Type 2 diabetes screening 1-3 years postpartum in individuals with gestational diabetes

SUMMARY

In a retrospective cohort of 5,801 births from individuals with gestational diabetes, 363 (6%) completed 4-12-week postpartum screening for GDM and had evidence of ongoing clinical care in our system. Among these, only half completed recommended type 2 diabetes screening 1–3 years postpartum. Screening rates did not vary by demographic or clinical factors and were primarily performed by primary care clinicians, highlighting persistent gaps in long-term diabetes surveillance after gestational diabetes.

INTRODUCTION

Gestational diabetes (GDM) is one of the most common pregnancy complications affecting 6% of pregnancies in the US.[1] Individuals with GDM in pregnancy have up to a 60% lifetime risk of developing type 2 diabetes (T2DM).[2] Given this increased risk, the American College of Obstetricians and Gynecologists (ACOG) and the American Diabetes Association (ADA) recommend that individuals with GDM have enhanced lifelong screening for T2DM following their pregnancy. Guidelines state that individuals with GDM should be screened for diabetes between 4-12 weeks postpartum and if negative, every 1–3 years thereafter.[3]

While it is known that overall adherence to T2DM screening at 4-12 weeks is low,[4] little is known about provider adherence to diabetes screening beyond 1 year postpartum among US populations. We sought to quantify adherence to this guideline and identify factors associated with screening completion.

MATERIAL AND METHODS

This is a retrospective cohort study using an established obstetric outcomes database within an academic-community healthcare system that includes an urban tertiary care academic hospital and five suburban and rural satellite hospitals[5]. This study received IRB approval (# STUDY00016861) and is reported using STROBE guidelines.[6]

We identified index, singleton, non-anomalous pregnancies affected by GDM with birth between 2014-2019. Individuals were diagnosed with GDM in pregnancy using a two-step approach in accordance with ACOG guidelines[7]. Because many individuals receive obstetric and primary care in private practices, laboratory data are not uniformly integrated into the system electronic health record (EHR), limiting the ability to ascertain longitudinal follow-up. To ensure data completeness, we restricted our cohort to individuals with documented ongoing engagement in our health system postpartum. Specifically, we only included those who (1) completed a 4–12-week postpartum 75-gram OGTT that ruled out T2DM AND (2) had documented outpatient clinical care within our system 1–3 years after birth.

Exclusion criteria included: multifetal gestation, major fetal anomalies, pregestational diabetes, individuals who did not receive care within our health system within specified postpartum time frame, and subsequent births by the same individual. We defined major fetal anomalies as birth defects diagnosed prenatally that may result in a neonatal intensive care unit admission or neonatal surgery. This study timeframe was chosen because the ADA first recommended this enhanced T2DM screening in 2013, and we expected a 1-year grace period to allow for its incorporation into care. We chose 2019 as the endpoint so as to capture individuals undergoing 3-year screening at the time this study was initiated.

The electronic health record was used to assess whether individuals had recommended screening using hemoglobin A1c (A1c), fasting plasma glucose, or 75-gram OGTT. Clinical characteristics and clinician types were compared between groups using Student's t-tests, Chi-square or Fisher's exact tests as indicated. Reported p-values are two-sided and a significance level of 0.05 was used. Statistical analyses were performed using R (version 4.2.2, R Core Team).

RESULTS

Between 2014-2019, 5,801 of 53,659 births from individuals with GDM were identified. We identified 363 that completed 4–12-week postpartum screening, of which, 297 had ongoing clinical care within our system and no evidence of T2DM on this immediate postpartum screening (Figure 1). 152 (51.2%) were screened for diabetes 1-3 years postpartum. There was no difference in 1–3-year screening rates based on age, parity, race, BMI, insurance status, GDM type, GDM medication use, or number of previous pregnancies impacted by GDM (Table 1). Of those that received screening, 4.6% met criteria for T2DM.

Figure 1:

Flow Chart

Table 1:

