Objective
The purpose of this study was to assess an early hemoglobin A1c (HgbA1c) value from 5.7-6.4% as an early predictor of progression to gestational diabetes (GDM).
Study Design
A retrospective cohort study was performed on all women who delivered at a single institution over 2 years who had an early screening HgbA1c test performed at ≤20 weeks of gestation. Women with known preexisting diabetes mellitus or HgbA1c values ≥6.5% were excluded. The primary outcome was GDM development. Secondary outcomes included delivery route, maternal weight gain, birthweight, and neonatal morbidities. Women with an HgbA1c value of 5.7-6.4% were compared with those with an HgbA1c level of <5.7%.
Results
Nearly one-third of those patients in the HgbA1c 5.7-6.4% group (27.3%) experience the development of GDM compared with only 8.7% in the HgbA1c <5.7% group (odds ratio, 3.9; 95% confidence level, 2.0–7.7). This 3-fold increase remained significant (adjusted odds ratio, 2.4) after adjustment for age, prepregnancy body mass index, gestational age at HgbA1c collection, gestational age at screening, ethnicity, and method of screening. There were no significant differences in the need for medical treatment, weight gain, delivery route, birthweight, macrosomia, or neonatal morbidities.
Conclusion
More than 10% of patients in our cohort had an early screening HgbA1c value of 5.7-6.4%. Women in this group have a significantly higher risk of progression to GDM compared with women with normal HgbA1c values and should be considered for closer GDM surveillance and possible intervention.
Gestational diabetes mellitus (GDM) is an increasingly common maternal condition with proven maternal and fetal morbidity and a prevalence of approximately 6-8% in the United States. The prevalence of GDM is even higher in obese pregnancies, ranging from 7-14% Associated fetal risks of GDM include fetal death, macrosomia, shoulder dystocia, hypoglycemia, respiratory distress syndrome, and childhood obesity. Maternal risks include preeclampsia, cesarean delivery, and, importantly, increased risk of the development of type 2 diabetes mellitus later in life. GDM also carries an economic burden that results in an increase of 25-34% in maternity care costs and a 49% increase in neonatal intensive care unit costs, compared with those pregnancies without GDM. Identification and treatment of even mild GDM may reduce adverse pregnancy outcome, which underscores the need to screen properly for and diagnose this important comorbidity.
Glycated hemoglobin (HgbA1c) is a form of hemoglobin that characterizes a patient’s plasma glucose over a prolonged period of time. It is typically used in the nonpregnant population as both a screening tool for diabetes mellitus and as a tool to assess the glycemic control of known diabetic patients. HgbA1c has limited efficacy in the diagnosis of GDM in the third trimester and has demonstrated, at best, moderate sensitivity (85.7%), but poor specificity (61.1%). Compared with maternal glucose levels, HgbA1c has been shown to be less predictive of certain adverse pregnancy outcomes. In the United States, its use has not been standardized for GDM management by the American College of Obstetrics and Gynecology (ACOG) or the United States Preventive Services Task Force. In contrast, organizations such as the American Diabetes Association (ADA) and the International Association of Diabetes in Pregnancy Study Group (IADPSG) support its use in pregnancy.
Our practice uses an early GDM screening strategy whereby all patients, regardless of prepregnancy GDM risk, undergo an early screening HgbA1c at their first prenatal visit. If patients have HgbA1c value of ≥6.5%, they are diagnosed automatically as having overt diabetes mellitus and are referred immediately for diabetic counseling and treatment. If their HgbA1c value is <6.5%, they do not typically undergo any additional GDM testing until 24-28 weeks of gestation, at which time they undergo routine oral glucose tolerance testing.
