Characterizing glucose in labor: Continuous glucose monitoring in parturients without diabetes

Abstract

Introduction

Current recommendations for intrapartum glycemic control in those with diabetes mellitus are based upon expert opinion. Our objective was to characterize intrapartum interstitial glucose values using continuous glucose monitors (CGMs) in term gravidas without diabetes.

Methods

A prospective observational cohort of 72 parturients without pre‐gestational or gestational diabetes. CGMs were placed in early labor and the following were obtained: glucose at delivery, mean glucose in active and second stage of labor, time in range defined as 70–120 mg/dL, time above range, and time below range. Glucose parameters were compared by body mass index (BMI) and by adequacy of gestational weight gain. Neonatal outcomes were assessed.

Results

The mean glucose value at delivery was 109 ± 27 mg/dL and over the 2 h prior to delivery was 99 ± 23 mg/dL. A total of 26% of parturients had a glucose value > 120 mg/dL at the time of delivery. Mean glucose in active labor was 101 ± 25 mg/dL and in the second stage was 99.5 ± 23 mg/dL. In labor, average time‐in‐range was 69.8% (10th–90th percentile 33.8%–95.8%), time‐above‐range was 19.9% (10th–90th percentile 0%–66.2%), and time‐below‐range was 10.3% (10th–90th percentile 0%–66%). Neither BMI nor gestational weight gain had a significant clinical impact on these parameters. A total of 28 infants had glucose checked for clinical indications and 5 required treatment for neonatal hypoglycemia; this was unrelated to mean glucose during labor or at delivery, time‐above‐range in labor and second stage, and time‐above‐range in the 4 h prior to delivery.

Conclusion

Interstitial glucose greater than 120 mg/dL was common, and neither BMI nor gestational weight gain appears to have a significant clinical impact on glucose in labor. These data challenge contemporary definitions of glucose goals in labor, and further study is needed to better define glucose in labor and identify risk factors for adverse outcomes.

1. INTRODUCTION

During pregnancy, blood glucose is tightly controlled in order to prevent adverse maternal, fetal, and neonatal outcomes in gravidas with gestational or pre‐gestational diabetes. During labor, strict glucose control is recommended primarily to avoid neonatal hypoglycemia, which is associated with neonatal intensive care admission, interrupted breastfeeding, and neurodevelopmental impairment in children [1, 2, 3]. The target glucose for parturients with diabetes is not based upon normalized values, as these do not currently exist. As a result, glucose targets during labor are variable and institution dependent. Most commonly, the recommended target glucose range is 70–110 mg/dL with some institutions liberalizing the upper limit to as high as 140 mg/dL [2, 4]. Most recently, a prospective randomized clinical trial involving liberalized control up to 180 mg/dL demonstrated no differences in neonatal outcomes compared to tight control for gravidas with diabetes in labor [5]. To our knowledge, there have been no large studies characterizing physiologic glucose values in women without diabetes in labor, and risk factors for adverse outcomes related to glucose are not well studied in this population.

Continuous glucose monitors (CGMs) measure interstitial glucose every 1–5 min and have been demonstrated to correlate well with plasma glucose [5, 6, 7, 8]. CGM glucose values may lag behind serum glucose by 9–12 min [8, 9, 10, 11]. Although primarily utilized for the clinical management of type 1 diabetes, CGMs are being utilized in other clinical and research settings.

The primary aim of the study was to characterize CGM glucose values during labor for nondiabetic parturients. Secondary aims included comparison of glucose in labor by pre‐gestational body mass index (BMI) categories and by adequacy of gestational weight gain and comparison of mean glucose values in distinct time periods: total labor duration, active labor (≥6 cm dilation), and second stage. We hypothesized that glucose would be higher among individuals with obesity (BMI > 30) and those with excessive gestational weight gain and that we would observe a transiently increased mean glucose value during the second stage of labor compared to labor overall. This is based upon data demonstrating a transient hyperglycemia during the onset of intense exercise, similar to the physiologic stress of the second stage of labor [12, 13].

