before 20 weeks gestation	2.1 (1.01–4.5)	0.045
Use of >20 units/day insulin during pregnancy	1.9 (0.94–3.7)	0.074
Macrosomia	2.8 (0.28–27.4)	0.38
ppOGTT
ppOGTT Fasting plasma glucose ≥5.2 mmol/L (93 mg/dL)	2.1 (1.1–3.9)	0.027
ppOGTT 1-h plasma glucose ≥9.2 mmol/L (165 mg/dL)	4.6 (2.5–9.7)	<0.001
ppOGTT 2-h plasma glucose ≥7.6 mmol/L (136 mg/dL)	8.6 (4.2–17.6)	<0.001
ppOGTT IFG (Fasting plasma glucose ≥5.6 mmol/L (100 mg/dL)	3.2 (1.5–6.8)	0.002
ppOGTT IGT (2-h plasma glucose ≥7.8 mmol/L (140 mg/dL)	6.5 (3.3–12.8)	<0.001
HbA1c ≥5.7% (71 mmol/mol)	2.9 (1.5–5.5)	0.001
IGI ≥0.48	0.53 (0.27–1.03)	0.057
HOMA-IR ≥1.003	3.5 (1.8–7.1)	<0.001

(b) Multivariate logistic regression analysis, using a cut off value for fasting and 2-h plasma glucose concentrations 12 weeks postpartum, after adjustment for facility
Variable	Adjusted Odds Ratio (95% confidence interval)	p value
Pre-pregnancy BMI ≥25 kg/m2	5.2 (2.4–11.1)	<0.001
ppOGTT 2-h plasma glucose ≥7.6 mmol/L (136 mg/dL)	6.4 (3.0–13.8)	<0.001
ppOGTT HbA1c ≥5.7% (71 mmol/mol)	2.6 (1.2–5.7)	0.016

(c) Multivariate logistic regression analysis, using the conventional cut off glucose values for IFG and IGT in fasting and 120-min plasma samples 12 weeks postpartum, after adjustment for facility
Variable	Adjusted Odds Ratio (95% confidence interval)	p value
Pre-pregnancy BMI ≥25 kg/m2	4.3 (1.9–9.7)	<0.001
ppOGTT 2-h plasma glucose ≥7.8 mmol/L (140 mg/dL (IGT))	3.9 (1.8–8.3)	<0.001
ppOGTT HbA1c ≥5.7% (71 mmol/mol)	2.6 (1.2–5.7)	0.015

Odds Ratios for subsequent T2DM in women with GDM
(d) Sensitivity analysis
	Total		Nagasaki		Osaka
Variable	Adjusted OR (95% CI)	p value	Adjusted OR (95% CI)	p value	Adjusted OR (95% CI)	p value
Pre-pregnancy BMI ≥25 kg/m2	4.3 (1.9–9.7)	<0.001	5.8 (2.1–15.8)	0.001	1.9 (0.50–6.8)	0.35
ppOGTT 2-h plasma glucose ≥7.8 mmol/L (140 mg/dL (IGT))	3.9 (1.8–8.3)	<0.001	4.5 (1.7–11.7)	0.002	5.4 (1.5–19.7)	0.011
ppOGTT HbA1c ≥5.7% (71 mmol/mol)	2.6 (1.2–5.7)	0.015	1.9 (0.70–4.9)	0.21	3.4 (0.86–13.5)	0.082

Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus; OGTT, oral glucose tolerance test; pp, postpartum; OGTT, oral glucose tolerance test; IGI, insulinogenic index.

*Because the cut-off value for pre-pregnancy BMI using the receiver operating characteristic curve was 24.7 kg/m², we used 25 kg/m² as the cut-off value, which is commonly used.

