Chapter 60. Diabetes Mellitus. Will Obs

 Diabetes Mellitus

BS. Nguyễn Hồng Anh

According to the Centers or Disease Control and Prevention

(2020), nearly 27 million adults in the United States have been

diagnosed with diabetes. Another 7.3 million are suspected to

be undiagnosed, and an estimated 88 million have prediabetes.

Reasons or these substantial rates include an aging population,

which is more likely to develop type 2 diabetes; population

growth within minority groups at particular risk or type 2

diabetes; and a dramatic rise in obesity rates. In 2019, almost

three in 10 women were considered obese prior to becoming

pregnant (Driscoll, 2020). Te strong relationship between diabetes and the current obesity epidemic in the United States

underlines the critical need or diet and liestyle interventions

to change the trajectory o both.

TYPES OF DIABETES

In nonpregnant individuals, the type o diabetes is based on its

presumed pathogenesis and its maniestations. Absolute insulin

deciency, which generally is autoimmune in etiology, characterizes type 1 diabetes. In contrast, insulin resistance, relative

insulin deciency, or elevated glucose production characterizes

type 2 diabetes (Table 60-1). Both types are generally preceded

by a period o abnormal glucose homeostasis oten reerred to

as prediabetes. Pancreatic β-cell destruction can begin at any

age, but type 1 diabetes is clinically apparent most oten beore

age 30. ype 2 diabetes usually develops with advancing age

but is increasingly identied in younger obese adolescents.

Etiological overlap in diabetes subtypes is well established and

has led to the proposal o a single classication system centered on β-cell unction along with the concept o individualized treatment strategies (World Health Organization, 2020).

Other orms o diabetes include maturity-onset diabetes of the

young (MODY). Te more common MODY type is in obese

adolescents (ODAY Study Group, 2021). Te less common

orm is an autosomal dominant condition and characterized

by mild diabetes diagnosed in adolescence or young adulthood

(Udler, 2020).

■ Classification During Pregnancy

Diabetes is the most common medical complication o pregnancy. Women can be separated into those diagnosed with

diabetes beore pregnancy—pregestational or overt diabetes, and

those diagnosed during pregnancy—gestational diabetes.

Te proportion o pregnancies complicated by diabetes more

than doubled between 1994 and 2008, ater which rates have

leveled (Deputy, 2018; Jovanovič, 2015). In 2018, 7.6 percent

o pregnant women in this country had some orm o diabetes (Martin, 2019). Prevalence is highest among non-Hispanic

TABLE 60-1. Etiological Classification of Diabetes Mellitus

Type 1: β-Cell destruction, usually absolute insulin deficiency

Immune-mediated

Idiopathic

Type 2: Ranges from predominantly insulin resistance to predominantly an insulin secretory

defect with insulin resistance

Other types

Genetic mutations of β-cell function: MODY 1–6, others

Genetic defects in insulin action

Genetic syndromes: Down, Klinefelter, Turner, others

Diseases of the exocrine pancreas: pancreatitis, cystic fibrosis

Endocrinopathies: Cushing syndrome, pheochromocytoma, others

Drug or chemical induced: glucocorticosteroids, thiazides, β-adrenergic agonists, others

Congenital infections: rubella, cytomegalovirus, coxsackievirus

Gestational diabetes (GDM)

MODY = maturity-onset diabetes of the young.

Adapted from American Diabetes Association, 2020; Powers, 2018

blacks, Mexican-Americans, Puerto Rican-Americans, and Native

Americans (Powers, 2018).

■ White Classification

Until the mid-1990s, the classication system o Priscilla White

(1978) or diabetic pregnant women was the linchpin o management. oday, the White classication is used less oten but

still provides simple, useul inormation on pregnancy risks and

prognosis (Bennett, 2015). Because most currently cited literature also contains data rom these older classications, the one

previously recommended by the American College o Obstetricians and Gynecologists (1986) is shown in Table 60-2.

Te American College o Obstetricians and Gynecologists

(2019a) no longer recommends the White classication. Te

current ocus is whether diabetes antedates pregnancy or is rst

diagnosed during pregnancy. Many now recommend the system

proposed by the American Diabetes Association (Table 60-3).

TABLE 60-1. Etiological Classification of Diabetes Mellitus

Type 1: β-Cell destruction, usually absolute insulin deficiency

Immune-mediated

Idiopathic

Type 2: Ranges from predominantly insulin resistance to predominantly an insulin secretory

defect with insulin resistance

Other types

Genetic mutations of β-cell function: MODY 1–6, others

Genetic defects in insulin action

Genetic syndromes: Down, Klinefelter, Turner, others

Diseases of the exocrine pancreas: pancreatitis, cystic fibrosis

Endocrinopathies: Cushing syndrome, pheochromocytoma, others

Drug or chemical induced: glucocorticosteroids, thiazides, β-adrenergic agonists, others

Congenital infections: rubella, cytomegalovirus, coxsackievirus

Gestational diabetes (GDM)

MODY = maturity-onset diabetes of the young.

Adapted from American Diabetes Association, 2020; Powers, 2018.

TABLE 60-2. Modified White Classification Scheme Used from 1986 Through

1994 for Diabetes Complicating Pregnancy

Plasma Glucose Level

Class Onset Fasting 2-Hour Postprandial Therapy

A1 Gestational <105 mg/dL <120 mg/dL Diet

A2 Gestational >105 mg/dL >120 mg/dL Insulin

Class Age of Onset (yr) Duration (yr) Vascular Disease Therapy

B Over 20 <10 None Insulin

C 10 to 19 10 to 19 None Insulin

D Before 10 >20 Benign retinopathy Insulin

F Any Any Nephropathya Insulin

R Any Any Proliferative retinopathy Insulin

H Any Any Heart Insulin

aWhen diagnosed during pregnancy: proteinuria ≥500 mg/24 hr before

20 weeks’ gestation.

PREGESTATIONAL DIABETES

Considering the previously mentioned high percentage o

undiagnosed diabetes, many women identied with gestational diabetes likely have previously unrecognized type 2

diabetes. Indeed, 5 to 10 percent o women with gestational

diabetes are diagnosed with overt diabetes immediately ater

pregnancy.

■ Diagnosis

Women with high plasma glucose levels, glucosuria, and ketoacidosis present no diagnostic challenge. Women with a random

plasma glucose level >200 mg/dL plus classic signs and symptoms such as polydipsia, polyuria, and unexplained weight loss,

those with a asting glucose level >125 mg/dL, or those with

a rst-trimester glycosylated hemoglobin (HbA1c) level o ≥6.5

percent are considered by the American Diabetes Association

blacks, Mexican-Americans, Puerto Rican-Americans, and Native

Americans (Powers, 2018).

■ White Classification

Until the mid-1990s, the classication system o Priscilla White

(1978) or diabetic pregnant women was the linchpin o management. oday, the White classication is used less oten but

still provides simple, useul inormation on pregnancy risks and

prognosis (Bennett, 2015). Because most currently cited literature also contains data rom these older classications, the one

previously recommended by the American College o Obstetricians and Gynecologists (1986) is shown in Table 60-2.

Te American College o Obstetricians and Gynecologists

(2019a) no longer recommends the White classication. Te

current ocus is whether diabetes antedates pregnancy or is rst

diagnosed during pregnancy. Many now recommend the system

proposed by the American Diabetes Association (Table 60-3).

TABLE 60-1. Etiological Classification of Diabetes Mellitus

Type 1: β-Cell destruction, usually absolute insulin deficiency

Immune-mediated

Idiopathic

Type 2: Ranges from predominantly insulin resistance to predominantly an insulin secretory

defect with insulin resistance

Other types

Genetic mutations of β-cell function: MODY 1–6, others

Genetic defects in insulin action

Genetic syndromes: Down, Klinefelter, Turner, others

Diseases of the exocrine pancreas: pancreatitis, cystic fibrosis

Endocrinopathies: Cushing syndrome, pheochromocytoma, others

Drug or chemical induced: glucocorticosteroids, thiazides, β-adrenergic agonists, others

Congenital infections: rubella, cytomegalovirus, coxsackievirus

Gestational diabetes (GDM)

MODY = maturity-onset diabetes of the young.

Adapted from American Diabetes Association, 2020; Powers, 2018.

TABLE 60-2. Modified White Classification Scheme Used from 1986 Through

1994 for Diabetes Complicating Pregnancy

Plasma Glucose Level

Class Onset Fasting 2-Hour Postprandial Therapy

A1 Gestational <105 mg/dL <120 mg/dL Diet

A2 Gestational >105 mg/dL >120 mg/dL Insulin

Class Age of Onset (yr) Duration (yr) Vascular Disease Therapy

B Over 20 <10 None Insulin

C 10 to 19 10 to 19 None Insulin

D Before 10 >20 Benign retinopathy Insulin

F Any Any Nephropathya Insulin

R Any Any Proliferative retinopathy Insulin

H Any Any Heart Insulin

aWhen diagnosed during pregnancy: proteinuria ≥500 mg/24 hr before

20 weeks’ gestation.1070

Section 12

Medical and Surgical Complications

(2019) and the World Health Organization (2019) to have overt

diabetes rst detected in pregnancy.