2. MATERIALS AND METHODS

This prospective cohort study included term parturients who had neither gestational nor pre‐gestational diabetes admitted to labor and delivery in early labor between July 2022 and June 2023. All enrollments occurred in one academic center and was approved by the Institutional Review Board (IRB).

Parturients were eligible for inclusion if they were age 18 or older, spoke English or Spanish, had a singleton pregnancy between 37 weeks 0 days and 42 weeks gestational age, were admitted for induction of labor or early labor, and had a negative 1 h 50 g glucose challenge test (GCT) with a cutoff of 135 mg/dL or, if they had one‐step gestational diabetes (GDM) screening, a completely negative 2 h 75 g glucose tolerance test (GTT) between 24 and 28 weeks gestational age.

Exclusion criteria included: planned cesarean delivery, pre‐gestational diabetes, abnormal or inadequate screening for GDM, major fetal anomaly, or fetal demise. Parturients receiving therapeutic magnesium for preeclampsia were excluded due to strict diet and fluid restrictions in this group.

Parturients were screened for eligibility and approached for inclusion while they were in early labor or upon admission for induction of labor. Informed consent was obtained prior to commencement of any study procedures. Dexcom G6 CGM were placed on the posterior upper arm per manufacturer's instructions [14]. The Dexcom G6 CGM has a 2‐h warm‐up period during which no glucose values were obtained and glucose values were subsequently obtained every 5 min until device removal. All glucose values recorded by the CGM were blinded to both the parturient and care providers throughout their admission, and all alarms on the device were disabled. Previously published data have demonstrated the concordance of Dexcom G6 CGM and point‐of‐care glucose testing in both pregnancy and intense physiologic stress [6, 7], and thus point‐of‐care glucose was checked with a glucometer only if indicated per standard institutional indications. The Dexcom G6 CGM does not require calibration after placement [14]. The CGM was removed at parturient request, the day of discharge, or at 10 days, whichever occurred first.

All labor management, including intravenous fluids and diet, were managed according to the discretion of the primary obstetric provider. At our institution, parturients are typically restricted to a clear liquid diet if they have an epidural or are in active labor. There are no carbohydrate restrictions for those without diabetes and the default intravenous fluid of choice is lactated ringers. Medications, including penicillin or other antibiotics, terbutaline, and oxytocin were given according to the treating obstetrician's discretion. Penicillin was the only medication given during labor that was administered in a dextrose solution.

For the purposes of data collection and analysis, “labor” was defined as the time of placement of the CGM until the delivery of the neonate. “Early labor” was defined as the time spent at <6 cm dilation. “Active labor” was defined as the time from the first active labor exam (≥6 cm dilation) until complete dilation (10 cm). The “second stage” was defined as the time from complete dilation until the time of delivery.

Our primary outcomes were the mean glucose value at delivery and during each labor stage and the time‐in‐range during labor. Time‐in‐range was defined as glucose 70–120 mg/dL based on institutional policy for glycemic control targets for pregnancies complicated by diabetes. Secondary outcomes included time‐above‐range (glucose > 120 mg/dL), time‐below‐range (glucose < 70 mg/dL), and time‐in‐range during the different stages of labor and incidence of neonatal hypoglycemia. Neonatal hypoglycemia was defined as neonatal glucose < 47 mg/dL within 24 h of birth [15].

As participants entered the study at different time points and had differing labor lengths, additional analyses were performed examining glucose parameters in the 2 and 4 h prior to delivery in an effort to better characterize the time period approaching delivery for all participants. These glycemic control parameters were analyzed descriptively for the entire cohort and were compared between groups of parturients based upon pre‐pregnancy BMI and adequacy of gestational weight gain.

