The Diagnosis of Gestational Diabetes

From Medscape Diabetes & Endocrinology
Change Is in the Air
Laura A. Stokowski, RN, MS

08/04/10

A Parallel Epidemic

The word "epidemic" is so overused that it has lost its undercurrent of urgency. We are experiencing epidemics of obesity, high cholesterol, cardiovascular disease, and diabetes. An epidemic has become the norm. Even the word "pandemic," thanks to swine flu, no longer conveys a sense of gravity. New words are needed to describe the overarching implications of a society in which type 2 diabetes afflicts at least 1 in 10 people and, quite possibly, many more.

The prevalence of gestational diabetes mellitus (GDM) will likely grow, as it has in the past, in direct proportion to that of type 2 diabetes.[1] Indications are that GDM already parallels the rapid increase in type 2 diabetes. In a US medical center where the screening method and diagnostic criteria for GDM have remained constant, the prevalence of this complication of pregnancy doubled in 8 years -- a 12% increase per year that cannot be explained by changes in age, ethnic distribution, or previous history of GDM among screened pregnancies.[2]

Arguably more disturbing than the number of people diagnosed with type 2 diabetes or GDM is the number of people who have prediabetes (ie, impaired fasting glucose and/or impaired glucose tolerance).
Currently, an estimated 19% of people over the age of 20 have prediabetes,[3] and this is the pool from which childbearing women are drawn.

Gestational Diabetes Mellitus
Gestational diabetes is glucose intolerance with onset or first recognition during pregnancy.
Pregnancy is already a diabetogenic state, with progressive deterioration of insulin resistance and glucose tolerance that become more significant in the third trimester.

Neonatal problems of offspring of frankly diabetic mothers (eg, congenital defects, spontaneous abortion, fetal macrosomia, birth injury, hypoglycemia, polycythemia, and hyperbilirubinemia) are well described. Exposure to diabetes during gestation also increases the risk for childhood and adult obesity, diabetes, and cardiovascular disease. However, risk for adverse outcomes associated with degrees of maternal hyperglycemia that are short of overt diabetes remains controversial.

In the United States, GDM is commonly diagnosed using either the World Health Organization's criteria (the same as those used to diagnose diabetes in nonpregnant women), or on the basis of a woman's risk of developing diabetes in the future. Neither of these methods links the key metabolic aberration of GDM (ie, maternal hyperglycemia) with the risk for adverse outcomes in the fetus and newborn. Most experts agree that new, more clinically relevant, risk-based criteria for the diagnosis of GDM are needed, especially considering the similarities between GDM and prediabetes. However, levels of maternal glucose intolerance that correlate with poor neonatal outcomes had not been sufficiently ascertained until the results of the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study were reported.

The HAPO Study
The HAPO epidemiologic study was the first to conclusively establish a relationship between elevated maternal glucose concentrations and undesirable perinatal outcomes in women not previously diagnosed with diabetes.

The HAPO study clarified the association between multiple adverse outcomes of pregnancy and degrees of hyperglycemia less severe than those diagnostic of diabetes. Data for establishing internationally agreed-upon diagnostic criteria for GDM were also provided. This well-controlled study involved 25,000 women in 9 countries -- a multicultural, ethnically diverse cohort. All women underwent a 2-hour oral glucose tolerance test (OGTT) with a 75-g glucose load at 24-32 weeks' gestation, and random plasma glucose testing at 34-37 weeks. Results were blinded unless a woman's fasting, 2-hour, or random glucose values were elevated to a level that mandated immediate treatment.

Primary outcomes were macrosomia (ie, birth weight > 90th percentile), primary cesarean delivery, and clinical neonatal hypoglycemia and hyperinsulinemia (ie, cord serum C-peptide > 90th percentile). The analysis aimed to determine whether threshold levels of maternal glucose -- for any of the 1-hour, 2-hour, or fasting plasma glucose (FPG) levels -- could be identified for any of the negative outcomes. They found that each of the primary outcomes was associated not only with extremely high maternal glucose concentrations, but in a continuous and graded manner across the full range of observed glucose levels,[9] which precluded easy identification of threshold levels where risk for adverse outcomes rose. The relationship between maternal glucose levels and fetal growth and neonatal outcome seemed to be a basic biologic phenomenon, and not a clearly demarcated disease state, as had previously been thought.

