HAPO results leading to gestational diabetes recommendations

Experts are hopeful that a consensus will be reached.

Physicians who treat women with gestational diabetes marked an important milestone in May. The long-awaited Hyperglycemia and Adverse Pregnancy Outcomes results, first presented at the 2007 Annual Meeting of the American Diabetes Association, were published in The New England Journal of Medicine. Since then, HAPO researchers from around the world have sought a consensus in diagnosing the disease.

The HAPO results showed strong associations between the mother’s glucose level and specific adverse outcomes, including birth weight above the 90th percentile and primary cesarean section across a range of glucose levels from normal to diabetic, according to Boyd E. Metzger, MD, professor in the division of endocrinology at Northwestern Medical School and lead researcher on the HAPO study.

“There’s clear evidence now that outcomes we see in women with gestational diabetes, which are similar to those seen in people with known diabetes, though not as frequently, are really related to the mother’s glucose level and not to other factors, like her degree of overweight or the fact that she might be older than the average patient who’s giving birth or having high blood pressure,” Metzger told Endocrine Today. “All of those were adjusted for and these associations are taking all of these other factors into account.”

The publishing of the HAPO results brought the issue to the international stage. In June, researchers from professional organizations around the world gathered to discuss the HAPO findings, hear specifics from the researchers and begin the long process of bringing agreement to the recommended diagnostic cutpoints and the optimal diagnostic test for practitioners the world over.

Endocrine Today spoke with Metzger and other attendees of the consensus conference to find out how the HAPO results will affect clinical practice, how they will bring agreement to the guidelines, and what remains to be done in the area of gestational diabetes management.

Why consensus is necessary

Boyd E. Metzger, MD, is a Professor in the Division of Endocrinology at Northwestern Medical School and Lead Researcher of HAPO. Photo by Jim Ziv

In February, Endocrine Today reported on the various differences among the guidelines for the diagnosis and management of gestational diabetes. Professional societies, including the ADA, American College of Obstetricians and Gynecologists, Joslin Diabetes Center, World Health Organization, the International Diabetes Center and the U.S. Preventive Services Task Force, each offer recommended cutpoints and diagnostic tests with differing details. This has caused confusion among practitioners and has complicated research in the area.

Jennifer Williams, MSN, MPH, APRN-BC, and her colleagues analyzed the differences in the guidelines between the professional organizations and spoke with Endocrine Today in February. She explained that the differences in the guidelines are minimal, but that they may lead to confusion for women with gestational diabetes who see multiple providers and who could potentially receive conflicting information.

According to the study results, published in the Journal of Women’s Health in 2007, most of the guidelines recommended screening women at risk for developing gestational diabetes, but three of the professional societies suggested that low-risk women may be excluded from screening. Screening and diagnostic tests also differed: ACOG, ADA, the International Diabetes Center and Joslin guidelines recommended the 50-g one-hour OGTT and the 100-g three-hour OGTT, but WHO guidelines recommended the one-step 75-g OGTT.

Diagnostic cutpoints also varied with the WHO guidelines, recommending a cutpoint of 126 mg/dL for the fasting value and 200 mg/dL for the two-hour value — both using the 75-g OGTT. Using the 100-g OGTT, ACOG, ADA and the International Diabetes Center recommended diagnosis if a woman meets two of the following cutoffs, based on the Carpenter Coustan testing method: 95 mg/dL for fasting, 180 mg/dL for one-hour, 155 mg/dL for the two-hour and 140 mg/dL for the three-hour OGTT. ACOG, along with Joslin, also recommended diagnosis based on the National Diabetes Data Group testing method: 105 mg/dL for fasting, 190 mg/dL for one-hour OGTT, 165 mg/dL for the two-hour OGTT and 145 mg/dL for the three-hour OGTT.

Lois Jovanovic, MD, who attended the consensus conference in June as a representative for the International Diabetes Federation, said that practitioners can expect an international consensus on each of these points based on the HAPO results.

“I can guarantee you, they’ll all agree on universal screening. We won’t have, ‘only if you’re high risk’ or ‘only if you have risk factors,’" Jovanovic, CEO and chief scientific officer of the Sansum Diabetes Research Institute in Santa Barbara, Calif., told Endocrine Today.

M. Sue Kirkman

“Going into the meeting, there was a clear desire to have one worldwide standard for diagnosing gestational diabetes because not only is it slightly different from organization to organization but different countries do it differently,” M. Sue Kirkman, MD, vice president of Clinical Affairs, at the ADA, told Endocrine Today. “Certainly the ADA is very much in favor of [having a worldwide standard].”

According to Kirkman, who attended the consensus meeting as a representative for the ADA, while most U.S. physicians perform a glucose screening test and then an OGTT at 100 g, other countries skip the screening and perform 75-g OGTTs.

“It makes it very difficult to compare from country to country or from group to group things such as prevalence, how common is gestational diabetes in this country or this part of the country compared with that country or that part of the country,” she said. “It also makes it difficult to compare therapies if slightly different types of women are being treated or not treated.”