CGM data, including the calculated time‐in‐range, time‐above‐range, and time‐below‐range were analyzed in the defined labor stages as well as by parturients’ pre‐pregnancy BMI categories and adequacy of gestational weight gain. Gestational weight gain groups (inadequate, adequate, and excessive) were categorized based upon the Institute of Medicine guidelines based on pre‐pregnancy BMI [16, 17]. BMI was categorized based upon common cutoffs for normal weight, overweight, and obese individuals: 18–24.9, 25–29.9, and > 30 kg/m².

The sample size was determined based upon an estimated difference in mean glucose values of 10 mg/dL (approximately 10% difference with a predicted average glucose of 100 mg/dL) between BMI groups. Assuming a two‐tailed alpha error <0.05 and power of 80%, we calculated that we would need 16 parturients in each BMI category. To account for drop out due to parturient preference, device malfunction, unanticipated cesarean early in labor, etc., we aimed to enroll 20 parturients in each BMI category. Enrollment was complete once there were at least 20 eligible parturients enrolled in each BMI category.

Python 3.10.4 was used for analysis of CGM data and SPSS was used for chi square, ANOVA, and univariate analysis [18, 19].

3. RESULTS

A total of 72 parturients were included in the analysis out of 78 total recruited. Of those who were excluded from analysis, one parturient's CGM malfunctioned and did not collect any data, two CGMs were accidentally discarded prior to data transfer, two parturients rapidly labored and delivered during the CGM warmup period, and one parturient did not meet the inclusion criteria. Baseline characteristics and demographics are presented in Table 1. The study group represented a diverse cohort with a median parturient age of 31. The mean parturient glucose value on the standard 1 h GCT was 102 mg/dL (SD 19).

TABLE 1.

Demographic and delivery characteristics, N = 72.

Nulliparous	45 (63.5%)
Publicly insured	25 (35%)
Mean parturient age	31 ± 4.8 years
Gestational age (GA) at delivery (range)	37w2d–42w2d
Vaginal delivery	51 (71%)
1 h glucose challenge test	102 ± 19 mg/dL
BMI at intake < 25	25 (34.7%)
BMI at intake 25–29.9	27 (37.5%)
BMI at intake > 30	20 (27.8%)
Inadequate gestational weight gain	15 (20.8%)
Adequate gestational weight gain	22 (30.6%)
Excessive gestational weight gain	35 (48.6%)
Epidural during labor	53 (74%)
Mean neonatal birthweight	3457 g ± 400 g
Neonates with glucose checked for clinical indication	23/72 (32%)
Neonates treated for hypoglycemia	6/23 (26%)

Note: Categorical variables are reported as N (%) and continuous variables are reported as the mean ± standard deviation.

Abbreviation: BMI: body mass index.

If a patient had an intrapartum cesarean section, they were not included in the analysis for labor stages they did not reach. For example, participants who had a cesarean in early labor were excluded from analysis of active labor and the second stage. Overall, 51/72 (71%) delivered vaginally and had second‐stage data available for analysis. The average duration of CGM use for each parturient was 17.5 h or approximately 210 glucose measurements (SD 161.9).

The mean glucose value at the time of delivery (reported as the value closest to time of delivery) was 109 mg/dL (SD 27 mg/dL, 10th–90th percentile 79.9–143 mg/dL). Mean glucose in active labor was 101 mg/dL (SD 25 mg/dL) and in the second stage was 99.5 mg/dL (SD 23 mg/dL) (Table 2). Time‐in‐range from CGM placement until delivery was on average 69.8% (SD 23.6%, 10th–90th percentile 33.8%–95.8%) and in the 2 h prior to delivery was 68.7% (SD 34%, 10th–90th percentile 0.8%–100%) (Figure 1). Time‐above‐range from CGM placement until delivery was on average 19.9% (SD 25.7%, 10th–90th percentile 0%–66.2%) and was 22.8% (SD 34.6%, 10th–90th percentile 0%–99%) in the 2 h prior to delivery. Greater than 25% of the cohort had a blood glucose value > 120 mg/dL at the time of delivery.

TABLE 2.

Glucose parameters in the overall cohort.

