Abstract
Objective
Optimal strategies for intrapartum glycemic control among patients with gestational diabetes mellitus (GDM) and type 2 diabetes mellitus (T2DM) remain uncertain. In 2024, our labor and delivery unit implemented a subcutaneous (SC) insulin protocol to replace routine intravenous (IV) insulin initiation when glucose exceeded a treatment threshold. We hypothesized that this change would not worsen neonatal outcomes and would improve nursing (RN) satisfaction.
Methods
We conducted a retrospective cohort quality improvement study of patients with GDM or T2DM who delivered at ≥36 weeks between March 2021 and December 2025. July 2024, when the protocol was introduced, was excluded as a washout period. The intervention included both a shift to SC insulin‐first management and standardization of intrapartum glycemic treatment to a target of ≤ 120 mg/dL. The primary outcome was a composite of neonatal hypoglycemia or neonatal intensive care unit (NICU) admission for hypoglycemia within 24 h of life. Secondary outcomes included IV insulin duration and RN satisfaction. Outcomes were compared before and after implementation. Poisson regression with an interrupted time series framework assessed pre‐implementation trends, immediate level changes, and post‐implementation trends in the primary composite outcome. Multivariable logistic regression evaluated associations between insulin modality and the primary outcome. Surveys were analyzed by Wilcoxon rank‐sum tests.
Results
Of 1140 deliveries in the analytic cohort, the primary outcome occurred in 269 (23.6%). Rates were similar before and after protocol implementation (23.4% vs. 23.9%, p = 0.837). Interrupted time series analysis showed no significant pre‐implementation trend, but did demonstrate a significant immediate reduction in monthly event rates at implementation (incidence rate ratio [IRR], 0.20; 95% confidence interval [CI], 0.04–0.90), with no significant post‐implementation trend change (IRR, 1.03; 95% CI, 0.99–1.06). In adjusted analyses, insulin modality was not significantly associated with the primary outcome. RN satisfaction improved after implementation, with agreement or strong agreement increasing from 29.2% to 51.8% (p = 0.032).
Conclusion
Implementation of an SC insulin‐first intrapartum diabetes management protocol for patients with GDM and T2DM was not associated with increased neonatal hypoglycemia nor NICU admission for hypoglycemia. The protocol was associated with improved RN satisfaction. This approach may offer a pragmatic alternative to routine IV insulin initiation during labor.
Keywords: gestational diabetes mellitus, intrapartum glycemic control, labor and delivery management, neonatal hypoglycemia, quality improvement, subcutaneous insulin, type 2 diabetes in pregnancy
1. INTRODUCTION
Diabetes complicates an increasing proportion of pregnancies in the United States, including both pregestational type 2 diabetes mellitus (T2DM) and gestational diabetes mellitus (GDM). These conditions are associated with multiple perinatal complications, including neonatal hypoglycemia, which remains one of the most common metabolic abnormalities among newborns, and is particularly prevalent in infants of patients with diabetes [1, 2]. Neonatal hypoglycemia can lead to increased monitoring, separation of the birthing parent and infant, neonatal intensive care unit (NICU) admission, and, in severe cases, neurologic injury [1, 3]. Because fetal hyperinsulinemia is driven in part by maternal hyperglycemia during labor and delivery (L&D), intrapartum glucose management has long been considered an important strategy to reduce neonatal hypoglycemia and related morbidity [2, 3, 4, 5].
Despite the clinical importance of intrapartum glycemic control, optimal management strategies remain uncertain. Many institutions rely on protocols that initiate intravenous (IV) insulin infusions once maternal glucose levels exceed a predefined threshold [6, 7]. These approaches aim to maintain tight glycemic targets during labor, but require frequent monitoring, IV access, and continuous nursing management. Although IV insulin protocols are widely used and are recommended in many practice guidelines, the evidence supporting their superiority over other approaches is limited [2, 8].
Subcutaneous (SC) insulin administration represents a potential alternative strategy to intrapartum glucose management. Continuation of SC insulin, including insulin pump therapy, has been studied most extensively among individuals with type 1 diabetes [9, 10]. However, evidence guiding intrapartum insulin management remains limited for patients with gestational or type 2 diabetes, who represent the majority of pregnancies complicated by diabetes [5, 11]. Institutional protocols frequently default to IV insulin despite limited comparative evidence with SC insulin and the increased workflow burden associated with infusion‐based management [10].
In July 2024, our institution implemented a quality improvement (QI) project with the release of a protocol initiating SC insulin rather than automatically starting IV insulin when intrapartum glucose levels exceeded a treatment threshold for patients with T2DM or GDM. This change was designed to provide a simpler and more pragmatic approach to intrapartum glucose management while maintaining safe neonatal outcomes. Our study evaluated the use of an SC insulin protocol for intrapartum hyperglycemia. We hypothesized that this protocol change would not increase neonatal hypoglycemia or related NICU admissions and would improve nursing (RN) satisfaction with intrapartum diabetes management.