Women with only minimal metabolic derangement may be

more dicult to identiy. o diagnose overt diabetes in pregnancy, the International Association o Diabetes and Pregnancy

Study Groups (IADPSG) Consensus Panel (2010) recognizes the

threshold values ound in Table 60-4 or asting or random plasma

glucose and HbA1c levels at prenatal care initiation. Te American

Diabetes Association (2020) and the World Health Organization

(2019) now also consider a plasma glucose level >200 mg/dL

measured 2 hours ater a 75-g oral glucose load to be diagnostic.

No consensus has been reached as to whether such testing should

be universal or limited to those women classied as high risk.

Regardless, a presumed diagnosis o overt diabetes in pregnancy based on these thresholds should be conrmed postpartum. Risk actors or impaired carbohydrate metabolism

in pregnant women include a strong amilial history o diabetes, prior delivery o a large newborn, persistent glucosuria, or

unexplained etal losses.

■ Harm in Pregnancy

With overt diabetes, the embryo, etus, and mother requently

experience serious complications directly attributable to diabetes (Egan, 2020). Many o these complications might be

prevented each year by preconceptional care or improved glycemic control (Peterson, 2015). Using HbA1c values as objective risk quartiles, Finneran and Kieer (2020) reported that

glycemic control throughout pregnancy and a late-pregnancy

HbA1c level <6.5 percent leads to reduced rates o adverse

obstetrical and neonatal outcomes. Te likelihood o successul

outcomes with overt diabetes, however, is not simply related

to glucose control. Te degree o underlying cardiovascular or

renal disease may be more important. Tus, advancing stages

o the White classication, seen in able 60-2, are inversely

related to avorable pregnancy outcomes. Shown in Table 60-5

are data that chronicle the adverse pregnancy outcomes with

overt diabetes, and many are described next (Battarbee, 2020b).

■ Fetal Effects

Spontaneous Abortion

Up to 25 percent o overtly diabetic mothers have an early

miscarriage, and poor glycemic control is an associated actor.

In one study, those whose HbA1c concentrations were >12

percent or whose preprandial glucose concentrations persisted

above 120 mg/dL had an elevated miscarriage risk (Galindo,

2006). In a large Chinese population-based study, those with a

history o diabetes had an increased risk o miscarriage, and the

risk rose 8 percent or each 20-mg/dL incremental rise in asting glucose (Wei, 2019). In 89 pregnancies in women with the

monogenic orm o MODY, only women with the causative

glucose kinase (GCK) mutation were more likely to miscarry

(Bacon, 2015). Another analysis o 128 GCK-MODY pregnancies showed that the observed miscarriage rate was comparable to the background rate (Dickson, 2019). Tese women

have hyperglycemic variability that can be dicult to control.

Preterm Delivery

Overt diabetes is an undisputed risk actor or preterm birth.

In an analysis o more than 500,000 pregnancies in Ontario,

TABLE 60-3. Proposed Classification System for Diabetes in Pregnancy

Gestational diabetes: diabetes diagnosed during pregnancy that is not clearly overt (type 1 or type 2) diabetes

Type 1 Diabetes:

Diabetes resulting from β-cell destruction, usually leading to

absolute insulin deficiency

a. Without vascular complications

b. With vascular complications (specify which)

Type 2 Diabetes:

Diabetes from inadequate insulin secretion in the face of

increased insulin resistance

a. Without vascular complications

b. With vascular complications (specify which)

Other types of diabetes: genetic in origin, associated with pancreatic disease, drug-induced, or chemically induced

Data from American Diabetes Association, 2017a; Powers, 2018.

TABLE 60-4. Diagnosis of Overt Diabetes in Pregnancya

Measure of Glycemia Threshold

Fasting plasma glucose At least 7.0 mmol/L (126 mg/dL)

Hemoglobin A1c At least 6.5%

Random plasma glucose At least 11.1 mmol/L (200 mg/dL) plus confirmation

aApply to women without known diabetes antedating pregnancy. The decision to perform

blood testing for evaluation of glycemia on all pregnant women or only on women with

characteristics indicating a high risk for diabetes is based on the background frequency of

abnormal glucose metabolism in the population and on local circumstances.

Data from International Association of Diabetes and Pregnancy Study Groups Consensus

Panel, 2010.CHAPTER 60

Diabetes Mellitus 1071

almost 20 percent o women with diabetes were delivered preterm compared with 5.6 percent o women without diabetes,

obesity, or hypertension (Berger, 2020). More than 60 percent were indicated preterm births, that is, due to obstetrical

or medical complications. Notably, more than 37 percent o

women with diabetes and chronic hypertension delivered preterm. In a review o Caliornia births, 19 percent o women

with pregestational diabetes delivered beore 37 weeks’ gestation compared with 9 percent o controls (Yanit, 2012). In a

study rom the United Kingdom, the preterm delivery rate was

42 percent or 8690 type 1 diabetic women and 3 percent or

those with type 2 diabetes (Murphy, 2021).

Malformations

Te incidence o major malormations in etuses o women

with type 1 diabetes approximates 11 percent and is at least

double the rate in etuses o nondiabetic mothers (Jovanovič,

2015). Tis risk is present or all women with pregestational

diabetes, including those with type 2 disease (inker, 2020).

Congenital anomalies constitute almost hal o perinatal deaths

in diabetic pregnancies.

Cardiovascular malormations account or more than hal

o the anomalies, and Table 60-6 lists selected malormations

reported by the National Birth Deects Prevention Study (inker, 2020). In a study o more than 2 million births in Canada, the risk o an isolated cardiac deect was veold higher

in women with type 1 diabetes compared with nondiabetic

mothers (Liu, 2013). Te caudal regression sequence, described

in Chapter 15 (p. 280), is a rare malormation that, according

to the National Birth Deects Prevention Study, is 80 times

more likely in women with pregestational diabetes.

Poorly controlled diabetes, both preconceptionally and

early in pregnancy, is thought to underlie this elevated risk

or major malormation. In one study o mothers with type

1 diabetes, a periconceptional HbA1c <6.5 percent carried a

twoold risk or a etal major cardiac deect, and a HbA1c >9

percent carried a sixold risk (Ludvigsson, 2018). Others cite

similar requencies (Murphy, 2021). Figure 60-1 shows the

association between HbA1c levels and etal congenital malormation rates in a Parkland Hospital cohort o women with

overt diabetes (Martin, 2021).

Te etiological mechanisms that explain this link include

excess production o toxic superoxide radicals, altered cell signaling pathways, upregulation o some genes by hyperglycemia, and

TABLE 60-5. Selected Maternal and Perinatal Outcomes

in Percent in Women with Gestational

and Overt Diabetes Compared with

Nondiabetic Womena

Normal GDM Overt DM

182,464b 10,549b 2993b

Factor (%) (%) (%)

Maternal

Chronic HTN 1.5 5.0 11.5

Renal disease 0.5 0.6 2.7

Gestational HTN 8.0 12.9 22.9

Chorioamnionitis 4.1 4.0 5.2

Cesarean birth 42 59 69

Neonatal

RDS 3.8 4.7 11.0

Ventilation 2.8 3.1 7.0

NICU admit 12.8 19.8 41.9

Macrosomia 7.7 18.1 29.9

Deathc 0.2 0.1 0.4

aUnless specified, all comparisons p <0.001.

bNumber of women in group.

c

p <0.05.

DM = diabetes mellitus; GDM = gestational diabetes mellitus;

HTN = hypertension; NICU = neonatal intensive care unit;

RDS = respiratory distress syndrome.

TABLE 60-6. Selected Congenital Malformations

in Pregnancies Complicated by Overt

Diabetes

Birth Defect OR (95% CI)

Cardiovascular

Fallot tetralogy 5.3 (3.5–8.0)

AV septal defect 10.5 (6.2–17.9)

Aortic coarctation 4.5 (2.8–7.1)

Neural-tube defect

Anencephaly 3.5 (1.9–6.4)

Encephalocoele 5.4 (2.5–11.7)

Hydrocephaly 8.2 (5.0–13.5)

Cleft palate 4.3 (2.9–6.5)

Esophageal 3.4 (1.9–6.1)

Hypospadias 2.8 (1.7–4.8)

Renal 8.1 (3.9–16.9)

Sacral agenesis 80.2 (46.1–139.3)

CI = confidence interval; OR = odds ratio.

