Correlation Between Third Trimester Glycemic Variability in Non-Insulin-Dependent Gestational Diabetes Mellitus and Adverse Pregnancy and Fetal Outcomes

Abstract

Background:

Gestational diabetes mellitus (GDM) is a pregnancy-related metabolic complication. Despite optimal glycemic control from self-monitoring blood glucose (SMBG) in non-insulin-dependent GDM, variations in pregnancy outcomes persist. Glycemic variability is believed to be a factor that causes adverse pregnancy outcomes. Continuous glucose monitoring system (CGMS) detects interstitial glucose values every 5 minutes, and glycemic variability data from CGMS during the third trimester may be a predictor of fetal birth weight and pregnancy outcomes. The aim of this study was to investigate correlation between third trimester glycemic variability in non-insulin-dependent GDM and fetal birth weight.

Method:

This prospective study was conducted in 55 pregnant volunteers with non-insulin-dependent GDM that were recruited at 28 to 32 weeks’ gestation from the outpatient clinic of the Department of Obstetrics and Gynecology, Faculty of Medicine Siriraj Hospital during the study period of August 1 to December 31, 2016. Patients had CGMS installed for at least 72 hours and glycemic variability data were analyzed.

Results:

Of 55 enrolled volunteers, the data from 47 women were included in the analysis. Mean CGMS duration was 85.5 ± 12.83 hours. No statistically significant correlation was identified between glycemic variability in third trimester and birth weight percentiles, or between third trimester CGMS parameters and pregnancy outcomes in the study.

Conclusion:

Based on these findings, third trimester glycemic variability data from CGMS are not a predictor of fetal birth weight percentile, and no significant association was found between CGMS parameters and adverse pregnancy outcomes; thus, CGMS is not necessary in non-insulin-dependent GDM.

Gestational diabetes mellitus (GDM) is defined as glucose intolerance that develops during pregnancy or that is diagnosed at an antenatal visit.¹ The global prevalence of GDM varies from 1% to 14%.¹ Nine hundred cases per year were diagnosed at Siriraj Hospital during 2013-2015.

Optimal glycemic control has a direct impact on the following maternal and fetal conditions: fetal death, intrauterine growth restriction (IUGR) or small-for-gestational-age (SGA), preterm labor (PTL), postpartum hemorrhage (PPH), pregnancy-induced hypertension (PIH), and large-for-gestational-age (LGA), which increases the chance of cesarean delivery or neonatal hypoglycemia.^1-4 Insulin resistance influences long-term maternal and fetal complications, with studies showing increasing risk of diabetes over a 20-year time period of up to 50%.⁵ Children born LGA from GDM mothers tend to develop metabolic syndrome by age 11 at a rate greater than those born with normal weight.⁶

In non-insulin-dependent GDM, self-monitoring blood glucose (SMBG) or intermittent glucose monitoring has been employed, and this group of GDM is presumed to have better neonatal and maternal outcomes than those who have additional insulin requirement. Treatment for the non-insulin-dependent GDM group includes diet control and lifestyle modifications. Despite optimal glycemic control that is achieved via the use of SMBG, variations in fetal and maternal outcomes still occur. A continuous glucose monitoring system (CGMS) was used in some studies to evaluate pregnant women with and without diabetes to obtain more data about blood glucose level, and to analyze glycemic variability and its impact on fetal birth weight percentile.⁷

CGMS is a system that records interstitial glycemic level every 5 minutes via a subcutaneous sensor.^3,7-12 Glycemic value derived from CGMS is highly correlated with blood glucose value (r = .92). Values derived from CGM are calculated into parameters that include mean amplitude of glycemic excursion (MAGE), percentage coefficient of variation (%CV), standard deviation (SD), and mean blood glucose (MBG). %CV is defined as the ratio of the SD to the mean. This parameter describes the magnitude sample values and the variation within them, and allows for standardized comparisons between patients with different levels of mean glycemia. MAGE and mean ± SD are the most popular parameters for assessing glycemic variability, and they are calculated based on the arithmetic mean of differences between consecutive peaks and nadirs of differences greater than one SD of mean glycemia.¹³ Glycemic variability increases oxidative stress and cellular apoptosis but decreases endothelial cell function which may cause adverse pregnancy outcomes.