2. MATERIALS AND METHODS
We conducted a retrospective cohort study evaluating a QI intervention designed to standardize intrapartum glucose management for patients with T2DM and GDM. The study was performed at a quaternary, university‐affiliated hospital with an L&D unit with approximately 2500 births per year. L&D and postpartum care were provided by teams of medical students, nurses, residents, fellows, certified nurse‐midwives, and attending physicians. The protocol change was developed and implemented with support from both physician and perinatal nursing leadership. The sample size was determined by all eligible deliveries during the available electronic medical record period and the post‐implementation plan‐do‐study‐act evaluation window; no a priori sample size calculation was performed. This study was conducted under an IRB‐approved department L&D QI umbrella protocol and followed SQUIRE 2.0 guidelines.
This retrospective cohort quality improvement study included patients with GDM or T2DM who delivered at ≥36 weeks of gestation or later between March 2021 and December 2025. Patients with type 1 diabetes mellitus or pancreaticogenic diabetes (including causes such as cystic fibrosis) were excluded. Implementation was preceded by L&D RN education in May 2024 and resident/provider orientation around the start of the academic year; July 2024 was excluded as a washout period to allow protocol adoption. The protocol was developed in response to emerging evidence supporting SC insulin strategies and substantial variability in intrapartum diabetes management within our institution, where no standardized approach had previously existed. Decisions regarding when to initiate insulin therapy, whether to use SC or IV insulin, and how frequently to monitor glucose were often individualized by provider. Multiple order sets and practice patterns were used across clinicians, contributing to variation in care and uncertainty among providers and nursing staff responsible for implementing intrapartum glucose management. This variability reflected the absence of clear, evidence‐based standards for managing intrapartum hyperglycemia in patients with T2DM or GDM. Before implementation, usual practice generally used a less stringent treatment threshold, commonly targeting a blood glucose ≤140 mg/dL.
The updated intrapartum diabetes protocol standardized glucose monitoring and insulin treatment (Appendix S1). The new protocol included a lower intrapartum glycemic target of ≤120 mg/dL and a change in the initial recommended insulin modality from routine IV insulin to SC correctional insulin. On admission, recommendations included discontinuation of oral hypoglycemic agents and continuation of basal insulin at 50% of the home dose, with prandial insulin continued in latent labor. Patients having a scheduled cesarean delivery were counseled to arrive fasting with clear liquid allowed until 2 h prior to the time of their scheduled surgery. All other patients were given a regular diet prior to delivery, with laboring patients transitioned to a clear liquid with neuraxial anesthesia placement per hospital protocol. Blood glucose was checked every 4 h in latent labor, every 2 h in active labor, and every 1 h during the second stage, with every 1 h checks for 2 h after any correctional insulin was given. The protocol recommended correctional rapid‐acting insulin for glucose values of 121–199 mg/dL, with initiation of an obstetric intravenous insulin infusion for persistent or severe hyperglycemia >199 mg/dL. Correctional insulin scales were adjusted based on stage of labor and nutritional status, except that established antepartum correction factors were preferentially used.
The primary outcome was a composite of neonatal hypoglycemia or NICU admission for hypoglycemia within 24 h of life. The institutional newborn glucose screening protocol for infants born to patients with diabetes defines neonatal hypoglycemia as a glucose concentration <40 mg/dL within the first 4 h of life, or <45 mg/dL between 4 and 24 h of life (Appendix S2). Secondary outcomes included IV insulin duration, maternal hypo‐ and hyperglycemia, and L&D RN satisfaction with the intrapartum diabetes management protocol. Maternal hypoglycemia was defined as BG <70 mg/dL, and maternal hyperglycemia was defined as BG >120 mg/dL.
L&D RNs were surveyed before protocol implementation and again approximately 10 months after implementation. Recruitment was performed using flyers with QR codes posted on the perinatal unit and via email. Survey responses were collected anonymously using Microsoft Office Forms. Participants were required to log in to ensure that each individual could respond only once, but identifying information was not retained in the analytic dataset. The survey included three Likert‐scale items and responses were analyzed in Stata Statistical Software: Release 18 (StataCorp LLC, 2023).
Maternal demographic characteristics, intrapartum glucose values, insulin use, and neonatal outcomes were abstracted from the electronic medical record. Patients with missing intrapartum glucose values were excluded from secondary maternal glycemic analyses; primary outcome and insulin exposure data were hand‐abstracted and re‐reviewed when discrepancies were identified. All NICU admissions were reviewed by a board‐certified obstetrician‐gynecologist to confirm the reason for admission. Neonatal glucose testing was performed according to institutional protocol for infants born to patients with diabetes, with the first glucose assessment obtained within 30 min after the first feeding or within 2 h of birth if feeding could not be initiated (Appendix S2). Study data outside of the RN survey were collected and managed using REDCap electronic data capture tools hosted at the University of Washington (grant UL1 TR002319) [12, 13].