FIGURE 60-1 Association between fetal malformation rates and

HbA1c values determined at initiation of prenatal care in 1573 pregnancies in women with pregestational diabetes.1072

Section 12

Medical and Surgical Complications

activation o programmed cell death (Basu, 2018; Reece, 2012).

One review o potential molecular mechanisms suggests that cellular responses to oxidative stress represent possible therapeutic

targets to prevent diabetes-induced embryopathy (Yang, 2015).

Altered Fetal Growth

Diminished etal growth may result rom congenital malormations or rom substrate deprivation due to advanced maternal

vascular disease. Tat said, etal overgrowth is more typical o

pregestational diabetes. Maternal hyperglycemia prompts etal

hyperinsulinemia, and this in turn stimulates excessive somatic

growth. Except or the brain, most etal organs are aected by

the macrosomia that characterizes the etus o diabetic women.

Newborns are described as anthropometrically dierent rom

other large-or-gestational age (LGA) neonates (Catalano,

2003; Durnwald, 2004). Specically, those whose mothers

are diabetic have excessive at deposition on the shoulders and

trunk, which predisposes to shoulder dystocia or etopelvic disproportion (Fig. 60-2).

Te incidence o macrosomia rises signicantly when mean

maternal blood glucose concentrations chronically exceed 130

mg/dL (Hay, 2012). In addition, the overall birthweight distribution o neonates o diabetic mothers is skewed toward consistently heavier birthweights. In one sonographic study, the

macrosomia rates or Nordic women with type 1, type 2, or

gestational diabetes were 35, 28, and 24 percent, respectively

(Hammoud, 2013). Moreover, the abdominal circumerence

grew disproportionately larger in the diabetic groups. Analysis o

head circumerence/abdominal circumerence (HC/AC) ratios

shows that this disproportionate growth occurs mainly in diabetic pregnancies that ultimately yield macrosomic newborns.

Unexplained Fetal Demise

Te risk o etal death is three to our times higher in women

with pregestational diabetes (Gardosi, 2013; Patel, 2015). In

the United Kingdom study o more than 17,000 pregnancies in

women with pregestational diabetes, the stillbirth incidence was

similar in those with type 1 or type 2 diabetes—10.4 versus 13.5

per 1000 live and stillbirths, respectively (Murphy, 2021). Stillbirth without an identiable cause is a phenomenon relatively

limited to pregnancies complicated by overt diabetes. Tese stillbirths are “unexplained” because common actors such as obvious

placental insuciency, placental abruption, etal-growth restriction, or oligohydramnios are not identied. Tese etuses are typically LGA and die beore labor, usually later in the third trimester.

Tese unexplained stillbirths are usually associated with poor

glycemic control. In one study, suboptimal glycemic control was

identied in two thirds o unexplained stillbirths between 1990

and 2000 (Lauenborg, 2003). Fetuses o diabetic mothers also

oten have elevated lactic acid levels. Salvesen and colleagues

(1992, 1993) analyzed etal blood samples and reported that

mean umbilical venous blood pH was lower in diabetic pregnancies and was signicantly related to etal insulin levels. Tese

ndings support a hypothesis that hyperglycemic-mediated

chronic aberrations in oxygen and etal metabolite transport

may underlie these unexplained etal deaths (Pedersen, 1977).

Maternal ketoacidosis also can cause etal death (Bryant,

2017). Moreover, stillbirths rom placental insuciency occur

with increased requency in women with overt diabetes and are

usually associated with severe preeclampsia. In the prior cited

Caliornia study, the etal death risk was sevenold higher in

women with hypertension and pregestational diabetes compared with the threeold increased risk associated with diabetes

alone (Yanit, 2012). Stillbirth rates are also elevated in women

with advanced diabetes and vascular complications.

Hydramnios

Diabetic pregnancies are oten complicated by excess amnionic

fuid. In one sonographic study, 18 percent o 314 pregnancies complicated by pregestational diabetes were ound to have

hydramnios in the third trimester, which was dened as an

amnionic fuid index (AFI) >24 cm. Women with elevated

HbA1c values in the third trimester were more likely to have

hydramnios (Idris, 2010). In a similar study, poor maternal

glucose control was associated with macrosomia and hydramnios (Vink, 2006). A likely—albeit unproven—explanation

is that etal hyperglycemia causes etal polyuria (Chap. 14, p.

256). In a study rom Parkland Hospital, Dashe and coworkers

(2000) ound that the AFI parallels the amnionic fuid glucose

level among women with diabetes.

Neonatal Effects

Beore tests o etal health and maturity became available, delivery beore term was deliberately selected or women with diabetes to avoid unexplained stillbirth. Although this practice has

been largely abandoned, a higher requency o preterm delivery

in women with diabetes persists. One analysis o early-term

delivery (370/7 to 386/7 weeks) showed a 13-percent reduction

in such deliveries in women with diabetes between 2005 and

2011 (Little, 2015). Most indicated deliveries are prompted by

advanced diabetes with superimposed preeclampsia.

Although modern neonatal care has reduced neonatal death

rates due to immaturity, neonatal morbidity due to preterm

birth continues to be a serious consequence. In one Neonatal Research Network study o 10,781 extremely preterm neonates, those born to diabetic women treated with insulin prior

FIGURE 60-2 This 6050-g macrosomic neonate was born to a

woman with gestational diabetes.CHAPTER 60

Diabetes Mellitus 1073

to pregnancy were at greater risk or necrotizing enterocolitis

and late-onset sepsis than similarly aged neonates o mothers

without diabetes (Boghossian, 2016).

Respiratory Distress Syndrome

Gestational age rather than overt diabetes is likely the most

signicant actor associated with respiratory distress syndrome

(Chap. 34, p. 615). Indeed, in one analysis o 19,399 verylow-birthweight neonates delivered between 24 and 33 weeks’

gestation, rates o respiratory distress syndrome in newborns o

diabetic mothers were not higher compared with rates in neonates o nondiabetic mothers (Bental, 2011). Diabetes does not

appear to alter the benecial eects o antenatal corticosteroids

to hasten lung maturity (Battarbee, 2020a).

Hypoglycemia and Hypocalcemia

Newborns o a diabetic mother experience a rapid drop in

plasma glucose concentration ater delivery. Tis is attributed

to the hyperplasia o etal β-islet cells that is induced by chronic

maternal hyperglycemia. Low glucose concentrations—dened

as <45 mg/dL—are particularly common in newborns o

women with unstable glucose concentrations during labor (van

Kempen, 2020). Preliminary observations showed that earlyterm etuses exposed to antenatal corticosteroids have a high

rate o hypoglycemia (Gupta, 2020). Frequent blood glucose

measurements in the newborn and active early eeding practices

can mitigate this complication.

Hypocalcemia is dened as a total serum calcium concentration <8 mg/dL in term newborns. Early-onset hypocalcemia

is one o the potential metabolic derangements in neonates o

diabetic mothers. Its cause is unclear. Teories include altered

magnesium–calcium economy, asphyxia, and preterm birth.

In one randomized study, 137 pregnant women with type 1

diabetes were managed with strict versus customary glucose

control (DeMarini, 1994). Almost a third o neonates in the

customary control group developed hypocalcemia compared

with only 18 percent o those in the strict-control group.

Hyperbilirubinemia and Polycythemia

Te pathogenesis o hyperbilirubinemia in neonates o diabetic

mothers is uncertain. A major contributing actor is newborn

polycythemia, which raises the bilirubin load (Chap. 33, p.

606). Polycythemia is thought to be a etal response to relative hypoxia. According to Hay (2012), the sources o this etal

hypoxia are hyperglycemia-mediated elevations in maternal

anity or oxygen and etal oxygen consumption. ogether

with insulin-like growth actors, this hypoxia leads to elevated

etal erythropoietin levels and red cell production.

Cardiomyopathy

Newborns o pregestational and gestational diabetic pregnancies requently have hypertrophic cardiomyopathy. Tis

remodeling can be associated with cardiac dysunction (Aguilera, 2020; Depla, 2021). Huang and coworkers (2013) propose that pathological ventricular hypertrophy in neonates

born to women with diabetes is due to insulin excess. In severe

cases, this cardiomyopathy may lead to obstructive cardiac ailure. In one study, the etuses o 26 women with pregestational

diabetes underwent serial echocardiographic evaluation. In the

rst trimester, etal diastolic dysunction was already evident in

some. In the third trimester, the etal interventricular septum

and right ventricular wall were thicker in etuses o diabetic

mothers (Russell, 2008). Others report similar observations

(Aguilera, 2020). Most aected newborns are asymptomatic

ollowing birth, and hypertrophy usually resolves in the months

ater delivery. Any long-term adverse sequelae remain to be

determined (Depla, 2021).

