Preexisting (pregestational) diabetes mellitus: Antenatal glycemic management

INTRODUCTION — 

Preexisting (also called pregestational)diabetes refers to type 1 or type 2 diabetes mellitus that has been diagnosed before pregnancy. The goal of medical management of patients with preexisting diabetes is to maintain blood glucose concentrations at or near normoglycemic levels at the time of conception and throughout the entire pregnancy, taking into account that "normoglycemia" in pregnant patients without diabetes is lower than in the nonpregnant state. Therefore, glycemic targets when preparing for and during pregnancy are lower than during usual care. Maintaining target glucose levels during pregnancy decreases the likelihood of adverse maternal, fetal, and newborn outcomes (eg, congenital anomalies, preeclampsia, macrosomia, neonatal hypoglycemia). Since patients often do not know that they are pregnant until after fetal organogenesis is well underway, ideally, good glucose management should be achieved before conception to reduce the risk of congenital anomalies.

The medical management of preexisting diabetes mellitus in the antenatal period will be discussed here. Additional aspects of pregnancy complicated by preexisting diabetes are reviewed separately:

●Optimizing glucose management before pregnancy, intrapartum, and postpartum (see "Preexisting (pregestational) diabetes: Preconception counseling, evaluation, and management", section on 'Management of hyperglycemia' and "Preexisting (pregestational) and gestational diabetes: Intrapartum and postpartum glucose management")

●Fetal and maternal risks of pregnancy (see "Preexisting (pregestational) diabetes: Preconception counseling, evaluation, and management", section on 'Fetal and neonatal risks' and "Preexisting (pregestational) diabetes: Preconception counseling, evaluation, and management", section on 'Maternal medical risks')

●Obstetric management of the pregnancy (see "Preexisting (pregestational) diabetes mellitus: Obstetric issues and pregnancy management")

●Diagnosis and management of diabetic ketoacidosis (See "Diabetic ketoacidosis in pregnancy".)

●Medical care, complications, and long-term prognosis of the infant (see "Infants of mothers with diabetes (IMD)")

ASSESSING GLYCEMIC MANAGEMENT — 

Assessment of glucose levels during pregnancy may involve blood glucose monitoring (BGM) with fingersticks for capillary blood sampling plus use of a glucose meter with or without use of a device for continuous glucose monitoring (CGM), in combination with periodic measurement of glycated hemoglobin (also called A1C, hemoglobin A1C, glycohemoglobin, or HbA1C). BGM and CGM results are reviewed with a clinician during ambulatory or telemedicine visits.

Blood glucose monitoring — BGM enables timely recognition of hyperglycemia and hypoglycemia patterns so that dietary and insulin adjustments may be made. High preprandial and high postprandial glucose levels have been correlated with excessive fetal size and newborn complications [1-7]. (See "Glucose monitoring in the ambulatory management of nonpregnant adults with diabetes mellitus", section on 'BGM systems'.)

Timing/frequency of BGM — We suggest that all pregnant patients with preexisting diabetes check blood glucose levels daily at the following times:

●Fasting

●Before each meal

●One or two hours after the first bite of each meal

●Bedtime

Monitoring fasting, preprandial, and bedtime glucose levels facilitates calculation of basal insulin doses. Monitoring postprandial glucose levels is necessary to titrate mealtime bolus dosing.

Additional testing is indicated in selected patients, such as those with:

●Suspected nocturnal hypoglycemia – Glucose assessment midway through sleep, typically between 2 and 4 AM, is useful in patients in whom nocturnal hypoglycemia is suspected, particularly those with type 1 diabetes mellitus. Individualized timing is necessary for people with atypical schedules, such as those who work overnight.

Confirming suspected hypoglycemia is critical so that timely interventions to prevent subsequent hypoglycemic episodes can be initiated. CGM may identify such patients and alert them to nocturnal hypoglycemia. (See 'Continuous glucose monitoring systems' below.)

Possible etiologies for hypoglycemia during sleep include a basal insulin dose that is too high, a presleep snack with insufficient calories, or overcorrection of a high glucose level with bolus insulin.

●Prebreakfast hyperglycemia – Glucose assessment between 3 and 5 AM with review of insulin administration and dietary habits is useful for evaluating patients with prebreakfast hyperglycemia. Possible etiologies include:

•Insufficient basal insulin overnight or too large a carbohydrate snack at bedtime

•"Dawn phenomenon" (rising blood glucose levels prebreakfast following stable nighttime levels), which has been attributed to the normal overnight release of the counterregulatory hormone growth hormone

Target blood glucose values — There are no randomized trials of different glucose targets in pregnancy complicated by preexisting diabetes. The upper thresholds recommended by the American College of Obstetricians and Gynecologists (ACOG) and the American Diabetes Association (ADA) for target blood glucose values in capillary blood are based on consensus expert opinion [8-12]. The lower thresholds during pregnancy compared with the nonpregnant state are based on a systematic review of glycemic parameters in nondiabetic pregnancies [10] and should be applied to those on antihyperglycemic drugs (eg, insulin).

●Fasting, preprandial, and nocturnal glucose: 70 to 95 mg/dL (3.9 to 5.3 mmol/L) and

•One-hour postprandial glucose 110 to 140 mg/dL (6.1 to 7.8 mmol/L) or

•Two-hour postprandial glucose 100 to 120 mg/dL (5.6 to 6.7 mmol/L)

●Mean capillary glucose 100 mg/dL (5.6 mmol/L)

These proposed target ranges may be difficult to achieve in pregnancies complicated by preexisting diabetes given the risk of severe hypoglycemia at lower glucose targets in this population.

We generally recommend a one-hour postprandial glucose concentration target <130 mg/dL (7.3 mmol/L) based on a prior ADA technical review of glycemic targets in preexisting diabetes in pregnancy [13] and normative data derived from pooling data from 11 studies using BGM and CGM in normal pregnancies [10]. These data demonstrated the following mean (+1 standard deviation [SD]) blood glucose levels:

●Fasting 71 mg/dL (+1 SD 79 mg/dL) (3.9 mmol/L [+1 SD 4.3 mmol/L] )

●One-hour postprandial 109 mg/dL (+1 SD 122 mg/dL)(6.0 mmol/L [+1 SD 6.8 mmol/L])

●Two-hour postprandial 99 mg/dL (+1 SD 109 mg/dL)(5.5 mmol/L [+1 SD 6.0 mmol/L])

●24-hour mean blood glucose 88 mg/dL (+1 SD 98 mg/dL ) (4.9 mmol/L [+1 SD 5.4 mmol/L])

Continuous glucose monitoring systems

Overview — CGM may be used as an adjunct to BGM, especially in patients with type 1 diabetes, but no data support using CGM alone for glucose assessment in pregnancy. CGM systems involve placement of a subcutaneous sensor that measures the glucose content of interstitial fluid, which lags plasma glucose by 15 minutes but otherwise correlates well. The sensor is connected to a transmitter that sends this information to a receiver (eg, receiver device or smart phone), which provides a digital display of blood glucose every 5 to 10 minutes.

Real-time CGMs (rtCGMs) have replaced intermittently-scanning (isCGMs) and are the preferred technology. rtCGMs transmit data continuously to the receiver and have low- and high-glucose alarms. rtCGMs with an FDA pregnancy indication include Dexcom G7, and FreeStyle Libre 3 [14,15]. The Dexcom G7 CGM has been found to be accurate in pregnancy when compared with a laboratory-quality Yellow Springs Instrument (YSI) analyzer [16]. A prospective, observational study including 20 patients with type 1 diabetes simultaneously monitored with isCGM and rtCGM for seven days in early pregnancy reported a higher percentage of time below range (TBR) measured by isCGM than by rtCGM. Based on these findings, asymptomatic hypoglycemia by isCGM should be confirmed by BGM prior to a change in clinical management [17]. (See "Glucose monitoring in the ambulatory management of nonpregnant adults with diabetes mellitus", section on 'CGM systems'.)

In a 2019 systematic review of BGM and CGM for monitoring blood glucose levels during pregnancy in patients with preexisting diabetes (12 trials, 863 participants), CGM appeared to reduce the risk for pregnancy-associated hypertensive disorders [18]. Nonsignificant reductions in cesarean birth, large for gestational age (LGA) infant, and perinatal mortality were also observed.

In the largest trial in the analysis, the Continuous Glucose Monitoring in Women with Type 1 Diabetes in Pregnancy Trial (CONCEPTT) Collaborative Group ran parallel open label trials in which 325 patients with type 1 diabetes receiving intensive insulin therapy were randomly assigned to a CGM group that used CGM plus capillary glucose monitoring or a control group that used capillary glucose monitoring alone [19]. One trial enrolled pregnant patients at ≤13+6 weeks of gestation (n = 215), and the second trial enrolled patients planning pregnancy (n = 110). Both groups were stratified by insulin delivery schedule (ie, multiple daily injections [MDI] or continuous subcutaneous insulin infusion [insulin pump]). Major findings were:

●In patients planning pregnancy, the CGM and control groups had similar outcomes, including change in A1C at conception or at 24 weeks follow-up if not pregnant; A1C at pregnancy confirmation; and percentage achieving target A1C ≤7.0 percent (53 mmol/mol) before pregnancy.

