Obstetrics & Gynecology:
The Relationship Between Maternal Glycemia and Perinatal Outcome
Landon, Mark B. MD; Mele, Lisa ScM; Spong, Catherine Y. MD; Carpenter, Marshall W. MD; Ramin, Susan M. MD; Casey, Brian MD; Wapner, Ronald J. MD; Varner, Michael W. MD; Rouse, Dwight J. MD; Thorp, John M. Jr MD; Sciscione, Anthony DO; Catalano, Patrick MD; Harper, Margaret MD, MSc; Saade, George MD; Caritis, Steve N. MD; Sorokin, Yoram MD; Peaceman, Alan M. MD; Tolosa, Jorge E. MD, MSCE; Anderson, Garland D. MD; for the Eunice Kennedy Shriver National Institute of Child Health, and Human Development (NICHD) Maternal–Fetal Medicine Units (MFMU) Network
From the Department of Obstetrics and Gynecology, The Ohio State University, Columbus, Ohio; The George Washington University Biostatistics Center, Washington, DC; the Eunice Kennedy Shriver National Institute of Child Health and Human Development, Bethesda, Maryland; the Departments of Obstetrics and Gynecology at Brown University, Providence, Rhode Island; The University of Texas Health Science Center at Houston, Houston, Texas; The University of Texas Southwestern Medical Center, Dallas, Texas; Columbia University, New York, New York; University of Utah, Salt Lake City, Utah; University of Alabama at Birmingham, Birmingham, Alabama; University of North Carolina, Chapel Hill, North Carolina; Drexel University, Philadelphia, Pennsylvania; Case Western Reserve University–MetroHealth Medical Center, Cleveland, Ohio; Wake Forest University Health Sciences, Winston-Salem, North Carolina; University of Texas Medical Branch, Galveston, Texas; University of Pittsburgh, Pittsburgh, Pennsylvania; Wayne State University, Detroit, Michigan; Northwestern University, Chicago, Illinois; and Oregon Health and Science University, Portland, Oregon.
*For a list of other members of the NICHD MFMU, see the Appendix online at http://links.lww.com/AOG/A215.
Supported by grants from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (HD27915, HD34116, HD40485, HD34208, HD27869, HD40500, HD40560, HD34136, HD40544, HD27860, HD40545, HD53097, HD21410, HD27917, HD40512, HD53118, HD36801), General Clinical Research Centers Grant (M01-RR00034), and the National Center for Research Resources (UL1-RR024989, M01-RR00080, UL1-RR025764, C06-RR11234) and does not necessarily represent the official views of the National Institute of Child Health and Human Development or the National Institutes of Health.
The authors thank Julia Zachary for data management; Francee Johnson, RN, Joanne Tillinghast, RN, and Susan Tolivaisa for coordination between clinical research centers; Elizabeth Thom, PhD, for study design, data management, statistical analysis, and manuscript development; Steven G. Gabbe, MD, and Menachem Miodovnik, MD, for study design and protocol development; and Roger Newman, MD, for study design.
Presented in part at the 30th annual meeting of the Society for Maternal–Fetal Medicine, February 1–6, 2010, Chicago, Illinois.
Dr. Spong, Associate Editor of Obstetrics & Gynecology, was not involved in the review or decision to publish this article.
Corresponding author: Mark B. Landon, MD, The Ohio State University College of Medicine, Department of Obstetrics and Gynecology, 395 W 12th Avenue, Suite 572, Columbus, OH 43210; e-mail: Mark.Landon@osumc.edu.
Financial Disclosure The authors did not report any potential conflicts of interest.
OBJECTIVE: To examine the relationship between varying degrees of maternal hyperglycemia and pregnancy outcomes.
