Arguments in favor

The diagnosis of GDM has traditionally been made by using OGTTs that were originally designed in such a way that about 2.5% of the pregnant population (>2 SD) had values exceeding the normal range or were simply based on criteria used in nonpregnant individuals. GDM was therefore a laboratory-based diagnosis and not one that was related to perinatal outcome characteristics. The HAPO results provide data on the relation between OGTT values and perinatal outcome and are therefore better suited to identify abnormalities in metabolism than the previous criteria, although the exact cut-off levels remain uncertain.

A recent survey of prevalence of GDM by country showed a median incidence of GDM of about 5% in all world regions except southeast Asia (median 8%). Obesity rates have doubled in the last 15 years in many countries, including the United States (presently about 30% ). The odds ratio (OR) for GDM in obese women with a body mass index (BMI) >30 is 3.5-4. Therefore a higher incidence of GDM than the reported 5% seems likely.

Two randomized studies have shown that treatment of GDM improves outcome, by lowering the incidence of fetal macrosomia, mortality, birth trauma, and–in one study–cesarean deliveries. These results make treatment and screening programs for GDM mandatory. However, criteria for GDM differed from the new proposed diagnostic criteria and were likely to include the more severe cases. Benefits for diagnosing and treating minor glucose abnormalities have, therefore, not yet been proven. In October 2012, Bodmer-Roy et al published data indicating that women classified as nondiabetic by the Canadian Diabetes Association Criteria but considered to have GDM according to the IADPSG criteria (study group) had similar pregnancy outcomes as women without GDM. However, the data were not conclusive since there was a higher incidence of fetal macrosomia and preeclampsia in the study group, which was, however, not significant (the series reported only 186 cases and 372 controls).

Treatment of GDM is usually relatively easy, with insulin requirement in only 8-20% of women. This argument would favor identification of all women with minor hyperglycemia, but it does not take into account that overdiagnosis might result in other actions like labor inductions or cesarean deliveries without proven benefit (see below).

Thus far 2 studies have addressed cost-effectiveness of GDM screening according to the IADPSG criteria using decision analysis models. One study showed that the IADPSG recommendations are cost-effective only when postdelivery care would reduce the frequency of diabetes in these women. In a second study, the cost-effectiveness of the current US screening, 1-hour glucose challenge test followed by a 3-hour OGTT, was compared with the new IADPSG guidelines for the 2-hour OGTT. In the latter study, treatment according to the new guidelines would be effective if treatment would result in decreased preeclampsia >0.55% and decreased cesarean delivery >2.7%.

The arguments against

The OGTT has poor reproducibility, with about 30% of patients having a negative test result after a previous positive result. Reproducibility is even poorer with minor degrees of glucose elevations. In other words, the stricter the threshold values the more patients will be diagnosed by chance as having GDM.

It is questionable if a birthweight >90th centile is an adequate endpoint for defining OGTT threshold values. Even with the strict IADPSG threshold criteria, only about 22% of all large-for-gestational-age pregnancies occur in patients with GDM. It is unknown if this will be the subgroup of infants with an increased risk for obesity and metabolic syndrome later in life. Growth characteristic for type 1 diabetes, type 2 diabetes, and GDM is characterized by disproportionate fetal growth with a large abdominal circumference as compared to the head circumference, both in appropriate- and large-for-gestational-age infants. A better description/definition of macrosomia will therefore be necessary.

Moreover and as has been pointed out before, GDM is related to childhood obesity, but mainly in case of coexisting maternal obesity. A Finnish study has recently shown that overweight and abdominal obesity in 16-year-old adolescents was related to maternal obesity and especially to the combination of maternal obesity and GDM, but not to GDM alone. In a systematic review on the association between GDM and childhood overweight and obesity it was found that there is currently inconsistent evidence of an association due to methodological limitations of existing studies. In another recent metaanalysis it was found that there is an association between all types of diabetes in pregnancy with childhood obesity, but that this association disappeared after adjustment for maternal BMI in the 3 studies in which this was reported. The Finnish study by Pirkola et al had not yet been included in this analysis. So there is considerable doubt as to the independent effect of GDM on obesity in offspring, but data corrected for maternal BMI are limited. Absence of long-term risks of GDM for the offspring would weaken the need for diagnosing very mild cases.

From the HAPO data, we know that both obesity and GDM are independent risk factors with synergistic effects regarding preeclampsia, primary cesarean deliveries, macrosomia, increased cord C-peptide levels, and newborn body fat. Follow-up studies of these infants will be important to assess the risk factors for metabolic syndrome during childhood and thereafter, ie, maternal obesity and/or GDM. Given the synergistic effects of GDM and obesity on direct and later outcome in the offspring, it might prove useful to use stricter threshold criteria for obese women than for nonobese women, instead of using the same thresholds for everyone.