●In pregnant patients, the CGM group had better glycemic outcomes and some better newborn outcomes, but no significant difference in pregnancy outcomes.

•Glycemic management – Compared with the control group, CGM users achieved a greater reduction in A1C at 34 weeks (6.35 versus 6.53 percent, mean difference -0.19 percent, 95% CI -0.34 to -0.03), spent more time in the glucose target range of 63 to 140 mg/dL (3.5 to 7.8 mmol/L; 68 versus 61 percent), and spent less time above the target glucose range (27 versus 32 percent), with similar rates of maternal hypoglycemia (3 versus 4 percent). Over 50 percent of patients in both groups achieved A1C ≤6.5 percent (48 mmol/mol) at 34 weeks (66 versus 52 percent). Glucose levels in CGM users was similar for patients using insulin pumps and those using MDI.

•Neonatal outcomes – Fewer newborns in the CGM group were treated for hypoglycemia with intravenous glucose (15 versus 28 percent). Mean birth weight was nearly identical in both groups, and approximately 25 percent of newborns in both groups were macrosomic (≥4000 grams). The prevalence of LGA was high in both groups, although significantly fewer newborns in the CGM group were LGA (53 versus 69 percent). In a secondary analysis, mothers of newborns with neonatal hypoglycemia had higher mean A1Cs in the second and third trimesters (6.6±0.6 versus 6.2±0.6 percent and 6.7±0.6 versus 6.3±0.6 percent, respectively) and lower CGM time in range (46 versus 53 percent and 60 versus 66 percent, respectively) [20].

•Pregnancy outcomes – CGM users had similar rates of miscarriage, hypertensive disorders, preeclampsia, cesarean birth, and preterm birth as the control group. Of note, the overall incidence of preterm birth <37 weeks was unusually high (approximately 40 percent) despite the baseline population characteristics (60 percent of participants had uncomplicated diabetes) and rates of hypertensive disorders that were within the expected range.

Because treatment was not blinded to either patients or providers, several of the neonatal outcomes that were significantly different between groups may have been influenced by provider bias regarding treatment group assignment. These outcomes included those dependent on provider discretion, such as initiation of intravenous glucose for hypoglycemia, length of stay, and high level of neonatal care. The study population may also not be generalizable to other settings.

Target blood glucose values — The International Consensus on Time in Range [21] recommends pregnancy target ranges and goals for time in range for patients using CGM. These data are displayed on the Ambulatory Glucose Profile.

Target range 63 to 140 mg/dL (3.5 to 7.8 mmol/L):

●Time in range, goal >70 percent (ie, >16.8 hours)

●Time below range (<63 mg/dL [3.5 mmol/L]), goal <4 percent (ie, <1 hour)

●Time below range (<54 mg/dL [3.0 mmol/L]), goal <1 percent (ie, <0.24 hour)

●Time above range (>140 mg/dL [7.8 mmol/L]), goal <25 percent (ie, <6 hours)

These goals were derived from data in patients with type 1 diabetes. More data are needed to determine whether the goals should be different for pregnant people with type 2 diabetes as they spend one-third less time hyperglycemic than those with type 1 diabetes and 90 percent achieve the time in range goal [21,22]. In addition, more normative data form pregnant patients without diabetes is needed [10]

Glycated hemoglobin (A1C) — Although BGM with or without CGM should be used as the primary measure of glycemic management in pregnancy, A1C levels are also monitored during pregnancy because, despite the physiologic reduction in A1C levels that occur in pregnancy discussed below, higher values are predictive of increased risks for first-trimester miscarriage, congenital anomalies, and high birth weight [23]. A1C also provides a laboratory-based method for assessing the patient's mean blood glucose level over recent weeks for comparison with BGM records. (See "Measurements of chronic glycemia in diabetes mellitus", section on 'Glycated hemoglobin (A1C)' and "Preexisting (pregestational) diabetes: Preconception counseling, evaluation, and management".)

Target A1C level — The measured A1C value is lower in pregnant patients because physiologic expansion of the red blood cell mass and decreased erythrocyte life span in pregnancy result in a "younger" population of red blood cells [24-29]. Since pregnancy itself reduces A1C levels, A1C targets are lower in pregnant patients with diabetes compared with nonpregnant patients with diabetes. In addition, because A1C represents an average of glucose levels over time, it does not identify patterns of marked swings in glucose from hyperglycemia to hypoglycemia. Thus, a low A1C level, while desirable, may be falsely reassuring if it reflects frequent hypoglycemia, which is unsafe and undesirable.

●ACOG clinical guidelines recommend target A1C levels of <6.0 percent (42 mmol/mol) throughout pregnancy, if safely achievable [11]. However, in practice, achieving this goal, especially in patients with type 1 diabetes, is challenging and only may be accomplished safely in a minority of patients.

●The ADA clinical guidelines also recommend a target A1C of <6.0 percent (42 mmol/mol) in pregnancy, if safely achievable without causing frequent or severe episodes of hypoglycemia that interfere with daily life [12], because observational studies have demonstrated the lowest rates of adverse fetal outcomes (including miscarriage and congenital anomalies) with levels <6.0 to 6.5 percent (42 to 48 mmol/mol) in early gestation [30-32]. As pregnancy progresses, maintaining an A1C <6.0 percent (42 mmol/mol) is associated with lower risk for an LGA infant and preeclampsia [33]. The target may be relaxed to <6.5 percent (48 mmol/mol) or <7.0 percent (53 mmol/mol) if necessary to prevent significant hypoglycemia. (See 'Hypoglycemia' below.)

Since clinical trials have not evaluated the comparative risks and benefits of achieving targets of <6.0 versus <6.5 percent (<42 versus <48 mmol/mol, respectively), treatment goals should be individualized and account for the risk of maternal hypoglycemia.

Assay — A Diabetes Control and Complications Trial-aligned assay should be used for measurement of A1C levels. (See "Measurements of chronic glycemia in diabetes mellitus", section on 'Standardization of the assay'.)

Frequency of monitoring — The frequency of A1C monitoring in pregnancy varies among clinicians. In nonpregnant patients, A1C is generally monitored two to four times per year, but the frequency is sometimes increased in pregnancy because of the alteration in red blood cell kinetics and the physiologic changes in glycemic parameters described above.

●The authors of this topic usually monitor A1C at four- to eight-week intervals to detect any drift in glycemic management, particularly in patients who cannot provide accurate BGM logs. More frequent monitoring with timely detection of higher-than-expected results enables prompt recognition and correction of otherwise inapparent dietary and insulin issues.

●Other UpToDate contributors report checking an A1C at the first prenatal visit and repeat the test only in patients who do not provide their BGM records, some obtain an A1C once per trimester, and some check it every 6 to 10 weeks.

●ACOG and ADA guidelines state that A1C measurement may occur as frequently as monthly [11,12].

When to test for ketonuria — Cells produce ketones when they are deprived of glucose and must switch to fatty acids as an energy source. Generally, this can occur in one of two ways: starvation, as inadequate food intake results in an inadequate circulating glucose supply, or adequate glucose in the circulation but inability of that glucose to enter cells due to inadequate serum insulin levels. Due to the accelerated transition to lipolysis that occurs in normal pregnancy, even an overnight fast can cause ketonuria, particularly in the third trimester [34]. Patients with nausea and vomiting may develop ketonuria; however, nausea and vomiting may also be symptoms of diabetic ketoacidosis (DKA), which is also characterized by ketonuria due to inadequate insulin levels. In patients with nausea and vomiting, it is important to be aware of both possibilities; checking serum ketones may be useful to make the correct diagnosis as quickly as possible so the correct treatment can be initiated.

In pregnant patients with preexisting diabetes (type 1 or 2), testing for ketonuria is reasonable if blood glucose values exceed 200 mg/dL (11.1 mmol/L) [35]; 10 to 30 percent of cases of DKA in pregnancy have been observed with blood glucose levels <250 mg/dL (13.9 mmol/L) [35]. Testing for ketonuria should also be performed during periods of illness or stress or if there are symptoms compatible with ketoacidosis, such as nausea, vomiting, and abdominal pain. In some countries, meters are available to measure capillary blood beta-hydroxybutyrate directly. Routine daily urine testing for ketones is not recommended unless there is a concern that patients are restricting carbohydrates beyond recommended targets.