O’Connor et al investigated normal values for HgbA1c in pregnancy in nondiabetic women and identified a normal reference interval of 4.3-5.4% in the first trimester. In the nonobstetric population, it has been shown that an HbgA1C value of 5.7-6.4% indicates impaired glucose tolerance and a high risk of future diabetes. HgbA1c levels of 5.5-6.0% have been associated with a 5-year cumulative incidence of diabetes mellitus that range from 12-25%. This study aims to correlate early HgbA1c levels and the subsequent development of GDM. Given the evidence that an HgbA1C level of 5.7-6.4% indicates impaired glucose tolerance and is associated with development of diabetes mellitus in nonpregnant patients, we hypothesized that those patients with HgbA1C levels of 5.7-6.4% are at increased risk for the development of GDM when compared with those patients with an HgbA1c level of <5.7%. We also sought to assess differences in pregnancy and neonatal outcomes between these 2 groups.
Materials and Methods
This was a retrospective study of a cohort of patients from MemorialCare Center for Women at Miller Children’s Hospital in Long Beach, CA. Women were included if they had a screening HgbA1c test at ≤20 0/7 weeks of gestation and had been delivered by our practice from January 2011 to January 2013. Women were excluded if they had known pregestational diabetes mellitus, their HgbA1c result was ≥6.5% (which indicates overt diabetes mellitus), they never underwent GDM screening or assessment, or ultimately did not deliver at our institution. The MemorialCare Health Systems Institutional Review Board approved the study. A comprehensive prenatal chart review was performed with the use of both inpatient and outpatient data.
The 2 groups that were used for our comparative analyses were patients with HgbA1c levels of 5.7-6.4% vs those with HgbA1c levels of <5.7%. The primary outcome was the development of GDM. Secondary outcomes included other maternal and delivery characteristics (route of delivery, weight gain, and glucose values at time of GDM screening). Neonatal outcomes included a composite of adverse outcomes such as neonatal intensive care unit admission, neonatal hypoglycemia, hyperbilirubinemia, transient tachypnea, or acute respiratory distress. Other neonatal outcomes included 5-minute Apgar score <7, birthweight, macrosomia (defined as birthweight >4 kg), small for gestational age (defined as birthweight <10th percentile), and intrauterine fetal death.
Recent Los Angeles and Orange County healthcare surveys have demonstrated the prevalence of GDM to be 7-12% in our area. Based on these figures, we estimated that patients with an HgbA1c level of <5.7% would have a lower GDM prevalence of approximately 5%, which is slightly lower than the local GDM prevalence. To detect a clinically significant increase in GDM prevalence from 5% in the group with an HgbA1c level of <5.7% to 15% in the group with an HgbA1c level of 5.7-6.4%, a sample size of at least 55 women in each group would be required when we used an alpha error of .05 and 80% power.
We used the Mann-Whitney U test and the Student t test for comparison of continuous variables and Pearson’s chi-square or Fisher exact test for discrete variables. We also performed multivariable logistic regression analysis and analysis of covariance to perform adjustment for confounders in determining maternal outcomes. Adjustments were made for age, race/ethnicity, prepregnancy body mass index (BMI), gestational age at HgbA1c sample collection, gestational age at GDM screening/diagnosis, and method of GDM screening (2-step vs 1-step screening). Two-step screening entailed a 1-hour 50-g glucose test, which, if >140 mg/dL, was followed by a 3-hour 100-g glucose according to Carpenter-Coustan criteria. Women screened by the 1-step screening were given a 2-hour 75-g glucose test. Results were expressed in odds ratios (ORs), 95% confidence intervals (CIs), and mean ± standard deviation. Significance was defined as a probability value of < .05. SPSS software (version 20.0; IBM Corporation, Armonk, NY) was used for statistical analysis.
Results
There were 526 women who met inclusion criteria during the study period: 471 of the women had an HgbA1c level of <5.7%; 55 women had an HgbA1c level of 5.7-6.4%. An HgbA1c sample was drawn at gestational ages that ranged from 2 5/7 to 20 0/7 weeks of gestation, with a mean gestational age of 9.6 weeks.