In the multivariate logistic regression analysis, the following risk factors were included: maternal age at childbirth ≥35 y, pre-pregnancy BMI ≥25 kg/m², a family history of diabetes, a diagnosis of GDM before 20 weeks of pregnancy, use of >20 units/day insulin during pregnancy, early postpartum FPG concentration ≥5.2 mmol/L (93 mg/dL), 1-h plasma glucose concentration ≥9.2 mmol/L ( 165 mg/dL), and 2-h plasma glucose concentration ≥7.6 mmol/L (136 mg/dL), early postpartum HbA1c ≥5.7 %, IGI ≥0.43, and HOMA-IR ≥1.003. We also adjusted for the facility as a standard categorical variable, because we found differences in the postpartum follow-up method, follow-up rate, and the prevalence of T2DM among the three facilities. As shown in Table 3b, we found three risk factors that were independently associated with the development of diabetes: pre-pregnancy BMI ≥25 kg/m² (OR 5.2, 95% CI: 2.4–11.1), 2-h plasma glucose during the early postpartum OGTT ≥136 mg/dL (OR 6.4, 3.0–13.8), and HbA1c at the early postpartum assessment ≥5.7% (OR 2.6, 1.2–5.7). We also conducted multivariate logistic regression analysis using cut-off values for fasting and 2-h plasma glucose concentrations of ≥5.6 mmol/L (100 mg/dL (IFG)) and ≥7.8 mmol/L (140 mg/dL) (IGT), and identified three independent risk factors for the development of postpartum diabetes: pre-pregnancy BMI ≥25 kg/m² (OR 4.3, 95% CI: 1.9–9.7), 2-h plasma glucose during the early postpartum OGTT ≥140 mg/dL (IGT; OR 3.9, 1.8–8.3), and HbA1c at the early postpartum assessment ≥5.7% (OR 2.6, 1.2–5.7) (Table 3c). For clinical practice, we decided to use IGT, defined by a plasma glucose concentration ≥140 mg/dL, as a categorical variable in the further analysis, instead of using the cut-off value of 136 mg/dL.

Then as shown in Table 3d, a sensitivity analysis was performed for each center, and the results of two of these (Nagasaki and Osaka) analyses were similar to the overall results. However, it was not possible to perform such an analysis on the Tokyo participants because the incidence of T2DM was too low in this group.

Fig. 2a shows the cumulative prevalence of T2DM in women with a pre-pregnancy BMI ≥25 kg/m² and early postpartum IGT. We categorized women with pre-pregnancy BMI ≥25 kg/m² and IGT as high-risk, those with a pre-pregnancy BMI ≥25 kg/m² but with NGT, or those with a pre-pregnancy BMI <25 kg/m² and IGT as moderate-risk, and those with pre-pregnancy BMI <25 kg/m² and NGT as low-risk, when women with GDM underwent a 75-g OGTT within 12 weeks of giving birth. We combined those with a pre-pregnancy BMI ≥25 kg/m² and NGT with those with a pre-pregnancy BMI <25 kg/m² and IGT because their cumulative incidences of diabetes were similar. Furthermore, by adding HbA1c ≥5.7% (71 mmol/mol) at the early postpartum assessment to the classification, the high-risk group (BMI ≥25 kg/m² and early postpartum IGT) could be further divided into two groups: those with early postpartum HbA1c <5.7% (71 mmol/mol) (moderately high-risk group) and those with early postpartum HbA1c ≥5.7% (71 mmol/mol) (extremely high-risk group) (Fig. 2b). Thus, women with GDM can be classified into low, moderate, moderately high, and extremely high diabetic risk groups using pre-pregnancy BMI, 2-h plasma glucose concentration during a 75-g OGTT, and HbA1c within 12 weeks of delivery. The proportions of participants in each group were 56.3%, 32.9%, 5.7%, and 5.1%, respectively. Compared with the low risk group, the ORs for the development of T2DM after adjustment for the facility was 3.2 (95% CI: 1.3–7.8, p = 0.01), 9.2 (95% CI: 3.0–28.3, p < 0.00), and 51.4 (95% CI: 16.1–163.9, p < 0.00) for the moderate risk group, the moderately high risk group, and the extremely high risk group, respectively.

Fig. 2

Cumulative incidence of T2DM in women with GDM, according to their risk categorization

(a) According to a pre-pregnancy BMI ≥25 kg/m² and IGT in the early postpartum period

(b) According to pre-pregnancy BMI ≥25 kg/m², and IGT and HbA1c ≥5.7% in the early postpartum period

By adding the parameter HbA_1c ≥5.7% in the early postpartum period to the analysis shown in Fig. 2a, the highest risk group could be further subdivided into two groups. Thus, we obtained four groups of postpartum women with GDM. HbA_1c = glycated hemoglobin; IGT = impaired glucose tolerance (2-h glucose of 140–199 mg/dL).