Patient demographics, maternal outcomes, and neonatal outcomes were compared between the pre‐ and post‐implementation periods. Poisson regression with an interrupted time series framework was applied to monthly outcome rates to assess pre‐implementation trends, immediate changes at the time of protocol implementation, and post‐implementation trends in the primary outcome. Multivariable logistic regression was used to evaluate associations between insulin modality and the primary outcome. Insulin modality was classified as any SC insulin use and any IV insulin use; some patients received both insulin modalities. Four models were evaluated: model 1, unadjusted; model 2, adjusted for maternal age at delivery and insurance type; model 3, additionally adjusted for gestational age; and model 4, additionally adjusted for diabetes type, plurality, and mode of delivery. A sensitivity analysis was performed excluding patients with diet‐controlled GDM (A1GDM) given the lower likelihood of needing insulin treatment during these patients’ hospitalizations. Survey responses were compared between the pre‐ and post‐implementation periods using Wilcoxon rank‐sum tests. Statistical analyses were performed using Stata Statistical Software: Release 18 (StataCorp LLC, 2023), and a two‐sided p value < 0.05 was considered statistically significant.
3. RESULTS
A total of 1149 deliveries met inclusion criteria during the study period, including 718 (62.5%) in the pre‐implementation period, 422 (36.7%) in the post‐implementation period, and nine (0.8%) during the July 2024 washout period. The analytic cohort therefore included 1140 deliveries (Table 1). Within the analytic cohort, median maternal age at delivery was similar before and after implementation (34.5 vs. 34.5 years). The cohort was racially and ethnically diverse, and most patients had either private insurance or Medicaid. Overall, 477 (41.8%) had diet‐controlled GDM (A1GDM), 447 (39.2%) had medication‐controlled GDM (A2GDM), and 216 (18.9%) had T2DM. Baseline characteristics were generally similar between periods, although private insurance was more common in the post‐implementation period (61.0% pre vs. 69.9% post), and Medicaid was less common (36.5% pre vs. 27.7% post).
TABLE 1.
Patient demographics.
| Pre | Post | Washout | Total | |
|---|---|---|---|---|
| N | 718 (62.5%) | 422 (36.7%) | 9 (0.8%) | 1149 |
| Age at delivery | 34.5 (5.0) | 34.5 (5.1) | 35.3 (4.9) | 34.4 (5.1) |
| Self‐reported race | ||||
| American Indian or Alaska Native | 12 (1.7%) | 5 (1.2%) | 0 (0.0%) | 17 (1.5%) |
| Asian | 193 (26.9%) | 138 (32.7%) | 2 (22.2%) | 333 (29.9%) |
| Black or African American | 82 (11.4%) | 50 (11.9%) | 4 (44.4%) | 136 (11.8%) |
| Native Hawaiian or Other Pacific Islander | 10 (1.4%) | 10 (2.4%) | 0 (0.0%) | 20 (1.7%) |
| White | 349 (48.6%) | 165 (39.1%) | 3 (33.3%) | 517 (45.0%) |
| Other * | 38 (5.3%) | 43 (10.2%) | 0 (0.0%) | 81 (7.1%) |
| Declined/Unknown | 34 (4.7%) | 11 (2.6%) | 0 (0.0%) | 45 (3.9%) |
| Ethnicity | ||||
| Hispanic | 118 (16.4%) | 67 (15.9%) | 1 (11.1%) | 186 (16.2%) |
| Non‐Hispanic | 582 (81.1%) | 340 (80.6%) | 8 (88.9%) | 930 (80.9%) |
| Unknown | 18 (2.5) | 15 (3.6) | 0 (0.0%) | 33 (2.9%) |
| Insurance type | ||||
| Medicaid | 262 (36.5%) | 117 (27.7%) | 3 (33.3%) | 382 (33.3%) |
| Private insurance | 438 (61.0%) | 295 (69.9%) | 6 (66.7%) | 739 (64.3%) |
| Tricare | 12 (1.7%) | 7 (1.7%) | 0 (0.0%) | 19 (1.7%) |
| None | 0 (0.0%) | 1 (0.2%) | 0 (0.0%) | 1 (0.1%) |
| Medicare | 6 (0.8%) | 2 (0.5%) | 0 (0.0%) | 8 (0.7%) |
| Diabetes type | ||||
| Diet‐controlled gestational diabetes (A1GDM) | 289 (40.3%) | 188 (44.6%) | 4 (44.4%) | 481 (41.9%) |
| Medication‐controlled gestational diabetes (A2GDM) | 274 (38.2%) | 173 (41.0%) | 4 (44.4%) | 451 (39.3%) |
| Type 2 diabetes | 155 (21.6%) | 61 (14.5%) | 1 (11.1%) | 217 (18.9%) |
| Plurality | ||||
| Singleton | 703 (97.9%) | 419 (99.3%) | 9 (100.0%) | 1131 (98.4%) |
| Twins | 15 (2.1%) | 3 (0.7%) | 0 (0.0%) | 18 (1.6%) |
| Gestational age (weeks) | 38.08 (1.18) | 38.25 (1.23) | 37.78 (1.09) | 38.14 (1.20) |
| BMI | 33.7 [28.8, 39.7] | 32.8 [28.6, 38.8] | 30.8 [26.7, 38.5] | 33.4 [28.7, 39.3] |
| Type of insulin administered | ||||
| Neither | 556 (77.4%) | 288 (68.3%) | 7 (77.8%) | 851 (74.1%) |
| Subcutaneous only | 95 (13.2%) | 102 (24.2%) | 0 (0.0%) | 197 (17.2%) |
| IV only | 25 (3.5%) | 2 (0.5%) | 0 (0.0%) | 27 (2.4%) |
| Both | 42 (5.8%) | 30 (7.1%) | 2 (22.2%) | 74 (6.4%) |
| Mode of delivery | ||||
| Vaginal delivery | 349 (48.6%) | 205 (48.6%) | 3 (33.3%) | 557 (48.5%) |
| Cesarean delivery | 369 (51.4%) | 217 (51.4%) | 6 (66.7%) | 592 (51.5%) |
Abbreviations: BMI, body mass index; IV, intravenous.
