Exploring the Optimal Diagnostic Thresholds of Hyperglycemia During Pregnancy

Gestational diabetes mellitus (GDM) is a specific disease during pregnancy that is characterized by the onset of abnormal maternal glucose tolerance,¹ which was first discovered by O’Sullivan and Mahan² in 1964. Since then, the gradual understanding of the impact of GDM on the long-term risk of maternal diabetes and other metabolic diseases has led to modifications in the diagnostic criteria of GDM by different associations and organizations. The American National Diabetes Data Group³ and the American Diabetes Association⁴ both recommended a 50-g glucose challenge test before the 100 g/75 g oral glucose tolerance test (OGTT), forming the rudiment of a “two-step” screening strategy for GDM. A revolutionary transformation in GDM diagnostic criteria occurred in 2008, following the publication of the hyperglycemia and adverse pregnancy outcome (HAPO) study. This long-term observational study of 25,505 nonovert diabetes singleton pregnancies from 15 centers of nine countries explored the relationship between hyperglycemia and the adverse perinatal outcomes. A continuous positive linear association between hyperglycemia and adverse perinatal outcomes was observed, whether the glucose level was measured categorically or continuously. This demonstrated that, besides the long-term impact of GDM on maternal metabolism, GDM directly impacted the gestational process as well, especially in increasing the fetal weight and the C-peptide levels of the umbilical cord. The fetal birth-weight increased from 240 to 300 g with the lowest to the highest glucose category, while the adjusted odds ratio increased from 1.37 to 5.01 and 1.41 to 7.65, respectively, for the risk of large-for-gestational-age (LGA) infant and cord-blood serum levels of C-peptide.⁵ In view of these findings, the International Association of Diabetes and Pregnancy Study Group (IADPSG) proposed a new GDM diagnostic criteria. The IADPSG recommended a “one-step” approach for screening, with the threshold based on a 1.75-fold risk of developing neonatal macrosomia, which was much lower than the previous diagnostic standards.⁶ This proposal was accepted by many global organizations, including the World Health Organizations⁷ and the International Federation of Obstetrics and Gynecology,⁸ and was widely used in different countries including China, aiming to reduce the risk of LGA, cesarean delivery (CD), and other adverse perinatal outcomes. However, other experts expressed concern over adopting this approach, arguing that it was not well understood how the additionally identified patients with GDM would benefit from the treatment, especially weighing the consequent financial costs, workload of the health service (more intensive maternal and fetal assessments), and the psychological burden when labeling individuals as GDM patients.⁹ Furthermore, there were concerns as to whether there was sufficient high-quality evidence proving that one approach could replace all the other approaches and be popularized worldwide.

Recently, Professor Crowther and her team¹⁰ shared their latest findings from a 5-year randomized controlled trial (RCT) exploring the diagnostic efficacy over lower and higher GDM diagnostic criterion in New Zealand. Both of the standards were based on the 75-g “one-step” approach OGTT taken between 24 and 32 weeks. The lower criterion was the IADPSG criterion, where GDM was defined when any of the glucose measurements reached 92 mg/dL (5.1 mmol/L), 180 mg/dL (10.0 mmol/L), or 153 mg/dL (8.5 mmol/L) for fasting, 1-hour, and 2-hour levels, respectively. The threshold for the higher criteria was 99 mg (5.5 mmol/L) for the fasting glucose level and 162 mg/dL (9.0 mmol/L) for the 2-hour glucose level. A total of 4061 singleton pregnancies were randomized 1:1 to these groups according to their body mass index (<25 and ≥25 kg/m²). The study found that, in addition to a 2.5 times higher incidence of GDM in the lower criteria group (15.3% vs. 6.1%), both groups showed a similar risk of LGA (8.8% in the lower criteria group vs. 8.9% in the higher criteria group; adjusted relative risk, 0.98; 95% confidence interval, 0.80–1.19; P = 0.91). However, the risk of induction of labor, pharmacologic agent usage, neonatal hypoglycemia, and the demand of health services were significantly higher in the lower criteria group (relative risk, 1.12, 2.40, 1.27, and 2.42, respectively; P < 0.05). The authors concluded that even when treated positively, a lower threshold of criteria for GDM diagnosis did not essentially improve the perinatal outcomes.

