Planned Section Structure
This article is organized to match the clinical importance of intrapartum and postpartum glucose management in type 1 diabetes. Sections include: clinical background and unmet need; study design and prespecified analysis framework; intrapartum results; early postpartum results; safety outcomes; clinical interpretation and workflow implications; strengths and limitations; implications for practice and research; and funding and reference details.
Highlights
In the 24 hours before delivery, automated insulin delivery with Control-IQ increased time in the pregnancy-specific glucose range of 63-140 mg/dL compared with standard care, with an adjusted mean difference of 13.2 percentage points (95% CI 5.2 to 21.2).
Closed-loop users were far less likely to require intrapartum intravenous insulin: 2% versus 45% with standard care.
During the first postpartum week, closed-loop therapy reduced time spent below 70 mg/dL, with an adjusted mean difference of -1.8 percentage points (95% CI -2.7 to -0.9).
No diabetic ketoacidosis occurred in either group, and no maternal severe hypoglycemia occurred in the closed-loop group.
Background
Pregnancy in women with type 1 diabetes remains one of the most demanding settings for insulin management. Glycemic targets are tighter than in the nonpregnant state because maternal hyperglycemia is associated with adverse perinatal outcomes, including large-for-gestational-age birthweight, neonatal hypoglycemia, and preterm delivery. At the same time, physiologic variability increases as pregnancy advances, and insulin requirements can change rapidly during labor and then fall abruptly after placental delivery. This combination creates a narrow therapeutic margin: clinicians must avoid maternal hyperglycemia and ketosis while also minimizing hypoglycemia during a period of fluctuating caloric intake, changing stress hormone exposure, and transitions in insulin dosing.
In routine practice, intrapartum glucose management often relies on frequent capillary testing and, in many institutions, variable-rate intravenous insulin protocols. These approaches can be labor-intensive for both staff and patients. The early postpartum period introduces an additional challenge. Insulin resistance declines quickly after delivery, breastfeeding may further alter glucose patterns, and women are often exhausted, sleep-deprived, and adjusting to newborn care. As a result, early postpartum hypoglycemia is a common concern.
Automated insulin delivery, also called closed-loop insulin delivery, offers a biologically plausible solution. By integrating continuous glucose monitoring with algorithm-driven insulin dosing, these systems can respond dynamically to rapidly changing insulin needs. The parent CIRCUIT trial previously demonstrated improved glycemia in pregnancy with the Control-IQ closed-loop system. The present prespecified analysis addresses a particularly consequential question for clinicians: can those benefits be maintained during labor and immediately after birth, when glucose management is notoriously difficult?
Study Design
This report is a prespecified analysis of the CIRCUIT randomized controlled trial, titled Closed loop Insulin delivery by glucose Responsive Computer algorithms In Type 1 diabetes pregnancies. The analysis compared intrapartum and early postpartum glycemic effectiveness and safety between participants assigned to Control-IQ closed-loop insulin delivery and those assigned to standard care.
The primary outcome for this analysis was the percentage of time spent in the pregnancy-specific continuous glucose monitoring range of 63-140 mg/dL during the 24 hours before childbirth. This is a clinically appropriate endpoint because intrapartum glycemia is directly relevant to maternal safety, labor management, and neonatal metabolic adaptation.
A key secondary outcome was the percentage of time spent below 70 mg/dL during the first postpartum week. This endpoint is especially important because postpartum hypoglycemia is common and can be dangerous in the context of fatigue, breastfeeding, and reduced insulin requirements. Additional secondary outcomes included other continuous glucose monitoring metrics through 6 weeks postpartum.
The analyses were adjusted for baseline measure, insulin delivery mode, and site, which strengthens internal validity by accounting for pre-existing glycemic differences and practice variation across centers.
Key Results
Intrapartum glycemic control
During the 24 hours before delivery, 39 of 44 participants in the intervention arm, or 89%, continued to use closed loop intrapartum. This high continuation rate is clinically notable. It suggests that use of the system during labor was feasible in real-world obstetric care rather than being restricted to idealized trial conditions.
Closed-loop therapy outperformed standard care for the primary endpoint. Participants using Control-IQ spent 79.6% of time in the pregnancy-specific target range of 63-140 mg/dL, compared with 64.8% in the standard-care group. The adjusted mean difference was 13.2 percentage points, with a 95% confidence interval of 5.2 to 21.2. This effect size is both statistically persuasive and clinically meaningful. In practical terms, it represents more than 3 additional hours in target range over the 24-hour intrapartum period.