Discussion

Using OGTT data collected during the first 12 weeks after childbirth, we have shown that a pre-pregnancy BMI ≥25 kg/m², early postpartum IGT, and a HbA1c ≥5.7% (71 mmol/mol) are independent risk factors for subsequent T2DM in women with GDM. A pre-pregnancy BMI ≥25 kg/m² carries almost the same risk as IGT in the early postpartum period, and the risk of developing T2DM can be categorized using a combination of these three risk factors.

The present study is the first to categorize women with GDM diagnosed using the IADPSG criteria according to the risk factors for developing T2DM, identified using OGTT data obtained within 12 weeks of delivery. Only two studies conducted many years ago had previously demonstrated a subsequent risk of T2DM in women with GDM by evaluating glucose tolerance in the early postpartum period [15, 16], but the diagnostic criteria used for GDM and postpartum diabetes differed from those used in the present study. In the study conducted in Denmark [15], an OGTT was performed 8 weeks postpartum, and the independent risk factors for T2DM between 2 and 11 years postpartum were high fasting blood glucose at the time GDM was diagnosed, a high 2-h glucose concentration at the same time, a higher area under the OGTT glucose curve postpartum, and maternal pre-pregnancy obesity (BMI ≥25 kg/m²). Thus, the results were partially consistent with those of the present study. In the other study, which was of Latina women in North America [16], an OGTT was performed 4–16 weeks postpartum, and the independent risk factors for T2DM 5–7 years postpartum were an early diagnosis of GDM, a higher fasting blood glucose during pregnancy, and larger areas under the OGTT glucose curve during pregnancy and postpartum. However, pre-pregnancy BMI was not found to be an independent risk factor. The study population consisted of GDM women with poor glucose tolerance, indicated, for example, by a FPG ≥140 mg/dL before hospital discharge, and moreover, ~70% of the participants had a pre-pregnancy BMI ≥27 kg/m². In this instance, the selection of the study objectives may have led to pre-pregnancy BMI not being among the independent risk factors identified.

Other previous studies have investigated the relationship between glucose tolerance 1 year postpartum and the risk factors for subsequent T2DM. In these studies, 75-g OGTTs and intravenous glucose tolerance tests were conducted between 15 and 30 months after Hispanic women living in North America had given birth [21] and 12 months after delivery by Korean women [22]. Both studies showed that hyperglycemia and poor pancreatic β-cell function 12–30 months postpartum are associated with the subsequent development of T2DM. In the Korean study [22], high glucose concentrations during the OGTT were a predictor of future diabetes, suggesting that glucose tolerance testing is informative even a year after childbirth in an Asian population.

We have also shown that early postpartum IGT, but not IFG, is associated with a similar risk of T2DM as pre-pregnancy obesity in East Asian women with GDM. Women with postpartum IGT after GDM are reported to have a lower insulin sensitivity index and a weaker insulin response to glucose than women with NGT after GDM [23, 24]. In addition, in obese Latina women, early postpartum IGT was associated with an 84% probability of T2DM 5 years after delivery [16]. Thus, in ethnic groups with a high risk of T2DM, the postpartum evaluation of women with GDM should include glucose tolerance testing.

We have also shown pre-pregnancy obesity to be a potent risk factor for T2DM in women with GDM. A recent meta-analysis found the same, because pre-pregnancy BMI ≥25 kg/m² increased the risk of T2DM (relative risk 3.18 [95% CI: 1.96–5.2]) [10]. In addition, postpartum obesity [25] and postpartum weight gain [21, 25, 26] have been reported to be associated with T2DM after GDM. Therefore, it is important to recognize pre-pregnancy obesity as being associated with a high risk of subsequent T2DM in women with GDM, and to encourage obese women to minimize their weight gain or lose weight postpartum. Maternal BMI at the time of postpartum OGTT is a risk factor for T2DM (OR 4.2, 95% CI: 2.0–8.6) as same as he pre-pregnancy BMI. However, at the time the OGTT was performed, within 12 weeks of giving birth, maternal body weight is still returning to normal, rather than being stable. Therefore, we analyzed the pre-pregnancy BMI data.

We have also identified an early postpartum HbA1c ≥5.7 % (71 mmol/mol) as an independent risk factor for T2DM in women with GDM. A similar HbA1c value at the time of GDM diagnosis during pregnancy has also been previously reported to be a risk factor for T2DM [9]. This threshold is similar to that in the general population with regard to T2DM [19, 27].