Willman et al reported that the incidence of fetal macrosomia was not found to significantly increase until the mean glucose concentration reached 130 mg/dl.¹⁴ However, adverse fetal and maternal impact has been reported in patients with non-insulin-dependent GDM that had optimal SMBG or intermittent glucose monitoring and whose MBG were within optimal range of 130 mg/dl. A study in pregnant women showed a correlation between glycemic excursion in the third trimester and birth weight percentiles (r = .29).⁷ Another study in pregnant women with diabetes found the highest correlation between blood glucose and fetal fat at 27-28 weeks, regardless of BMI.^15,16 Many studies have investigated CGMS or glycemic excursion in GDM patients, but no studies have established a cutoff value for glycemic excursion to standardize antenatal care in non-insulin-dependent GDM.

The primary objective of this study was to investigate association between third trimester glycemic variability in non-insulin-dependent GDM and fetal birth weight percentile. The secondary objective was to identify associations between third trimester CGMS parameters and adverse pregnancy and fetal outcomes.

Materials and Methods

This prospective study was conducted in 55 pregnant volunteers with non-insulin-dependent gestational diabetes that were recruited at 28 to 32 weeks’ gestation from the outpatient clinic of the Department of Obstetrics and Gynecology, Faculty of Medicine Siriraj Hospital during the study period of August 1 to December 31, 2016. Siriraj Hospital is Thailand’s largest national tertiary referral center. Written informed consent was obtained from all participants prior to their inclusion in this study. The protocol for this study was approved by the Siriraj Institutional Review Board, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand.

Included volunteers were diagnosed with gestational diabetes following the two-step protocol for diagnosis of GDM, which includes a 50 g 1-hour glucose challenge test (GCT) and a 100 g 3-hour oral glucose tolerance test (OGTT). If glucose challenge test results exceed the selected threshold of ≥140 mg/dl, the OGTT is performed. Two abnormal values meeting or exceeding 95, 180, 155, and 140 mg/dl at fasting, 1 hour, 2 hours, and 3 hours, respectively, after glucose loading are required for diagnosis. This diagnostic protocol is used for early pregnancy screening of women at high risk for gestational diabetes, and at 24 to 28 weeks’ gestation.^17-19

All enrolled volunteers were Southeast Asian, over 18 years old, without other underlying diseases, and currently on diet control and lifestyle modification according to instructions provided by the antenatal care unit without insulin therapy requirement. Gestational age was confirmed by last menstrual period (LMP) and ultrasound early in the first trimester. Anomaly scan revealed no fetal anomaly prior to enrollment. Glycemic data were collected with a CGMS that was installed for at least 72 hours (iPro2® CGM System; Medtronic, Minneapolis, MN, USA). This device has a subcutaneous sensor that measures and records the patient’s mean interstitial glucose level every 5 minutes. CGMS devices were calibrated twice daily using a finger stick capillary blood glucose measurement system (Accu-Chek®; Roche Diagnostics, Basel, Switzerland). Each CGMS monitor was installed for at least 72 hours (range: 72-168), and CGMS data were downloaded into website-based program provided by the CGMS manufacturer. Glycemic variability was assessed using Glyculator, a program designed to calculate glycemic variability.¹³ Standard deviation, percentage coefficient of variation (%CV), MAGE, mean glucose level, AUC of glucose >120, and AUC >130 mg/dl were calculated and electronically recorded. Study participants were blinded to CGMS data. Routine antenatal care for GDM was given to all patients for the remaining duration of their pregnancies. All patients remained on diet control and lifestyle modification with intermittent glucose monitoring. Maternal and neonatal outcomes were assessed after delivery. Pregnancy outcomes included fetal birth weight and adverse outcomes, including neonatal hypoglycemia, respiratory distress syndrome (RDS), neonatal intensive care unit admission (NICU admission), birth asphyxia, PPH, PIH, PTL, and PTB. Patients who received insulin therapy later in pregnancy, who were monitored for less than 72 hours, who were lost to follow-up, and/or who underwent magnetic resonance imaging or computerized tomograghic scan were excluded. Included patients were informed that they could reverse their decision to participate and opt out of the study at any time.

Statistical analysis was performed using SPSS Statistics version 16 (SPSS, Inc, Chicago, IL, USA). Pearson correlation coefficient, chi-square test, and Fisher’s exact test were used to compare variables and to test for association. Data are presented as number, number and percentage, or mean ± standard deviation. A P value < .05 was regarded as being statistically significant.

