Optimizing Patient Outcomes in Pregnancy: What Continuous Glucose Monitoring Metrics Should We Be Aiming For?

As self-monitoring of blood glucose revolutionized diabetes care in the 1980s, continuous glucose monitoring (CGM) is transforming management today. The use of CGM has improved glycemic control (1) and quality of life for many patients and their families (2). Increasing data also show the usefulness of CGM during pregnancy. The use of masked, retrospective, intermittent use of CGM was first shown to be beneficial during pregnancy in a U.K. randomized trial showing an improvement in HbA_1c at 32–36 weeks’ gestation and a reduction in macrosomia in the intervention group (3). Further benefit was shown with real-time CGM in the Continuous Glucose Monitoring in Women With Type 1 Diabetes in Pregnancy Trial (CONCEPTT), a randomized trial of real-time CGM versus standard care with self-monitoring of blood glucose in women with type 1 diabetes during pregnancy. In this trial, the use of CGM was associated with improvement in glycemic parameters and reduction in adverse neonatal outcomes (4). Participants in the CGM arm spent 68% of time in the pregnancy target range of 63–140 mg/dL at 34 weeks’ gestation, compared with 61% in the control group. CGM use was associated with a reduction in large-for-gestational-age (LGA) infants, neonatal hypoglycemia, and neonatal intensive care >24 h.

Since this randomized trial, observational studies have shown CGM metrics to be predictive of pregnancy outcomes. In an observational study of 186 pregnancies in Sweden, one-half of which included use of real time-CGM and one-half use of intermittently scanned CGM, several parameters (mean glucose, standard deviation [SD], time in range [TIR], time above range [TAR]) were significantly associated with LGA and a composite of neonatal outcomes (5). In a further analysis of the CONCEPTT data, Scott et al. (6), using functional data analysis, showed that while women with LGA infants had only slightly higher mean glucose levels, they remained at these levels for several hours in the day. Glucose levels of women with LGA infants were ∼7–16 mg/dL higher for 4.5 h/day at 12 weeks’ gestation, for 16 h/day at 24 weeks’ gestation, and for 14 h/day at 34 weeks’ gestation in comparison with women with non-LGA infants. In addition, a study by Meek et al. (7) looking at CGM metrics and biomarkers in CONCEPTT demonstrated that TIR and TAR were as predictive of adverse pregnancy outcomes as HbA_1c.

Based on this evidence, the International Consensus on Time in Range made the following recommendation for women with type 1 diabetes in pregnancy (8):

1. 1.

   That they spend >70% of time with glucose levels within the pregnancy time in range (63–140 mg/dL)

2. 2.

   That they spend <25% of time with glucose levels above range (above 140 mg/dL)

3. 3.

   That they spend <4% of time with glucose levels below range (below 63 mg/dL)

Given the relative paucity of data with regard to other CGM metrics, and the paucity of data in women with type 2 diabetes, more research was needed.

As reported in this issue of Diabetes Care, Sanusi et al. (9) performed a retrospective cohort study of 117 patients with type 1 and type 2 diabetes who delivered in a U.S. tertiary center and who used CGM during pregnancy. The aim was to assess the association of various CGM metrics with maternal and neonatal outcomes and identify evidence-based targets for use during pregnancy. They found that all CGM metrics, except time below range, were associated with the adverse neonatal composite outcome of fetal or neonatal mortality, LGA, small for gestational age, neonatal intensive care admission, neonatal hypoglycemia, shoulder dystocia or birth trauma, and hyperbilirubinemia. These metrics included TIR, TAR, glucose variability, average glucose, and glucose management indicator. The neonatal composite was reduced by 28% for every 5%-point increase in the pregnancy TIR, and a 5% increase in TAR was associated with >40% increase in this composite outcome. They also found that the statistically optimal TIR was 66–71%, giving more credence to the current recommendation of achieving >70% TIR throughout pregnancy. In addition, individual components of the composite were assessed and found to be associated with the majority of CGM metrics, giving more credibility to the use of these metrics for assessment and guidance during pregnancy. Interestingly, average glucose and glucose management indicator were found to have similar if not slightly better predictive properties in comparison with TIR and TAR.

This study has many strengths. The study provides important information regarding CGM metrics and their relationship with pregnancy outcomes, which is very relevant for patients and clinicians. Participants wore the CGM 83% of the time, for at least 4 months, and the sample size was sufficient to draw interesting and valuable conclusions. This is one of the first studies to assess CGM metrics from women with type 2 diabetes and show a positive association of these metrics with pregnancy outcomes in these women as well.

Some limitations of the study bear consideration. Because of the limited number of women in the study, receiver operating characteristic curves for assessment of evidence-based targets could not be used for type 1 and type 2 populations separately. Data were only obtained from 19 weeks onward, so limited conclusions can be drawn from the first trimester, and individual trimester-specific targets cannot be inferred. Interestingly, in a study combining CONCEPTT and Swedish data and looking at glycemic profiles weekly during pregnancy, Scott et al. (10) found that CGM metrics diverged as early as 10 weeks, in those with and without LGA infants, suggesting that glucose targets need to be achieved very early on to prevent complications. Further data from early in pregnancy, or even preconception, would be welcome. Although the optimal statistical TIR was found to be 66–71%, whether a higher TIR may be of further benefit remains unclear. Larger prospective studies will be required to better elucidate this question.

What further questions remain unanswered? This study suggests that, as with women with type 1 diabetes, the CGM metrics of women with type 2 diabetes are predictive of adverse perinatal outcomes. However, more data are needed to determine more precisely what the CGM targets should be in patients with type 2 diabetes and indeed GDM and whether they should be the same for all women with diabetes. The CGM summary metrics associated with having LGA compared with normal birth weight have been defined weekly across each trimester of pregnancy, and the metrics change progressively from beginning to end of pregnancy (10). We do not yet know whether the CGM metrics associated with composite adverse perinatal outcomes likewise vary across each trimester. Several studies have looked at what “normal” mean glucose is during pregnancy, and further research will be needed to elucidate how close women with diabetes need to get to these levels to achieve good outcomes. In addition, further data on whether the use of mean glucose plus TIR is better than use of TIR alone will be helpful. Furthermore, whether nocturnal targets should be different from daytime targets remains to be determined. It is worth noting that glucose variability was the only CGM metric associated with preeclampsia and preterm birth in this study. Whether this will be found in future studies remains to be determined.

While CGM has transformed the management of preexisting diabetes in pregnancy, it is notable that the CGM targets are still difficult to achieve for a majority of women and there remains room for improvement to get pregnancy outcomes closer to those seen for women without diabetes. In the CONCEPTT trial, only 7.7% of women achieved >70% TIR in the first trimester compared with 10.2% in the second trimester and 35.5% in the third trimester (11). In this study by Sanusi et al. (9), only 8.8% of women with type 1 and 40.8% of women with type 2 diabetes achieved this goal over the course of pregnancy. Further data from closed loop technology trials in pregnancy should inform patients and caregivers of the added benefits of this technology, over and above CGM alone (12), and are thus eagerly awaited.