The HAPO findings

HAPO researchers looked at gestational diabetes in an entirely new light. Metzger and colleagues asked practitioners to consider pregnancy outcome risks based on glucose level. They asked at each cutpoint what is the risk for birth weight above the 90th percentile, primary cesarean section, clinical neonatal hypoglycemia and C-peptide level above the 90th percentile?

They obtained blinded data for 23,316 women from 15 centers in nine countries. Each woman underwent a 75-g OGTT during 24 to 32 weeks of gestation. Data were blinded if women were not diagnosed with diabetes (two-hour plasma glucose level >200 mg/dL) and if fasting plasma glucose was <105 mg/dL. Women were stratified into seven categories based on glycemic measure: >100 mg/dL, 95-99 mg/dL, 90-94 mg/dL, 85-89 mg/dL, 80-84 mg/dL, 75-79 mg/dL and <75 mg/dL.

Findings revealed that the frequency of the primary outcomes increased as the maternal glucose level increased. Just looking at fasting plasma glucose levels, birth weight above the 90th percentile occurred in 26.3% of pregnancies of women in the highest category of fasting plasma glucose compared with 5.3% of women in the lowest category. Similarly, primary cesarean section occurred in 27.9% of women in the highest category and in 13.3% in the lowest; and C-peptide level above the 90th percentile occurred in 32.4% of women in the highest category compared with 3.7% in the lowest.

Lois Jovanovic

Birth weight differed by as much as 305 g between the lowest and highest categories of the fasting, one-hour and two-hour glucose measurements when birth weight was modeled as a continuous variable, according to the researchers.

“These findings really do mean that any elevation in glucose increases the risk of a major variable that’s associated with complications of diabetes,” said Jovanovic, who is also an Endocrine Today Editorial Board member.

The researchers also calculated adjusted ORs for each of the pregnancy outcomes associated with an increase of 1 standard deviation in the fasting plasma glucose level (6.9 mg/dL), the one-hour plasma glucose level (30.9 mg/dL), and the two-hour plasma glucose level (23.5 mg/dL).

The ORs for birth weight above the 90th percentile were 1.38 for both the fasting and two-hour (95% CI, 1.32-1.44) and 1.46 for one-hour (95% CI, 1.39-1.53). For cord-blood serum C-peptide levels above the 90th percentile, the researchers calculated the following ORs: 1.55 for fasting (95% CI, 1.47-1.64), 1.46 for one-hour (95% CI, 1.38-1.54) and 1.37 for two-hour (95% CI, 1.30-1.44).

The OR for cord-blood serum C-peptide level above the 90th percentile and maternal glycemia was 7.65 for the highest category of maternal glucose (>100 mg/dL; 95% CI, 5.17-11.32) (See charts for a complete list of ORs at each glucose level).

’There is a direct relationship between the mother’s glucose values and the primary outcome variables,” Donald R. Coustan, MD, professor of obstetrics and gynecology at Brown University in Providence, R.I., and a HAPO researcher, told Endocrine Today. “The surprising finding was that the relationship held even well down into the normal range of blood sugars, so that it’s not, if you‘r blood sugar is below the 90th percentile, you’re normal; and above that, you have all these problems.’ It’s that these issues relate to blood sugar no matter what the blood sugar is. And that means it’s not a disease, it’s a basic biologic phenomenon.”

Source: N Engl J Med. 2008;358:1991-2002

Secondary outcome findings

Although weaker than these primary outcome findings, the researchers also observed associations with the mother’s glucose level and secondary outcomes, including delivery before 37 weeks of gestation, shoulder dystocia or birth injury, need for intensive neonatal care, hyperbilirubinemia and preeclampsia. According to Metzger, these outcomes were considered secondary because the researchers did not design the study to have a large enough sample size to consider them primary. The researchers found the strongest association for preeclampsia with the OR for each 1 standard deviation increase in glucose level ranging from 1.21 to 1.28.

They also found that no single measure of glucose level – fasting, one-hour or two-hour OGTTs — stood out from the rest as superior in predicting the primary outcomes. Slight differences were seen in predicting clinical neonatal hypoglycemia, with the one-hour test being the only significant predictor, according to the researchers. However, this outcome was infrequent with only a 2.1% overall incidence. Fasting plasma glucose appeared to be the strongest predictor for cord-blood serum C-peptide levels, although all three tests were “highly predictive.”

“It’s very clear that the fasting plasma glucose predicts the risk as well as the one-hour and two-hour tests or even better than those two,” Jovanovic said. “Any blood sugar above normal – whether it’s the fasting, the one-hour or the two-hour OGTT – is associated with increased risk of ‘big baby’ and ‘sick baby.’”

In the first of several planned studies looking at different areas of the HAPO results, the researchers published data on the association between the mother’s glucose levels and cord-serum C-peptide with neonatal adiposity, essentially testing the Pederson hypothesis. They used multiple logistic regression analyses, adjusting for maternal age, parity, BMI, mean arterial pressure, height, gestational age at delivery and the baby’s gender.