Management of ketonuria and DKA — Patients with moderate to large ketonuria should alert their physician immediately. Additional insulin should be prescribed to prevent or reverse DKA, which is both a medical and an obstetric emergency since it is associated with risks to both the mother and the fetus; in a retrospective cohort study, the rate of fetal demise after DKA in pregnancy was 15 percent [36]. DKA is diagnosed when the triad of hyperglycemia, anion gap metabolic acidosis, and ketonemia is present. The most common precipitating factors are infection and inadequate insulin therapy, which can result from intentional or unintentional discontinuation as well as pump infusion set catheter occlusion or displacement or malfunction of an insulin pump. (See "Diabetic ketoacidosis in pregnancy".)

In the absence of hyperglycemia, ketonuria indicates a catabolic state and implies a negative caloric balance, which can happen after an overnight fast in late pregnancy; one study found 7 percent of participants without diabetes had fasting ketonuria in the third trimester [34]. This process may be accelerated in pregnant patients with diabetes [37]. In pregnant patients with diabetes, persistent ketonuria can be corrected by increasing caloric intake, particularly carbohydrates, matched to insulin dose adjustments. Ketonemia is not present when ketonuria occurs in this setting.

RISKS OF STRICT GLYCEMIC MANAGEMENT

Hypoglycemia — Normoglycemia during pregnancy reduces the frequency of maternal, obstetric, and neonatal complications; however, the risk of hypoglycemia is increased when normoglycemia is the therapeutic goal. On a case-by-case basis, providers should balance the potential improvement in some pregnancy outcomes with a glycated hemoglobin (A1C) target <6.0 percent (lower rates of preeclampsia and large for gestational age birth weight) against the higher risk for hypoglycemia at this target compared with an A1C target <6.5 or <7.0 percent.

Data to help providers with this decision are limited. In a 2016 systematic review, there were no clear maternal or neonatal benefits from attempting to achieve very strict fasting blood glucose targets (61 to 91 mg/dL [3.33 to 5.00 mmol/L]) versus moderate to strict fasting glucose targets (81 to 116 mg/dL [4.45 to 6.38 mmol/L]), and significantly more patients in the strict target groups had hypoglycemia [38]. However, only three small trials were included in the analysis and all were at high risk of bias.

The threshold for hypoglycemia in pregnant patients is also controversial; a value <63 mg/dL (3.5 mmol/L) has been proposed to avoid overclassification of hypoglycemia in asymptomatic patients since fasting blood glucose values are normally slightly lower in nondiabetic women during pregnancy [39,40]. (See "Hypoglycemia in adults without diabetes mellitus: Clinical manifestations, causes, and diagnosis".)

In contrast to hyperglycemia, there are no compelling data that hypoglycemia is teratogenic or otherwise harmful to the developing fetus [41,42]. Nevertheless, hypoglycemia poses a risk to the mother, and the potential for maternal injury could, in turn, injure the fetus.

Management — Treatment of symptomatic hypoglycemia with 15 grams of fast-acting carbohydrate should raise the blood glucose into the target range without inducing hyperglycemia. We prefer treatment with three to four glucose tablets to avoid overtreatment of hypoglycemia, but 4 ounces of fruit juice or 1 cup of milk may be used. Glucagon can be administered if the patient is unable to take carbohydrate orally. Users of predictive low glucose suspend and hybrid closed-loop (HCL) devices may only require 8 to 10 grams of carbohydrate to treat hypoglycemia because the predictive basal turn off feature should engage prior to hypoglycemia [43].

Patients should be instructed to retest their glucose level by blood glucose monitoring (BGM) after 15 minutes to ensure correction of hypoglycemia. Continuous glucose monitoring (CGM) should not be used to evaluate correction of hypoglycemia because of the 15-minute lag between plasma and interstitial glucose readings.

Worsening retinopathy — Improvement in glycemic management in pregnant (and nonpregnant) hyperglycemic patients can transiently worsen retinopathy, most commonly manifesting as increased formation of soft exudates [44,45]. The risk is related, in part, to the speed and magnitude of the reduction in chronic hyperglycemia [46]. Although intensive insulin therapy is associated with acute transient acceleration of retinopathy, maintaining good glycemic management slows the progression of retinopathy over time (figure 1).

Progression of retinopathy during pregnancy is strongly related to baseline prepregnancy retinal health. Retinopathy is common in patients with preexisting type 1 or type 2 diabetes. In a meta-analysis including 18 observational studies, in early pregnancy the prevalences of any diabetic retinopathy (DR) and proliferative DR were 52.3 and 6.1 percent, respectively; 15 percent of patients (95% CI 9.9-20.8) had a new diagnosis of DR [47]. The pooled rates of worsened nonproliferative and proliferative DR were 31.0 percent (95% CI 23.2-39.2) and 37.0 percent (95% CI 21.2-54.0), respectively; more than 1 in 20 patients experienced sight-threatening progression from nonproliferative DR to proliferative DR (pooled rate 6.3 percent; 95% CI 3.3-10.0). These results suggest close follow-up should be maintained during pregnancy to prevent vision loss. Retinopathy is discussed in more detail separately. (See "Diabetic retinopathy: Classification and clinical features", section on 'Worsening during pregnancy'.)

GENERAL APPROACH TO THERAPY — 

The approach to glucose management during pregnancy depends, in part, on the patient's prepregnancy antihyperglycemic regimen, as described below.

Patients on medical nutritional therapy prior to pregnancy — Patients with type 2 diabetes who have good glycemic management with medical nutritional therapy alone can remain on this therapy during pregnancy while closely monitoring glucose levels, as described above. (See 'Timing/frequency of BGM' above and 'Target blood glucose values' above.)

The majority, however, will not be able to achieve and maintain target glucose values without additional treatment. In these cases, we usually begin insulin therapy with a combination of lispro or aspart insulin and neutral protamine Hagedorn (NPH) or other long-acting insulin (see 'Insulin pharmacotherapy' below). We generally do not start noninsulin antihyperglycemic drugs for management of preexisting diabetes in pregnancy.

Patients on noninsulin antihyperglycemic agents prior to pregnancy

Overview — We usually discontinue noninsulin antihyperglycemic drugs and initiate insulin therapy to achieve adequate glucose levels. Ideally, this is done sufficiently in advance of conception to optimize glycemic management during the critical period of organogenesis early in the first trimester. In nonpregnant patients with type 2 diabetes, use of oral and injectable antihyperglycemic agents other than insulin is common; however, most experts believe that intensive insulin therapy is the only means of achieving the degree of glycemic management desirable throughout pregnancy in patients with type 1 and type 2 diabetes. The American Diabetes Association advises using insulin to manage type 1 diabetes in pregnancy and considers it the preferred agent for managing type 2 diabetes in pregnancy [12]. The American College of Obstetricians and Gynecologists (ACOG) also recommends insulin therapy and states that use of other agents (ie, metformin, glyburide) for glucose management of type 2 diabetes mellitus during pregnancy should be limited and individualized until more data confirming safety and efficacy become available [11].

The only noninsulin antihyperglycemic drugs used in pregnancy are metformin and glyburide; both cross the placenta, and some data suggest potential harm to offspring related to both agents (see 'Metformin' below and 'Glyburide' below). In a systematic review, periconception exposure to sulfonylureas, dipeptidyl peptidase-4 (DPP-4) inhibitors, sodium-glucose cotransport-2 (SGLT-2) inhibitors, and glucagon-like peptide-1 receptor agonists (GLP-1 RA) was not associated with a higher risk of congenital anomalies [48]. However, other than treatment with sulfonylureas, there are no or very limited data regarding these medications, alpha glucosidase inhibitors, meglitinides, thiazolidinediones, and amylin mimetics on pregnancy and neonatal outcomes.

The approach for switching patients from noninsulin antihyperglycemic agents to insulin is described separately. (See "Insulin therapy in type 2 diabetes mellitus", section on 'Designing an insulin regimen'.)

Patients on metformin or glyburide — Some patients with type 2 diabetes have excellent glycemic management on metformin at conception. The majority of these patients have overweight or obesity, often with insulin resistance or impaired insulin secretion related to polycystic ovary syndrome, which manifests as impaired glucose tolerance or type 2 diabetes. Metformin can be continued safely and effectively as the transition to insulin is initiated and until the dose of injected insulin is sufficient to achieve metabolic control [49-52]. An increased risk of congenital anomalies has not been observed following maternal use of metformin, but the majority of patients with type 2 diabetes will require supplemental insulin to meet glycemic targets. Concerns have been raised about continued use of metformin in pregnancy with regards to increased risks for small for gestational age (SGA) newborns and long-term impact on offspring (eg, childhood adiposity). (See 'Metformin' below.)

Similarly, patients with type 2 diabetes who are treated with glyburide should be transitioned to insulin as soon as feasible. We titrate insulin doses upward, while titrating glyburide downward to avoid hyperglycemia. An increased risk of congenital anomalies has not been observed following maternal use of glyburide. (See 'Glyburide' below.)

Metformin — We suggest not using metformin for the treatment of type 2 diabetes in pregnancy, given the lack of efficacy in improving pregnancy outcome in patients with type 2 diabetes, concerning animal data, and a signal suggesting long-term impact on offspring.

