Recently updated guidelines call for oral glucose tolerance testing (OGTT) to diagnose gestational diabetes, but questions remained about the feasibility of using HbA1c for this purpose. Investigators from the landmark Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study evaluated this issue, and their findings are the subject of this edition of Strategies.

Owing to evidence for the first time linking maternal glucose intolerance with adverse pregnancy outcomes, several professional organizations, including the International Association of Diabetes and Pregnancy Study Groups (IADGSP), the American Diabetes Association (ADA), and the National Association for Clinical Biochemistry (NACB), recently endorsed new diagnostic criteria for gestational diabetes. Each group advised that all pregnant women not previously known to have diabetes undergo a 75-g OGTT when they’re 24-28 weeks’ gestation. When thresholds for any one of three measurements—fasting, 1- or 2-hour—are exceeded, the patient would be diagnosed with gestational diabetes.

ADA and NACB also recently endorsed using HbA1c to diagnose diabetes in non-pregnant adults. However, both stopped short of extending this recommendation to include diagnosis of gestational diabetes. Well-described physiological changes associated with pregnancy, such as increased red cell mass, influence HbA1c values. In addition, previous, albeit small, studies have shown that HbA1c and fructosamine measurements inadequately differentiate women with normal pregnancy from those with gestational diabetes. These and other issues left questions about how well HbA1c might identify women with gestational diabetes. This led HAPO study researchers to examine the association between maternal glucose and HbA1c to determine whether HbA1c could be used to measure glucose intolerance in pregnant women (Diabetes Care 2012 35:574-580).

HAPO, which involved more than 25,000 pregnant women in nine countries, was designed to assess the risks of adverse outcomes associated with various degrees of maternal glucose intolerance less severe than in overt, pre-existing diabetes. Participants underwent OGTT when they were 24-32 weeks’ pregnant and also had HbA1c testing performed. After delivery, samples of babies’ cord blood were taken and analyzed for C-peptide and glucose values, and various anthropometric measurements were taken. The original report of the study, published in 2008, found significant graded relationships between increasing maternal glucose levels and the frequency of four primary and five secondary outcomes. In their recently published follow-up analysis involving HbA1c, HAPO researchers conducted numerous statistical analyses assessing associations between HbA1c and OGTT results and both primary and secondary outcomes.

“Even though there had been previous reports indicating that HbA1c was not an effective way of detecting gestational diabetes mellitus, it had never been examined in a large setting. HAPO gave us an opportunity to do that,” explained co-author Boyd Metzger, MD. “Our thinking was that if HbA1c could be used in this way, it would simplify the detection of gestational diabetes mellitus, at least in a percent of cases.” Metzger is Tom D. Spies professor of metabolism and nutrition at the Northwestern University Feinberg School of Medicine in Chicago.

The researchers found that among the 21,064 participants for whom HbA1c results were available and who did not have hemoglobinathies, the mean HbA1c value was 4.79 ±0.40%. Correlations between HbA1c and fasting, 1-hr, and 2-hr plasma glucose and a composite OGTT measure combining all three glucose measures were 0.29, 0.254, 0.227, and 0.324, respectively. The authors used a series of statistical models to assess the relationships between maternal HbA1c and primary outcomes and between maternal glucose, HbA1c, and primary and secondary outcomes. They found that associations with several outcomes measures, including birth weight, sum of skinfolds, and percent body fat >90th percentile, were significantly stronger for each of the glucose measures than for HbA1c. HbA1c also was not significantly related to sum of skinfolds >90th percentile. As well, associations between three glucose measurements—fasting, 1-hr, and composite—were significantly stronger for cord C-peptide >90th percentile than was HbA1c. The authors did not, however, observe a significant difference between the various glucose measurements and HbA1c for associations with primary C-section, clinical neonatal hypoglycemia, and preeclampsia.

Given these findings, the investigators concluded that HbA1c results are not a useful alternative to OGTT for diagnosing gestational diabetes. “While there were associations with HbA1c and outcomes, they were not as strong as associations with glucose tolerance test measurements. Therefore, we couldn’t make a good case that doing HbA1c as a screening test first would be an effective way of detecting gestational diabetes,” explained Metzger.

If not a total surprise, these findings demonstrate the limitations of HbA1c as a marker for glucose intolerance in pregnancy. “This study suggests we can’t rely on HbA1c to diagnose gestational diabetes,” said Allison Stuebe, MD. “The results provide empiric data that we loose information when we rely on HbA1c rather than glucose load testing.” Stuebe, who was not involved in HAPO, is an assistant professor of obstetrics and gynecology at the University of North Carolina at Chapel Hill.

Both she and Metzger acknowledged the legitimate interest among clinicians and patients in simplifying the gestational diabetes testing process. “From an implementation standpoint, getting every pregnant woman to show up fasting and complete a two-hour glucose challenge test is not trivial,” said Stuebe. “If we could do an HbA1c test and say, your value is less than X, we’re pretty sure your two-hour oral glucose tolerance test is going to be OK so there’s no need to do it, that would be really useful information. If we could get by with HbA1c alone that would be even better.”

Metzger agreed. “If HbA1c had been a strong way to identify gestational diabetes, that would have been a good way to go. Oral glucose tolerance testing is an added burden on ladies in pregnancy, so it would definitely be a great appeal to test them anytime of the day they have an appointment—not just first thing in the morning, and when they’re fasting,” he said. “It was definitely worth a careful evaluation, but unfortunately, HbA1c doesn’t have the strength we wanted if we were going to use it as a primary test.”

Stuebe wondered whether further analysis of the findings might reveal actionable information. “The authors show that HbA1c is associated with more adverse outcomes, but the key question will be, what is the most effective strategy in clinical practice that identifies moms at risk,” she said. “We need to look at how much information we gain with the fasting, one- and two-hour glucose tolerance tests and weigh this against the additional logistical and financial costs of universal glucose load screening to determine which strategy will identify the most at-risk moms and give the best opportunity for intervention.”

If the HAPO findings don’t support a role for using HbA1c to diagnose gestational diabetes, Metzger believes the test still has value in the management of pregnant women. “In the IADGSP paper, there was a recommendation that HbA1c could be used as an early test of overt diabetes and so far, bodies like ADA haven’t endorsed that recommendation,” he explained. “However, off-label there will be some people who use the measurements in early pregnancy to detect diabetes. I think that’s very useful because women who enter pregnancy with overt, pre-existing diabetes really need to be treated as we treat people with known diabetes. They need to be identified, treated intensively, and followed-up after pregnancy, because it’s likely to persist after pregnancy.”