Long-term Cognitive Development

Intrauterine metabolic conditions are linked to neurodevelopment in ospring. In a study o more than 700,000 Swedish

men, the intelligence quotient o those whose mothers had diabetes during pregnancy averaged 1 to 2 points lower (Fraser,

2014). DeBoer and associates (2005) demonstrated impaired

memory perormance at age 1 year in inants o diabetic mothers. Results rom the Childhood Autism Risks rom Genetics

and the Environment (CHARGE) study indicated that autism

spectrum disorders or developmental delay also were more

common in children o diabetic women (Krakowiak, 2012).

Adane and colleagues (2016) conrmed a consistent relationship between maternal diabetes and diminished cognitive and

language development in studies o younger children but not

older children. Because interpreting eects o the intrauterine

environment on neurodevelopment is conounded by postnatal

actors, the link between maternal diabetes, glycemic control,

and long-term neurocognitive outcome remains unconrmed.

Inheritance

Te risk o developing type 1 diabetes i either parent is aected

is 3 to 5 percent. ype 2 diabetes has a much stronger genetic

component. I both parents have type 2 diabetes, the risk o their

ospring developing it approaches 40 percent (Powers, 2018).

Both types o diabetes develop ater a complex interplay between

genetic predisposition and environmental actors. ype 1 diabetes is prompted by environmental triggers such as inection, diet,

or toxins and is heralded by the appearance o islet-cell autoantibodies in genetically vulnerable people (Pociot, 2016).

■ Maternal Effects

Diabetes and pregnancy can interact to an extent that maternal welare can be jeopardized. With the possible exception o

diabetic retinopathy, however, the long-term course o diabetes

does not appear to be aected by pregnancy.

In one analysis o more than 800,000 pregnancies, the 1125

mothers with type 1 diabetes had greater risks or hypertension and

respiratory complications than nondiabetic gravidas (Jovanovič,

2015). Te 10,126 mothers with type 2 diabetes had an elevated

risk or hypertension, inection, depression, and cardiac or respiratory complications compared with nondiabetic gravidas.

Maternal death is uncommon, but rates in women with

diabetes are still higher than those in unaected gravidas. In

one analysis o 972 women with type 1 diabetes, the maternal

mortality rate was 0.5 percent. Deaths stemmed rom diabetic

ketoacidosis, hypoglycemia, hypertension, and inection (Leinonen, 2001).1074

Section 12

Medical and Surgical Complications

Preeclampsia

Pregnancy-associated hypertension is the complication that

most oten orces preterm delivery in diabetic women. Te incidence o chronic and gestational hypertension—and especially

preeclampsia—is remarkably increased (Chap. 40, p. 690).

In one metaanalysis o 92 studies, women with pregestational

diabetes had a nearly ourold higher pooled relative risk or

preeclampsia (Bartsch, 2016). In the prior cited Caliornia

study, preeclampsia developed three to our times more oten

in women with overt diabetes. Moreover, those diabetics with

coexistent chronic hypertension were almost 12 times more

likely to develop preeclampsia (Yanit, 2012).

Women with type 1 diabetes in more advanced White classes

typically exhibit preexisting vascular complications and nephropathy. Tese women are more likely to develop preeclampsia

(Fig. 60-3). Tis rising risk may be related to oxidative stress,

which plays a key role in the pathogenesis o diabetic complications and preeclampsia. With this in mind, the Diabetes and

Preeclampsia Intervention rial (DAPI) randomly assigned

762 women with type 1 diabetes to antioxidant vitamin C and

E supplementation or placebo in the rst hal o pregnancy

(McCance, 2010). Preeclampsia rates were not improved except

in the ew with a low antioxidant status at baseline.

Preventively, low-dose aspirin prophylaxis is recommended

in women at high risk o preeclampsia, which includes those

with type 1 or type 2 diabetes (Chap. 40, p. 705). An 81-mg

oral daily dose is initiated between 12 and 28 weeks’ gestation

and is continued until delivery (American College o Obstetricians and Gynecologists, 2020a).

Diabetic Nephropathy

Diabetes is the leading cause o end-stage renal disease in

the United States (Chap. 56, p. 1003). Clinically detectable

nephropathy begins with microalbuminuria, recognized as

30 to 300 mg o protein in a 24-hour urine collection specimen. Tis may maniest as early as 5 years ater diabetes onset.

Macroalbuminuria—more than 300 mg in a 24-hour collection specimen—develops in patients destined to have end-stage

renal disease. Hypertension almost invariably develops during

this period, and renal ailure ensues typically in the next 5 to

10 years. Te incidence o overt proteinuria is nearly 30 percent

in individuals with type 1 diabetes and ranges rom 4 to 20

percent in those with type 2 diabetes (Reutens, 2013). Regression is common, presumably rom improved glucose control.

Approximately 5 percent o gravidas with overt diabetes

already have renal involvement (American College o Obstetricians and Gynecologists, 2020c). As many as 40 percent o

these will develop preeclampsia (Vidae, 2008). In addition,

Ambia and associates (2020) reported that the rates o preterm

delivery, preeclampsia, and etal-growth restriction were signi-

cantly higher in diabetic women with proteinuria >300 mg/d

compared with rates in diabetic gravidas whose 24-hour protein

excretion was <300 mg/d.

In general, pregnancy does not appear to worsen diabetic

nephropathy. In one prospective study o 43 women with diabetes, diabetic nephropathy did not progress through 12 months

ater delivery (Young, 2012). Most o these women had only

mild renal impairment. Conversely, pregnant women with

moderate to severe renal impairment may have accelerated progression o their disease (Vidae, 2008). As in women with glomerulopathies, hypertension or substantial proteinuria beore or

during pregnancy is a major predictive actor or ultimate progression to renal ailure in women with diabetic nephropathy.

Diabetic Retinopathy

Retinal vasculopathy is a highly specic complication o both

type 1 and type 2 diabetes. In the United States, diabetic retinopathy is the most important cause o visual impairment in

working-aged adults. Te rst and most common visible lesions

are small microaneurysms ollowed by blot hemorrhages that

orm when erythrocytes escape rom the aneurysms. Tese areas

leak serous fuid that creates hard exudates. Such eatures are

termed background or nonproliferative retinopathy.

With increasingly severe retinopathy, the abnormal vessels

o background eye disease become occluded. Tis leads to retinal ischemia and inarctions that appear as cotton wool exudates.

Tese are considered preproliferative retinopathy. In response to

ischemia, neovascularization begins on the retinal surace and

out into the vitreous cavity. Vision is obscured when these vessels bleed. Laser photocoagulation beore hemorrhage reduces

the rate o visual loss progression and blindness by hal. Te

procedure may be perormed during pregnancy when indicated.

In one study o nearly 500 pregnancies in women with type

1 diabetes, a third had prepregnancy retinopathy and 16 percent o these showed worsening (Bourry, 2021). Almost 25

percent o those without prepregnancy retinopathy developed

disease during pregnancy. Fortunately, development o sightthreatening retinopathy was rare, and only 4 percent o patients

ollowed postpartum had progression o disease.

Another group o investigators evaluated 80 women with

type 2 diabetes and identied retinopathy, mostly mild, in 14

percent during early pregnancy (Rasmussen, 2010). Progression was identied in 14 percent.

Other risk actors that have been associated with progression o retinopathy include hypertension, higher levels o insulinlike growth actor 1, placental growth actor, and macular

edema identied in early pregnancy (Huang, 2015; Vestgaard,

2010). Te American Academy o Ophthalmology (2019)

15

B

(n=164)

Diabetic class C

(n=129)

Preeclampsia incidence

(percent)

D

(n=172)

F

(n=26)

Total

(n=491)

30

45

60

Swedena

United Statesb

FIGURE 60-3 Incidence of preeclampsia in 491 women with type

1 diabetes stratefied by diabetic class in Sweden and the United

States. (Data from Hansona, 1993; Sibaib, 2000.)CHAPTER 60

Diabetes Mellitus 1075

recommends that pregnant women with overt diabetes should

be oered retinal assessment ater the rst prenatal visit. Subsequent eye examinations depend on severity o retinopathy

and level o glucose control. Currently, most agree that laser

photocoagulation and good glycemic control during pregnancy

minimize the potential or deleterious eects o pregnancy.

Ironically, “acute” rigorous metabolic control during pregnancy has been linked to acute worsening o retinopathy. In a

study o 201 women with retinopathy, almost 30 percent su-

ered eye disease progression during pregnancy despite intensive

glucose control (McElvy, 2001). Tat said, Wang and coworkers (1993) observed that although retinopathy worsened during the critical months o rigorous glucose control, long-term

progression o eye disease actually slowed.