Similar circumstances were observed in another large RCT study conducted in the United States of America.¹¹ A total of 23,792 women participated in this study, with the team comparing the one-step IADPSG criterion with the “two-step” prevalent Carpenter-Coustan criterion that is currently recommended by the American Diabetes Association¹² and the American College of Obstetricians and Gynecologists.¹³ In the “two-step” approach, the participants first took a 50-g glucose challenge test, and if the 1-hour plasma glucose level was within 130/140 mg/dL (7.2/7.8 mmol/L in two hospitals, respectively) to 180 mg/dL, a 100-g OGTT was applied with the thresholds for fasting, 1-hour, 2-hour, and 3-hour glucose levels at 95 mg/dL (5.3 mmol/L), 180 mg/dL, 155 mg/dL (8.6 mmol/L), and 140 mg/dL, respectively; GDM was diagnosed if at least two values reached the threshold. The researchers also found twice as much GDM morbidity in the IADPSG group (one step vs. two step:16.5% vs. 8.5%). The risk of LGA and gestational hypertensive disorders, CD rate, and other composite perinatal outcomes (stillbirth, neonatal birth, and birth injury) were comparable between the groups; and this remains after the intention-to-treat analysis (relative risk, 0.95, 1.00, 0.98, and 1.04, respectively; P > 0.05). This result is in accordance with the findings from another RCT study, which also compared the efficacy between these two diagnostic criteria.¹⁴ In this third RCT, the same results were observed whether in a higher rate of GDM in the IADPSG group (14.4% vs. 4.5%, P < 0.001) or a similar risk of LGA, CD, small for gestational age, and the composite adverse maternal outcomes in both groups (relative risk, 0.85, 1.03, 0.93, and 0.99, respectively; P > 0.05), in both the intention-to-treat analysis and the subgroup analysis of the participants who strictly followed all procedures. Moreover, this result remained the same in the non-GDM pregnancies of each group. Collectively, these three studies shared the opinion that, although a lower threshold for GDM diagnostic criteria is sensitive for detecting more women at risk of hyperglycemia, it does not efficiently reduce the risk of LGA, CD, and other complications compared with the higher standard; thus, the value is still debatable.

The surveillance of hyperglycemia during pregnancy is of great significance to the national maternity care system. Geographic and demographic divergences mean there is considerable variation in lifestyle and dietary habits among global regions; thus, establishing one unified universal GDM diagnostic criterion may be challenging, and specific considerations must be taken into account so as to benefit the majority of the population. For instance, China is still faced with a deficiency in screening preconceptional diabetes (with an awareness rate of only 1.2%)¹⁵; hence more women are at risk of hyperglycemia when they are pregnant. GDM “one-day care” clinics have been established in China since 2011. However, the uneven distribution of medical resources, especially the lack of proper guidance from the healthworkers, means the concept of maintaining a healthy lifestyle including moderate exercise and balanced nutrition intake —a fundamental part of GDM management—is not widely recognized.^16,17 Therefore, considering the heavier GDM burden of China,¹⁸ a lower threshold of diagnostic criteria may be more applicable from a broader perspective. In addition, in Crowther’s study, the subgroup analysis of the women whose blood glucose level were within the two criteria revealed that those who were considered to have GDM and received the relevant treatment in the lower criteria group (n = 195) had a significantly lower risk of LGA compared with the women who were considered euglycemia with no intervention based on the higher GDM criteria (n = 178; rate of LGA, 6.2% vs. 18.0%; adjusted relative risk, 0.33; 95% confidence interval, 0.18–0.62).¹⁰ This proved the effect of a positive intervention based on the lower criteria. Moreover, as mentioned in the Pittsburgh study, the limited sample sizes means the perinatal outcomes can only be compared with the whole population between the two groups, instead of with the GDM patients of each group.¹⁴ This illustrates the need for additional solid evidence to fill this blank in the future.

In conclusion, although recent studies have revealed that a lower threshold of GDM criteria may not decrease the risk of the relevant adverse perinatal outcomes overall, whether it can be applied in other countries warrants further discussion. We advocate for further high-quality research to explore the optimum diagnostic thresholds of hyperglycemia during pregnancy.