Researchers obtained cord-serum C-peptide results for 19,885 babies and skin fold measurements for 19,389 babies. ORs for the two measures of adiposity – skin folds greater than the 90th percentile or body fat greater than the 90th percentile – ranged from 1.35 to 1.44. As in the main study, the ORs were based on the fasting, one-hour and two-hour plasma glucose levels higher than 1 standard deviation.

Consensus meeting

Immediately following the ADA meeting in June, 250 delegates from around the world representing roughly 40 to 50 countries, attended the three-day consensus conference to discuss the HAPO findings and to begin building a consensus, according to Metzger. On the third day, 50 of those representatives met to transfer the discussion into specific guidelines.

“What is a challenge is to decide how much increase in risk is the point at which treatment should be initiated and what is the hope and expectation for the treatment to reduce those risks and that’s where the effort to gain a consensus is headed,” Metzger said.

Representatives at the consensus conference are being asked to consider the risks that they are willing to take before intervening. In other words, how much of an increase in risk for outcomes, such as large birth weight or high insulin in the baby, would they consider the level at which to diagnose a woman with gestational diabetes?

According to Jovanovic, the process began with a survey after the meeting. The representatives answered several questions, including, at what risk ratio should the cutoffs be made? At press time, the group of representatives was being asked to weigh in and vote again after the initial tallies from that survey.

“The world is coming to the conclusion that the cutoff numbers should be a level that increases the risk above the general population, or more than 1.5 to two times the general population,” Jovanovic said. “The consensus will be what numbers identify women who have a RR of 1, and probably they’ll come to a middle row, which is 1.75 increased risk of macrosomia, hypoglycemia, cesarean section and neonatal C-peptide elevation of greater than 1.75 above the general population.”

Once the group reaches a consensus, Metzger said, the tentative recommendations will be circulated to representatives of various organizations, such as the ADA and ACOG. After the organizations’ feedback and critiques, more discussion if necessary and any modifications, Metzger said they hope to publish recommendations with broad support.

What’s next for the HAPO data?

Metzger and colleagues plan to publish several more papers analyzing the HAPO data in different ways.

“We’re examining in more detail the relationship between the increase in mother’s weight and what influence that has on outcomes. We do know from other data that the heavier the mother, the higher risk she has for some of the same outcomes that are seen with higher blood sugar,” Metzger said. “We have a unique opportunity to study that in detail in this population in whom – below a certain level – there has been no treatment of the blood sugar levels, so the parallel findings of the mother’s weight and her glucose levels can be examined in the same people.”

Metzger said they are also considering following the children and examining the influence that the findings in pregnancy may have on the risk for overweight in children.

“We do hope to have some follow-up at some of our centers to address this question. There is information from other sources that, just in the same way that blood sugar levels that aren’t high enough are called diabetes and increase the pregnancy outcome risk, they also increase the likelihood of overweight in the children and abnormalities in glucose and blood pressure,” he said.

“Reaching decisions about these questions remains important because of the high prevalence that gestational diabetes has reached in our population as we now diagnose it. If we have to identify gestational diabetes in more people than treated presently, this is the time to get those questions answered because of the current concerns about the increasing prevalence of diabetes and overweight in our whole population.” — by Tina DiMarcantonio

Michelle D. Owens-Gary

The CDC’s current efforts to improve gestational diabetes research and care

The Centers for Disease Control and Prevention’s Division of Diabetes Translation and the Division of Reproductive Health established a collaborative project, Better Data, Better Care, with the National Association of Chronic Disease Directors to address the quality of gestational diabetes data and care. By working with five states (Michigan, North Carolina, Oklahoma, Utah and West Virginia) and several programs — maternal and child health programs, diabetes, chronic disease and vital records — this project will compare and validate data on gestational diabetes from various sources (the Pregnancy Risk Assessment Monitoring System, hospital discharge data, medical records, the Behavioral Risk Factor Surveillance System and the Pregnancy Nutrition Surveillance System).

The aim is to identify gaps in data currently available, and determine to what extent these data can address the prevalence of gestational diabetes and the quality of diabetes care among women of childbearing age. The results from this collaborative project will allow CDC and NACDD to develop recommendations for improving data quality and care for gestational diabetes. Initial results from a pilot program in Utah indicate a need for follow-up activities to improve the quality of the data collected through documentation of gestational diabetes screening, treatment and follow-up during prenatal and postpartum care, and accurate translation of hospital charting information to birth certificates.

— Michelle D. Owens-Gary, PhD
Behavioral Scientist, Division of Diabetes Translation, CDC

For more information:

• Endocrine Today. "Waiting for HAPO: Why the results are so anticipated." February 23, 2008;6(3):cover.
• Metzger BE, Lowe LP, Dyer AR, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med. 2008;358:1991-2002.
• Metzger BE, Lowe LP, Dyer AR, et al. Hyperglycemia and adverse pregnancy outcome (HAPO) study: Associations with neonatal anthropometrics. Diabetes. 2008;doi:10.2337/db08-1112. Accessed on Nov. 25, 2008.
• Mulholland C, Njoroge T, Mersereau P, Williams J. Comparison of guidelines available in the United States for diagnosis and management of diabetes before, during and after pregnancy. J Womens Health. 2007;16:790-801.