Diabetic Neuropathy

Peripheral, symmetrical sensorimotor diabetic neuropathy is

uncommon in pregnant women. However, one orm o this,

diabetic gastropathy, can be troublesome during pregnancy. It

causes nausea and vomiting, nutritional problems, and dicult

glucose control. Aected women are advised that this complication is associated with a high risk o morbidity and poor

perinatal outcome (Kitzmiller, 2008). reatment with metoclopramide and dopamine D2 receptor antagonists is sometimes successul. Gastric neurostimulators have shown success

during pregnancy (Fuglsang, 2015). reatment o hyperemesis

gravidarum can be challenging, and we routinely provide insulin initially by continuous inusion or diabetic women who are

admitted with this condition (Chap. 57, p. 1015).

Diabetic Ketoacidosis

Tis serious complication develops in approximately 1 percent o

diabetic pregnancies and is most oten encountered in women with

type 1 diabetes (Ehrmann, 2020). It is increasingly being reported

in women with type 2 or even those with gestational diabetes (Bryant, 2017; Sibai, 2014). Diabetic ketoacidosis (DKA) may develop

with hyperemesis gravidarum, inection, insulin noncompliance,

insulin pump ailures, β-mimetic drugs given or tocolysis, and

corticosteroids given to induce etal lung maturation. DKA results

rom an insulin deciency combined with an excess in counter-regulatory hormones such as glucagon. Tis leads to gluconeogenesis

and ketone body ormation. Te ketone body β-hydroxybutyrate

is synthesized at a much greater rate than acetoacetate, which is

preerentially detected by commonly used ketosis-detection methods. Tus, serum or plasma assays or β-hydroxybutyrate more

accurately refect true ketone body levels.

Te maternal mortality rate with DKA is <1 percent, but

perinatal mortality rates associated with a single episode o DKA

may reach 35 percent (Bryant, 2017; Guntupalli, 2015). Noncompliance is a prominent actor, and this and DKA were historically considered prognostically bad signs in pregnancy (Pedersen,

1974). Importantly, pregnant women usually develop DKA at

lower blood glucose thresholds than when nonpregnant. In a

study rom Parkland Hospital, the mean glucose level or pregnant

women with DKA was 380 mg/dL, and the mean HbA1C value

was 10 percent (Bryant, 2017). Euglycemic ketoacidosis during

pregnancy also is possible but is rare (Sibai, 2014; Smati, 2020).

One management protocol or diabetic ketoacidosis is shown

in Table 60-7. An important cornerstone o management is

TABLE 60-7. Management of Diabetic Ketoacidosis During Pregnancy

Laboratory Assessment

Obtain arterial blood gases to document degree of acidosis present; measure glucose,

creatinine, ketones, and electrolyte levels at 1- to 2-hour intervals

Insulin

Low-dose, intravenous

Loading dose: 0.2–0.4 U/kg

Maintenance: 2–10 U/hr

Fluids

Isotonic sodium chloride for 3L, then 0.45% saline

Total replacement in first 12 hours of 4–6 L

1 L in first hour

500–1000 mL/hr for 2–4 hours

250 mL/hr until 80 percent replaced

Glucose

Begin 5-percent dextrose in 0.45% saline when glucose plasma level reaches 250 mg/dL

(14 mmol/L)

Potassium

If initially normal or reduced, an infusion rate up to 15–20 mEq/hr may be required; if

elevated, wait until levels decrease into the normal range, then add to intravenous

solution in a concentration of 20–30 mEq/L

Bicarbonate

Add one ampule (44 mEq) to 1 L of 0.45 normal saline if serum pH is <7.1

Data from Bryant, 2017; Powers, 2018; Sibai, 2014.1076

Section 12

Medical and Surgical Complications

vigorous rehydration with crystalloid solutions o normal saline

or Ringer lactate.

Infections

Te rates o many inections are higher in diabetic pregnant

women. Common ones include candidal vulvovaginitis, bacterial urinary and respiratory tract inections, and puerperal pelvic sepsis. However, in one study o more than 1250 diabetic

gravidas screened beore 16 weeks’ gestation, rates o bacterial

vaginosis or vaginal colonization with Candida or Trichomonas

species were not increased (Marschalek, 2016).

In one population-based study o almost 200,000 pregnancies,

the risk o asymptomatic bacteriuria in women with diabetes was

increased twoold (Sheiner, 2009). Another study ound positive

urine culture results in 25 percent o diabetic women (Alvarez,

2010). In a 2-year analysis o pyelonephritis in pregnant women

at Parkland Hospital, 5 percent with diabetes developed pyelonephritis compared with 1.3 percent without diabetes (Hill, 2005).

Rates o urinary tract inections can be minimized by screening

and eradicating asymptomatic bacteriuria (Chap. 56, p. 995).

Last, Johnston and colleagues (2017) reported that 16.5 percent o women with pregestational diabetes had postoperative

wound complications ollowing cesarean delivery. Prevention

options are described in Chapter 30 (p. 550).

■ Preconceptional Care

Because o the close relationship between pregnancy complications and maternal glycemic control, eorts to achieve glucose

targets are typically more aggressive during pregnancy. Unortunately, nearly hal o pregnancies in the United States are

unplanned, and diabetic women requently begin pregnancy

with suboptimal glucose control (Finer, 2016; Kim, 2005).

Management preerably should begin beore pregnancy and

then include specic goals during each trimester.

o minimize early pregnancy loss and congenital malormations, the National Preconception Health and Healthcare Initiative Clinical Workgroup established values or optimal glycemic

control beore conception (Frayne, 2016). Tis was dened as a

HbA1c level <6.5 percent in women with pregestational diabetes. Te American Diabetes Association (2017b) has also dened

optimal preconceptional glucose control in those using insulin.

Refective values are sel-monitored preprandial glucose levels

o 70 to 100 mg/dL, peak 2-hour postprandial values o 100 to

120 mg/dL, and mean daily glucose concentrations <110 mg/dL.

In one prospective population-based study o 933 pregnant

women with type 1 diabetes, the risk o congenital malormations was not demonstrably higher with HbA1c levels <6.9 percent compared with the risk in more than 70,000 nondiabetic

gravidas (Jensen, 2010). However, a substantial ourold greater

risk or malormations was ound or HbA1c levels >10 percent

(see Fig. 60-1). Another study ound ewer adverse outcomes

with improved HbA1c values rom conception to midpregnancy

(Davidson, 2020).

I indicated, evaluation and treatment or diabetic complications such as retinopathy or nephropathy should be instituted

beore pregnancy. Last, olate, 400 μg/d orally, is given periconceptionally and during early pregnancy to decrease the risk

o neural-tube deects (Egan, 2020).

■ Firsttrimester Care

Careul monitoring o glucose control is essential. For this reason, we routinely hospitalize women with overt diabetes during early pregnancy to initiate an individualized glucose control

program and provide education. Tis also aords an opportunity to assess the extent o diabetic vascular complications

and precisely establish gestational age. Te checklist provided

by the Society or Maternal-Fetal Medicine (2020) is helpul.

Some initial evaluations done at Parkland Hospital include

assessment o 24-hour urine protein excretion and serum creatinine level, retinal examination, and echocardiogram i chronic

hypertension is comorbid.

First-trimester screening or aneuploidy may include measurement o maternal serum pregnancy-associated plasma protein A (PAPP-A), β-human chorionic gonadotropin (hCG),

and ultrasound measurement o etal nuchal translucency

(Chap. 17, p. 333). Noninvasive prenatal testing with cell-ree

DNA also is suitable. Although initially thought not to be di-

erent in women with pregestational diabetes, one analysis o

more than 100 insulin-treated women identied reductions

in median PAPP-A and β-hCG levels compared with gravidas without diabetes (Gurram, 2014). Not surprisingly, they

detected no dierence in nuchal translucency measurements.

When calculating aneuploidy risks in women with diabetes,

these dierences should be considered.

■ Insulin Treatment

Te gravida with overt diabetes is best treated with insulin.

Although oral hypoglycemic agents have been used successully

or gestational diabetes (p. 1083), these agents are not currently recommended or overt diabetes, despite some controversy (American College o Obstetricians and Gynecologists,

2020c). In an international, placebo-controlled trial including

502 insulin-treated women with type 2 diabetes, adjunctive

metormin therapy was associated with improved glycemic

control, reduced maternal weight gain, lower cesarean delivery

rate, and less neonatal adiposity (Feig, 2020). Importantly, the

proportion o small-or-gestational age newborns was higher in

the insulin plus metormin group, but the rate o composite

serious neonatal outcomes was not increased compared with

those receiving insulin plus placebo.

Maternal glycemic control can usually be achieved with multiple daily insulin injections and adjustment o dietary intake.