Several secondary outcomes were also evaluated in the HAPO study, including preeclampsia, preterm delivery (ie, delivery at < 37 weeks' gestation), shoulder dystocia and/or birth injury, hyperbilirubinemia, and admission to neonatal intensive care. Shoulder dystocia or birth injury, preterm delivery, and preeclampsia were significantly associated with ≥ 1 elevated glucose values.[9] The blinding of maternal glucose (except when overt diabetes was suggested) is a strength of the HAPO study, because maternal glucose was not a factor in obstetric management. It is also promising that the study's findings did not vary by medical center or country. The results are therefore applicable globally and can be used to develop criteria for classifying gestational diabetes world-wide. Next Step: New Diagnostic Criteria The HAPO investigators did not attempt to translate their findings into new criteria for the diagnosis of gestational diabetes.[8] This task fell to a committee of experts, the International Association of Diabetes and Pregnancy Study Groups (IADPSG), who met to review the data, form a consensus, and make recommendations. A pivotal decision involved the threshold for diagnosing GDM using data from fasting glucose and 2-hour OGTT. This threshold would be somewhat arbitrary, because no inflection points were apparent in the linear relationships between maternal glucose concentrations and outcomes.[8] The IADPSG examined the strong linear associations between risk for neonatal outcomes and the 3 measures of maternal glucose (FPG, 1-hour OGTT, and 2-hour OGTT). The threshold was set at an odds ratio of 1.75, which identified 16.1% of the pregnant population as having GDM (Table 1). Table 1. Proposed and Current Thresholds for the Diagnosis of GDM Maternal glucose test Proposed diagnostic thresholda Above threshold (cumulative %) Current thresholda Fasting plasma glucose 92 mg/dl (5.1 mmol/L) 8.3 95 mg/dl (5.3 mmol/L) 1 hour plasma glucose 180 mg/dl (10 mmol/L) 14.0 180 mg/dl (10 mmol/L) 2 hour plasma glucose 153 mg/dl (8.5 mmol/L) 16.1 155 mg/dl (8.6 mmol/L) aWith 75 g OGTT Only 1 of these cut-offs (FPG, 1-hour OGTT, or 2-hour OGTT) must be met or exceeded to diagnose GDM, unlike current American Diabetes Association (ADA) criteria, which require 2 elevated glucose levels to diagnose GDM. The proposed thresholds are those for which the odds of having a baby with birth weight, cord C-peptide, or neonatal body fat in the > 90th percentiles are 1.75 times the estimated odds of these outcomes at mean glucose values. Setting the threshold higher would decrease the number of women diagnosed with GDM, but would also fail to identify many women whose glucose concentrations place them at the same risk for adverse pregnancy outcomes, and who might benefit from treatment. In this study, 1.7% of patients were unblinded because of elevated FPG or OGTT results. When these patients are included, the total incidence of gestational diabetes in pregnant women rises to 17.8%.

Glucose Testing Considerations
The finding that slight differences in maternal glucose levels are associated with marked differences in outcomes throws intoclear relief the importance of precision in glucose testing. Odds ratios and frequencies for outcomes increase substantially over relatively small changes in glucose. While the handling of blood samples in research is tightly controlled, the real world typically introduces extensive variability. Even a small error in test results caused by poor handling or analytic technique could result in the misclassification of a patient.

For reliable diagnosis and classification of hyperglycemia in pregnancy, venous plasma or serum glucose must be analyzed with a highly accurate enzymatic method. The collection, handling, and transport of blood samples to minimize pre-analytic glycolysis are extremely important. The IADPSG recommends that only venous samples be used for glucose determination, emphasizing that capillary and venous samples are not interchangeable. If the plasma is not separated promptly, the blood sample should be kept cold, because glycolysis will continue in the presence of red and white blood cells, falsely reducing the patient's blood glucose level by 5-15%. It is often mistakenly believed that as soon as the blood is placed in sodium fluoride (a glycolysis inhibitor), glycolysis will stop, but in fact, sodium fluoride has little effect on glycolysis in the first 1-2 hours after sample collection.Point-of-care glucose testing with handheld glucose meters is not appropriate for the diagnosis of GDM.

Significance of Proposed Guidelines
Lowering the diagnostic threshold will undoubtedly raise the frequency of hyperglycemic disorders seen in clinical practice. This would matter less if it could be shown that a benefit can be derived from improving even mild aberrations in glucose metabolism during pregnancy. The longstanding debate about the value of screening pregnant women for hyperglycemia has centered on uncertainties about the treatment benefit.