This option was self‐reported by patients as noted in the medical record.
The primary composite outcome of neonatal hypoglycemia or NICU admission for hypoglycemia within 24 h occurred in 269 of 1140 (23.6%) births. Rates were similar before and after protocol implementation, occurring in 168 of 718 (23.4%) births in the pre‐implementation period and 101 of 422 (23.9%) of births in the post‐implementation period (p = 0.837; Table 2). Of the 269 primary outcome events, 219 were neonatal hypoglycemia and 50 other neonates required NICU admission for hypoglycemia.
TABLE 2.
Primary outcome by insulin protocol period and insulin type.
| Preintervention | Postintervention | |||
|---|---|---|---|---|
| No | Yes | No | Yes | |
| Neonatal hypoglycemia or NICU admission for hypoglycemia | 550 (77%) | 168 (23%) * | 321 (76%) | 101 (24%) * |
| Type of insulin administered | ||||
| Neither | 446 (80%) | 110 (20%) | 230 (80%) | 58 (20%) |
| Subcutaneous only | 61 (64%) | 34 (36%) | 69 (68%) | 33 (32%) |
| IV only | 20 (80%) | 5 (20%) | 1 (50%) | 1 (50%) |
| Both | 23 (55%) | 19 (45%) | 21 (70%) | 9 (30%) |
| Total | 550 | 168 | 321 | 101 |
Abbreviations: IV, intravenous; NICU, neonatal intensive care unit.
There is no statistically significant difference in the proportion of infants with the primary outcome before and after protocol implementation (p = 0.837).
When stratified by insulin modality, primary outcome rates varied numerically but did not demonstrate a consistent pattern. In univariate analysis, in the pre‐implementation period the primary outcome occurred in 36% of those who received SC insulin only, 20% of those who received IV insulin only, and 45% of those who received both modalities (Table 2). In the post‐implementation period, the primary outcome occurred in 32% of those who received SC insulin only, 50% of those who received IV insulin only, and 30% of those who received both SC and IV insulin; however only two patients received IV insulin alone in the post‐implementation period, limiting interpretation of this subgroup. In multivariable logistic regression, implementation period was not significantly associated with the primary outcome in either unadjusted or adjusted analyses. In the fully adjusted model, the post‐implementation period was associated with an odds ratio of 1.09 (95% confidence interval [CI], 0.80–1.48) relative to the pre‐implementation period. Compared with IV insulin only, insulin modality was not significantly associated with the primary outcome in adjusted models. In the fully adjusted model, the odds ratio (OR) was 1.72 (95% CI, 0.56–5.27) for neither insulin modality, 2.62 (95% CI, 0.84–8.21) for SC insulin only, and 2.55 (95% CI, 0.76–8.53) for receipt of both SC and IV insulin. Greater gestational age was associated with lower odds of the primary outcome (adjusted OR, 0.71; 95% CI, 0.63–0.82), whereas maternal age and insurance type were not associated in the fully adjusted model (Table 3). In sensitivity analysis excluding patients with A1GDM, results were similar to those observed in the primary analysis. In the fully adjusted model, the odds ratio for the post‐implementation period was 0.85 (95% CI, 0.58–1.25). Estimates for insulin modality were also similar to those observed in the primary analysis.
TABLE 3.
