Despite widespread acceptance of the value of screening for and treating gestational diabetes mellitus (GDM) in the United States, evidence concerning a benefit to this approach has been lacking. Confounding variables have specifically raised the question as to whether mild forms of carbohydrate intolerance are associated with increased perinatal and maternal risks. The findings of 2 recently conducted large-scale randomized clinical trials both indicate significant reductions in fetal overgrowth, shoulder dystocia, and preeclampsia with treatment. Thus, compelling evidence now exists that the benefit outweighs risks associated with the treatment of mild GDM.
The frequency of gestational diabetes mellitus (GDM) is increasing worldwide with 1-14% of pregnancies being affected. The current US economic burden of treatment for GDM assuming a 4.5% incidence translates to estimated maternal and newborn costs of $636 million annually. With both increased incidence and proposed lowering of the thresholds for diagnosis, the health care cost of GDM can be expected to rise proportionately. It follows that the debate as to whether a benefit exists to the treatment of GDM assumes even greater importance now than in the past.
Although it has long been recognized that women with preexisting type 1 and type 2 diabetes are at increased risk for adverse maternal and fetal outcomes, the relationship of GDM to various perinatal risks has been less clear. The seminal work of O’Sullivan and Mahan to develop glucose tolerance test criteria for the diagnosis of GDM was performed nearly 50 years ago. Using a longitudinal study design in following up 752 unselected pregnant women, the oral glucose tolerance test (OGTT) diagnostic thresholds selected were based on the subsequent likelihood of developing adult-onset diabetes and were not predicated on any specific pregnancy outcome.
Women with GDM and elevated fasting glucose levels appear to be at risk for fetal overgrowth and perinatal morbidity if optimal care is not provided. Whether a significant association of milder forms of carbohydrate intolerance exists with macrosomia and adverse perinatal outcomes has been questioned and debated for several decades. Much of the problem with interpreting older studies focusing on this issue is that confounding variables including parity and maternal obesity may not have been considered in the analyses or that treatment was in fact applied to the population described. Thus, the effect of glycemia on various outcomes may have been incorrectly estimated.
The recent multicenter observational Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study has settled much of the debate concerning the relationship of various degrees of maternal glycemia to specific pregnancy outcomes. In this large-scale international study, which included enrollment of women with fasting glucose up to 105 mg/dL, the authors described the continuous association between maternal glucose concentrations (utilizing a 75 g blinded 2 hour OGTT) and increasing birthweight, cord blood serum C-peptide levels and other adverse pregnancy outcomes. These associations were present at glucose levels currently lower than those used to diagnose GDM.
Screening for GDM has been recommended for most pregnant women despite challenges by some and little evidence to support a treatment benefit to the identification and treatment of mild carbohydrate intolerance during pregnancy. The practice of universal screening for GDM has therefore been adopted by most US obstetricians for nearly 15 years. However, lack of international consensus regarding diagnostic criteria has continued to add to the problem of assessing the value of treatment.
At present, a 3 hour 100 g diagnostic test is utilized predominantly in the United States, whereas much of the world uses a 75 g, 2 hour test. Nearly 5 years ago, the Australian Carbohydrate Intolerance Study in Pregnant women (ACHOIS), the first large-scale randomized treatment trial for GDM, concluded that treatment reduces serious morbidity and may also improve the quality of life.
Despite these findings, the 2008 guidelines of the US Preventive Services Task Force concluded that there is insufficient evidence to assess the benefits and harms of screening and treatment of GDM. This group, however, acknowledged that their review failed to consider intermediate outcomes such as macrosomia, cesarean delivery, and shoulder dystocia.
The recent update to the Cochrane review on this subject identified 8 randomized trials concerning treatment of GDM. Of the 1418 women enrolled in these studies, 1000 were actually represented by the ACHOIS study. In fact, ACHOIS was the only treatment trial that included a large untreated control group because the remainder of the studies described were trials comparing various treatment modalities for GDM. The author’s conclusions were that women with GDM should be considered only for specific treatment in addition to routine obstetric care.
This recommendation was followed by the acknowledgement that the decision about whether to offer specific dietary advice or more intensive treatment is not yet clear. With a disconnection between current practice, the findings of a well-designed randomized trial, and these conclusions, further analysis of existing and recent data concerning potential treatment benefit of GDM is warranted.
Retrospective studies
Langer et al provided the only large-scale case control study describing perinatal outcomes in treated vs untreated GDM. These authors described pregnancy outcomes in a largely Hispanic population of 555 women with untreated GDM (diagnosed after 37 weeks’ gestation and including women with the entire spectrum of glucose abnormalities meeting the criteria for diagnosis) who were matched with 1110 women treated for GDM as well as 1110 normal subjects. A composite adverse outcome consisting of large for gestational age (LGA), neonatal respiratory disease, hypoglycemia, hyperbilirubinemia, shoulder dystocia, and stillbirth was present in 59% untreated compared with 18% treated and 11% for the nondiabetic population. The rates of LGA infants were 29% in untreated vs 11% with treatment that was similar to the rate in the control population. A 2-4 fold increase in metabolic complications in the offspring of untreated women was found compared with controls.