Benefit of treating GDM. In May 2008, the US Preventive Services Task Force concluded that there was inadequate evidence to recommend treatment of GDM, largely because of inadequate prospective studies.[14] A subsequent review concluded that treatment of GDM after 24 weeks of pregnancy improves some maternal and neonatal outcomes; however, evidence for screening before 24 weeks' gestation is more sparse.[15] The gestational time at which hyperglycemia screening should be initiated, and the level of hyperglycemia that warrants aggressive intervention remain controversial.[16]

In 2005, the results of ACHOIS (Australian Carbohydrate Intolerance Study in Pregnant Women), a 10-year multicenter randomized trial, were published.[17] ACHOIS assessed whether treating mild GDM would reduce perinatal morbidity and mortality. Treatment with dietary counseling, self glucose monitoring, and insulin when indicated, significantly reduced adverse primary outcomes (ie, perinatal death, shoulder dystocia, and birth trauma), neonatal adipoinsular macrosomia, maternal preeclampsia, and labor induction. Landon and colleagues conducted a randomized controlled trial sponsored by the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units (NICHD-MFMU) Network, compared untreated women with mild GDM with those receiving treatment (nutritional counseling, diet therapy, insulin if needed, etc.).[18] Significant reductions in macrosomia, neonatal fat mass, shoulder dystocia, preeclampsia, and cesarean section were seen in the treated cohort.

Both treatment trials revealed a positive effect of treatment in preventing large for gestational age (LGA) births, macrosomia, and shoulder dystocia.
Together, these 2 studies argue convincingly for a treatment benefit for mild GDM.While neither study found significant effects of treatment on neonatal morbidities such as hypoglycemia or hyperbilirubinemia, their findings of reduced neonatal fat mass, LGA, and macrosomia have important implications for long-term child and adult health. Excess neonatal fat and adipoinsular macrosomia are linked to childhood obesity and later development of diabetes. If these findings are real, then the successful treatment of maternal GDM, even mild GDM, could positively influence the health of the next generation.

Health system burden.
Although the new criteria are expected to double the number of women diagnosed with GDM, the rate will be consistent with the high prevalence of obesity and glucose intolerance in the general population. Donald R. Coustan, MD, Professor of Obstetrics and Gynecology, Warren Alpert Medical School of Brown University, Division of Maternal-Fetal Medicine, Women & Infants Hospital of Rhode Island, Providence, Rhode Island, explained the implications of classifying such a large slice of the pregnant population as having GDM and the potential burden on high-risk maternity care:

"Currently (as of 2007), the Centers for Disease Control and Prevention tells us that 10.7% of adult Americans have diabetes; furthermore, the American Diabetes Association states that 19% of adult Americans have prediabetes. So 17% is not such an extreme prevalence for GDM, which is quite similar to prediabetes in its severity. In the 2 randomized trials of identification and treatment of mild GDM (ACHOIS in Australia and NICHD-MFMU in the US), only 20% and 8%, respectively, of treated patients required insulin; the rest were managed successfully with diet. The burden will be in having adequate numbers of dietitians, and there will likely be creative approaches to educating newly diagnosed women, such as group counseling sessions. Self glucose monitoring could also increase the burden, but these milder forms may not require testing every day."

Others have raised concerns that the higher-risk GDM diagnostic label, irrespective of a woman's degree of glucose control, could stimulate an increase in perinatal interventions, earlier deliveries, caesarean section rates, babies admitted to special care nurseries, healthcare costs,[19] and psychological distress and anxiety related to the diagnosis of GDM.[15]

Implications for Clinical Practice
The IADPSG Consensus Panel recommendations are currently being considered for adoption by leading consumer and professional organizations such as the ADA and the American Congress of Obstetricians and Gynecologists. In a recent update of their position statement on diagnosis and classification of diabetes, the ADA said the following about the consensus panel's recommendations:

"At the time of publication of this update, ADA is planning to work with US obstetrical organizations to consider adoption of the IADPSG diagnostic criteria and to discuss the implications of this change. While this change will significantly increase the prevalence of GDM, there is mounting evidence that treating even mild GDM reduces morbidity for both mother and baby."[11]

If adopted, the new guidelines are expected to have immediate, widespread clinical implications.

Changes in screening for GDM. The detection strategy recommended by the IADPSG has 2 phases:

1.Testing for overt diabetes at the initial prenatal visit. Universal early testing in populations with a high prevalence of type 2 diabetes is recommended; others may choose to test only high-risk groups (Table 2)
2.A 75-g OGTT after an overnight fast at 24-28 weeks' gestation in all women not previously diagnosed with overt diabetes or GDM.