Logistic regression for association of primary event with insulin type stratified by pre/post policy change and insulin type.
| Model 1a | Model 1b | Model 2 | Model 3 | Model 4 | Model 1a (excluding A1 GDM) | Model 4 (excluding A1GDM) | |
|---|---|---|---|---|---|---|---|
| OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |
| Pre/post‐policy | |||||||
| Pre‐period | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | |
| Post‐period | 1.03 (0.78, 1.37) | 0.96 (0.72, 1.29) | 1.04 (0.77, 1.40) | 1.09 (0.80, 1.48) | 0.86 (0.60, 1.23) | 0.85 (0.58, 1.25) | |
| Type of insulin administered | |||||||
| Neither | 1.32 (0.45, 3.89) | 1.48 (0.50, 4.42) | 1.69 (0.56, 5.08) | 1.72 (0.56, 5.27) | 1.59 (0.50, 4.99) | ||
| Subcutaneous only | 2.69 (0.89, 8.14) | 2.90 (0.94, 8.93) | 2.94 (0.95, 9.13) | 2.62 (0.84, 8.21) | 2.57 (0.80, 8.26) | ||
| IV only | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | ||
| Both | 3.42 (1.06, 11.02) * | 3.60 (1.10, 11.76) * | 3.09 (0.94, 10.19) | 2.55 (0.76, 8.53) | 2.50 (0.73, 8.56) | ||
| Age at delivery | 1.03 (1.00, 1.06) * | 1.03 (1.00, 1.06) * | 1.02 (0.99, 1.05) | 1.02 (0.99, 1.06) | |||
| Insurance type | |||||||
| Private insurance | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | |||
| Medicaid | 1.46 (1.08, 1.96) * | 1.25 (0.92, 1.71) | 1.18 (0.86, 1.62) | 0.99 (0.67, 1.47) | |||
| Tricare/MediCARE/None | 2.21 (0.99, 4.96) | 1.84 (0.81, 4.21) | 1.81 (0.78, 4.19) | 1.84 (0.73, 4.66) | |||
| Gestational age (per week) | 0.67 (0.59, 0.76) * | 0.71 (0.63, 0.82) * | 0.70 (0.59, 0.84) * | ||||
| Diabetes type | |||||||
| Diet‐controlled gestational diabetes | 1.00 (ref) | – | |||||
| Medication‐controlled gestational diabetes | 1.18 (0.83, 1.67) | 1.00 (ref) | |||||
| Type 2 diabetes | 1.44 (0.90, 2.30) | 1.20 (0.78, 1.85) | |||||
| Plurality | |||||||
| Singleton | 1.00 (ref) | 1.00 (ref) | |||||
| Twins | 2.49 (0.93, 6.63) | 3.64 (0.95, 14.84) | |||||
| Mode of delivery | |||||||
| Vaginal delivery | 1.00 (ref) | 1.00 (ref) | |||||
| Cesarean delivery | 1.35 (0.99, 1.83) | 1.41 (0.95, 2.09) |
Abbreviations: CI, confidence interval; IV, intravenous; OR, odds ratio.
p < 0.05.
In interrupted time series analysis of monthly outcome rates, there was no significant pre‐implementation monthly trend in the primary outcome (incidence rate ratio [IRR], 1.01; 95% CI, 0.99–1.02; p = 0.209). At the time of protocol implementation, there was a significant immediate reduction in the monthly event rate (IRR, 0.20; 95% CI, 0.04–0.90; p = 0.036). However, there was no significant change in the post‐implementation monthly trend (IRR, 1.03; 95% CI, 0.99–1.06; p = 0.083) (Figure 1).
FIGURE 1.
Proportion of births with neonatal hypoglycemia or NICU admission plotted by month (March 2021–December 2025). The dashed vertical line indicates implementation of the new insulin policy in July 2024. Linear trends were fit separately for the pre‐protocol and post‐protocol periods. In interrupted time series analysis, protocol implementation was associated with a statistically significant immediate reduction in monthly event rates with no statistically significant change in the post‐protocol trend. NICU, neonatal intensive care unit.
Secondary maternal glycemic outcomes varied by insulin modality (Appendix S3). Maternal hypoglycemia occurred in 121 (10.8%) of 1121 patients with available intrapartum glucose data and was most frequent among patients who received both SC and IV insulin (27%) or IV insulin only (20%) compared with 17% among those receiving SC insulin only and 7% among those receiving neither modality. Maternal hyperglycemia occurred in 488 (43.5%) of those 1121 patients and was present in 100% of those who received both modalities of insulin, 96% of those receiving IV insulin only, 72% of those receiving SC insulin only, and 29% of those receiving neither modality.
Among patients who received IV insulin, 65 (66.3%) were pre‐intervention, 31 (31.6%) were post‐intervention, and 2 (2.0%) during the washout period (Figure S1). IV insulin duration prior to delivery was 14.75 h (IQR, 8.45–19.88) in the pre‐implementation period and 11.2 h (IQR, 2.88–20.21) in the post‐implementation period.