In meta-analyses, a significantly increased risk for major congenital anomalies was not observed when metformin was taken during the first trimester [53,54]; this finding was confirmed in a population-based study in the United States [55], a cohort study of nationwide registers in four Nordic countries [56], and the European congenital anomaly registry [57].

Metformin use improves maternal metabolic outcome; however, its effects on fetal growth are less clear.

●In a trial of 500 pregnant patients with type 2 diabetes randomly assigned to receive insulin alone or both insulin and metformin (1000 mg twice daily), combined treatment improved maternal glycemic management (A1C at 34 weeks 5.9 versus 6.1 percent [41.0 versus 43.2 mmol/mol]; mean glucose 109 versus 113 mg/dL [6.1 versus 6.3 mmol/L] ) and reduced insulin requirements (1.1 versus 1.5 units/kg/day at 34 weeks of gestation) and total gestational weight gain (7.2 versus 9.0 kg) [58].

When compared with newborns exposed in utero to insulin alone, newborns exposed to insulin and metformin had:

•Lower birth weight (3156 versus 3375 g)

•Lower rates of large for gestational age (LGA, >97^th percentile: 9 versus 15 percent) and macrosomia (≥4000 g: 12 versus 19 percent)

•Less neonatal adiposity

•A higher rate of small for gestational age (SGA, 13 versus 7 percent).

Other pregnancy and newborn outcomes (including rates of primary cesarean birth, hypertensive disorders of pregnancy, perinatal mortality, and serious neonatal morbidity) were similar for both groups. Thus, improvements in maternal glycemic management and lower rates of both LGA and macrosomia with metformin did not translate to improvements in rates of cesarean birth, preeclampsia, morbidity from shoulder dystocia, or neonatal hypoglycemia.

●A subsequent trial evaluated insulin plus placebo versus insulin plus metformin in nearly 800 adult pregnant patients with either preexisting type 2 diabetes or gestational diabetes diagnosed in early pregnancy [59]. Mothers in the metformin group had lower third-trimester mean A1C compared with the placebo group (5.97 versus 6.22; mean ratio 0.96, 95% CI 0.93-1.00) but testing was only performed in 39 percent of individuals for clinical care and was not included in the study design.

The composite adverse neonatal outcome was similar and over 70 percent in both groups. When compared with newborns exposed in utero to insulin alone, newborns exposed to insulin and metformin had:

•Lower birth weight (3089 versus 3244 g)

•Less LGA (>90^th percentile: 26 versus 36 percent)

Other pregnancy and newborn outcomes were similar in the two groups: SGA (8 versus 7 percent), cesarean birth (overall rate: 63 percent in both groups), preeclampsia (35 versus 30 percent), neonatal hypoglycemia (39 versus 42 percent), and mean neonatal fat mass (0.46 versus 0.50 kg). The trial was not powered to detect differences in the rate of cesarean birth performed for suspected macrosomia.

A possible association between metformin exposure and SGA needs to be explored further given the discordant results of the two randomized trials and an observational study [60]. An increased risk for SGA would be a concern, if real. SGA in other settings has been associated with adverse effects on neurodevelopment and cardiometabolic health, and in severe SGA, complete catch-up growth during childhood/adolescence may not occur (see "Fetal growth restriction (FGR) and small for gestational age (SGA) newborns", section on 'Long-term outcome'). Study results on the effect of metformin on neonatal adiposity have also been discordant (reduced risk versus no difference) and needs further study. A meta-analysis of use of metformin, insulin, or glyburide in patients with GDM (33 trials, nearly 5000 participants) demonstrated lower birth weight neonates with reduced lean mass in pregnancies treated with metformin (plus insulin if needed) compared with insulin alone [61]. Animal data suggest metformin may decrease skeletal muscle mass [62].

Long-term follow-up of offspring of mothers randomly assigned to receive metformin versus insulin or placebo for treatment of GDM (third trimester) or polycystic ovary syndrome (PCOS; throughout pregnancy) demonstrated that those exposed to metformin had increased body mass index (BMI) and childhood adiposity, which have long-term implications for adult health [63-65].

A study of rhesus monkeys treated with the equivalent of therapeutic doses of metformin throughout pregnancy demonstrated fetal serum levels that were bioequivalent to maternal levels. There was fetal bioaccumulation of metformin in multiple organs; of particular concern, the gastrocnemius muscle had more than 50 percent lower cross-sectional area and sections of the fetal kidney demonstrated delayed maturation and glomerular disorganization [62].

Additional data on the safety of metformin use in the third trimester are reviewed in detail separately. (See "Gestational diabetes mellitus: Glucose management, maternal prognosis, and follow-up", section on 'Oral antihyperglycemic medications'.)

Glyburide — While small observational studies in patients with preexisting diabetes and more extensive data in patients with gestational diabetes have reported effective glycemic management with glyburide, subsequent data have highlighted several limitations [66-68]. Meta-analyses of randomized trials comparing glyburide, metformin, and insulin in patients with gestational diabetes presented convincing evidence that glyburide performed less well than either metformin or insulin, with higher rates of both macrosomia and neonatal hypoglycemia [61,69]. A subsequent noninferiority randomized trial that compared glyburide with insulin in 900 patients with gestational diabetes failed to demonstrate noninferiority (ie, glyburide may be inferior to insulin). Glyburide performed less well on all three neonatal measures in the composite outcome (morbidity including macrosomia, hypoglycemia, and hyperbilirubinemia) [70]. These outcomes may be attributable to poor glycemic management as well as placental transfer of glyburide into fetal circulation. Older studies found minimal measurable glyburide in cord blood, but a contemporary study that used a more sensitive assay found cord blood glyburide levels to be highly variable and, on average, one-half of maternal glyburide levels [71].

Additional data on the safety of glyburide use in the third trimester are reviewed in detail separately. (See "Gestational diabetes mellitus: Glucose management, maternal prognosis, and follow-up", section on 'Oral antihyperglycemic medications'.)

Patients on multiple daily injection therapy prior to pregnancy — In the absence of other indications for therapeutic change, patients on multiple daily injection (MDI) therapy should continue this approach during pregnancy. We suggest a combination of lispro or aspart insulin and NPH insulin or insulin glargine during pregnancy. These insulins are safe and effective, and dosing can be adjusted frequently and quickly in response to variable caloric intake and the insulin resistance related to pregnancy. We prefer the pharmacokinetics of glargine for patients with type 1 diabetes and NPH for patients with type 2 diabetes. (See 'Type of insulin' below.)

Randomized trials to this point have provided no compelling evidence to support switching patients from MDI to an insulin pump before or during pregnancy. As an example, a prespecified analysis of data from the Continuous Glucose Monitoring in Women with Type 1 Diabetes in Pregnancy Trial (CONCEPTT) found that MDI users had better glycemic outcomes and lower rates of gestational hypertension, neonatal hypoglycemia, and neonatal intensive care unit admissions than pump users [72]. However, the trial participants chose their method of insulin delivery; thus, this difference may have been related to unmeasured confounders.

Patients on a continuous subcutaneous insulin infusion (insulin pump) prior to pregnancy — Patients using continuous subcutaneous insulin infusion (insulin pump) effectively prepregnancy can continue this therapy. Insulin requirements will increase in pregnancy. (See 'Continuous subcutaneous insulin infusion (insulin pump)' below.)

The use of a predictive low-glucose suspend insulin delivery system, also known as sensor-augmented insulin pump, is associated with reduction in hypoglycemia without increase in mean glucose in nonpregnant adults [73], adolescents, and children [74]. As yet, there are limited data in pregnancy to evaluate the impact on pregnancy outcomes. (See 'Continuous subcutaneous insulin infusion (insulin pump)' below.)

Evidence regarding pump use in pregnancy is limited:

●In a retrospective study of 646 pregnancies in 478 patients with type 1 diabetes, the use of an insulin pump was associated with lower A1C in the first and second trimester, but higher odds of large for gestational age (OR 1.65, 95% CI 1.06-2.58) and macrosomia (OR 1.81, 95% CI 1.03-3.18) when compared with MDI [75]. This study did not include patients who used insulin pumps with predictive low-glucose suspend or hybrid closed-loop (HCL) technology.

●A multicentered prospective cohort study in 112 pregnancies complicated by type 1 diabetes compared glycemic management and pregnancy outcomes in those using hybrid closed-loop (HCL) devices versus MDI plus CGM [76]. The HCL systems used were mostly Medtronic 780 G (80 percent); the remainder included Diabeloop and Tandem control IQ. Major findings were:

•There were no significant differences in A1C, time in range (TIR; 62 to 140 mg/dL [3.4 to 7.8 mmol/L]), or time above range (TAR; >140 mg/dL [7.8 mmol/L]) between the groups.