Table 1 shows the baseline characteristics between women with HgbA1c levels of <5.7% vs 5.7-6.4%. Women in the HgbA1c 5.7-6.4% group trended towards being older, to have a higher prepregnancy BMI, to have higher HbgA1C values, and to be at a slightly earlier gestational age at testing. A lower proportion was screened by the 2-step method, and a higher proportion had a previous pregnancy that had been complicated by GDM.
| Baseline characteristic | Hemoglobin A1c value | P value | |
|---|---|---|---|
| <5.7% (n = 471) | 5.7-6.4% (n = 55) | ||
| Age, y a | 28.8 ± 6.4 | 30.8 ± 6.8 | .06 |
| Prepregnancy body mass index, kg/m 2 b | 26.5 (22.5–30.8) | 30.3 (23.9–35.4) | .002 |
| Body mass index ≥30 kg/m 2 , n/N (%) | 129/465 (27.7) | 30/55 (54.5) | < .001 |
| Gestational age at delivery, wk a | 39.2 ± 1.9 | 39.0 ± 2.1 | .95 |
| Hemoglobin A1c value a | 5.2 ± 0.3 | 5.9 ± 0.2 | < .001 |
| Gestational age at hemoglobin A1c testing, wk a | 11.4 ± 3.7 | 9.7 ± 2.7 | .003 |
| Gestational age at diabetes screen/diagnosis, wk a | 26.2 ± 1.9 | 24.3 ± 5.4 | .08 |
| Race/ethnicity, n (%) | .03 | ||
| African American | 69 (14.6) | 17 (30.9) | |
| Hispanic | 264 (56.1) | 29 (52.7) | |
| White | 73 (15.5) | 3 (5.5) | |
| Asian | 58 (12.3) | 5 (9.1) | |
| Other | 7 (1.5) | 1 (1.8) | |
| Parity, n (%) | .68 | ||
| 0 | 206 (43.7) | 23 (41.8) | |
| 1-4 | 250 (53.1) | 29 (52.7) | |
| >4 | 15 (3.2) | 3 (5.5) | |
| Method of screening (2-step), n/N (%) | 366/471 (77.7) | 30/52 (57.7) | .001 |
| Previous pregnancy complicated by gestational diabetes mellitus, n/N (%) | 14/265 (5.3) | 5/32 (15.6) | .02 |
| Pregnancy conceived with assisted reproductive technology, n (%) | 7 (1.5) | 0 | — |
a Data expressed as mean ± standard deviation
Table 2 demonstrates the maternal outcomes between the groups. After adjustments were made for age, prepregnancy BMI, gestational age at HgbA1c sample collection, gestational age at screening/diagnosis, race/ethnicity, and method of screening, women with HgbA1c levels of 5.7-6.4% were found to have a 2.4-fold higher odds of the development of GDM when compared with their HgbA1c <5.7% counterparts (27.3% vs 8.7%; adjusted OR, 2.38; 95% CI, 1.01–5.63). There was no statistically significant difference between the groups in terms of those who needed medical treatment or the type of medical treatment (insulin vs oral hypoglycemic). There were no differences in cesarean delivery rates or weight gain. The mean 1-hour 50-g glucose tolerance value was significantly higher in the HgbA1c 5.7-6.4% group (130.2 vs 114.3 mg/dL; P = .02), as was the fasting glucose tolerance value (90.6 vs 81.1 mg/dL; P < .001).