For the nursing survey, 65 responses were collected before protocol implementation and 56 after implementation. Satisfaction with the current GDM and T2DM management protocol improved after implementation, with the proportion of RNs who “agreed” or “strongly agreed” increasing from 29.2% before implementation to 51.8% after implementation (p = 0.032). Perceived consistency of the management protocol did not significantly differ between periods, with “agreement” or “strong agreement” reported by 57.8% of respondents before implementation and 55.4% after implementation (p = 0.731). Similarly, the proportion of RNs who “agreed” or “strongly agreed” that management of GDM and T2DM required a significant amount of effort remained high and did not significantly change after implementation (47.7% before vs. 41.1% after implementation, p = 0.806) (Table 4).
TABLE 4.
Perinatal RN satisfaction with diabetes protocol.
| Pre‐intervention | Post‐intervention | p value | |
|---|---|---|---|
| N | 65 (53.7%) | 56 (46.3%) | |
| Happy with current management | p = 0.032 | ||
| Agree | 19 (29.2%) | 29 (51.8%) | |
| Neutral | 31 (47.7%) | 17 (30.4%) | |
| Disagree | 15 (23.1%) | 10 (17.9%) | |
| Management is consistent | p = 0.731 | ||
| Agree | 37 (57.8%) | 31 (55.4%) | |
| Neutral | 16 (25.0%) | 15 (26.8%) | |
| Disagree | 11 (17.2%) | 10 (17.9%) | |
| Management requires significant effort | p = 0.806 | ||
| Agree | 31 (47.7%) | 23 (41.1%) | |
| Neutral | 16 (24.6%) | 19 (33.9%) | |
| Disagree | 18 (27.7%) | 14 (25.0%) | |
Abbreviation: RN, nursing.
4. DISCUSSION
Implementation of an SC insulin–based intrapartum glucose management protocol for patients with GDM and T2DM did not increase the primary composite outcome of neonatal hypoglycemia nor NICU admission for hypoglycemia. In interrupted time series analysis, there was a significant immediate reduction in the monthly event rate at the time of implementation, although there was no significant change in the post‐implementation monthly trend. In multivariable analysis, insulin modality was not significantly associated with the primary outcome. Together, these findings demonstrate that implementation of an SC insulin‐first protocol was not associated with worse neonatal outcomes.
These findings add to a limited and evolving literature on intrapartum glucose management. Existing evidence has largely focused on patients with type 1 diabetes or has combined multiple diabetes types, limiting its applicability to the larger population of patients with GDM and T2DM [4, 5, 14]. In a randomized controlled trial among individuals with type 1 diabetes using insulin pumps, there was no difference in neonatal glucose outcomes between continuation of continuous SC insulin infusion and transition to IV insulin infusion [9]. Similarly, another randomized controlled trial reported no difference in neonatal glucose outcomes between rotating IV fluids and continuous insulin infusion among patients with A2GDM or T2DM, although that study was underpowered because of recruitment and crossover challenges [10]. Our findings are consistent with prior studies in suggesting that less invasive intrapartum management approaches may achieve neonatal outcomes comparable to IV insulin‐based strategies. Our study also extends this question to a pragmatic protocol change in a real‐world L&D setting.
Observational studies demonstrate substantial variation in intrapartum glucose management, and systematic reviews suggest that the relationship between maternal intrapartum glycemia and neonatal outcomes remains incompletely defined [4, 5, 15]. Some studies further suggest that tighter intrapartum glycemic control does not necessarily reduce rates of neonatal hypoglycemia, and that many existing protocols are based more on tradition than on robust comparative evidence [2, 4, 14, 15]. Together, these data highlight ongoing uncertainty regarding the benefit of intensive glucose control strategies during labor. In this context, our findings support the possibility that routine escalation to IV insulin may not be necessary for many patients with GDM or T2DM, particularly when balanced against the clinical burden of infusion‐based protocols.
Our protocol also represented a change in glycemic targets. ACOG recommends a threshold of 110 mg/dL or less during labor [6]. Historically, ≤140 mg/dL was used in practice at our institution before implementation, but this shifted to a target of ≤120 mg/dL with implementation of the protocol based on literature suggesting that this target would be able to achieve safe neonatal outcomes and without good evidence at the time of protocol creation that a target of ≤140 was appropriate [16]. The Joint British Diabetes Society for Inpatient Care guideline and recent narrative reviews have increasingly acknowledged that more pragmatic intrapartum glycemic targets may be reasonable, particularly when weighed against the burden and risks of intensive insulin infusion protocols [2, 17]. In that context, our findings support the possibility that automatically escalating to IV insulin may not be necessary for many patients with GDM or T2DM.