•HCL users spent less time below range (TBR; hypoglycemia <63 mg/dL [3.5 mmol/L] ) in the second and third trimester

•HCL users received higher total insulin dose in the second trimester (+0.13 international units/Kg/dL), and HCL therapy was associated with increased maternal weight gain during pregnancy (beta adjusted 3.20 kg, 95% CI 0.90-5.50).

•Newborns of HCL users had higher birthweight (beta adjusted 279.0 g, 95% CI 39.5-518.5) and higher risk for macrosomia (adjusted OR 3.18, 95% CI 1.05-9.67) when compared with MDI users; however, these associations were explained by differences in gestational weight gain and glycemic management.

No other pregnancy outcomes were significantly different between the groups.

MEDICAL NUTRITION THERAPY — 

Consensus recommendations for management of diabetes in pregnancy advise individualized medical nutrition therapy (MNT) supervised by a registered dietician with expertise in MNT during pregnancy [12,13,77]. A common approach focuses on dietary quality and can be achieved by consuming nutrient-dense whole foods (fruits, vegetables, legumes, whole grains) and healthy fats (ie, foods with omega-3 fatty acids such as nuts, seeds, and fish) [12,77]. These foods are less likely than processed foods to promote excessive gestational weight gain.

Goals of therapy — The optimal diet to manage glucose levels during pregnancy takes into account caloric intake, macronutrient distribution, and frequency of meals throughout the day. The goals of MNT are to:

●Provide adequate nutrient intake for maternal and fetal/neonatal health.

●Achieve and maintain normoglycemia.

●Provide adequate nutrition to achieve gestational weight gain within National Academy of Medicine (formerly the Institute of Medicine [IOM]) targets specific to prepregnancy body mass index (BMI).

●Provide appropriate food, physical activity, and behavioral education.

A synopsis of nutritional therapy for pregnant patients with diabetes is provided below. A detailed review of MNT for nonpregnant individuals with diabetes can be found separately. (See "Nutritional considerations in type 1 diabetes mellitus" and "Medical nutrition therapy for type 2 diabetes mellitus".)

Calorie requirements — We suggest caloric intake to achieve total gestational weight gain in the range recommended by the National Academy of Medicine (table 1) [78]. These recommendations are based on prepregnancy BMI. Excessive weight gain should be avoided as it contributes independently to risk for large for gestational age birth weight [79-81]. (See "Gestational weight gain".)

Caloric requirements for a singleton pregnancy are increased by an average of approximately 300 kcal/day above basal daily needs in nonpregnant patients [82]. Caloric recommendations in pregnancy may be based on total energy expenditure calculations using the Harris-Benedict equation for patients with a healthy BMI (calculator 1) [83] or Mifflin-St. Jeor equation for BMI in the overweight or obese categories [84]. The Joslin Diabetes Center uses the following approach as a starting point to estimate daily caloric requirements and also takes into account the patient's activity level, age, and gestational weight gain over time.

●Underweight – 30 kcal/kg in the first trimester and 36 to 40 kcal/kg in the second and third trimesters.

●Normal weight – 30 kcal/kg in the first trimester, 36 kcal/kg in the second trimester, and 36 to 38 kcal/kg in the third trimester.

●Overweight and obese – 24 kcal/kg throughout pregnancy.

Maternal obesity is associated with excessive fetal growth, independent of diabetes, as well as insulin resistance in both type 1 and type 2 diabetes. Weight loss during pregnancy in patients with obesity is not recommended; the National Academy of Medicine guidelines for gestational weight gain for BMI >30 kg/m² account for obligate increased mass due to placenta, amniotic fluid, fetus, uterus, breast tissue, and plasma volume expansion with minimal increase in adiposity [78]. Gestational weight gain and loss are discussed in detail separately. (See "Gestational weight gain".)

Macronutrient composition — Nutritional intake is divided among the macronutrient components to promote optimal glucose levels and avoid hypoglycemia and ketonemia. There are no randomized trials to guide optimal macronutrient composition during pregnancy complicated by diabetes; data in nonpregnant individuals suggest that dietary quality is more important for long-term health outcomes than macronutrient composition [85]. We agree with the American College of Obstetricians and Gynecologists' suggested guidelines [11]:

●Complex, high-fiber carbohydrates – 40 to 50 percent of total calories.

●Protein – 15 to 30 percent of total calories.

●Fats, primarily unsaturated – 20 to 35 percent of total calories.

Carbohydrates and protein provide 4 kcal/gram; fat provides 9 kcal/gram.

The food plan should be based on a nutrition assessment with guidance from the Dietary Reference Intakes. In absolute terms, the recommended dietary allowances and adequate intakes for all pregnant patients are 175 g of carbohydrate, 71 g of protein, and 28 g of fiber daily [12]. The diet should emphasize monounsaturated and polyunsaturated fats while limiting saturated fats and avoiding trans fats, and it should minimize simple carbohydrates, which will result in higher postprandial excursions. We support these recommendations for preexisting diabetes in pregnancy.

Postprandial blood glucose levels are directly dependent upon the quality and quantity of the carbohydrate content of a meal [86]. The postprandial glucose rise, therefore, can be blunted if the diet is carbohydrate controlled. Complex carbohydrates, such as those in legumes (peas, beans, lentils), whole grains, and vegetables, are more nutrient dense and raise postprandial blood glucose levels more slowly and less than simple sugars. The importance of nutritional counseling before and during pregnancy was highlighted by findings from a secondary analysis of the CONCEPTT study, which found that 46 percent of total daily carbohydrate intake in preconception and pregnant patients with type 1 diabetes came from simple and/or processed sources (eg, sugars, preserves, confectionery, biscuits, cakes) [87]. (See "Nutritional considerations in type 1 diabetes mellitus" and "Medical nutrition therapy for type 2 diabetes mellitus", section on 'Macronutrient composition'.)

We recommend carbohydrate intake of no less than 40 percent of daily calories and emphasize high-quality carbohydrate intake from minimally processed sources. Low-carbohydrate diets, including those designed to induce ketosis, have not been well studied in pregnancy. Carbohydrate restriction may impact embryonic and fetal development; two case-control studies have demonstrated an association with low-carbohydrate intake prior to pregnancy and an increased risk for neural tube defects, which was not completely attributable to reduced intake of folate-fortified grain products [88,89]. Animal data suggest ketogenic diets may be associated with changes in brain development [90,91]. Another concern is that carbohydrate restriction by necessity implies increased protein and/or fat intake; increased dietary fat intake from animal sources is, in turn, associated with increased insulin resistance [92].

Existing guidelines suggest a daily intake of at least 175 grams of carbohydrate to support maternal metabolic needs and fetal brain development as well as the placental energy requirements. Further study is needed to ensure this is sufficient for all patients [93].

Calorie distribution — For patients on a fixed daily insulin dose, a consistent pattern of carbohydrate intake with respect to time and amount improves glycemic management and reduces the risk of hypoglycemia. Macronutrient distribution should be individualized, but the following distribution works well for many patients:

●Breakfast – 10 to 20 percent of total calories. Because insulin resistance is greatest in the morning, this is the smallest meal, and carbohydrate intake is limited to maintain postprandial normoglycemia.

●Lunch – 20 to 30 percent of total calories.

●Dinner – 30 to 40 percent of total calories.

●Snacks – Up to 30 percent of total calories two to three hours after each meal to prevent hypoglycemia. The need for snacking is based on caloric needs. Bedtime snacks are often needed to minimize nocturnal hypoglycemia and should be mostly protein and fat, rather than carbohydrate, to minimize hyperglycemia. A meal plan that includes frequent snacking will need to be matched by insulin delivery, likely requiring more boluses with a pump and potentially more injections of very rapid-acting insulin with multiple daily injection regimens.

Alternatively, carbohydrate allocation can be the dietary focus [94]:

●30 to 45 grams at breakfast

●45 to 60 grams at lunch and dinner

●15 grams for each snack

We recommend avoiding intermittent fasting, also known as time-restricted eating, during pregnancy.

Micronutrient composition — A prenatal multivitamin is prudent given some data that nutritional intake in pregnant patients with diabetes may be insufficient in calcium, vitamins C and E, copper, magnesium, and zinc [13]. Furthermore, a prenatal vitamin containing iron and folic acid appeared to have favorable effects on several birth outcomes in a meta-analysis of randomized trials [95]. Iodine requirements increase during pregnancy and lactation; in patients who are not taking levothyroxine, prenatal vitamins should also include iodine (150 to 250 mcg) [96]. (See "Nutrition in pregnancy: Dietary requirements and supplements", section on 'Iodine'.)

The prenatal multivitamin should contain at least 0.4 mg of folic acid and preferably 0.8 to 1 mg if the patient has other risk factors for having a child with a neural tube defect [11,12]. Some guidelines recommend high-dose (4 or 5 mg) folic acid supplements at least one month prior to conception and continuing through the first trimester for patients with preexisting diabetes because these individuals are at increased risk of offspring with an open neural tube defect [97,98]. However, there are no data to support these pharmacologic doses, and high levels of folic acid may have adverse effects. (See "Preconception and prenatal folic acid supplementation", section on 'Preexisting diabetes' and "Preconception and prenatal folic acid supplementation", section on 'Potential risks'.)