The need for intravenous insulin was also dramatically lower with closed loop. Only 1 participant in the closed-loop group, or 2%, required intrapartum intravenous insulin, compared with 20 participants, or 45%, in the standard-care group (P < 0.001). This finding has substantial workflow implications. Reduced dependence on intravenous insulin protocols may lessen nursing burden, simplify labor management, and improve patient autonomy, provided institutional protocols support safe continued use of insulin pumps and continuous glucose monitoring.
Early postpartum glycemia
The first postpartum week is often one of the highest-risk windows for hypoglycemia in women with type 1 diabetes. In this analysis, closed-loop therapy performed favorably on the key secondary outcome. Participants in the closed-loop group spent 1.7% of time below 70 mg/dL, compared with 3.2% in the standard-care group. The adjusted mean difference was -1.8 percentage points, with a 95% confidence interval of -2.7 to -0.9.
Although the absolute percentages may appear small, even modest reductions in time below range can matter in the postpartum period, when severe sleep disruption and caregiving demands can amplify the consequences of hypoglycemia. A reduction of about 1.5 to 2 percentage points corresponds to roughly 20 to 30 fewer minutes per day spent below 70 mg/dL. Across a week, that may translate into a meaningful improvement in safety and quality of life.
The abstract also notes that continuous glucose monitoring metrics were assessed through 6 weeks postpartum, although detailed numerical values beyond the first week are not included in the provided summary. Even so, the short-term signal is important because the first postpartum days are precisely when insulin requirements shift most abruptly.
Safety
Safety outcomes were reassuring. No maternal severe hypoglycemia occurred in the closed-loop group, whereas one postpartum severe hypoglycemic episode occurred in the standard-care group. No diabetic ketoacidosis occurred in either group.
These findings support the conclusion that better glycemic outcomes were not achieved at the expense of increased acute metabolic risk. That point is essential when evaluating automated insulin delivery in pregnancy-related settings, where clinicians may understandably worry that algorithm-driven dosing could be challenged by the physiologic turbulence of labor and immediate postpartum recovery.
Clinical Interpretation
This analysis reinforces an increasingly important concept: automated insulin delivery may be useful not only during routine ambulatory pregnancy care, but also during the transitional periods of labor and postpartum recovery. These periods have traditionally been viewed as too complex or too unstable for routine pump-based diabetes management, leading many centers to default to intravenous insulin or highly manual protocols. The CIRCUIT findings challenge that assumption.
Several aspects of the data are especially relevant to bedside practice. First, the magnitude of intrapartum improvement in time in range is clinically credible and likely meaningful for neonatal risk reduction, though this analysis was not presented as a neonatal-outcomes study. Better maternal glucose control around delivery has long been considered important because fetal exposure to maternal hyperglycemia may increase the risk of neonatal hypoglycemia after birth. While one should be cautious not to overinterpret surrogate endpoints, the intrapartum glucose benefit is directionally aligned with established pathophysiology.
Second, the sharply reduced use of intravenous insulin may be one of the most practice-changing findings. In hospitals with appropriate expertise, allowing women already established on automated insulin delivery to continue their system through labor may reduce treatment disruption and preserve continuity of care. This could be particularly valuable for patients who are highly skilled pump users and for centers aiming to standardize patient-centered obstetric diabetes pathways.
Third, the postpartum reduction in hypoglycemia is highly relevant clinically. After delivery, clinicians and patients often struggle to rapidly recalibrate insulin dosing. An adaptive system that reduces time below range may offer an important safety buffer during this unstable phase.
Strengths of the Analysis
The study has several notable strengths. It is based on a randomized controlled trial rather than a purely observational dataset, which strengthens causal inference. The intrapartum and postpartum analysis was prespecified, reducing concern that these findings arose from data-driven post hoc exploration. Continuous glucose monitoring outcomes are particularly appropriate in this setting because they capture the full temporal profile of glycemia rather than relying on sparse point measurements. Adjustment for baseline measure, insulin delivery mode, and site also enhances the rigor of the comparative analysis.
Another strength is the pragmatic nature of the intrapartum continuation data. Nearly 9 in 10 participants in the intervention arm remained on closed loop during labor, suggesting that implementation was feasible across actual clinical settings participating in the trial.