An important finding in our study was the improvement in RN satisfaction after protocol implementation. The proportion of RNs who reported satisfaction with the current GDM and T2DM management increased substantially after implementation, although perceptions of management consistency and required effort did not significantly change. This suggests that while intrapartum diabetes management remained labor‐intensive, the updated protocol may have been perceived as more usable or preferable overall. Given that IV insulin infusions require continuous titration, IV access, and close RN surveillance, a burden described in both older reviews and more recent guidance documents, it is plausible that an SC insulin‐first strategy was experienced as more usable or less disruptive, even without fully eliminating the workload associated with managing diabetes during labor [2, 8, 17]. In an L&D environment where safe protocol execution depends heavily on RN implementation, this is a meaningful finding.
Our study has several strengths. It evaluates a standardized clinical practice change in a large cohort of patients with GDM and T2DM, a population that is underrepresented in prior studies of intrapartum glucose management [4, 5]. Our study also incorporates multiple complementary analytic approaches, including pre‐ and post‐comparisons, interrupted time series, and multivariable logistic regression. Because this was a real‐world implementation, the findings may be more generalizable to routine obstetric practice. Inclusion of RN survey data also provides insight into an implementation outcome that is clinically important but rarely reported in studies of intrapartum diabetes management.
This study should be interpreted with consideration of several limitations. First, this was a retrospective observational evaluation of a QI intervention at a single institution, and residual confounding is likely. Patients were not randomized to insulin modality, and receipt of IV insulin likely reflected greater intrapartum glycemic severity. Although we adjusted for selected maternal and obstetric covariates, we did not include an aggregate measure of intrapartum glucose burden in the regression models; therefore, residual confounding by indication remains possible. Second, the post‐implementation period was shorter than the pre‐implementation period, which may have limited power to detect changes in trend over time. Third, although the interrupted time series showed an immediate reduction in the monthly event rate, this finding should be interpreted cautiously given the absence of a significant sustained post‐implementation slope change and the potential for temporal instability in monthly event rates. Fourth, because the protocol changed both insulin modality and the intrapartum glycemic target, the observed findings cannot be attributed to insulin modality alone. Fifth, some secondary analyses, particularly those involving IV insulin alone in the post‐implementation period, were limited by the small number of patients in these groups. Finally, the RN survey was anonymous and voluntary, with changes in personnel during the study period, which may introduce response bias.
5. CONCLUSION
Overall, implementation of an SC insulin‐based intrapartum glucose management protocol for patients with GDM and T2DM was not associated with an increase in neonatal hypoglycemia nor NICU admission for hypoglycemia. The new protocol was associated with improved RN satisfaction. These findings support SC insulin as a pragmatic alternative to IV insulin‐based protocols in appropriate patients. Additional multicenter studies are needed to clarify optimal intrapartum glycemic targets and insulin strategies for patients with GDM and T2DM [4, 14, 17].
AUTHOR CONTRIBUTIONS
Leah M. Savitsky: Conceptualization; investigation; writing—original draft; methodology; validation; visualization; writing—review & editing; project administration; formal analysis; data curation. Leilani White: Investigation; writing—review and editing. Juliet Henderson: Investigation; writing—review and editing. Tara Saleh: Investigation; writing—review and editing. Nadine Martinez: Conceptualization; writing—review and editing. Chantelle Barr: Investigation; writing—review and editing. LaVone Simmons: Conceptualization; methodology; validation; visualization; writing—review and editing; project administration; supervision; resources. Ronit Katz: Conceptualization; methodology; writing—review and editing; formal analysis; data curation.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
FUNDING INFORMATION
The authors received no specific funding for this work.
Supporting information
Supplemental Figure 1. IV insulin utilization rates plotted by month (March 2021–December 2025). The dashed vertical line indicates implementation of the new insulin policy in July 2024.