Non-nutritive sweeteners — All non-nutritive sweeteners should be used in moderation. Aspartame should not be used in patients with high phenylalanine levels or phenylketonuria. (See "Nutrition in pregnancy: Assessment and counseling", section on 'Use of non-nutritive sweeteners'.)

ROLE OF PHYSICAL ACTIVITY — 

The patient and provider should discuss the role of physical activity during pregnancy [13]. In patients with gestational diabetes mellitus, physical activity may provide several maternal benefits, including reduced risk for excessive gestational weight gain, improved glucose management [99], and reduced severity of low back and pelvic pain [100]. More studies are needed to evaluate the impact of exercise in pregnant patients with preexisting diabetes. Overall data are insufficient to recommend a specific type, volume, or timing of activity in pregnant patients with preexisting diabetes [101]. Types of exercise and strategies to avoid hypoglycemia should be informed by the patient's prepregnancy exercise preferences and patterns [13,94,102,103]. There are rare absolute and relative contraindications to physical activity, which are addressed separately. (See "Exercise during pregnancy and the postpartum period".)

Hypoglycemia may be more severe and more frequent in pregnancy compared with the nonpregnant state, so a snack including carbohydrates prior to exercise may be needed [13].

INSULIN PHARMACOTHERAPY — 

Effective use of insulin requires an understanding of the major variables that affect glucose management: the insulin preparation, the size of the subcutaneous depot, injection technique, the site of injection, and subcutaneous blood flow. Issues with respect to insulin therapy in pregnant patients are addressed below. A detailed discussion on insulin management in nonpregnant patients with diabetes can be found separately. (See "General principles of insulin therapy in diabetes mellitus" and "Management of blood glucose in adults with type 1 diabetes mellitus" and "Insulin therapy in type 2 diabetes mellitus".)

Insulin requirements in pregnancy — Total daily insulin requirements vary by gestational age in observational studies:

●Between gestational weeks 3 and 9, insulin requirements rise [104].

●From gestational weeks 9 to 16 [105], a significant decline in insulin requirements can occur.

●From weeks 16 to 36, insulin requirements rise by approximately 5 percent/week [104-106].

●At the end of the third trimester, insulin requirements plateau or even decrease somewhat. (See 'Implications of falling insulin requirement' below.)

The doubling of total daily insulin dose by late pregnancy compared with preconception doses is manifested by increases in basal insulin and bolus insulin, including strengthening of both the insulin to carbohydrate ratio and sensitivity factor [107].

In early pregnancy, the changes in insulin requirements are multifactorial and reflect decreased caloric intake in patients with nausea and vomiting of pregnancy, pregnancy-related changes in glucose homeostasis and insulin sensitivity, as well as provider and patient efforts to improve glucose management. In the second and third trimesters, glucose uptake by the fetus and placenta decreases fasting glucose levels, while placental hormones increase insulin resistance and promote postprandial hyperglycemia.

Type of insulin — The approximate times of onset, peak activity, and duration of action of various insulins are shown in the table (table 2).

Use of insulin preparations with low antigenicity will minimize the transplacental transport of metabolically active insulin/anti-insulin antibody complexes: Human insulin is the least immunogenic of the commercially available preparations.

●Rapid- or short-acting insulins – We use the rapid-acting lispro and aspart in pregnancy. The three rapid-acting insulin analogs (lispro, aspart, glulisine) are comparable in immunogenicity to the short-acting regular human insulin (clear zinc insulin), but only lispro and aspart have been investigated in pregnancy and shown to have acceptable safety profiles, minimal transfer across the placenta, and no evidence of teratogenesis [108-112]. These two insulin analogs also reduce the risk of postprandial glycemic excursion and delayed postprandial hypoglycemia compared with regular human insulin [111]. In an observational study, use of lispro resulted in similar pregnancy outcomes as regular human insulin but with increased patient satisfaction [112].

●Intermediate-acting insulin – The safety and efficacy of neutral protamine Hagedorn (NPH) in pregnancy are supported by abundant observational data published over decades. Importantly, as an intermediate-acting insulin, doses can be adjusted frequently and quickly in response to variable caloric intake and insulin sensitivity in pregnant patients.

We prefer to use the intermediate-acting NPH insulin during pregnancy for patients with type 2 diabetes. However, if a patient on a long-acting insulin has good glycemic management prepregnancy or at their first prenatal visit, we do not switch them to NPH just because they are planning to conceive or are pregnant.

●Long-acting insulins – Long-acting insulin analogs can be useful for management of patients with type 1 diabetes who need stable basal coverage (see "General principles of insulin therapy in diabetes mellitus"). A disadvantage of the longer acting insulin analogs is that their activity remains fairly constant over approximately 24 hours, and this level of activity may not be ideal to optimize both daytime and nighttime basal needs. Slow pharmacokinetics, in contrast with intermediate-acting insulin, may also be problematic during the third trimester when frequent, relatively large changes in insulin dose may be required.

•Glargine – There are no randomized trials comparing glargine with NPH or degludec. A meta-analysis of observational data from 331 pregnancies with glargine exposure during the first, second, and/or third trimester showed no statistical increase in any maternal or neonatal adverse events compared with the use of NPH [113], providing some reassurance about its safety. Placental transfer is probably absent to minimal [114-117]. Glargine has a 12-hour half-life, so basal adjustments can be made every three days.

•Degludec – Insulin degludec is an ultralong-acting insulin analog that lasts up to 42 hours and is administered once daily. A noninferiority trial of degludec versus detemir found no increase in maternal safety events during pregnancy [118]. Although the trial was not powered to look at pregnancy and neonatal outcomes and did not report on delayed postpartum hypoglycemia, the degludec group had higher rates of preeclampsia, severe maternal hypoglycemia, preterm birth, and large for gestational age, which were not statistically significant. Because dose titration of degludec takes longer than with shorter-acting insulins and there are minimal safety and efficacy data, we are less likely to recommend insulin degludec in pregnancy. However, in patients who present to their initial prenatal visit already using insulin degludec, we suggest continuing it in pregnancy to avoid destabilization of glucose control while transitioning to a different basal insulin. The long half-life of 25 hours precludes adjustment more frequently than every five to six days. Detemir is no longer available.

Insulin dosing

Multiple daily injection dosing — Multiple daily injection (MDI) regimens provide basal insulin coverage with intermediate- or long-acting insulin and boluses with rapid-acting insulin to cover prandial insulin needs. Depending on the type of diabetes and insulins used, three to five injections daily are usually required to achieve glucose and A1C goals. Prandial insulin should be injected 10 to 15 minutes before the start of the meal to blunt the rapid rise in blood glucoses after meal ingestion. In the event of hypoglycemia prior to the meal, treatment of hypoglycemia (see 'Hypoglycemia' above) should be initiated, and prandial insulin and meal consumption should be delayed until hypoglycemia has resolved.

●Type 1 diabetes – Insulin requirements during the first trimester are similar to those prior to pregnancy in patients with type 1 diabetes. Dosing is continually adjusted based on blood glucose monitoring (BGM) and continuous glucose monitoring (CGM),values. (See 'Assessing glycemic management' above.)

The average total daily insulin requirement in pregnant patients with type 1 diabetes is 0.7 units/kg in the first trimester, often increasing to 0.8 units/kg for weeks 13 to 28, 0.9 units/kg for weeks 29 to 34, and 1 unit/kg for weeks 35 to term; however, the range of change in insulin requirements is broad. In one study of patients with type 1 diabetes, the mean daily insulin requirement increased by 52 units from prepregnancy to the end of the third trimester, related in part to weight gain [106]. Patients with obesity may need initial doses as high as 1.5 to 2 units/kg to overcome the insulin resistance that results from the combination of pregnancy and adiposity [119].

Approximately 50 percent of the total insulin dose is administered as a rapid-acting insulin (lispro or aspart) before each meal, and the other 50 percent is administered as basal coverage using an intermediate- or long-acting insulin (eg, NPH or glargine) twice daily.

A weight-based strategy can be used to estimate insulin requirements, with each premeal dose approximately 0.15 times the current weight in kg [120]. As an example, an 80 kg pregnant patient with diabetes in the third trimester would take 12 units of lispro or aspart before each meal. The basal dose is calculated as 0.45 times the patient's weight in kg, so this 80 kg patient would take 18 units of NPH twice daily. The first NPH dose is given before breakfast, and the second dose is given either before dinner with a rapid-acting insulin or at bedtime: whichever works best for maintaining overnight glucose goals and avoiding nocturnal hypoglycemia.