Mean glucose at delivery	109 ± 27 mg/dL
Mean glucose over the 2 h prior to delivery	99 ± 23 mg/dL
Mean glucose in active labor	101 ± 25 mg/dL
Mean glucose in the 2nd stage	99.5 ± 23 mg/dL
Mean time‐in‐range from CGM placement to delivery	69.8% (10th–90th percentile, 33.8%–95.8%)
Mean time‐in‐range over the 2 h prior to delivery	68.7% (10th–90th percentile, 0.8%–100%)
Mean time‐above‐range from CGM placement to delivery	19.9% (10th–90th percentile, 0%–66.2%)
Mean time‐above‐range in the 2 h prior to delivery	22.8% (10th–90th percentile, 0%–99%)
Parturients with glucose > 120 mg/dL at the time of delivery	26.0%
Mean time‐below‐range from CGM placement to delivery	10.3% (10th–90th percentile, 0%–66%)
Mean time‐below‐range in the 2 h prior to delivery	8.5% (10th–90th percentile, 0%–36%)

Note: Continuous variables are reported as the mean ± standard deviation. Percentages are reported as the mean and 10th–90th percentile range. Time‐in‐range = glucose 70–120 mg/dL, Time‐above‐range = glucose > 120 mg/dL, Time‐below‐range = glucose < 70 mg/dL.

Abbreviation: CGM, continuous glucose monitor.

FIGURE 1.

Mean glucose over the 2 h prior to delivery for all participants. Mean glucose values by participant over the 2 h prior to delivery ± SD. Participants have been sorted by mean glucose in ascending order. Highlighted area reflects the glucose target range of 70–120 mg/dL.

There were no differences in the glycemic parameters by parturient BMI (Table 3, Figure 2). There were small statistically significant differences in the mean glucose in active labor and the mean glucose in the second stage by gestational weight gain group, with the highest mean glucose in both time periods in the excessive gestational weight gain group; however, these differences are not likely to be clinically significant (Table 3, Figure 3).

TABLE 3.

Glucose parameters by BMI and gestational weight gain (GWG) groups.

	BMI 18–24.9 kg/m2 (N = 25)	BMI 25–29.9 kg/m2 (N = 27)	BMI > 30 kg/m2 (N = 20)	p value	Inadequate GWG (N = 15)	Adequate GWG (N = 23)	Excessive GWG (N = 35)	p value
Mean glucose at delivery (mg/dL)	110	113	103	0.47	107	102	115	0.18
Mean glucose in labor (mg/dL)	98	103	99	0.56	101	95	103	0.32
Mean glucose in active labor (mg/dL)	99	104	100	0.76	100	89	108	<0.04 *
Mean glucose 2 h prior to delivery (mg/dL)	104	106	99	0.54	103	99	106	0.49
Mean glucose in 2nd stage (mg/dL)	99	103	94	0.57	95	91	108	<0.05 *
Overall TIR	68 ± 24%	66 ± 25%	76 ± 19%	0.34	72 ± 21%	73 ± 19%	67 ± 27%	0.65
Overall TAR	20 ± 26%	24 ± 29%	14 ± 21%	0.43	18 ± 24%	14 ± 17%	24 ± 30%	0.32
Overall TBR	12 ± 15%	10 ± 14%	10 ± 12%	0.84	10 ± 11%	13 ± 16%	8 ± 13%	0.44

Note: Intake BMI and gestational weight gain (GWG)’s impact on peripartum glucose. Inadequate, adequate, and excessive GWGs were determined as per Institute of Medicine guidelines based on pre‐pregnancy BMI. Time‐in‐range (TIR) = glucose 70–120 mg/dL, time‐above‐range (TAR) = glucose > 120 mg/dL, and time‐below‐range (TBR) = glucose < 70 mg/dL.

Abbreviation: BMI: body mass index.

^*indicates statistically significant difference between GWG groups.

FIGURE 2.