Supporting Information
Supporting Information
Supporting Information
This study was also presented as a poster presentation at SMFM's The Pregnancy Meeting 2026, Las Vegas, NV, February 12, 2026.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
REFERENCES
- 1. Thornton, P. S. , Stanley C. A., De Leon D. D., Harris D., Haymond M. W., Hussain K., Levitsky L. L., et al. 2015. “Recommendations From the Pediatric Endocrine Society for Evaluation and Management of Persistent Hypoglycemia in Neonates, Infants, and Children.” Jornal De Pediatria 167(2): 238–45. 10.1016/j.jpeds.2015.03.057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Bartal, M. F. 2025. “Intrapartum Care for People With Diabetes—Working Towards Evidence‐Based Management.” American Journal of Perinatology 42(09): 1103–8. 10.1055/a-2405-1846. [DOI] [PubMed] [Google Scholar]
- 3. van Kempen, A. , Eskes P. F., Nuytemans D., van der Lee J. H., Dijksman L. M., van Veenendaal N. R., van der Hulst F., et al. 2020. “Lower Versus Traditional Treatment Threshold for Neonatal Hypoglycemia.” New England Journal of Medicine 382(6): 534–44. 10.1056/NEJMoa1905593. [DOI] [PubMed] [Google Scholar]
- 4. Yamamoto, J. M. , Benham J., Mohammad K., Donovan L. E., and Wood S.. 2018. “Intrapartum Glycaemic Control and Neonatal Hypoglycaemia in Pregnancies Complicated by Diabetes: A Systematic Review.” Diabetic Medicine 35(2): 173–83. 10.1111/dme.13546. [DOI] [PubMed] [Google Scholar]
- 5. Ulyatt, C. M. , Roberts L. F., Crowther C. A., Harding J. E., and Lin L.. 2024. “Intrapartum Maternal Glycaemic Control for the Prevention of Neonatal Hypoglycaemia: A Systematic Review and Meta‐Analysis.” BMC Pregnancy and Childbirth 24: 423. 10.1186/s12884-024-06615-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. 2018. “ACOG Practice Bulletin No. 201: Pregestational Diabetes Mellitus.” Obstetrics and Gynecology 132(6): e228–48. 10.1097/AOG.0000000000002960. [DOI] [PubMed] [Google Scholar]
- 7. American Diabetes Association Professional Practice Committee for Diabetes . 2025. “15. Management of Diabetes in Pregnancy: Standards of Care in Diabetes—2026.” Diabetes Care 49(Suppl 1): S321–38. 10.2337/dc26-S015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Garrison, E. A. , and Jagasia S.. 2014. “Inpatient Management of Women With Gestational and Pregestational Diabetes in Pregnancy.” Current Diabetes Reports 14(2): 457. 10.1007/s11892-013-0457-x. [DOI] [PubMed] [Google Scholar]
- 9. Wilkie, G. L. , Delpapa E., and Leftwich H. K.. 2023. “Intrapartum Continuous Subcutaneous Insulin Infusion vs Intravenous Insulin Infusion Among Pregnant Individuals With Type 1 Diabetes Mellitus: A Randomized Controlled Trial.” American Journal of Obstetrics and Gynecology 229(6): 680.e1–8. 10.1016/j.ajog.2023.07.003. [DOI] [PubMed] [Google Scholar]
- 10. Sullivan, M. , Angras K., Boyd V., Young A. J., Mackeen A. D., and Paglia M. J.. 2025. “Intrapartum Glycemic Control With Insulin Infusion Versus Rotating Fluids: A Randomized Controlled Trial.” American Journal of Perinatology 42(11): 1491–9. 10.1055/a-2510-4906. [DOI] [PubMed] [Google Scholar]
- 11. Dude, A. M. , Niznik C., Peaceman A. M., and Yee L. M.. 2020. “Evaluation of an Intrapartum Insulin Regimen for Women With Diabetes.” Obstetrics and Gynecology 136(2): 411–6. 10.1097/AOG.0000000000003940. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Harris, P. A. , Taylor R., Thielke R., Payne J., Gonzalez N., and Conde J. G.. 2009. “Research Electronic Data Capture (REDCap)—A Metadata‐Driven Methodology and Workflow Process for Providing Translational Research Informatics Support.” Journal of Biomedical Informatics 42(2): 377–81. 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Harris, P. A. , Taylor R., Minor B. L., Elliott V., Fernandez M., O'Neal L., McLeod L., et al. 2019. “The REDCap Consortium: Building an International Community of Software Platform Partners.” Journal of Biomedical Informatics 95: 103208. 10.1016/j.jbi.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bitar, G. , and Fishel Bartal M.. 2025. “Intrapartum Glycemic Control and Clinical Outcomes.” Diabetes Spectrum: A Publication of the American Diabetes Association 38(4): 400–6. 10.2337/dsi25-0010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Ringholm, L. , Søholm J. C., Pedersen B. W., Clausen T. D., Damm P., and Mathiesen E. R.. 2024. “Glucose Control During Labour and Delivery in Type 1 Diabetes—An Update on Current Evidence.” Current Diabetes Reports 25(1): 7. 10.1007/s11892-024-01563-1. [DOI] [PubMed] [Google Scholar]
- 16. Hamel, M. S. , Kanno L. M., Has P., Beninati M. J., Rouse D. J., and Werner E. F.. 2019. “Intrapartum Glucose Management in Women With Gestational Diabetes Mellitus: A Randomized Controlled Trial.” Obstetrics and Gynecology 133(6): 1171–7. 10.1097/AOG.0000000000003257. [DOI] [PubMed] [Google Scholar]
- 17. Dashora, U. , Levy N., Dhatariya K., Willer N., Castro E., and Murphy H. R., 2022. “Managing Hyperglycaemia During Antenatal Steroid Administration, Labour and Birth in Pregnant Women With Diabetes—An Updated Guideline From the Joint British Diabetes Society for Inpatient Care.” Diabetic Medicine 39(2): e14744. 10.1111/dme.14744. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1. IV insulin utilization rates plotted by month (March 2021–December 2025). The dashed vertical line indicates implementation of the new insulin policy in July 2024.
Supporting Information
Supporting Information
Supporting Information
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.