For patients who do not maintain a fixed carbohydrate intake, carbohydrate counting can help guide prandial insulin dosing. In a study of 101 patients with type 1 diabetes over the course of pregnancy, the mean insulin to carbohydrate ratio strengthened over the course of gestation from approximately 1:10 to 1:5 at breakfast, 1:10 to 1:8 at lunch, and 1:13 to 1:6 at dinner [107]. Others have suggested an increased insulin to carbohydrate ratio of 20 to 30 percent over the course of pregnancy [121].

●Type 2 diabetes – Insulin requirements during the first trimester are similar to those prior to pregnancy in patients with type 2 diabetes, but then requirements increase. During the second half of pregnancy, insulin requirements increase disproportionately in patients with type 2 diabetes compared with those with type 1 diabetes, likely due to additional insulin resistance at baseline. In one study, for example, insulin doses in the third trimester were 1.6 units/kg/day in type 2 diabetes compared with 1.2 units/kg/day in type 1 diabetes [122]. Dosing is continuously adjusted based on BGM values. (See 'Assessing glycemic management' above.)

Continuous subcutaneous insulin infusion (insulin pump) — In general, if a patient is using continuous subcutaneous insulin infusion (CSII with an insulin pump) effectively prepregnancy, there is no need to discontinue this approach. However, even patients with hemoglobin A1C in the preconception target range (<6.5 percent) can expect a two-fold increase in total insulin requirements from preconception to 36 weeks of gestation [104]. Increasing insulin requirements can affect the frequency of infusion set and reservoir changes; therefore, patients using an insulin pump may need assistance navigating insurance coverage.

We generally do not start patients on insulin pumps during pregnancy because they have not been proven to provide superior pregnancy outcomes, there are logistical challenges to transitioning from MDI to pumps, and the glucose targets/thresholds set by most semi-automated pumps are not appropriate for pregnant patients. Although some clinicians believe use of a pump achieves optimal glycemic management during pregnancy, meta-analyses of randomized trials and cohort studies have not demonstrated a clear improvement in maternal or fetal outcome [123-125]. In one such meta-analysis, which included data from 7824 pregnancies, pump use was associated with modest increases in gestational weight gain (weighted mean difference 1.02 kg, 95% CI 0.41-1.62) and risk for large for gestational age (LGA) birth weight (relative risk 1.16, 95% CI 1.07-1.24) [126]. Other potential risks of pump therapy include technological complications, such as accidental catheter disconnection with risk for hyperglycemia and ketoacidosis.

If CSII by pump is used, most pregnant patients require at least two to three basal infusion rates in a 24-hour period, including an increased rate in the early morning hours (5 to 9 AM) to counteract the increased release of the anti-insulin hormones cortisol and growth hormone ("dawn phenomenon"). Insulin to carbohydrate ratios for preprandial dosing may need to be strengthened by 30 to 50 percent over gestation. The sensitivity factor may also need to be strengthened over gestation, although this change may not be as marked [121,127]. (See "Management of blood glucose in adults with type 1 diabetes mellitus", section on 'Choice of insulin delivery' and "Continuous subcutaneous insulin infusion (insulin pump)".)

CSII may be used with intermittent BGM with or without a CGM device (sensor-augmented insulin pump). A multicenter randomized trial (AiDAPT) of hybrid closed-loop (HCL) therapy beginning at 11 weeks found that this approach improved maternal glycemic management compared with standard insulin therapy in 124 pregnant patients with type 1 diabetes and baseline A1C ≥6.5 percent [128]. Unlike many commercially available systems in the United States, the predictive control algorithm (CamAPS) used in the trial allowed for customization of glycemic targets appropriate to pregnancy. The recommended targets were 100 mg/dL (5.5 mmol/L) in early pregnancy and 81 to 90 mg/dL (4.5-5.0 mmol/L) beginning at 16 to 20 weeks and continuing through delivery. Major findings were:

●The closed-loop group had higher mean percentage of time with maternal glucose level in the target range (68 versus 56 percent; mean adjusted difference: 10.5 percentage points, 95% CI 7.0-14.0) and a substantial reduction in the mean percentage of time above range (29 versus 41 percent; mean adjusted difference: -10.2 percentage points, 95% CI -13.8 to -6.60).

●The percentage of time below range and below 54 mg/dL (3 mmol/L) was similar for both groups. Safety events were not significantly different between groups, with a similar number of episodes of severe hypoglycemia (six versus five) and diabetic ketoacidosis (one in each group). The rate of device-related adverse events in the closed-loop system was 24.3 per 100 person-years, with seven events related to closed-loop use and seven to the continuous glucose monitor.

●Although fewer cases of hypertensive disorders were observed in the treatment group (12 versus 25), the sample size was insufficient to detect statistically significant differences in the effect of the intervention on maternal and neonatal outcomes. A larger trial is needed to determine whether a true improvement in pregnancy and neonatal outcomes exists.

These data support the use of HCL systems with pregnancy-specific targets to improve glycemic management during pregnancy. The CamAPs FX algorithm has been approved by the US Food and Drug Administration (FDA), but a compatible pump remains under review. No HCL pump in the US has FDA approval for use in pregnancy. Other commercially available HCL pumps in the United States, which do not have pregnancy-specific therapeutic targets, may not achieve similar results.

The CRISTAL Study was an open label randomized trial comparing the Minimed 780 G advanced hybrid closed-loop (AHCL) pump set to target 100 mg/dl (5.5 mmol/L) versus standard therapy (MDI, standalone insulin pump or sensor augmented insulin pump with predictive suspension of insulin infusion at or before a low glucose threshold), with both treatment groups using CGM [129]. Major findings included:

●No significant difference in overall TIR (primary outcome), which increased from 60.5 percent to 66.5 percent in the AHCL pump therapy group and from 57.6 percent to 63.2 percent in the standard insulin therapy group (adjusted mean difference 1.88 percentage points, 95% CI –0.82 to 4.58).

●Higher overnight TIR in the AHCL pump therapy group (adjusted mean difference 6.58 percentage points, 95% CI 2.31-10.85).

●Lower TBR overall (adjusted mean difference –1.34 percentage points, 95% CI –2.19 to –0.49) and overnight (adjusted mean difference –1.86 percentage points, 95% CI –2.90 to –0.81) in the AHCL pump group.

●The proportion of participants with severe hypoglycemia (participant-reported and defined as requiring third-party assistance) (13 versus 10.2 percent) and the A1C achieved (6.2 versus 6.1 percent) was similar in both groups.

●NICU admissions due to neonatal hypoglycemia occurred less frequently in the AHCL group (14.3 versus 63.6 percent), but there were no significant differences in preterm birth rates, cesarean births, or other neonatal complications (birth trauma, shoulder dystocia, hyperbilirubinemia with need for phototherapy, respiratory distress).

These data support the safety of ACHL therapy.

In our experience, the use of other HCL-loop pumps that have therapeutic targets above the goals for pregnancy may be helpful in reducing hypoglycemia and correcting nocturnal hyperglycemia, thus providing safety and stability compared with use of the pump in open loop. As with all forms of insulin delivery, close monitoring is required to ensure adequacy of diabetes management. Insulin pumps with predictive low-glucose suspend technology that are able to achieve overnight and fasting targets of 80 mg/dL (4.4 mmol/L) while reducing hypoglycemia continue to be a preferred option.

Popular HCL pumps available in the US require special considerations for use during pregnancy. The Tandem Tslim X2 (or Mobi) with Control IQ targets should be programmed for use in sleep activity mode 24 hours/day to establish a target glucose range of 112 to 120 mg/dL (6.2 to 6.7 mmol/L). The Control IQ algorithm is unable to deliver more than 3 units/hour of basal insulin while in the target range. If higher basal rates are needed, lispro U200 insulin can be used to double the dosing provided by the algorithm. The Omnipod 5 (OP5) algorithm uses data from the last three pod changes to make automated basal insulin dose adjustments, which may not keep up with the increasing insulin requirements of pregnancy; in our experience, this can frequently result in nocturnal hyperglycemia. Some individuals may need to switch from closed (automated) mode to open (manual) mode, particularly in the second half of pregnancy.

Assessing barriers in patients with suboptimal glucose levels — Patients with difficulty achieving glycemic targets during pregnancy should have careful individualized assessment for financial, social, and educational barriers with connection to services, as appropriate. Diabetes-related outcomes are known to be associated with multiple social determinants of health, including socioeconomic status, physical and food environment, education, and access to health care as poverty has been associated with poor glycemic management in several retrospective studies [130]. In many US states, people with low income do not have access to health insurance and therefore do not have optimal diabetes management prior to pregnancy. In one US state with near-universal insurance coverage, pregnant people with type 1 or 2 diabetes covered by Medicaid (public health insurance program for people with low income) were still less likely than those with private insurance to achieve A1C <6 percent in early pregnancy (37.2 versus 58.4 percent) [131]. Similarly, a retrospective analysis observed that patients living in a census tract with a high social vulnerability index (SVI) were less likely to achieve hemoglobin A1C target <6 percent in the third trimester, with each 0.1 increase in SVI associated with a nearly 50 percent decreased likelihood of A1C <6 percent (adjusted risk ratio [aRR] 0.53; 95% CI 0.36-0.77) [132]. While data are limited in pregnancy, there are also well-characterized racial inequities in access to technology outside of pregnancy, with data from one large quality improvement collaborative demonstrating lower rates of both CGM and insulin pump use in non-Hispanic Black youth with type 1 diabetes when compared with other racial groups [133]. These gaps were associated with a lower likelihood of achieving A1C <7 percent among minoritized individuals when compared with non-Hispanic white individuals (19 versus 24 percent, respectively).