	Screened 1-3 years after birth  N=152)	Not screened 1-3 years after birth (N=145)	P-Value1
Age, mean (SD)	33.4 (4.6)	33.0 (4.1)	0.426
Multiparous, n (%)	87 (57.2%)	77 (53.1%)	0.474
Race2, n (%)			0.330
White	72 (51.4%)	73 (54.5%)
Black	32 (22.9%)	22 (16.4%)
Asian	29 (20.7%)	36 (26.9%)
American Indian or Alaska Native	2 (1.4%)	0 (0.0%)
More than 1 race	5 (3.6%)	3 (2.2%)
Pre-pregnancy BMI3 (kg/m2), mean (SD)	29.2 (7.1)	28.3 (5.8)	0.449
Insurance type, n (%)			0.943
Uninsured	10 (6.6%)	11 (7.6%)
Public	41 (27.0%)	39 (26.9%)
Private	101 (66.4%)	95 (65.5%)
Diabetes type, n (%)			0.404
GDM without medication	65 (42.1%)	69 (46.9%)
GDM requiring medication	87 (57.9%)	76 (53.1%)
Diabetes medication type, n (%)			0.165
None	65 (42.8%)	69 (47.6%)
Insulin	56 (36.8%)	52 (35.2%)
Metformin	3 (2.0%)	0 (0.0%)
Glyburide	27 (17.8%)	21 (14.5%)
Insulin and metformin	1 (0.7%)	0 (0.0%)
Other	0 (0.0%)	3 (2.1%)
Number of pregnancies affected by GDM, n (%)			0.197
1	120 (78.9%)	119 (82.1%)
2	26 (17.1%)	24 (16.6%)
3	6 (3.9%)	1 (0.7%)
4	0 (0.0%)	1 (0.7%)

BMI: body mass index; GDM: gestational diabetes, SD: standard deviation

¹To investigate the association between 1-3 years postpartum screening and continuous variables, Student's t-tests were used. Chi-square or Fisher's exact tests were used for categorical variables.

²Race was missing for 12 individuals who were screened and 11 who were not screened.

³BMI was missing for 1 individual who was screened and 1 individual who was not screened.

The postpartum screening was completed by a primary care clinician (43.4%), obstetrician-gynecologist (OBGYN) (21.7%), midwife (12.5%), endocrinologist (0.7%), unknown provider type (4.6%), or multiple clinician types (15.1%). A1c (88.2%) was the primary screening test, with 75-gram OGTT (1.3%) and multiple screening tests (9.2%) used less frequently (Supplemental Table 1).

DISCUSSION

Approximately 50% of individuals who complete a 4–12-week postpartum OGTT and have ongoing clinical care in our system were screened for T2DM 1-3 years postpartum. We did not identify differences in clinical characteristics between the two cohorts. Primary care clinicians ordered screening tests most often, and A1c was the most frequently ordered test. This highlights opportunities to improve screening beyond the 1^st postpartum year and demonstrates that primary care clinicians can play an important role in performing this screening.

Strengths of this study include the use of a large clinical obstetric outcomes database and defined criteria to identify individuals who remain engaged with care within the system (completing both the 2-hour OGTT at 4–12-weeks and having evidence of ongoing clinical care in the system). By examining pregnancies between 2014-2019, we ensured that sufficient time was available to examine relevant testing outcomes in the system. Several limitations are noted. Restriction to individuals with ongoing care within our system and completion of a 4–12-week postpartum oral glucose tolerance test—which is acknowledged to be completed by only a subset of eligible individuals[4, 8]—likely enriched the cohort for patients with higher healthcare engagement, limiting generalizability. Additionally, laboratory testing performed outside the system may not have been captured, although this was partially mitigated by the requirement for continued in-system care. While this study focuses on guidelines from the ADA and ACOG, other international organizations suggest additional screening for overt diabetes for those with a history of GDM and therefore has generalizability to other populations.[9, 10]

Given that up to 60% of individuals with GDM develop T2DM in their lifetime,[2] ACOG and ADA guidelines recommend that individuals with a history of GDM have lifelong screening for T2DM. Studies have previously demonstrated poor adherence to T2DM screening in the first 1-3 months postpartum, however, less is known about how often screening happens in the 1-3 years following pregnancy. One study identified annual screening rates of 23% following pregnancy with GDM,[8] and a recent analysis of claims data found that only 50% of individuals with GDM access primary care services within one year of birth.[11] Here, we find that in a population demonstrating ongoing engagement within a health system, only 51% receive indicated screening within 1-3 years postpartum. We suspect that rates are overall lower when accounting for those who do not engage in care.

This work highlights opportunities to improve rates of screening for individuals with GDM. Timely diagnosis of diabetes could allow for earlier interventions to reduce risk of diabetes-related complications. Further work is needed to evaluate effective interventions to improve this guideline-directed screening.

Highlights.

• 51% of individuals with gestational diabetes had type 2 diabetes screening 1–3 years postpartum.

• Screening did not differ by age, insurance status, gestational diabetes type or history.

• Screening was most frequently completed by primary care clinicians.

• Hemoglobin A1c was the most commonly used screening test.

Abbreviations:

GDM

Gestational Diabetes

T2DM

Type 2 Diabetes Mellitus

Data Availability:

All de-identified data supporting the findings of this study are available upon reasonable request.