Limitations and Cautions
Several limitations deserve attention. The sample size in this prespecified subgroup analysis appears modest, which means estimates for uncommon safety events and some secondary outcomes remain imprecise. The abstract does not provide detailed information here on mode of delivery, anesthesia, steroid exposure, nutritional intake, breastfeeding intensity, or exact postpartum insulin-adjustment protocols, all of which could influence glucose trajectories.
Generalizability also requires careful thought. The participants were enrolled in a clinical trial and likely received close support from specialized teams. Outcomes in less experienced centers may differ, particularly where labor-and-delivery staff have limited familiarity with continuous glucose monitoring, pump troubleshooting, or hospital policies for personal diabetes devices. It is also unclear from the abstract whether all participants were experienced users of similar technologies before randomization, which may affect real-world transferability.
Another limitation is that the report focuses primarily on glycemic metrics and acute safety. Those are appropriate endpoints, but clinicians will also want to know whether these intrapartum advantages translate into differences in neonatal hypoglycemia, neonatal intensive care use, maternal satisfaction, nursing workload, and length of stay. Such outcomes are highly relevant to policy and implementation decisions.
Implications for Practice
For diabetes-in-pregnancy programs, the CIRCUIT analysis provides meaningful support for considering continued automated insulin delivery during labor and in the first postpartum week in carefully selected women with type 1 diabetes. The data suggest that Control-IQ can improve time in range before birth and reduce postpartum hypoglycemia without introducing new safety signals.
However, implementation should be protocolized. Hospitals considering this approach need clear criteria for which patients can remain on closed loop, when to transition to intravenous insulin if needed, how to document sensor and pump data in the medical record, and how to adjust settings immediately postpartum. Staff training is essential, especially for obstetric nurses, endocrinology teams, and anesthesia personnel. Backup procedures for device failure, prolonged labor, emergency cesarean delivery, and perioperative fasting remain necessary.
From a patient-centered perspective, the findings may be encouraging. Many women prefer to maintain familiar diabetes technology during labor rather than switching to intravenous protocols. If this can be done safely and effectively, it may reduce stress during an already intense clinical experience.
Research Priorities
Future work should address several remaining questions. Larger studies could better define rare safety events and assess neonatal outcomes directly. Comparative effectiveness research could evaluate different automated insulin delivery systems, as algorithm performance and user interface vary across platforms. Implementation studies in broader obstetric settings, including community hospitals, would help determine how reproducible these benefits are outside specialized centers.
There is also a need for practical evidence on postpartum settings. Studies should examine how breastfeeding, nocturnal feeding schedules, cesarean recovery, and hospital discharge timing interact with automated insulin delivery performance. In addition, cost and workflow analyses would be useful for health systems weighing the tradeoffs between device-supported labor management and conventional intravenous insulin pathways.
Conclusion
This prespecified analysis of the CIRCUIT randomized controlled trial extends the case for automated insulin delivery in type 1 diabetes beyond antenatal outpatient management. Use of the Control-IQ closed-loop system during labor improved time in the pregnancy-specific glucose range and markedly reduced the need for intravenous insulin. In the first postpartum week, it lowered time spent in hypoglycemia. Safety outcomes were reassuring, with no diabetic ketoacidosis in either group and no severe hypoglycemia in the closed-loop arm.
For clinicians, the message is straightforward: in appropriately supported settings, automated insulin delivery appears capable of navigating two of the most unstable phases of diabetes management in pregnancy, labor and the early postpartum period. The findings are not the final word, but they are a strong step toward integrating closed-loop therapy across the full perinatal continuum.
Funding and ClinicalTrials.gov
Funding details and a ClinicalTrials.gov registration number are not reported in the provided abstract. Readers should consult the full Diabetes Care publication and the trial registry record for formal funding disclosures, registration details, and protocol information.
Reference
Donovan LE, Lemieux P, Yamamoto JM, Dunlop AD, Murphy HR, Liu SL, Chaput KH, Simmons D, Bell RC, Benham JL, Ross GP, Nerenberg KA, Booth JE, Mohammad K, Perkins BA, Crawford S, Tomlinson G, Feig DS; CIRCUIT Collaborative Group. Intrapartum and Early Postpartum Use of Automated Insulin Delivery in Type 1 Diabetes: A Prespecified Analysis of the CIRCUIT Randomized Controlled Trial. Diabetes Care. 2026-05-27. PMID: 42201836. Available at: https://pubmed.ncbi.nlm.nih.gov/42201836/