Table 60-8 lists the action proles o common short- and longacting insulins (Powers, 2018). Subcutaneous insulin inusion

by a calibrated pump does not yield better pregnancy outcomes

compared with multiple daily injections. However, an inusion pump is a sae alternative in appropriately selected patients

(Farrar, 2016). With the advent o sensor-augmented insulin pumps and closed-loop insulin delivery systems, improved

glycemic control with either manual or computer-generated

insulin adjustments based on continuous glucose monitoring is now possible (Bergenstal, 2021; Stewart, 2016). In one

study o insulin pump use in women with type 1 diabetes, total

daily insulin doses declined in the rst trimester but later rose

more than threeold. Postprandial glucose elevations prompted

most o the required daily-dose increases (Roeder, 2012). I aCHAPTER 60

Diabetes Mellitus 1077

continuous-inusion insulin pump is elected, initiation prepregnancy may help avoid hypoglycemia and ketoacidosis risks associated with the device's learning curve (Sibai, 2014).

Monitoring

Sel-monitoring o capillary glucose levels using a glucometer is recommended because this involves the woman in her

own care (Dong, 2020). Te American Diabetes Association

(2017b) recommends asting and postprandial glucose monitoring, and Table 60-9 lists glucose goals recommended in

pregnancy. Currently not standard care, one study showed that

longitudinal HbA1c values could be used or risk stratication

(Finneran, 2020). Advances in noninvasive glucose monitoring

will undoubtedly render intermittent capillary glucose monitoring obsolete. Subcutaneous continuous glucose monitoring devices have shown that pregnant women with diabetes

experience signicant periods o daytime hyperglycemia and

nocturnal hypoglycemia that are undetected by traditional

monitoring (Combs, 2012). Such glucose monitoring systems,

coupled with a continuous insulin pump, oer the potential o

an “articial pancreas” to avoid hypo- or hyperglycemia during

pregnancy (Bergenstal, 2021).

Diet

Nutritional planning includes appropriate weight gain through

carbohydrate and caloric modications based on height, weight,

and degree o glucose intolerance (American Diabetes Association, 2017b; Egan, 2020). Te mix o carbohydrate, protein,

and at is adjusted to meet the metabolic goals and individual

patient preerences. A minimum o 175 g/d o carbohydrates

ideally is provided. In one analysis o more than 200 obese

pregnant women with glucose intolerance, a lower carbohydrate intake, particularly late in pregnancy, was associated with

lower at mass in ospring at birth (Renault, 2015). Allotted carbohydrates are distributed throughout the day in three

small- to moderate-sized meals and two to our snacks. Weight

loss is not recommended, but modest caloric restriction may be

appropriate or overweight or obese women. An ideal dietary

composition is 55 percent carbohydrate, 20 percent protein,

and 25 percent at, o which <10 percent is saturated at.

Hypoglycemia

Diabetic control can be unstable in the rst hal o pregnancy,

and the incidence o hypoglycemia peaks during the rst trimester. One study ound hypoglycemic events—blood glucose

values <40 mg/dL—in 37 o 60 women with type 1 diabetes.

A ourth o these were considered severe because the women

were unable to treat their own symptoms and required assistance rom another person (Chen, 2007). Caution is recommended when attempting to achieve euglycemia in women with

recurrent episodes of hypoglycemia.

From one Cochrane database review, loose glycemic control,

dened as asting glucose values >120 mg/dL, was associated

with greater risks or preeclampsia, cesarean delivery, and birthweight >90th percentile compared with women with tight or

moderate control (Middleton, 2016). With very tight control,

dened by asting values <90 mg/dL, no obvious benets were

gained, but the number o cases o hypoglycemia increased.

■ Secondtrimester Care

Maternal serum alpha-etoprotein determination at 16 to 20

weeks’ gestation is used in association with targeted sonographic

examination to detect neural-tube deects and other anomalies (Chap. 17, p. 338). Tese serum levels may be lower in

diabetic pregnancies, and interpretation is altered accordingly.

Because the incidence o congenital cardiac anomalies shown

in able 60-6 is increased veold in mothers with diabetes,

etal echocardiography is an important part o second-trimester

sonographic evaluation (Society or Maternal-Fetal Medicine,

2020). Dashe and coworkers (2009) cautioned that detection

o etal anomalies in obese diabetic women is more dicult

than in similarly sized women without diabetes.

Regarding second-trimester glucose control, normoglycemia

with sel-monitoring continues to be the goal. Ater the rst

trimester's glucose instability, a stable period ensues. Tis is

then ollowed by higher insulin requirements. Tese stem rom

the enhanced peripheral resistance to insulin that is related to

pregnancy and is described in Chapter 4 (p. 57).

Although most women with pregestational diabetes require

a higher total daily insulin dose with advancing gestational age,

a small proportion will experience a reduction in their daily

dose later in pregnancy. Te signicance o this drop remains

uncertain. One Australian study evaluated women with alling

TABLE 60-8. Action Profiles of Commonly Used Insulins

Insulin Type Onset Peak (hr) Duration (hr)

Short-acting (SC)

Lispro <15 min 0.5–1.5 2–4

Glulisine

Aspart

<15 min

<15 min

0.5–1.5

0.5–1.5

2–4

2–4

Regular (SC) 30–60 min 2–3 3–6

Regular inhaled 30–60 min 2–3 3

Long-acting (SC)

Degludec 1–9 hr — <12

Detemir 1–4 hr Minimala 12–24

Glargine 1–4 hr Minimala 20–24

NPH 1–4 hr 4–10 10–16

aMinimal peak activity.

NPH = neutral protamine Hagedorn; SC = subcutaneous.

TABLE 60-9. Self-Monitored Capillary Blood Glucose

Goals

Specimen Level (mg/dL)

Fasting ≤95

Premeal ≤100

1-hr postprandial ≤140

2-hr postprandial ≤120

0200–0600 ≥60

Mean (average) 100

Hemoglobin A1c ≤6%1078

Section 12

Medical and Surgical Complications

insulin requirements (FIR), dened as a ≥15 percent drop

in the peak total daily insulin dose ater 20 weeks’ gestation

(Padmanabhan, 2017). In 158 women with type 1 and type 2

diabetes and FIR, the risk or a composite o adverse outcomes

indicative o placental insuciency was ourold greater than in

those without FIR. Te composite primary outcome included

preeclampsia, etal-growth disorders, delivery beore 30 weeks’

gestation, placental abruption, and etal death beore 20 weeks.

FIR was more common in women with type 1 diabetes.

■ Thirdtrimester Care and Delivery

During the past several decades, the threat o late-pregnancy stillbirth in women with diabetes prompted recommendations or

various etal surveillance programs beginning in the third trimester. Such protocols include etal movement counting, periodic

etal heart rate monitoring, intermittent biophysical prole evaluation, and contraction stress testing (Chap. 20, p. 384). None o

these techniques has been subjected to prospective randomized

clinical trials, and their primary value seems related to their low

alse-negative rates. Te American College o Obstetricians and

Gynecologists (2020c, 2021) suggests initiating such testing at 32

to 34 weeks’ gestation. With reassuring testing and no other complications, delivery is anticipated between 390/7 and 396/7 weeks.

At Parkland Hospital, women with diabetes are seen in a specialized Maternal–Fetal Medicine clinic every 2 weeks. During

these visits, glycemic control records are evaluated, and insulin

doses are adjusted. Women are routinely instructed to perorm

etal kick counts beginning early in the third trimester. At 34

weeks’ gestation, admission to our High-Risk Pregnancy Unit is

oered to all insulin-treated women. Approximately hal o these

women choose admission. While in the hospital, they continue

daily etal movement counts and undergo etal heart rate monitoring three times a week. With no other complications, delivery

at Parkland is typically planned or 38 weeks.

At the University o Alabama at Birmingham, weekly antenatal testing is initiated no later than 34 weeks’ gestation.

wice-weekly testing is reserved or those with poorly controlled diabetes or supervening medical or obstetrical complications. Delivery is planned or 39 weeks’ gestation in those

with good glycemic control and reassuring antenatal testing.

Earlier delivery is planned or those with poor glycemic control

or signicant comorbidities.

For vaginal delivery, labor induction may be attempted

when the etus is not excessively large and the cervix is considered avorable (Chap. 26, p. 486). Cesarean delivery at or near

term has requently been used to avoid the traumatic birth o

a large etus in a woman with diabetes. In women with more

advanced diabetes, especially those with vascular disease, the

reduced likelihood o successul labor induction remote rom

term has also contributed to an increased cesarean delivery rate.

In an analysis o pregnancy outcomes o diabetic women rom

University o Alabama at Birmingham according to the White

classication, the rates o cesarean delivery and preeclampsia

escalated with White class (Bennett, 2015). In another study, a

HbA1c level >6.4 percent at delivery was independently associated with urgent cesarean delivery (Miailhe, 2013). Tis suggests

that tighter glycemic control during the third trimester might

reduce late etal compromise and cesarean delivery or etal indications. Many o these women have undergone prior cesarean

delivery. Somewhat related to this, the cesarean delivery rate or

women with overt diabetes has remained at approximately 80

percent or the past 40 years at Parkland Hospital.