It is important to prevent or delay the onset of T2DM after GDM, because if T2DM develops before 40 years of age there is a higher risk of diabetic complications than if it develops later [28-30]. Therefore, women diagnosed with GDM during pregnancy should always be followed up postpartum, although various barriers to this may exist, including an underdeveloped follow-up system [31-35].

If all women who suffer from GDM were to be assessed annually or biannually using an OGTT, significant costs and complexity would be involved. To ameliorate these problems, it is very important to categorize women according to their risk of developing T2DM, follow up and receive interventions aimed at preventing T2DM. In the present study, our adopted policy is to calculate pre-pregnancy BMI and assess 2-h plasma glucose concentration during a 75-g OGTT in the early postpartum period in all women with GDM, but if they have both of the risk factors mentioned above, HbA1c should also be measured, to predict their risk of developing T2DM. So women with GDM can be classified into four groups (low, moderate, moderately high, and extremely high diabetic risk groups), and who classified into extremely high diabetic risk groups should be very carefully followed up and receive appropriate interventions, because most will develop T2DM in the near future. A sophisticated and practical method of clinical risk categorization for the follow-up of women with GDM has been described for the first time in this report.

Our study had a number of limitations. First, this was a retrospective study, and the study period, follow-up rate, and the method used to follow up women postpartum differed among the three perinatal centers. This may explain the differences in the incidence of progression of T2DM among these centers. Approximately 70% of the women were in the low risk group in both Tokyo and Nagasaki, while only 50% were in this group in Osaka, which may be the result of differences in socio-economic background in the area where each facility is located. Second, the participants in this study may not have been representative of every woman with GDM diagnosed using the IADPSG criteria, because the participants were regularly followed up after an early postpartum OGTT, were more likely to use insulin during pregnancy, and had higher 1-h glucose concentrations during the early postpartum 75-g OGTT than the women who were excluded from the study. Therefore, individuals who could not mount a compensatory insulin secretory response during pregnancy may have been more likely to be followed up after the early postpartum OGTT. Third, pre-pregnancy BMI was calculated using self-reported weight, not the actual value. In Japan, self-reported weight has been reported to be slightly less than actual weight in obese people [36]. Fourth, the present study participants were East Asian women, who are considered to be a high-risk ethnic group for T2DM [37], meaning that the results may not be universally applicable. Finally, the presence or absence of lactation could not be analyzed as an explanatory variable, because we obtained data regarding lactation from only a quarter of the women. Recently, it has been shown that lactation may be protective against the development of T2DM [38].

In conclusion, the risk of T2DM in women with GDM should be stratified according to the presence of pre-pregnancy obesity and early postpartum IGT, and if both are present, HbA1c should also be taken into account. According to their risk categorization, women with GDM should be followed up and receive appropriate interventions to prevent T2DM.

Acknowledgments

The authors thank Ms. Sie Saito and Ms. Chiharu Matsuzawa, National Center for Child Health and Development, for helping with data management, and Mark Cleasby, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.

Disclosure

The authors declare no conflicts of interest. This work has not previously been published in any other form.

Supporting grants and fellowships

This work was supported by a grant from Health Labor Science Research (grant number H24-Cardiovascular etc. (Lifestyle habit)—General-017) and a grant from Research Promotion and Practical Use for Women’s Health, Japan Agency for Medical Research and Development (AMED) (grant number 15gk0210010h0101/16gk02100012h0001).