Results

Mean ± standard deviation age of participants was 34.1 ± 5.3 years (range: 21-45) (n = 52). Of the 55 enrolled volunteers, data from 8 patients were not included in the final analysis for reasons described, as follows: 1 patient opted out citing being uncomfortable calibrating the device; 2 patients received CGMS monitoring for less than 72 hours; 3 patients were lost to follow-up; 1 patient had a device malfunction, so no data were recorded; and, the data from the CGMS of 1 patient could not be retrieved due to a download problem. Only one complication was reported – a minor pruritic rash caused by an adhesive patch. Mean CGMS duration was 85.5 ± 12.83 hours. Maternal characteristics and CGMS data are given in Tables 1 and 2. Pregnancy outcomes are shown in Table 3.

Table 1.

Descriptive Data From Maternal Characteristics and CGMS.

	n	Min	Max	Mean	SD
Age (years)	47	21	45	34.10	5.34
Parity	47	1	6	1.85	1.02
GA at enrollment (weeks)	47	28	32	30.08	1.63
GA at diagnosis (weeks)	47	6	32	16.50	8.13
GA at delivery (weeks)	47	33.71	40.43	38.10	1.65
Maternal height (m)	47	1.42	1.70	1.56	0.063
Prepregnancy BMI (kg/m2)	47	17.31	32.04	23.72	3.95
Gestational weight gain (kg)	47	2.30	24.00	12.12	4.86
Mean glycemic value (mg/dl)	47	92.24	128.30	107.71	10.19
%CV	47	9.51	37.31	19.28	6.09
MAGE	47	26.36	124.25	63.27	21.60
AUC >130 mg/dl	47	0.0	14.0	3.12	3.16
AUC >120 mg/dl	47	0.0	18.0	5.05	4.31
HbA1C (%)	47	4.8	6.1	5.37	0.36
Birth weight (grams)	47	1720	4120	3114.68	523.95
Placenta weight (grams)	34	300	900	666.76	151.45
Birth weight percentiles (Olsen)	47	3.20	94.50	44.83	23.55
2-hour postprandial BG	47	72.00	157.00	106.10	20.26

Table 2.

Glycemic Variability Derived From CGMS Monitoring.

		Mean (mg/dl)	SD (mg/dl)	%CV	MAGE (mg/dl)	AUC >130 (mg/dl)	AUC >120 (mg/dl)
Mean		107.7115	20.8846	19.2821	63.2706	3.117	5.050
SD		10.19039	7.24991	6.08546	21.59666	3.1608	4.3079
Min		92.24	8.79	9.51	26.36	0.0	0.0
Max		128.30	43.81	37.31	124.25	14.0	18.0
Percentiles	50th	106.6200	19.8300	18.0450	61.6750	1.950	3.550
	90th	125.8210	31.5160	26.2260	92.4620	8.240	11.700

Table 3.

Pregnancy Outcomes.

Outcome	Yes, n (%)	No, n (%)
LGA (>90th percentile for Asian)	8 (17)	39 (83)
Birth weight (>50th percentile for Asian)	25 (53.2)	22 (46.8)
Cesarean section (excludes prior C-section [n = 10])	14 (37.8)	23 (62.2)
Cephalopelvic disproportion (n = 37)	10 (26)	27 (74)
PIH	6 (12.8)	41 (87.2)
PPH	1 (2)	46 (98)
PTL <37 weeks	10 (21.3)	37 (78.7)
PTB <37 weeks	8 (17)	39 (83)
Neonatal hypoglycemia	3 (6.3)	44 (94.7)
RDS	5 (10.6)	42 (89.4)
NICU admission	7 (14.9)	40 (85.1)
Neonatal complication	9 (19.1)	38 (80.9)

The primary objective was to investigate for correlation between glycemic excursion in the third trimester and birth weight percentiles using the preterm infant growth chart calculator by Olsen.²⁰ The scatter plot diagram shown in Figure 1 revealed no correlation.

Figure 1.

Scatter plot diagram showing relationship between MAGE and birth weight percentiles.

No correlation was found between birth weight percentiles and any glycemic variability parameter, including %CV, MAGE, SD, AUC of glucose >120, and AUC of glucose >130 mg/dl (r²∼0) but significant correlations were identified between gestational weight gain and birth weight percentiles (r = .437, P = .002), and between maternal height and birth weight percentiles (r = .369, P = .011) as shown in Table 4.

Table 4.

Correlations Between Birth Weight Percentiles (Olsen) and Glycemic Variability Derived From CGMS.

	Mean glycemic value	%CV	MAGE	AUC > 130 mg/dl	AUC > 120 mg/dl	Gestational weight gain	Maternal height
Pearson correlation	0.214	−0.081	0.007	0.066	0.121	0.437	0.369
P valuea	0.149	0.589	0.962	0.661	0.417	0.002	0.011

^aP < .05 indicates statistical significance.