Multiple logistic regression analysis confirmed fasting glucose as the most predictive factor for the composite perinatal outcome in untreated GDM. Controlling for maternal weight, parity, and disease severity according to fasting glucose level also revealed a 2-3 fold decrease in adverse outcomes with treatment.
The authors also examined whether a treatment benefit existed across the spectrum of maternal glycemia. Treatment of women with mild GDM, defined as a fasting glucose level less than 95 mg/dL, was associated with a 10% rate of LGA compared with 20% without treatment. Thus, the findings of this study suggested that even mild degrees of glucose intolerance discovered during pregnancy can benefit from medical intervention.
Randomized treatment trials
The US Preventive Services Task Force identified only 2 randomized controlled trials that tested treatment vs no treatment of gestational diabetes in universal screening programs. The fair quality randomized controlled trial of O’Sullivan et al found that treatment of women at high risk for GDM reduced macrosomia but not perinatal death. Macrosomia was reduced from 13.1% in controls to a rate of 4.3% in treated subjects. Enrollment criteria in this study included a fasting glucose of 110 mg/dL such that women with mild postprandial hyperglycemia were not considered. Insulin was prescribed in the treatment arm of the study. Because this study was performed in an era preceeding the advent of self–blood glucose monitoring, O’Sullivan et al hypothesized that the treatment effect observed in this study was likely related to significant lowering of postabsorptive blood glucose levels.
Nearly 40 years passed before the findings of the long-awaited ACHOIS study were reported. This study was, a multicenter, 10 year, randomized-treatment trial of 1000 women conducted at 14 centers in Australia. The study was designed to determine whether treatment of mild GDM would reduce the rate of perinatal complications and whether it would affect maternal mood and quality of life. Enrollment was accomplished between 24 and 34 weeks’ gestation in women with a 75 g OGTT result of a 2 hour glucose level between 7.8 and 11.0 mmol/L (140-198 mg/dL) and a fasting plasma glucose level less than 7.8 mmol/L (<140 mg/dL).
The treatment group received individualized dietary counseling, and subjects were instructed to perform daily self–blood glucose monitoring 4 times daily with a target fasting value of 3.5-5.3 mmmol/L (63-99 mg/dL). The authors did not report glycemic data from the treatment arm; however, insulin was required to achieve glycemic control in 20% of women.
Treatment was associated with a significant reduction in the rate of the primary outcome, a composite of serious perinatal complications (perinatal death, shoulder dystocia, birth trauma including fracture or nerve palsy) (adjusted relative risk, 0.33; confidence interval, 0.14–0.75). Overall, 7 infants in the treatment group had a serious complication (all shoulder dystocia) compared with 23 infants in the untreated group of whom 5 experienced perinatal death, 4 had birth trauma, and there were 16 cases of shoulder dystocia.
Among secondary neonatal outcomes, there were no significant differences in the rates of hypoglycemia requiring intravenous therapy, jaundice requiring phototherapy, or respiratory disease requiring supplemental oxygen. Neonatal intensive care unit admissions were remarkably high in both groups; 71% in treated vs 61% in untreated ( P = .01). Importantly, treatment did reduce the frequency of LGA infants from 22% to 13% and birthweight greater than 4000 g from 21% to 10%. A secondary analysis also demonstrated that treatment was associated with lower cord serum leptin concentrations. Among maternal outcomes, preeclampsia was significantly reduced with treatment (12% vs 18%). The rate of induction of labor was higher in treated GDM, yet cesarean rates were nearly equivalent between groups (31% in treated vs 32% in untreated women).
The ACHOIS study, hailed by many as level 1 evidence now supplying proof that a benefit exists to treating mild GDM, has also been criticized. Most of the concerns have centered on certain elements of the trial including the choice of some of the components of the primary outcome. First is the fact that shoulder dystocia is a potentially subjective diagnosis and that resultant birth trauma is actually the outcome of interest attached to shoulder dystocia, which is most significant.
The inclusion of shoulder dystocia as a component of the primary outcome is particularly concerning as is the fact that the difference in this outcome among study groups provided the overall statistical difference observed between treatment and nontreatment groups. The details of the 5 perinatal deaths in the untreated group also have been cited as a concern. The report included 2 stillbirths of appropriately grown fetuses at term as well as inclusion of a lethal anomaly in this group in the untreated arm of the study, all of which seem to be unrelated to mild gestational diabetes.
In spite of the criticisms concerning the primary outcome choice, the findings of decreased fetal macrosomia and effects on the adipoinsular axis of the offspring associated with treatment in the ACHOIS trial are noteworthy. Although some consider macrosomia to be merely a secondary outcome and not a true health outcome, the increasing evidence that exists that excess neonatal fat poses a risk for both future diabetes and obesity is compelling. The ACHOIS study finding of a reduction in preeclampsia is similarly compelling because this complication increases both maternal morbidity and preterm delivery in the GDM population.
Did the ACHOIS trial, however, really evaluate women with mild GDM? The fairly wide range in fasting glucose criteria for entry would suggest this is not the case. Some enrolled women, although a minority, actually met criteria for overt diabetes with fasting glucose levels exceeding 125 mg/dL. The inclusion of women with this degree of carbohydrate intolerance limits the generalizability of the findings of this study as they relate to the mildest forms of GDM.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