METHODS: This was a secondary analysis of a treatment trial for mild gestational diabetes including four cohorts: 1) 473 women with untreated mild gestational diabetes; 2) 256 women with a positive 50-g screen and one abnormal oral glucose tolerance test (OGTT) value; 3) 675 women with a positive screen and no abnormal OGTT values; and 4) 437 women with a normal 50-g screen. Groups were compared by test of trend for a composite perinatal outcome (neonatal hypoglycemia, hyperbilirubinemia, elevated cord C-peptide level, and perinatal trauma or death), frequency of large for gestational age neonates, shoulder dystocia, and pregnancy-related hypertension. Three-hour OGTT levels (fasting, 1-, 2-, and 3-hour) levels were divided into categories and analyzed for their relationship to perinatal and maternal outcomes.
RESULTS: There were significant trends by glycemic status among the four cohorts for the composite and all other outcomes (P<.001). Analysis for trend according to OGTT categories showed an increasing relationship between fasting and all postload levels and the various outcomes (P<.05). Fasting glucose 90 mg/dL or greater and 1 hour 165 mg/dL or greater were associated with an increased risk for the composite outcome (odds ratios and 95% confidence intervals of 2.0 [1.03–4.15] and 1.46 [1.02–2.11] to 1.52 [1.08–2.15] for the fasting and 1 hour, respectively). A 1 hour glucose 150 mg/dL or greater was associated with an increased risk for large for gestational age (odds ratios, 1.8 [1.02–3.18] to 2.35 [1.35–4.14]); however, 2- and 3-hour glucose levels did not increase the risk for the composite or large for gestational age until well beyond current gestational diabetes diagnostic thresholds.
CONCLUSION: A monotonic relationship exists between increasing maternal glycemia and perinatal morbidity. Current OGTT criteria require reevaluation in determining thresholds for the diagnosis and treatment of gestational diabetes.
LEVEL OF EVIDENCE: II
Gestational diabetes mellitus represents a healthcare burden that can be expected to rise as the frequency of obesity increases worldwide.1 The lack of uniform criteria for diagnosis and the reliance on observational data drawing on historic controls has limited the accurate determination of the relationship between mild degrees of hyperglycemia and perinatal outcomes.2 The recent Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study confirmed that a strong continuous relationship exists between maternal glucose concentrations and increasing birth weight, cord blood serum C-peptide levels, and other markers of perinatal complications.3 However, inconsistent results have been reported with respect to postglucose load levels and the risk for outcomes such as macrosomia.4 Using patients enrolled in a blinded trial of treatment for mild gestational diabetes, we sought to examine the relationship between varying degrees of maternal glycemia and pregnancy outcomes.
MATERIALS AND METHODS
This study is a secondary analysis of the Maternal–Fetal Medicine Units Network randomized clinical trial for the treatment of mild gestational diabetes.5 The study group consists of 1,841 women enrolled between 24 weeks 0 days and 30 weeks 6 days of gestation. For this analysis, three groups are considered, consisting of 1) 473 women in the untreated arm of the randomized trial with mild gestational diabetes defined as a fasting glucose less than 95 mg/dL with two or more post 100-g glucose load measurements meeting or exceeding established cutoffs6; 2) an observational cohort of 931 women matched one-to-one to the randomized patients with respect to clinical center, race or ethnicity, and body mass index (BMI, calculated as weight (kg)/[height (m)]2) classified as less than 27 or 27 or greater, with an elevated 50-g screen (135 mg/dL or greater) but normal oral glucose tolerance test (OGTT) results (for some analyses, this group is further divided into those with one abnormal value [n=256] compared with those with no abnormal OGTT values [n=675]); and 3) an observational cohort of 437 women matched one-to-one to the untreated group with respect to clinical center, race or ethnicity, and BMI classified as less than 27 or 27 or greater, with a screening value less than 120 mg/dL and thus unlikely to have gestational diabetes. A 3-hour OGTT was not performed for this group. To analyze for trend between glycemia and outcomes for patients who underwent an OGTT test, five categories were created according to 5-mg increments for the fasting values and in 15-mg increments for the 1-, 2-, and 3-hour time points of the OGTT. Fasting cutoffs were chosen based on the HAPO study,3 whereas postprandial cutoffs were based on accepted OGTT thresholds with 15-g increments selected to provide for adequate sample size for analysis.