Relationship between glycemic excursion and birth weight of more than 90th percentile²¹ is described in Table 5. No significant difference was found for birth weight percentiles between the high excursion and low excursion groups.

Table 5.

Distribution of Fetal Birth Weight Percentiles.²¹

Percentiles		n (%)
Birth weight	<10th	5 (10.6)
	10th-50th	17 (36.2)
	50th-90th	17 (36.2)
	>90th	8 (17)
	Total	47 (100)

The cutoff values for high and low glycemic excursion were estimated from the 50th and 90th percentiles, and no association was identified between LGA infants and glycemic variability data derived from single period of CGMS monitoring during weeks 28-32 in patients with non-insulin-dependent GDM. Analysis to identify association between adverse pregnancy outcomes and glycemic excursion is shown in Tables 6.1 to 6.8. After excluding previous cesarean section, the cesarean rate was higher in the low excursion group, with MAGE lower than the 50th percentile. Nine of 19 patients in the low excursion group underwent cesarean delivery and 5 of 26 patients in high excursion group (P = .057). One of 4 patients in the high glycemic excursion group had PPH, whereas no PPH was found among low excursion group patients (P = .087). All 3 neonates that were born with neonatal hypoglycemia were born from mothers with MAGE lower than the 50th percentile (14%), but no significant difference was observed between MAGE lower and higher than the 50th percentiles (P = .082). From our study, no significant correlations were found between fetal birth weight and glycemic excursion, and between glycemic excursion and adverse pregnancy outcomes.

Table 6.1.

Difference in Perinatal Outcomes Between MAGE <90th Percentile and MAGE >90th Percentile.

Pregnancy outcomes	MAGE <90th percentile (n = 42)	MAGE >90th percentile (n = 5)	P value
LGA (>90th percentile for Asian)	8	0	.571
Birth weight (>50th percentile for Asian)	22	3	.747
Cesarean section (excludes prior C-section)	13	1	.569
Cephalopelvic disproportion	9	1	.941
PIH	5	1	.608
PPH	0	1	.087
PTL <37 weeks	10	0	.219
PTB <37 weeks	7	1	.851
Neonatal hypoglycemia	3	0	.537
RDS	5	0	.414
NICU admission	5	0	.322

A P value < .05 indicates statistical significance.

Table 6.8.

Difference in Perinatal Outcomes Between AUC >130 mg/dl at <50th Percentile and AUC >130 mg/dl at >50th Percentile.

Pregnancy outcomes	AUC >130 mg/dl at <50th percentile (n = 29)	AUC >130 mg/dl at >50th percentile (n = 18)	P value
LGA (>90th percentile for Asian)	6	2	.692
Birth weight (>50th percentile for Asian)	16	9	.730
Cesarean section (excludes prior C-section)	11	3	.087
Cephalopelvic disproportion	8	2	.180
PIH	3	3	.528
PPH	0	1	.187
PTL <37 weeks	6	4	.901
PTB <37 weeks	5	3	.959
Neonatal hypoglycemia	3	0	.158
RDS	3	2	.934
NICU admission	3	4	.266

A P value < .05 indicates statistical significance.

Table 6.2.

Difference in Perinatal Outcomes Between MAGE <50th Percentile and MAGE >50th Percentile.

Pregnancy outcomes	MAGE <50th percentile (n = 21)	MAGE >50th percentile (n = 26)	P value
LGA (>90th percentile for Asian)	4	4	.740
Birth weight (>50th percentile for Asian)	12	13	.626
Cesarean section (excludes prior C-section)	9	5	.057
Cephalopelvic disproportion	6	4	.272
PIH	2	4	.549
PPH	0	1	.354
PTL <37 weeks	3	7	.293
PTB <37 weeks	3	5	.654
Neonatal hypoglycemia	3	0	.082
RDS	2	3	.824
NICU admission	3	4	.916

A P value < .05 indicates statistical significance.

Table 6.3.

Difference in Perinatal Outcomes Between %CV <90th Percentile and %CV >90th Percentile.