| Outcome | Hemoglobin A1c value | Crude odds ratio (95% confidence interval) | P value | Adjusted odds ratio a (95% confidence interval) | P value | |
|---|---|---|---|---|---|---|
| <5.7% (n = 471) | 5.7-6.4% (n = 55) | |||||
| Development of gestational diabetes mellitus, n (%) | 41 (8.7) | 15 (27.3) | 3.93 (2.00–7.72) | < .001 | 2.38 (1.01–5.63) | .048 |
| Development of class A2 diabetes mellitus, n (%) | 17 (41.5) | 8 (53.3) | 1.61 (0.49–5.30) | .43 | 1.43 (0.19–11.02) | .73 |
| Insulin treatment, n (%) | 12 (70.6) | 6 (75.0) | 1.25 (0.19–8.44) | .82 | 1.60 (0.09–29.71) | .75 |
| Oral hypoglycemic treatment, n (%) | 5 (29.4) | 2 (25.0) | 0.80 (0.12–5.40) | .82 | 1.23 (0.09–18.19) | .86 |
| Cesarean delivery (nonelective), n/N (%) | 107/443 (24.2) | 14/52 (26.9) | 1.16 (0.60–2.22) | .66 | 0.94 (0.43–2.06) | .88 |
| Weight gained, lb b | 29.7 ± 17.0 | 27.2 ± 15.1 | — | .30 | — | .70 |
| Weight gain exceeding Institute of Medicine guidelines, n/N (%) | 228/465 (49.0) | 24/55 (43.6) | 0.81 (0.46–1.41) | .45 | 0.90 (0.48–1.70) | .75 c |
| 1-hr 50-g glucose tolerance test result, mg/dL b | 114.5 ± 30.3 | 128.9 ± 28.9 | — | .01 | — | .04 c |
| Fasting value from glucose tolerance test result, mg/dL b | 81.2 ± 11.1 | 89.1 ± 19.4 | — | .004 | — | .003 c |
a Adjusted for age, prepregnancy body mass index, race/ethnicity, gestational age at HgbA1c collection, gestational age at screening/diagnosis, and method of screening
b Data are expressed as mean ± standard deviation
c From an analysis of covariance model adjusted for age and race.
Three of the women in our cohort (early screening HgbA1c values of 6.1%-6.3%) were diagnosed immediately with GDM by their providers after their HgbA1c values (samples drawn at 9 0/7 to 11 6/7 weeks of gestation) were reviewed. These patients were initiated immediately on self-blood glucose monitoring and thus never underwent glucose screening at 24-28 weeks of gestation. All 3 women demonstrated evidence of hyperglycemia. Dietary and exercise modifications were initiated, but 2 of the 3 women ultimately required medical treatment.
Neonatal outcomes are listed in Table 3 . There were no differences between the 2 groups with respect to composite adverse neonatal outcome (including neonatal intensive care admission, hypoglycemia, hyperbilirubinemia, transient tachypnea, or acute respiratory distress), low 5-minute Apgar score, birthweight, macrosomia, small for gestational age, or intrauterine fetal death.
| Outcome | Hemoglobin A1c value | Odds ratio (95% confidence interval) | P value | |
|---|---|---|---|---|
| <5.7% (n = 471) | 5.7-6.4% (n = 55) | |||
| Composite adverse neonatal outcome, n (%) | 108 (22.9) | 11 (20.0) | 0.84 (0.42–1.68) | .62 |
| Neonatal intensive care admission, n (%) | 77 (71.3) | 6 (54.5) | 0.48 (0.14–1.70) | .48 |
| Neonatal hypoglycemia, n (%) | 5 (4.6) | 1 (9.1) | 2.06 (0.22–19.41) | .52 |
| Neonatal hyperbilirubinemia, n (%) | 56 (51.9) | 8 (72.7) | 2.48 (0.62–9.84) | .19 |
| Transient tachypnea, n (%) | 4 (3.7) | 0 | — | — |
| Acute respiratory distress, n (%) | 5 (4.6) | 0 | — | — |
| 5-minute Apgar score <7, n (%) | 3 (0.6) | 1 (1.8) | 2.89 (0.30–28.26) | .36 |
| Birthweight, g a | 3325 ± 526 | 3306 ± 629 | – | .81 |
| Macrosomic infant (>4 kg birthweight), n (%) | 40 (8.5) | 5 (9.1) | 1.08 (0.41–2.86) | .88 |
| Small-for-gestational-age infant, n (%) | 8 (1.7) | 2 (3.6) | 2.18 (0.45–10.55) | .28 |
| Intrauterine fetal death, n (%) | 1 (0.2) | 1 (1.8) | 8.70 (0.54–141.15) | .20 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