References

• 1   Metzger  BE,  Coustan  DR (1998) Summary and recommendations of the Fourth International Workshop-Conference on Gestational Diabetes Mellitus. The Organizing Committee. Diabetes Care 21 Suppl 2: B161–B167.
• 2   Jovanovic  L,  Pettitt  DJ (2001) Gestational diabetes mellitus. JAMA 286: 2516–2518.
• 3   O’Sullivan  JB (1991) Diabetes mellitus after GDM. Diabetes 40 Suppl 2: 131–135.
• 4   Bellamy  L,  Casas  JP,  Hingorani  AD,  Williams  D (2009) Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis. Lancet 373: 1773–1779.
• 5   Metzger  BE,  Gabbe  SG,  Persson  B,  Buchanan  TA,  Catalano  PA, et al. (2010) International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 33: 676–682.
• 6   Lowe  WL Jr.,  Scholtens  DM,  Lowe  LP,  Kuang  A,  Nodzenski  M, et al. (2018) Association of gestational diabetes with maternal disorders of glucose metabolism and childhood adiposity. JAMA 320: 1005–1016.
• 7   Brown  FM,  Wyckoff  J (2017) Application of one-step IADPSG versus two-step diagnostic criteria for gestational diabetes in the real world: impact on health services, clinical care, and outcomes. Curr Diab Rep 17: 85.
• 8   Bartakova  V,  Maluskova  D,  Muzik  J,  Belobradkova  J,  Kankova  K (2015) Possibility to predict early postpartum glucose abnormality following gestational diabetes mellitus based on the results of routine mid-gestational screening. Biochem Med (Zagreb) 25: 460–468.
• 9   Kugishima  Y,  Yasuhi  I,  Yamashita  H,  Sugimi  S,  Umezaki  Y, et al. (2018) Risk factors associated with the development of postpartum diabetes in Japanese women with gestational diabetes. BMC Pregnancy Childbirth 18: 19.
• 10   Rayanagoudar  G,  Hashi  AA,  Zamora  J,  Khan  KS,  Hitman  GA, et al. (2016) Quantification of the type 2 diabetes risk in women with gestational diabetes: a systematic review and meta-analysis of 95,750 women. Diabetologia 59: 1403–1411.
• 11   Blumer  I,  Hadar  E,  Hadden  DR,  Jovanovic  L,  Mestman  JH, et al. (2013) Diabetes and pregnancy: an endocrine society clinical practice guideline. J Clin Endorcinol Metab 98: 4227–4249.
• 12  American Diabetes Association (2018) 13. Management of diabetes in pregnancy: standards of medical care in diabetes-2018. Diabetes Care 41: S137–S143.
• 13  Diabetes Canada Clinical Practice Guideline Expert Committee,  Feig  DS,  Berger  H,  Donovan  L,  Godbout  A,  Kader  T, et al. (2018) Diabetes and pregnancy. Can J Diabetes 42 Suppl 1: S255–S282.
• 14  National Collaborating Centre for Women’s and Children’s Health (2015) National Institute for Health and Care Excellence: Clinical Guidelines. Diabetes in Pregnancy: Management of Diabetes and Its Complications from Preconception to the Postnatal Period. National Institute for Health and Care Excellence (UK) Copyright (c) 2015 National Collaborating Centre for Women’s and Children’s Health., London: 506–547.
• 15   Damm  P,  Kuhl  C,  Bertelsen  A,  Molsted-Pedersen  L (1992) Predictive factors for the development of diabetes in women with previous gestational diabetes mellitus. Am J Obstet Gynecol 167: 607–616.
• 16   Kjos  SL,  Peters  RK,  Xiang  A,  Henry  OA,  Montoro  M, et al. (1995) Predicting future diabetes in Latino women with gestational diabetes. Utility of early postpartum glucose tolerance testing. Diabetes 44: 586–591.
• 17  Screening for Diabetes Mellitus After Pregnancy. https://www.womenspreventivehealth.org/recommendations/screening-for-diabetes-mellitus-after-pregnancy/.
• 18  The Committee on Nutrition and Metabolism of the Japan Society of Obstetrics and Gynaecology (1984) The committee report. The Committee Report Acta Obstet Gynaecol Jpn 36: 2055–2058 (In Japanese).
• 19  American Diabetes Association (2019) 2. Classification and diagnosis of diabetes: standards of medical care in diabetes-2019. Diabetes Care 42: S13–S28.
• 20   Yamashita  H,  Yasuhi  I,  Kugishima  Y,  Sugimi  S,  Umezaki  Y, et al. (2018) Factors associated with patients with gestational diabetes in Japan being at increased risk of requiring intensive care. Int J Gynaecol Obstet 140: 170–174.