The primary study outcome consisted of a composite outcome that included perinatal mortality, hypoglycemia, hyperbilirubinemia, elevated cord blood C-peptide level, and birth trauma.5 Other outcomes assessed in this analysis included the frequency of large for gestational age neonates (birth weight greater than the 90th centile of a U.S. reference population), shoulder dystocia, gestational hypertension, preeclampsia, or gestational hypertension coupled with preeclampsia.5 Trained study personnel collected antepartum, intrapartum, and postdelivery data for enrolled women and their newborns at the time of discharge from the hospital. All cases of hypertensive disorders and shoulder dystocia underwent masked central review by two of the authors to ensure accurate diagnosis.
For statistical analysis, baseline categorical variables were analyzed using chi square or Fisher's exact test. Continuous variables were analyzed using the Wilcoxon rank sum test or the Kruskal-Wallis test. The Cochrane Armitage test was used to test for trends between OGTT category as defined previously and the outcomes of interest.7 Multiple logistic regression analysis was performed examining the relationship of higher OGTT glucose categories compared with the lowest OGTT glucose category to assess whether a trend or threshold was associated with increasing glucose levels and the outcome of interest while adjusting for other potential confounders. In addition to glucose categories, regression models included maternal age, gestational age at enrollment and at delivery, parity, BMI, and race or ethnicity. Adjusted odds ratios and profile likelihood 95% confidence intervals were calculated.
A nominal two-sided P value <.05 was considered to indicate statistical significance. This study was approved by the Institutional Review Boards of all participating centers.
The demographics of the study population are shown in Table 1. Among the three groups, there was a significant increase in maternal age with greater degrees of carbohydrate intolerance. There were no significant differences in parity, race, or BMI among the three groups.
Figure 1 illustrates a significant increasing trend (all P<.002) of the composite perinatal outcome as well as the frequency of large for gestational age neonates, elevated cord C-peptide level, shoulder dystocia, and hypertensive disorders of pregnancy among four groups (women with an abnormal screen have been subdivided into those with or without one abnormal OGTT value). For example, the frequency of large for gestational age increased from 6.7% in women with a normal screening value to 14.5% in untreated gestational diabetes. We did not observe a significant trend for either neonatal hypoglycemia or hyperbilirubinemia considered as a separate outcome.
Figure 2 demonstrates direct comparisons between groups. An abnormal glucose screen was associated with an increased risk for both the composite outcome and large for gestational age neonates compared with a normal glucose screening test. Women with gestational diabetes (two or more abnormal values on GTT) demonstrated an increased risk for large for gestational age, elevated C-peptide, and shoulder dystocia compared with those with an abnormal screening test only. There were no statistically significant differences in the frequency of any of the outcomes in women with one abnormal value compared with those with untreated gestational diabetes.
The distribution of subjects in the five glucose categories created from the OGTT is shown in Table 2. Significant trends for increasing frequency for large for gestational age neonates were present across the five glucose categories of the fasting, 1-hour, 2-hour, and 3-hour determinations (Fig. 3). Similar trends were apparent for the composite outcome (all P<.05), elevated cord C-peptide level (all P<.02) with the trends only being significant for the 1- and 2-hour glucose in relation to shoulder dystocia (P<.01). A significant trend was present for all three postglucose load times for hypertensive disorders (all P<.03).
Logistic regression analysis controlling for maternal age, gestational age at enrollment and at delivery, parity, BMI, and race and ethnicity for each glucose category was performed in which adjusted odds ratios were generated comparing higher glucose categories with the lowest to assess whether a trend or threshold was associated with increasing glucose levels and outcomes (Table 3). For fasting glucose, an increased risk for both elevated C-peptide and large for gestational age were evident at 85–89 mg/dL and for the composite at 90–94 mg/dL. For the 1-hour value, an increased risk for the composite, large for gestatational age, and hypertensive disorders was evident at levels less than 180 mg/dL. A 1-hour value of 165–179 mg/dL was associated with an increased risk for the composite outcome, 1-hour values of 150–164 mg/dL and 165–179 mg/dL was associated with an increased risk for large for gestational age, and a 1-hour value of 165–179 mg/dL was associated with an increased risk of hypertensive disorders. The highest 2-hour values were not associated with an increased risk for the composite, and the risk for large for gestational age was not apparent until glucose exceeded 170 mg/dL. Similarly, for the 3-hour value, an increased risk for several outcomes was present only in the highest glucose category.