Reducing or withholding the dose o long-acting insulin on

the day o delivery is recommended. Regular insulin should be

used to meet most or all o the insulin needs o the mother during this time, because insulin requirements typically drop markedly ater delivery. During labor, continuous insulin inusion by

calibrated intravenous pump is most satisactory (Table 60-10).

Te woman should be adequately hydrated intravenously and

given glucose in sucient amounts to maintain euglycemia.

Capillary or plasma glucose levels are checked hourly during

active labor, and regular insulin is administered accordingly.

■ Puerperium

Oten, many diabetic women may require virtually no insulin

or the rst 24 hours or more postpartum. Subsequently, insulin

requirements may fuctuate markedly during the next ew days.

Inection must be promptly detected and treated. When appropriate, oral agents can be restarted. For type 2 diabetic women

who will be transitioned to oral agents, this can be done on

postoperative day 1.

Counseling in the puerperium should include a discussion

o birth control. Eective contraception is especially important

in women with overt diabetes to allow optimal glucose control

beore subsequent conception (Chap. 38, p. 664).

GESTATIONAL DIABETES

In the United States in 2016, almost 6 percent o pregnancies

were complicated by gestational diabetes (Deputy, 2018). Worldwide, its prevalence diers according to race, ethnicity, age, and

body composition and by screening and diagnostic criteria. As

discussed in the ollowing sections, several persisting controversies pertain to the diagnosis and treatment o gestational diabetes.

TABLE 60-10. Insulin Management for Labor Induction or Scheduled Cesarean Delivery

Give evening dose insulin.

Withhold morning dose.

Infuse intravenous normal saline at 100–125 mL/hr.

Regular insulin is infused at 1–1.25 units/hr if glucose levels >100 mg/dL.

Measure glucose levels hourly.

With active labor or if glucose levels are >70 mg/dL, change from intravenous saline to 5% dextrose given at 100–150 mL/hr

with a target glucose level of ∼100mg/dL.CHAPTER 60

Diabetes Mellitus 1079

Te word gestational implies that diabetes is induced by pregnancy—ostensibly because o exaggerated physiological changes

in glucose metabolism (Chap. 4, p. 57). Gestational diabetes

is dened as carbohydrate intolerance o variable severity with

its onset or rst recognition during pregnancy (American College o Obstetricians and Gynecologists, 2019a). Tis denition applies whether or not insulin is used or treatment and

undoubtedly includes some women with previously unrecognized overt diabetes.

Te term gestational diabetes aims to communicate the need

or enhanced surveillance during pregnancy and to stimulate urther testing postpartum. Te most important perinatal correlate

is excessive etal growth, which may result in both maternal and

etal birth trauma. Te likelihood o etal death with appropriately treated gestational diabetes is not dierent rom that in the

general population. More than hal o women with gestational

diabetes ultimately develop overt diabetes in the ensuing 20 years.

Moreover, mounting evidence implicates long-range complications that include obesity and diabetes in their ospring.

■ Screening and Diagnosis

Despite almost 50 years o research, there is still no agreement

on the optimal screening method or gestational diabetes. Te

diculty in achieving consensus is underscored by the controversy ollowing publication o the single-step approach shown

in Table 60-11, which assesses the glucose values ollowing a

75-g oral glucose load. Tis strategy is espoused by the International Association o Diabetes and Pregnancy Study Groups

Consensus Panel (2010) and was greatly infuenced by results

o the Hypoglycemia and Pregnancy Outcomes (HAPO) study,

described subsequently. Although the American Diabetes Association (2019) supports this new scheme, the American College o Obstetricians and Gynecologists (2019a) continues to

recommend a two-step approach to screen and then diagnose

gestational diabetes. Similarly, the National Institutes o Health

(NIH) Consensus Development Conerence in 2013 concluded

that evidence is insucient to adopt a one-step approach.

Te recommended two-step approach begins with either

universal or risk-based selective screening using a 50-g, 1-hour

oral glucose challenge test. Participants in the Fith International

Workshop Conerences on Gestational Diabetes endorsed use o

selective screening criteria shown in Table 60-12 (Metzger, 1998).

TABLE 60-11. Threshold Values for Diagnosis of

Gestational Diabetes with 75-g OGTT

Plasma

Glucose

Glucose Concentration

Thresholda

Above Threshold

(%)

mmol/L mg/dL Cumulative

Fasting 5.1 92 8.3

1-hr OGTT 10.0 180 14.0

2-hr OGTT 8.5 153 16.1b

aOne or more of these values from a 75-g OGTT must be

equaled or exceeded for the diagnosis of gestational diabetes.

bIn addition, 1.7% of participants in the initial cohort were

unblinded because of fasting plasma glucose levels

>5.8 mmol/L (105 mg/dL) or 2-hr OGTT values

>11.1 mmol/L (200 mg/dL), bringing the total to 17.8%.

OGTT = oral glucose tolerance test.

TABLE 60-12. Risk-Based Recommended Screening Strategy for Detecting GDMa

GDM risk assessment: should be ascertained at the first prenatal visit

Low Risk: Blood glucose testing not routinely required if all the following are present:

Member of an ethnic group with a low prevalence of GDM

No known diabetes in first-degree relatives

Age <25 years

Weight normal before pregnancy

Weight normal at birth

No history of abnormal glucose metabolism

No history of poor obstetrical outcome

Average Risk: Perform blood glucose testing at 24 to 28 weeks using either:

Two-step procedure: 50-g oral glucose challenge test (GCT), followed by a

diagnostic 100-g OGTT for those meeting the threshold value in the GCT

One-step procedure: diagnostic 100-g OGTT performed on all subjects

High Risk: Perform blood glucose testing as soon as feasible, using the procedures

described above, if one or more of these are present:

Severe obesity

Strong family history of type 2 diabetes

Previous history of GDM, impaired glucose metabolism, or glucosuria

If GDM is not diagnosed, blood glucose testing should be repeated at 24 to 28

weeks’ gestation or at any time symptoms or signs suggest hyperglycemia

aCriteria of the Fifth International Workshop-Conference on Gestational Diabetes.

GDM = gestational diabetes mellitus; OGTT = oral glucose tolerance test.1080

Section 12

Medical and Surgical Complications

Conversely, the American College o Obstetricians and Gynecologists (2019a) recommends universal screening o pregnant

women with a protocol that provides a 50-g oral glucose load and

that is ollowed in 1 hour by a laboratory-based blood glucose test.

It is suggested that attempts to identiy the minority o women

who should not be screened would add unnecessary complexity.

Screening should be perormed between 24 and 28 weeks’ gestation in those women not known to have glucose intolerance earlier in pregnancy. Tis 50-g screening test is ollowed by a diagnostic

100-g, 3-hour oral glucose tolerance test (OGTT) i screening results

meet or exceed a predetermined plasma glucose concentration.

For the 50-g screening test, the plasma glucose level is measured 1 hour ater a 50-g oral glucose load without regard to

the time o day or time o last meal. In an earlier review, the

pooled sensitivity or a threshold o 140 mg/dL ranged rom 74

to 83 percent depending on 100-g thresholds used or diagnosis

(van Leeuwen, 2012). Sensitivity estimates or a 50-g screen

threshold o 135 mg/dL improved only slightly to 78 to 85 percent, but specicity dropped rom a range o 72 to 85 percent

or 140 mg/dL to 65 to 81 percent. Using a threshold o 130

mg/dL marginally improves sensitivity with a urther decline in

specicity (Donovan, 2013). In the absence o clear evidence

supporting one cuto value over another, the American College

o Obstetricians and Gynecologists (2019a) sanctions using any

one o these three 50-g screen thresholds. At Parkland Hospital,

we continue to use 140 mg/dL as the screening threshold to

prompt the 100-g test. Te threshold used at the University o

Alabama at Birmingham is 135 mg/dL.

Justication or screening and treatment o women with gestational diabetes was strengthened by a randomized treatment

trial in 1000 women (Crowther, 2005). Women diagnosed

with gestational diabetes ater a 75-g OG and assigned

to receive dietary advice with blood glucose monitoring plus

insulin therapy were compared with a cohort assigned to usual

prenatal care. Te ormer group had a signicantly lower risk

o a composite perinatal adverse outcome that included death,

shoulder dystocia, bone racture, and nerve palsy. Macrosomia

dened by birthweight ≥4000 g complicated 10 percent o

deliveries in the intervention group compared with 21 percent

in the routine prenatal care group. Cesarean delivery rates were

almost identical in the two study groups.