Pregnancy outcomes	%CV <90th percentile (n = 42)	%CV >90th percentile (n = 5)	P value
LGA (>90th percentile for Asian)	8	0	.571
Birth weight (>50th percentile for Asian)	22	3	.747
Cesarean section (excludes prior C-section)	13	1	.569
Cephalopelvic disproportion	9	1	.941
PIH	5	1	.608
PPH	0	1	.87
PTL <37 weeks	10	0	.219
PTB <37 weeks	7	1	.851
Neonatal hypoglycemia	3	0	.537
RDS	5	0	.414
NICU admission	7	0	.322

A P value < .05 indicates statistical significance.

Table 6.4.

Difference in Perinatal Outcomes Between %CV <50th Percentile and %CV >50th Percentile.

Pregnancy outcomes	%CV <50th percentile (n = 22)	%CV >50th percentile (n = 25)	P value
LGA (>90th percentile for Asian)	4	4	.843
Birth weight (>50th percentile for Asian)	12	13	.861
Cesarean section (excludes prior C-section)	9	5	.072
Cephalopelvic disproportion	6	4	.346
PIH	3	3	.867
PPH	0	1	.333
PTL <37 weeks	4	6	.627
PTB <37 weeks	3	5	.562
Neonatal hypoglycemia	3	0	.056
RDS	3	2	.532
NICU admission	4	3	.553

A P value < .05 indicates statistical significance.

Table 6.5.

Difference in Perinatal Outcomes Between AUC >120 mg/dl at <90th Percentile and AUC >120 mg/dl at >90th Percentile.

Pregnancy outcomes	AUC >120 mg/dl at <90th percentile (n = 43)	AUC >120 mg/dl at >90th percentile (n = 4)	P value
LGA (>90th percentile for Asian)	8	0	.344
Birth weight (>50th percentile for Asian)	24	1	.328
Cesarean section(excludes prior C-section)	13	1	.931
Cephalopelvic disproportion	9	1	.849
PIH	5	1	.443
PPH	0	1	.087
PTL <37 weeks	8	2	.142
PTB <37 weeks	7	1	.657
Neonatal hypoglycemia	3	0	.585
RDS	5	0	.471
NICU admission	6	1	.553

A P value < .05 indicates statistical significance.

Table 6.6.

Difference in Perinatal Outcomes Between AUC >120 mg/dl at <50th Percentile and AUC >120 mg/dl at >50th Percentile.

Pregnancy outcomes	AUC >120 mg/dl at <50th percentile (n = 29)	AUC >120 mg/dl at >50th percentile (n = 18)	P value
LGA (>90th percentile for Asian)	6	2	.692
Birth weight (>50th percentile for Asian)	16	9	.730
Cesarean section (excludes prior C-section)	11	3	.087
Cephalopelvic disproportion	8	2	.180
PIH	3	3	.528
PPH	0	1	.187
PTL <37 weeks	6	4	.901
PTB <37 weeks	5	3	.959
Neonatal hypoglycemia	3	0	.158
RDS	3	2	.934
NICU admission	3	4	.266

A P value < .05 indicates statistical significance.

Table 6.7.

Difference in Perinatal Outcomes Between AUC >130 mg/dl at <90th Percentile and AUC >130 mg/dl at >90th Percentile.

Pregnancy outcomes	AUC >130 mg/dl at <90th percentile (n = 42)	AUC >130 mg/dl at >90th percentile (n = 5)	P value
LGA (>90th percentile for Asian)	8	0	.685
Birth weight (>50th percentile for Asian)	23	2	.654
Cesarean section (excludes prior C-section)	13	1	.782
Cephalopelvic disproportion	9	1	.941
PIH	5	1	.608
PPH	0	1	.087
PTL <37 weeks	8	2	.279
PTB <37 weeks	7	1	.851
Neonatal hypoglycemia	3	0	.537
RDS	5	0	.414
NICU admission	6	1	.734

A P value < .05 indicates statistical significance.

Discussion

Studies in the relationship between glycemic variability and pregnancy outcomes are lacking and still controversy. We assumed that glycemic variability in third trimester is a stable factor in non-insulin-dependent GDM patients since the slightly deteriorated glycemic excursion has no significant difference over this period from the study of Dalfra et al²² and this should be a better predictor of pregnancy outcomes than intermittent blood glucose values.