Table 2. Low and High Risk Factors for GDM[4,10]

Low risk for GDM
•Age < 25 years •Normal body weight •No family history (1st degree) of DM •No history of abnormal glucose metabolism •Not of ethnic/racial group with high prevalence of DM (African-Americans, Asian-Americans, Hispanic- Americans, Native Americans, Pacific Islanders) High risk for GDM •Maternal age > 35 years
•Marked obesity
•Personal history of GDM
•Previous infant > 4 kg
•Pre-diabetes
•Glycosuria
•Strong family history of DM
•Hypertension before pregnancy
or in early pregnancy
•Ethnic/racial group with high prevalence of DM (African-Americans, Asian-Americans, Hispanic-Americans, Native Americans, Pacific Islanders)

The proposed screening strategy and cut-offs for the diagnosis of GDM are summarized in Table 3.

Table 3. Proposed Screening for GDM

When Diagnosis Test Cut-off for diagnosis
1st prenatal visit Overt diabetes FPG 126 mg/dL (7.0 mmol/L)
HbA1C ≥ 6.5%
Randoma 200 mg/dL (11.1 mmol/L)
24-28 weeks Gestational diabetes FPG 92 mg/dL (5.1 mmol/L)
75g OGTT-1 hr 180 mg/dL (10.0 mmol/L)
75g OGTT-2 hr 153 mg/dL (8.5 mmol/L)

aConfirmation required.

Women who exceed the threshold for GDM on FPG (92 mg/dl or 5.1 mmol/L) at the first prenatal visit are diagnosed as having gestational diabetes. Women whose FPG at first prenatal visit is below 92 mg/dL (5.1 mmol/L) are tested with a 2-hour OGTT at 24-28 weeks to rule out GDM.

Dr. Coustan addressed the issue of how these guideline changes might affect primary maternity care: "In some ways, life will be easier for clinicians if the new recommendations are adopted:

1.We will go from a 2-step screening procedure (50-g challenge, then 100-g 3-hour OGTT if challenge test is positive) to a 1-step procedure;
2.The glucose load used for the OGTT decreases from 100 g to 75 g, which may improve patient acceptance;
3.The 75-g, 2-hour test is the same as that used in nonpregnant adults, although diagnostic thresholds will be different, so less likelihood of error with the lab doing the wrong test;
4.A single elevated value (out of the 3 blood samples -- FPG, 1-hour, and 2-hour) will be sufficient to diagnose GDM, rather than the current requirement for 2 elevated values (out of 4) . . . this will eliminate the 'borderline' state of one abnormal value and the quandary as to how to treat these patients.
5.With the recommendation from IADPSG about how to diagnose pre-existing diabetes when patients present early in pregnancy with high glucose values, clinicians will be able to remove the ambiguity about how to manage such patients."
Prompt treatment of newly diagnosed women with GDM is important. Most fetal weight accretion occurs in the third trimester, so treatment should begin as soon as the diagnosis is made. Treatment strategies for GDM include dietary modifications, regulated exercise, and pharmacologic agents. Most women can be managed with diet and exercise, and will not require insulin.[20]

Preconception care. As many as half of pregnancies in the US are unplanned.[21] Thus, women with chronic medical conditions such as diabetes might not have the opportunity to take steps to optimize management of their diabetes before becoming pregnant. Adverse outcomes are more likely to occur in women with GDM who do not receive preconception counseling.[22]

When providers are able to do preconception assessment and counseling, women who have prediabetes should be taught how to improve their metabolic control prior to conception, in order to reduce the likelihood of birth defects if they progress to diabetes. Lifestyle interventions have been shown to prevent the progression of prediabetes to diabetes in a randomized trial.[23] With help, women may be able to avoid the risks of a diabetic pregnancy. Every healthcare provider who takes care of a woman of reproductive age has something to contribute to preconception care, by diagnosing prediabetes and helping that individual avoid progression to diabetes and its attendant risks during pregnancy.

Conclusion
If the IADPSG's proposed criteria for GDM diagnosis are adopted, we will be able to identify gravidas who have increased risks for adverse outcomes, such as large, fat, or hyperinsulinemic babies and cesarean section delivery.[8] Along with new evidence for a treatment benefit for GDM, the time may be right for a new approach to the screening of pregnant women for potentially correctable alterations of glucose metabolism. Clinicians should stay tuned for further developments, such as official adoption of the IADPSG recommendations, and be prepared for the changes to clinical practice that will inevitably follow.

References