Intake BMI influence on CGM parameters in labor. Intake BMI's impact on time in range, above range, and below range in labor. Normal BMI = 18–24.9 kg/m^2, overweight BMI = 25–29.9 kg/m², obese BMI ≥ 30 kg/m². BMI, body mass index; CGM, continuous glucose monitor; TAR, time‐above‐range; TBR, time below‐range; TIR, time‐in‐range.

FIGURE 3.

Gestational weight gain (GWG) influence on CGM parameters in labor. GWG's impact on time in range, above range, and below range in labor. Inadequate, adequate, and excessive GWG were determined as per Institute of Medicine (IOM) guidelines based on pre‐pregnancy BMI. BMI, body mass index; CGM, continuous glucose monitor; TAR, time‐above‐range; TBR, time below‐range; TIR, time‐in‐range.

Of the 28 infants who had their glucose checked for clinical indications (large or small‐for‐gestational age, risk of sepsis, symptoms, etc.), 5 (18%) required treatment for neonatal hypoglycemia. In this group of 28 infants, neonatal hypoglycemia was not associated with the parturient glucose value at delivery (p = 0.41), mean glucose in labor (p = 0.31) or second stage (p = 0.83), time‐above‐range in labor (p = 0.50) or second stage (p = 0.29), or the mean glucose in the 4 h prior to delivery (p = 0.55) although the small number limited statistical power to identify differences.

4. DISCUSSION

In this cohort of parturients without diabetes, the average glucose in labor was 109 mg/dL, but glucose higher than 120 mg/dL, a common threshold for treatment of “hyperglycemia” in parturients with diabetes in labor, was common. On average, parturients spent 20% of labor in time‐above‐range and more than 25% of gravidas exceeded glucose of 120 mg/dL at delivery. Intrapartum glucose did not clinically differ by stage of labor, maternal BMI, or gestational weight gain. Neonatal hypoglycemia, identified based on clinical factors alone, did not appear to be related to intrapartum glucose control; however, not all neonates were tested, and the overall number of neonates with hypoglycemia was low.

In the nonpregnant, nondiabetic adult population, hyperglycemia is best studied in the inpatient hospitalization and critical care setting. In these populations, hyperglycemia has been associated with increased mortality, increased risk of infection, and prolonged hospital stays [20]. Comparing these populations to the laboring pregnant patient has several downfalls. First, hyperglycemia in these populations is typically defined by blood glucose > 180 mg/dL, a target much higher than typical cutoffs used to define hyperglycemia in the pregnant population. Second, most patients studied in these populations have other significant physiologic aberrations, including critical illness, that are known to affect glycemic control and would not apply to most laboring individuals. Finally, these patients are not pregnant, which makes it impossible to extrapolate the effect of hyperglycemia in these populations to the potential effects on a fetus or neonate.

As noted, hyperglycemia is associated with adverse outcomes in the general population as well as the pregnant population with diabetes; however, normoglycemia and hyperglycemia during pregnancy are not well defined. In the context of gestational and pre‐gestational diabetes, intrapartum glucose is strictly monitored and controlled with the primary goal of preventing adverse neonatal outcomes including neonatal hypoglycemia. The current recommendations regarding glucose management are based primarily on expert opinion, and the normal physiology of intrapartum glucose is poorly understood. Our study begins to fill this knowledge gap.

Parturients in this cohort spent on average 20% of labor in time‐above‐range and more than 25% of gravidas exceeded glucose of 120 mg/dL at delivery. Based upon the glucose value at the time of delivery, greater than 25% of the cohort would have received insulin at our institution.

This may be physiologic and isn't entirely surprising, as transient hyperglycemia is commonly seen during periods of intense exercise as well as during periods of intense physiologic stress, both of which share similar physiology to certain aspects of labor [12, 13].

While there were statistically significant differences in intrapartum glucose by gestational weight gain groups, the absolute differences were small and not likely to be clinically significant. Intrapartum glucose did not statistically differ by BMI group. This suggests that neither pre‐gestational BMI nor adequacy of gestational weight gain leads to a significant clinical difference in blood sugar during labor and additional potential predictors of abnormal intrapartum glucose should be explored.