From our findings, glycemic excursion data derived from a single period of continuous glucose monitoring during weeks 28-32 of gestation is not significantly correlated with birth weight percentiles (r²≈0), and is not a predictor of neonatal outcomes in non-insulin-dependent GDM patients. A similar study by Taslimi et al²³ that investigated for association between maternal glucose excursion and fetal birth weight percentiles found no significant correlation in 20 GDM patients including insulin-dependent and non-insulin-dependent GDM patients at 27-28 weeks of gestation.²³ Our study analyzed data from 47 patients that had GDM in the third trimester, and we also found no significant correlation between glycemic excursion and fetal birth weight percentiles. Moreover, we found no correlation between high glycemic excursion and perinatal outcomes; preterm birth, RDS, NICU admission, neonatal hypoglycemia, PIH, PPH, LGA, rate of cesarean section. A pilot study in third trimester GDM by Porter et al⁴ found none of the following glycemic parameters to be predictive of fetal birth weight: average blood glucose level, preprandial blood glucose level, postprandial blood glucose level, or percentage of time blood glucose level exceeded 120 mg/dl.⁴ Differently, a recent study in glycemic excursion in pregnancy by Sung et al⁷ found slight correlation with fetal birth weight percentiles at AUCs above 110, 120, 130, and 140 mg/dl (r = .29); however, when the data were categorized into high and low groups by 90th percentile, there was no statistically significant difference in perinatal outcomes.⁷ In our study, we included parameters from CGMS that included %CV, MAGE, AUC of glucose above 120, and AUC of glucose above 130 mg/dl with cutoff values at both the 50th and 90th percentiles, and we found no significant difference in outcomes between the high and low excursion groups.

Our study in non-insulin-dependent GDM with well-controlled intermittent blood glucose level found a maximum glycemic value of 128 mg/dl and mean glycemic value of 107.7 mg/dl from CGMS monitoring data. All mean glycemic values in our study were lower than 130 mg/dl, which indicated good glycemic control. Mean 2-hour postprandial glucose was also below 130 mg/dl in our study population throughout pregnancy, and average gestational weight gain was 12 kg, which reflected optimal weight gain in most patients.

There was a larger study by Yu et al¹¹ about diabetes in pregnancy found MAGE was strongly associated with birth weight percentile at significant level of P < .001 but r² was only 0.133,¹¹ different from our study the study included all insulin-dependent DM and non-insulin-dependent DM while our study mainly focused on well-controlled non-insulin-dependent GDM.

Despite differences in glycemic excursion among patients, the mean SD of 10.2 ± 7.2 mg/dl proves that a very stable glucose regulation exists in these patients. As a result, no significant differences occur. This explains the result of no statistically differences in neonatal and maternal outcomes.

The incidence of adverse pregnancy outcomes in well-controlled GDM from our study maybe underpowered to detect the significant differences between high and low excursion groups.

Although our findings do not support the use of single period glycemic variability data from CGMS in the third trimester to predict fetal birth weight percentiles, gestational weight gain and maternal height were found to be significantly associated with birth weight percentiles and more predictive of fetal birth weight.

Our findings also suggest that the use of CGMS to evaluate glycemic variability in the third trimester in well-controlled non-insulin-dependent GDM patients may be unnecessary and ineffective for improving perinatal outcomes. It is also possible that glycemic factors in the first and/or second trimester(s) have more adverse impact on outcomes than factors in the third trimester. Most of our subjects were diagnosed as GDM in their late second trimester and, despite good glycemic control in the third trimester, the perinatal outcome may not change, because fetal growth may have been programmed earlier.^24,25

This study has some mentionable limitations. First, our study might contain bias as the patients could be strictly on diet control while being on CGMS. Dietary records of all patients had been reviewed and most of them were not restricted and intermittent glucose monitoring showed well-controlled blood sugar later during antenatal care, so the CGMS values derived from the period can represent the third trimester glycemic variability. Second, according to low incidence of adverse perinatal outcomes, this study could not find significant difference of perinatal outcomes between low and high excursion groups. Finally the study was only in third trimester.

Study of glycemic variability should be performed in the first and second trimesters to facilitate early diagnosis and improved maternal and fetal outcomes. A 2013 systematic review suggested that current evidence regarding the efficacy and effectiveness of CGMS for improving glycemic control during pregnancy is limited and that contradictory results have been reported.²⁶ Random controlled trials that study the effectiveness and cost-effectiveness of CGM in pregnancy are needed before wider use of CGM is implemented in daily clinical practice. Moreover, early assessment, diagnosis, and glycemic control in non-insulin-dependent GDM during the first and second trimesters may have more positive impact on maternal and fetal outcomes than assessment and treatment in the third trimester.

Conclusion

Based on these findings, third trimester glycemic variability data from CGMS is not a predictor of fetal birth weight percentile. CGMS is not necessary during the third trimester to improve perinatal outcomes in non-insulin-dependent GDM.