• 21   Xiang  AH,  Kawakubo  M,  Trigo  E,  Kjos  SL,  Buchanan  TA (2010) Declining beta-cell compensation for insulin resistance in Hispanic women with recent gestational diabetes mellitus: association with changes in weight, adiponectin, and C-reactive protein. Diabetes Care 33: 396–401.
• 22   Oh  TJ,  Kim  YG,  Kang  S,  Moon  JH,  Kwak  SH, et al. (2019) Oral glucose tolerance testing allows better prediction of diabetes in women with a history of gestational diabetes mellitus. Diabetes Metab J 43: 342–349.
• 23   Ignell  C,  Shaat  N,  Ekelund  M,  Berntorp  K (2013) The impact of ethnicity on glucose homeostasis after gestational diabetes mellitus. Acta Diabetol 50: 927–934.
• 24   Lim  S,  Choi  SH,  Park  YJ,  Park  KS,  Lee  HK, et al. (2007) Visceral fatness and insulin sensitivity in women with a previous history of gestational diabetes mellitus. Diabetes Care 30: 348–353.
• 25   Bao  W,  Yeung  E,  Tobias  DK,  Hu  FB,  Vaag  AA, et al. (2015) Long-term risk of type 2 diabetes mellitus in relation to BMI and weight change among women with a history of gestational diabetes mellitus: a prospective cohort study. Diabetologia 58: 1212–1219.
• 26   Moon  JH,  Kwak  SH,  Jung  HS,  Choi  SH,  Lim  S, et al. (2015) Weight gain and progression to type 2 diabetes in women with a history of gestational diabetes mellitus. J Clin Endorcinol Metab 100: 3548–3555.
• 27   Nanri  A,  Nakagawa  T,  Kuwahara  K,  Yamamoto  S,  Honda  T, et al. (2015) Development of risk score for predicting 3-year incidence of type 2 diabetes: Japan epidemiology collaboration on occupational health study. PloS One 10: e0142779.
• 28   Chan  JC,  Lau  ES,  Luk  AO,  Cheung  KK,  Kong  AP, et al. (2014) Premature mortality and comorbidities in young-onset diabetes: a 7-year prospective analysis. Am J Med 127: 616–624.
• 29   Dart  AB,  Martens  PJ,  Rigatto  C,  Brownell  MD,  Dean  HJ, et al. (2014) Earlier onset of complications in youth with type 2 diabetes. Diabetes Care 37: 436–443.
• 30   Huo  X,  Gao  L,  Guo  L,  Xu  W,  Wang  W, et al. (2016) Risk of non-fatal cardiovascular diseases in early-onset versus late-onset type 2 diabetes in China: a cross-sectional study. Lancet Diabetes Endocrinol 4: 115–124.
• 31   Aziz  S,  Munim  TF,  Fatima  SS (2018) Post-partum follow-up of women with gestational diabetes mellitus: effectiveness, determinants, and barriers. J Matern Fetal Neonatal Med 31: 1607–1612.
• 32   Bennett  WL,  Ennen  CS,  Carrese  JA,  Hill-Briggs  F,  Levine  DM, et al. (2011) Barriers to and facilitators of postpartum follow-up care in women with recent gestational diabetes mellitus: a qualitative study. J Womens Health (Larchmt) 20: 239–245.
• 33   McGovern  A,  Butler  L,  Jones  S,  van Vlymen  J,  Sadek  K, et al. (2014) Diabetes screening after gestational diabetes in England: a quantitative retrospective cohort study. Br J Gen Pract 64: e17–e23.
• 34   Nielsen  KK,  Kapur  A,  Damm  P,  de Courten  M,  Bygbjerg  IC (2014) From screening to postpartum follow-up—the determinants and barriers for gestational diabetes mellitus (GDM) services, a systematic review. BMC Pregnancy Childbirth 14: 41.
• 35   Rafii  F,  Rahimparvar  SFV,  Mehrdad  N,  Keramat  A (2017) Barriers to postpartum screening for type 2 diabetes: a qualitative study of women with previous gestational diabetes. Pan Afr Med J 26: 54.
• 36   Anai  A,  Ueda  K,  Harada  K,  Katoh  T,  Fukumoto  K, et al. (2015) Determinant factors of the difference between self-reported weight and measured weight among Japanese. Environ Health Prev Med 20: 447–454.
• 37   Yoon  KH,  Lee  JH,  Kim  JW,  Cho  JH,  Choi  YH, et al. (2006) Epidemic obesity and type 2 diabetes in Asia. Lancet 368: 1681–1688.
• 38   Gunderson  EP,  Hurston  SR,  Ning  X,  Lo  JC,  Crites  Y, et al. (2015) Lactation and progression to type 2 diabetes mellitus after gestational diabetes mellitus: a prospective cohort study. Ann Intern Med 163: 889–898.