Earlier this year Moynihan et al addressed the problem of overmedicalization in an article entitled “Preventing overdiagnosis: how to stop harming the healthy.” As drivers for overdiagnosis they mentioned: (1) technological changes detecting even smaller abnormalities; (2) commercial and professional vested interests; (3) conflicted panels producing expanded disease definitions and writing guidelines; (4) legal incentives that punish underdiagnosis but not overdiagnosis; (5) health system incentives favoring more tests and treatments; and (6) cultural beliefs that more is better, ie, faith in early detection unmodified by its risks. Many of these drivers may be present regarding GDM screening and GDM was indeed mentioned as an example. Quoting Moynihan et al and an earlier article by Black : “The ability to detect smaller abnormalities axiomatically tends to increase the prevalence of any given disease. In turn this leads to overestimation of the benefits of therapies, as milder forms of the disease are treated and improvements in health are wrongly ascribed to treatment success, creating a ‘false feedback’ loop fuelling a ‘cycle of increased testing and treatment, which may eventually cause more harm than benefit.’” Risks of overdiagnosing GDM may include earlier intervention and subsequent iatrogenic preterm delivery, increase in elective cesarean deliveries (presumed large baby), and inductions of labor and in secondary cesarean deliveries (failed inductions). Long-term risks of a slightly earlier delivery of the infant are largely unknown. Risks of an increasing cesarean delivery rate may include an increase in childhood autoimmune diseases like asthma and type 1 diabetes.

More GDM will generate a greater workload for obstetrician, endocrinologist, and dietician services. For 2 units in Australia it was conservatively estimated that overall workload would increase by 20-30% but possibly by 60%, which might require a structural change in services.

The arguments against

The OGTT has poor reproducibility, with about 30% of patients having a negative test result after a previous positive result. Reproducibility is even poorer with minor degrees of glucose elevations. In other words, the stricter the threshold values the more patients will be diagnosed by chance as having GDM.

It is questionable if a birthweight >90th centile is an adequate endpoint for defining OGTT threshold values. Even with the strict IADPSG threshold criteria, only about 22% of all large-for-gestational-age pregnancies occur in patients with GDM. It is unknown if this will be the subgroup of infants with an increased risk for obesity and metabolic syndrome later in life. Growth characteristic for type 1 diabetes, type 2 diabetes, and GDM is characterized by disproportionate fetal growth with a large abdominal circumference as compared to the head circumference, both in appropriate- and large-for-gestational-age infants. A better description/definition of macrosomia will therefore be necessary.

Moreover and as has been pointed out before, GDM is related to childhood obesity, but mainly in case of coexisting maternal obesity. A Finnish study has recently shown that overweight and abdominal obesity in 16-year-old adolescents was related to maternal obesity and especially to the combination of maternal obesity and GDM, but not to GDM alone. In a systematic review on the association between GDM and childhood overweight and obesity it was found that there is currently inconsistent evidence of an association due to methodological limitations of existing studies. In another recent metaanalysis it was found that there is an association between all types of diabetes in pregnancy with childhood obesity, but that this association disappeared after adjustment for maternal BMI in the 3 studies in which this was reported. The Finnish study by Pirkola et al had not yet been included in this analysis. So there is considerable doubt as to the independent effect of GDM on obesity in offspring, but data corrected for maternal BMI are limited. Absence of long-term risks of GDM for the offspring would weaken the need for diagnosing very mild cases.

From the HAPO data, we know that both obesity and GDM are independent risk factors with synergistic effects regarding preeclampsia, primary cesarean deliveries, macrosomia, increased cord C-peptide levels, and newborn body fat. Follow-up studies of these infants will be important to assess the risk factors for metabolic syndrome during childhood and thereafter, ie, maternal obesity and/or GDM. Given the synergistic effects of GDM and obesity on direct and later outcome in the offspring, it might prove useful to use stricter threshold criteria for obese women than for nonobese women, instead of using the same thresholds for everyone.

Earlier this year Moynihan et al addressed the problem of overmedicalization in an article entitled “Preventing overdiagnosis: how to stop harming the healthy.” As drivers for overdiagnosis they mentioned: (1) technological changes detecting even smaller abnormalities; (2) commercial and professional vested interests; (3) conflicted panels producing expanded disease definitions and writing guidelines; (4) legal incentives that punish underdiagnosis but not overdiagnosis; (5) health system incentives favoring more tests and treatments; and (6) cultural beliefs that more is better, ie, faith in early detection unmodified by its risks. Many of these drivers may be present regarding GDM screening and GDM was indeed mentioned as an example. Quoting Moynihan et al and an earlier article by Black : “The ability to detect smaller abnormalities axiomatically tends to increase the prevalence of any given disease. In turn this leads to overestimation of the benefits of therapies, as milder forms of the disease are treated and improvements in health are wrongly ascribed to treatment success, creating a ‘false feedback’ loop fuelling a ‘cycle of increased testing and treatment, which may eventually cause more harm than benefit.’” Risks of overdiagnosing GDM may include earlier intervention and subsequent iatrogenic preterm delivery, increase in elective cesarean deliveries (presumed large baby), and inductions of labor and in secondary cesarean deliveries (failed inductions). Long-term risks of a slightly earlier delivery of the infant are largely unknown. Risks of an increasing cesarean delivery rate may include an increase in childhood autoimmune diseases like asthma and type 1 diabetes.

More GDM will generate a greater workload for obstetrician, endocrinologist, and dietician services. For 2 units in Australia it was conservatively estimated that overall workload would increase by 20-30% but possibly by 60%, which might require a structural change in services.