We found a significant relationship between increasing levels of maternal glycemia and perinatal morbidity. This graded increase in adverse maternal–fetal outcomes is observed across the spectrum of carbohydrate intolerance and includes glucose values below current cutoffs for the diagnosis of gestational diabetes.2,3,8 This observation provides additional information as experts continue to debate what constitutes appropriate thresholds for the diagnosis of gestational diabetes.
Previous studies have documented that an abnormal glucose screening test alone is an independent predictor of macrosomia.9,10 In addition to an increased risk for large for gestational age neonates, we have also documented a risk exists for other neonatal morbidities as well as shoulder dystocia in this group of women. Because of a perceived intermediate risk for fetal macrosomia, some have advocated follow-up diagnostic testing or close surveillance for fetal overgrowth during the third trimester for women with an abnormal screening test (and normal OGTT) discovered between 24 and 28 weeks of gestation.11 Recently, a small randomized trial has shown a reduction in large neonates in women with abnormal screening results receiving nutritional intervention compared with control subjects.12
Although our sample size is relatively small, our findings support the hypothesis of an increased risk of perinatal outcomes with one abnormal OGTT value compared with the normal OGTT group both in the trend analysis and in the comparison with the untreated mild gestational diabetes (at least two abnormal OGTT values) group. The finding of similar frequencies of various perinatal outcomes among women with one abnormal OGTT value compared with untreated mild gestational diabetes is important because some previous studies have compared this group with treated gestational diabetes subjects and thus have been unable to estimate their true level of risk.13 Originally, O'Sullivan and Mahan14 required two values to be abnormal on the diagnostic OGTT to minimize misclassification resulting from rapid glucose absorption or laboratory error. An analysis from the HAPO study demonstrating independent risks at each timed OGTT value has suggested that the criteria for the diagnosis of gestational diabetes should be based on the presence of one abnormal OGTT value. In an effort to define gestational diabetes, Sacks and colleagues8 were among the first to attempt to quantitate the relationship between maternal glucose levels and birth weight centiles. However, within the ranges of glucose studied in over 3,000 women, these authors could not establish a “clinically meaningful” threshold for diagnosis. The HAPO study also failed to yield any apparent inflection points when comparing maternal glycemia with both perinatal and maternal outcomes. Thus, any proposed thresholds for diagnosis, as Sacks predicted, will rely on expert opinion and be based on a consideration of relative risk.8 Our analysis revealed that the risk for large for gestational age and other morbidities is present at fasting and 1-hour levels below the current criteria for the diagnosis of gestational diabetes. However, risks for large for gestational age did not become significant until 2- and 3-hour values exceeded the Carpenter-Coustan thresholds.6 After controlling for confounders, Sermer and colleagues reported that only fasting glucose could be related to macrosomia (birth weight greater than 4,000 g), whereas others have suggested that the 1-hour postprandial level, representing the peak of the glucose response curve, is most significant.2
We do note several limitations of this secondary analysis. First, we were able to only consider OGTT data from a cohort of women who failed the 50-g glucose screen. These individuals likely have an intermediate level of carbohydrate intolerance and also appear to have an increased risk for adverse perinatal outcomes compared with women with normal screening. We chose to generate odds ratios for outcomes in comparison to the lowest glucose categories for this reason. Thus, our finding of increased 2- and 3-hour OGTT thresholds for several outcomes among such women may actually be considered even more significant. Our study also used the 3-hour, 100-g OGTT, which, although standard for use in pregnancy in the United States, is not the diagnostic test common in many areas of the world. This limits direct comparisons to 75-g, 2-hour OGTT data found in other reports.8 Although we analyzed outcomes in relation to greater than 1,000 OGTTs, our treatment trial did not include a sample size sufficient to suggest precise cutoffs for the diagnosis of gestational diabetes. Current proposed recommendations for new diagnostic criteria for gestational diabetes, if adopted, will substantially increase the frequency of this diagnosis.15 Although two randomized controlled trials now demonstrate a treatment benefit to women meeting specific OGTT criteria, evidence is lacking concerning treating women with thresholds outside those we studied as well as one abnormal 75-g OGTT value.5,16 The results of treatment trials as well as the current analysis will hopefully further aid professional organizations as they consider the most cost-effective approach to the diagnosis and treatment of gestational diabetes.