Although this study starts to address a critical knowledge gap in the literature by beginning to characterize physiologic intrapartum glucose utilizing CGMs, there are several limitations to this study. First, diet, behavior, and labor practices are highly influenced by cultural background and while this study represents a diverse cohort, it may not reflect the general population in other regions of the United States. Furthermore, while parturients were managed according to our standard institutional protocols regarding diet and labor management, these protocols will differ from institution to institution and may not represent a broader population. Diet during labor was not strictly monitored, and thus, the relationship between glucose consumption during labor and intrapartum glucose could not be analyzed. While the population did not have strict diet or fluid monitoring, this allows the cohort to reflect a “real life” population of parturients rather than data derived from a highly structured setting or with dietary restrictions.

A high percentage (>70%) of this cohort had an epidural for pain management in labor. Epidurals are known to alter systemic catecholamine release and increases in catecholamine signaling have been shown to alter glucose metabolism and lead to increases in serum glucose [21, 22]. Blocking catecholamine release with an epidural may lower the serum glucose values in these parturients, which may, in turn, underestimate the true prevalence of hyperglycemia for unmedicated gravidas in labor.

Finally, there is potential for the Hawthorne effect as study participants may have, consciously or subconsciously, altered their diet or other behaviors during labor knowing that they are participating in a study regarding glycemic control.

In this cohort, few neonates developed neonatal hypoglycemia, however, only neonates with clinical factors were tested. As not all neonates had their glucose checked after birth, the relationship between intrapartum glucose and asymptomatic neonatal hypoglycemia could not be assessed. As symptomatic neonatal hypoglycemia in the nondiabetic population is a rare event, larger cohorts will be needed to assess whether intrapartum glucose alters risk for neonatal hypoglycemia in the nondiabetic population.

Despite these limitations, the study has several strengths. First, there is a significant knowledge gap regarding physiologic glucose values in labor in those without gestational or pre‐gestational diabetes. While we believe this study begins to fill this gap, this cohort still represents only a small sample of laboring people. The use of CGM allowed for the collection of large amounts of data and allowed us to characterize these glucose patterns in several different ways. Consensus goals already exist for percentage of time‐in‐range, time‐above‐range, and time‐below‐range for pregnant individuals in the antepartum setting treated for type 1 or type 2 diabetes [23]. Data collected during labor may allow clinicians in the future to define target goals for the percentages of time‐in‐range or time‐above‐range during this phase. Once normoglycemia in labor is better defined, further study may be conducted to understand the effects of hyperglycemia in the nondiabetic pregnant population, as these effects are currently understudied and poorly understood.

In summary, our study demonstrated that about 20% of parturients without diabetes had glucose in labor above what is recommended for management of parturients with diabetes, although this did not appear to be related to neonatal hypoglycemia identified based on clinical factors. The clinical significance of this finding is not yet known; however, a better understanding of physiologic glucose patterns in parturients without diabetes may help define evidence‐based guidelines for diabetes care in the future. A larger cohort will be needed to further characterize and define normoglycemia in labor and to determine if these data are representative of the general population outside of our institution. Larger studies will also allow for sufficient power to study peripartum outcomes, including abnormal labor and other maternal or neonatal morbidities, which may be related to intrapartum glycemic control.

5. CONCLUSION

Interstitial glucose above 120 mg/dL was common, and neither BMI nor gestational weight gain appears to have a significant clinical impact on glucose in labor. These data challenge contemporary definitions of physiologic glucose levels in labor, and further study is needed to better define goals for glucose in labor and identify risk factors for adverse outcomes.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

FUNDING INFORMATION

The authors received no specific funding for this work.

DISCLOSURE

The continuous glucose monitors (CGM), sensors, and transmitters were provided by DexCom, Inc. DexCom had no role in study design, data collection, data analysis, writing of the report, or the decision to submit the report for publication.

DATA AVAILABILITY STATEMENT

Research data are not shared.