Although data on the impact of health literacy on pregnancy outcomes in people with diabetes are limited, in a pilot study of patients given access to mobile technology to support diabetes care in pregnancy, digital literacy was associated with engagement [134].

Implications of falling insulin requirement — Insulin requirements normally decrease after 35 weeks of gestation, more commonly with preexisting than gestational diabetes mellitus [104,106,135-138].

Falling insulin requirements, specifically a ≥15 percent decrease, can be due to placental insufficiency, as well as decreased maternal consumption, vomiting, increased fetal demand for glucose, or increased maternal sensitivity to insulin in the fasting state [106,136,139-143]. Nonetheless, good pregnancy outcome has been reported with falls in insulin requirements of up to 30 percent [106,138,144,145]. A decreasing insulin requirement should prompt a thorough evaluation of maternal and fetal well-being because of the association with placental insufficiency in some studies, but it alone is not an indication for delivery.

INTRAPARTUM GLYCEMIC MANAGEMENT — 

(See "Preexisting (pregestational) and gestational diabetes: Intrapartum and postpartum glucose management".)

OBSTETRIC MANAGEMENT — 

Other important considerations of the care for pregnant patients with diabetes, including strategies (eg, low-dose aspirin) to reduce risk for obstetric complications such as preeclampsia, are addressed separately. (See "Preexisting (pregestational) diabetes mellitus: Obstetric issues and pregnancy management".)

SOCIETY GUIDELINE LINKS — 

Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diabetes mellitus in pregnancy".)

INFORMATION FOR PATIENTS — 

UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Care during pregnancy for people with type 1 or type 2 diabetes (The Basics)")

●Beyond the Basics topics (see "Patient education: Care during pregnancy for patients with type 1 or 2 diabetes (Beyond the Basics)" and "Patient education: Glucose monitoring in diabetes (Beyond the Basics)" and "Patient education: Hypoglycemia (low blood glucose) in people with diabetes (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Glucose monitoring

•Blood glucose monitoring (BGM) should be performed, as needed, to achieve glycemic targets. BGM is generally performed before each meal, one or two hours after each meal, and at bedtime. If nocturnal hypoglycemia is suspected, the patient should also monitor their blood glucose concentration during the night. (See 'Blood glucose monitoring' above.)

•Continuous glucose monitoring (CGM) should be used as an adjunct to BGM in individuals with type 1 diabetes because it improves outcomes. CGM may be considered in patients with type 2 diabetes. No data support CGM alone for glucose assessment in pregnancy. (See 'Continuous glucose monitoring systems' above.)

●Glucose targets

•We agree with the American College of Obstetricians and Gynecologists (ACOG) and the American Diabetes Association (ADA) glucose targets (see 'Target blood glucose values' above):

-Fasting, preprandial, and nocturnal glucose concentration 70 to 95 mg/dL (3.9 to 5.3 mmol/L) and

-One-hour postprandial glucose 110 to 140 mg/dL (6.1 to 7.8 mmol/L) or

-Two-hour postprandial glucose 100 to 120 mg/dL (5.6 to 6.7 mmol/L)

-Mean capillary glucose concentration 100 mg/dL (5.6 mmol/L)

•For patients using CGM, the target glucose range is 63 to 140 mg/dL (3.5 to 7.8 mmol/L), and the time in range goal is >70 percent. (See 'Continuous glucose monitoring systems' above.)

●A1C

•We suggest measuring glycated hemoglobin (A1C) at least each trimester, acknowledging that A1C levels vary based on changing red cell turnover, making second- and third-trimester values potentially less reliable indicators of true glycemic management. However, A1C levels are associated with risk for fetal outcomes and thus are important metrics before and during pregnancy. (See 'Glycated hemoglobin (A1C)' above.)

•We individualize the target A1C. Although a target A1C level of <6 percent (42 mmol/mol) throughout pregnancy is consistent with ACOG clinical guidelines, in clinical practice, this may be difficult to achieve without marked hypoglycemia; the ADA recommends target A1C of as high as <7 percent (53 mmol/mol) if necessary to prevent significant hypoglycemia. The provider must consider, on a case-by-case basis, whether the benefits (including lower risk for preeclampsia) with a target A1C <6.0 percent rather than <6.5 or <7.0 percent warrant the increased risk of hypoglycemia. (See 'Target A1C level' above.)

●Ketonuria – In patients with type 1 diabetes, urinary ketones should be measured during illness or when any blood glucose value is over 200 mg/dL (10 mmol/L). Ten to 30 percent of cases of diabetic ketoacidosis in pregnancy have been observed with blood glucose levels <250 mg/dL (13.9 mmol/L). (See 'When to test for ketonuria' above.)

●Medical nutritional therapy – Consensus recommendations for management of diabetes in pregnancy advise individualized medical nutrition therapy (MNT) supervised by a registered dietician with expertise in MNT during pregnancy. A balance of macronutrients is recommended with at least 175 grams of carbohydrate daily to meet glucose requirements of the maternal brain, fetal brain, and placenta. Caloric intake should enable total gestational weight gain in the range recommended by the National Academy of Medicine (table 1). (See 'Medical nutrition therapy' above.)

●Use of insulin and oral antihyperglycemic drugs

•Patients with type 1 diabetes require insulin therapy delivered as either continuous subcutaneous insulin infusion (CSII with an insulin pump) or multiple daily injections (MDI).

-For CSII, we use either lispro or aspart insulin. We generally do not initiate CSII with an insulin pump during pregnancy, but patients using a pump effectively prepregnancy can continue this therapy during pregnancy. Hybrid closed-loop (HCL) insulin pumps are evolving with some studies demonstrating reduced time below range on CGM compared with standard care. An FDA-approved CamAps algorithm that uses the glucose targets recommended for pregnancy improves CGM time in range, but usage is not widespread pending FDA approval of a compatible insulin pump. (See 'Patients on a continuous subcutaneous insulin infusion (insulin pump) prior to pregnancy' above and 'Continuous subcutaneous insulin infusion (insulin pump)' above.)

-For MDI, we use lispro or aspart as bolus insulin in combination with either neutral protamine Hagedorn (NPH) insulin or insulin glargine for basal requirements.

Prandial insulin should be injected 10 to 15 minutes before the start of the meal to blunt the rapid rise in blood glucose after meal ingestion; this may need to shift to 20 to 30 minutes before the start of the meal in later pregnancy. In the event of hypoglycemia prior to the meal, treatment of hypoglycemia should be initiated, and prandial insulin and meal consumption should be delayed until hypoglycemia has resolved. (See 'Patients on multiple daily injection therapy prior to pregnancy' above and 'Insulin pharmacotherapy' above.)

•For patients with type 2 diabetes who are not able to achieve and maintain target glycemic levels with medical nutritional therapy alone, we suggest insulin therapy rather than oral antihyperglycemic agents (Grade 2C). (See 'Patients on medical nutritional therapy prior to pregnancy' above and 'Insulin pharmacotherapy' above.)

•For patients with type 2 diabetes on metformin or glyburide before pregnancy with good glycemic management, we transition them to insulin as early as feasible in the first trimester. We discontinue other oral antihyperglycemic or non-insulin injectable agents and initiate insulin therapy, as needed, to achieve adequate metabolic control. (See 'Patients on noninsulin antihyperglycemic agents prior to pregnancy' above and 'Insulin pharmacotherapy' above.)

●Management of hypoglycemia – We treat symptomatic hypoglycemia with 15 grams of fast-acting carbohydrate as it should raise the blood glucose into the target range without inducing hyperglycemia. Alternatives include measured 4 ounces of fruit juice or 1 cup of milk. Glucagon can be administered if the patient is unable to take carbohydrate orally.

Patients should be instructed to retest their glucose level by BGM after 15 minutes to ensure correction of hypoglycemia. CGM should not be used to evaluate correction of hypoglycemia due to the delay between blood glucose and interstitial fluid glucose levels. (See 'Hypoglycemia' above.)

●Role of exercise – Moderate low-risk exercise has several benefits and can be performed during pregnancy in patients who have no medical or obstetric contraindications to this level of physical activity. (See 'Role of physical activity' above.)

ACKNOWLEDGMENTS — 

The UpToDate editorial staff acknowledges Michael F Greene, MD, Rhonda Bentley-Lewis, MD, and Vincenzo Berghella, MD, who contributed to earlier versions of this topic review.