1.Chen Y, Quick WW, Yang W, Zhang Y, Baldwin A, Moran J, et al. Cost of gestational diabetes mellitus in the United States in 2007. Popul Health Manag 2009;12:165–74.
2.Sermer M, Naylor CD, Gare DJ, Kenshole AB, Ritchie JW, Farine D, et al. Impact of increasing carbohydrate intolerance on maternal–fetal outcomes in 3637 women without gestational diabetes. The Toronto Tri-Hospital Gestational Diabetes Project. Am J Obstet Gynecol 1995;173:146–56.
3.The HAPO Study Cooperative Research Group; Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, Coustan DR, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 2008;358:1991–2002.
4.Dodd JM, Crowther CA, Antoniou G, Baghurst P, Robinson JS. Screening for gestational diabetes: the effect of varying blood glucose definitions in the prediction of adverse maternal and infant health outcomes. Aust N Z J Obstet Gynaecol 2007;47:307–12.
5.Landon MB, Spong CY, Thom E, Carpenter MW, Ramin SM, Casey B, et al. A multicenter, randomized trial of treatment for mild gestational diabetes. N Engl J Med 2009;361:1339–48.
6.Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol 1982;144:768–73.
7.Agresti A. Categorical Data Analysis. 1st ed. New York (NY): Wiley; 1990. p. 100–2.
8.Sacks DA, Greenspoon JS, Abu-Fadil S, Henry HM, Wolde-Tsadik G, Yao JF. Toward universal criteria for gestational diabetes: the 75-gram glucose tolerance test in pregnancy. Am J Obstet Gynecol 1995;172:607–14.
9.Leiken EL, Jenkins JH, Pomerantz GA, Klein L. Abnormal glucose screening tests in pregnancy: a risk factor for fetal macrosomia. Obstet Gynecol 1987;69:570–3.
10.Stamilio DM, Olsen T, Ratcliffe S, Sehdey HM, Macones GA. False-positive 1-hour glucose challenge test and adverse perinatal outcome. Obstet Gynecol 2004;103:148–56.
11.Neiger R, Coustan DR. The role of repeat glucose tolerance tests in the diagnosis of gestational diabetes. Am J Obstet Gynecol 1991;165:787–90.
12.Bonomo M, Corica D, Mion E, Gonçalves D, Motta G, Merati R, et al. Evaluating the therapeutic approach in pregnancies complicated by borderline glucose intolerance: a randomized clinical trial. Diabet Med 2005;22:1536–41.
13.Langer O, Brustman L, Anyaegbunam A, Mazze R. The significance of one abnormal glucose tolerance test value on adverse outcome in pregnancy. Am J Obstet Gynecol 1987;157:758–63.
14.O'Sullivan JB, Maban CM, Charles D, Dandrow RV. Screening criteria for high-risk gestational diabetic patients. Am J Obstet Gynecol 1973;116:895–900.
15.International Association of Diabetes and Pregnancy Study Groups Consensus Panel, Metzger BE, Gabbe SG, Persson B, Buchanan TA, Catalano PA, Damm P, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hypertension in pregnancy. Diabetes Care 2010;33: 676–82.
16.Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS, Robinson JS. Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 2005; 352:2477–86.
Supplemental Digital Content
© 2011 by The American College of Obstetricians and Gynecologists.