Slightly dierent results were reported by the Maternal–Fetal

Medicine Units Network randomized trial o 958 women identied with mild gestational diabetes using a two-step screening

and diagnosis approach (Landon, 2009). Tey reported no di-

erences in rates o composite morbidity that included stillbirth,

birth trauma, and neonatal hypoglycemia, hyperinsulinemia,

and hyperbilirubinemia. Secondary analyses did demonstrate a

50-percent reduction in macrosomia, ewer cesarean deliveries,

and a signicant decrease in shoulder dystocia rate—1.5 versus

4 percent—in treated versus routine care groups.

Based largely on these two landmark studies, the U.S. Preventive Services ask Force (2021) recommends universal

screening in low-risk women ater 24 weeks’ gestation. However, the ask Force concluded that evidence is insucient to

assess the balance o benets versus harms o screening beore

24 weeks. Earlier screening in obese women did not result in

better outcomes (Harper, 2020).

Te optimal OG to identiy gestational diabetes also is

debated. Proposed criteria or interpretation o the diagnostic 100-g OG are shown in Table 60-13. Criteria or the

75-g OG are shown in able 60-11. A secondary analysis o the Maternal-Fetal Medicine Units Network treatment

trial showed that women diagnosed with either the National

Diabetes Data Group (NDDG) or the Carpenter-Coustan

criteria beneted rom treatment (Harper, 2016). However,

the number needed to treat to prevent a shoulder dystocia

was higher or the Carpenter-Coustan criteria. Others report

similar observations (Ghaari, 2020). At Parkland Hospital,

we use the NDDG criteria or diagnosis, whereas CarpenterCoustan criteria are preerred at the University o Alabama at

Birmingham.

Controversies of Screening and Diagnosis

Te HAPO study was a 7-year international epidemiological

study o 23,325 pregnant women at 15 centers in nine countries analyzing the association o various levels o glucose intolerance during the third trimester with adverse perinatal outcomes

(HAPO Study Cooperative Research Group, 2008). Blood glucose levels were stratied into seven categories (Fig. 60-4). Tese

values were then correlated with rates or birthweight >90th

percentile, primary cesarean delivery, neonatal hypoglycemia,

and cord serum C-peptide levels >90th percentile. Cord serum

C-peptide is secreted in equimolar concentrations with insulin

and refects etal β-islet cell unction. Findings generally supported the supposition that increasing plasma glucose levels

were associated with increasing adverse outcomes.

The International Association of Diabetes and

Pregnancy Study Group

Tis workshop conerence in 2008 concerned the diagnosis and

classication o gestational diabetes. Ater reviewing the results

o the HAPO study, a panel developed recommendations or the

diagnosis and classication o hyperglycemia during pregnancy.

TABLE 60-13. Diagnosis of Gestational Diabetes Mellitus

Using Threshold Glucose Values from

100-g Oral Glucose Tolerance Testa,b

NDDGc Carpenter–Coustand

Time (mg/dL) (mmol/L) (mg/dL) (mmol/L)

Fasting 105 5.8 95 5.3

1-hr 190 10.6 180 10.0

2-hr 165 9.2 155 8.6

3-hr 145 8.0 140 7.8

aThe test should be performed when the patient is fasting.

bTwo or more of the venous plasma glucose concentrations listed are met or exceeded for a positive diagnosis.

cSerum glucose level.

dSerum or plasma glucose level.

NDDG = National Diabetes Data Group.

Data from American College of Obstetricians and

Gynecologists, 2019a; American Diabetes Association,

2019; Ferrara, 2002.CHAPTER 60

Diabetes Mellitus 1081

Tis panel allowed or the diagnosis o overt diabetes during

pregnancy as shown in able 60-4. It also recommended a

single-step approach to the diagnosis o gestational diabetes

using the 75-g, 2-hour OG and thresholds derived using an

arbitrary 1.75 odds ratio o outcomes such as LGA birthweight

and cord serum C-peptide levels >90th percentile. Only one

o these thresholds, shown in able 60-11, would need to be

met or exceeded to make the diagnosis o gestational diabetes.

During their deliberations, the IADPSG estimated that

implementation o these recommendations would raise the

prevalence o gestational diabetes in the United States to 17.8

percent! Said another way, using this approach, the number o

women with mild gestational diabetes would increase almost

threeold with no evidence o treatment benet (Cundy, 2012).

Feldman and coworkers (2016) evaluated the implementation o the IADPSG paradigm in a beore-ater analysis that

included more than 6000 women. Compared with a two-step

approach, the new strategy was associated with a signicant

rise in gestational diabetes diagnosis rates but not with reduced

macrosomia rates. Remarkably, they identied a higher primary cesarean delivery rate associated with adoption o the

IADPSG recommendations. Te American Diabetes Association (2013, 2019) initially recommended adopting this new

approach, based on benets inerred rom trials in women identied using a two-step approach described earlier. However, it

now concludes that data also support a two-step strategy.

NIH Consensus Development Conference

Prompted by these disparate recommendations, the NIH Consensus Development Conerence on Diagnosing Gestational

Diabetes Mellitus (2013) was convened. Te panel ound

insucient evidence to adopt the one-step diagnostic process

proposed by the IADPSG and recommended continuation o the two-step

approach. Te Consensus Conerence

panel did also suggest that urther studies resolving the benecial uncertainties

associated with the one-step approach

could warrant revision o their conclusions.

A recent trial compared the IADPSG

single-step approach to the more traditional two-step approach in almost

24,000 women (Hillier, 2021). Consistent with the IADPSG prediction, the

incidence o gestational diabetes was

16.5 percent in those who underwent

single-step testing compared with 8.5

percent in women assigned to two-step

testing. However, incidences o hypertensive disorders o pregnancy, primary

cesarean delivery, LGA newborns,

and/or a perinatal composite outcome

including measures o birth trauma and

perinatal death were not materially di-

erent between testing groups. Tus,

this large and pragmatic trial does not

show maternal or perinatal benet and

thus does not justiy the increased patient and healthcare costs

o broadening the diagnosis o gestational diabetes using the

proposed single-step approach (Casey, 2021).

■ Maternal and Fetal Effects

Adverse consequences o gestational diabetes dier rom those

o pregestational diabetes (see able 60-5). Similar to women

with overt diabetes, adverse maternal eects associated with

gestational diabetes include a higher requency o hypertension

and cesarean delivery.

Unlike women with overt diabetes, women with gestational

diabetes do not appear to have etuses with substantially higher

rates o anomalies than the general obstetrical population (American College o Obstetricians and Gynecologists, 2019a; Sheeld,

2002). In a study o more than 1 million women rom the Swedish

Medical Birth Registry, major malormation rates were marginally

elevated in etuses o gestational diabetics compared with those o

nondiabetic controls—2.3 versus 1.8 percent (Fadl, 2010).

From this Registry, the stillbirth rate was not higher. Additionally, the stillbirth rate was not increased in an analysis o

more than 800,000 pregnancies rom 2005 through 2011

(Jovanovič, 2015). Tat said, women with elevated asting glucose levels have elevated rates o unexplained stillbirths similar

to those o women with overt diabetes. Tis increasing risk with

progressive maternal hyperglycemia emphasizes the importance

o identiying women with evidence o preexisting diabetes

early in pregnancy (see able 60-4).

Fetal Macrosomia

Te primary eect attributed to gestational diabetes is excessive

etal size that is variably dened and discussed in Chapter 47

30

LGA frequency (percent)

20

25

15

10

5

Glucose levels

Fasting

2 hour

1 hour ≥212

≥178

≥100

158–177

194–211

95–99

140–157

172–193

90–94

126–139

156–171

85–89

109–125

133–155

80–84

91–108

106–132

75–79

≤90

≤105

≤75

FIGURE 60-4 Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study. The frequency of newborn birthweight ≥90th percentile for gestational age plotted against glucose levels fasting and at 1- and 2-hr intervals following a 75-g oral glucose load. LGA =

large for gestational age. (Reproduced with permission from HAPO Study Cooperative

Research Group, Metzger BE, Lowe LP, et al: Hyperglycemia and adverse pregnancy outcomes, N Engl J Med. 2008 May 8;358(19):1991–2002.)1082

Section 12

Medical and Surgical Complications

(p. 832). Te perinatal goal is to avoid dicult delivery rom

macrosomia and concomitant birth trauma associated with

shoulder dystocia. In one analysis o more than 80,000 vaginal

births in Chinese women, the calculated risk or shoulder dystocia in newborns weighing ≥4200 g was 76-old greater than

the risk in those weighing <3500 g (Cheng, 2013). However,

the odds ratio or shoulder dystocia in women with diabetes

was <2. Said another way, although gestational diabetes is

certainly a risk actor, it accounts or only a small number o

pregnancies complicated by shoulder dystocia.

One study identied shoulder dystocia in approximately

4 percent o women with mild gestational diabetes compared with <1 percent o women with a 50-g glucose screen

result <120 mg/dL (Landon, 2011). Te excessive shoulder

and trunk at that commonly characterizes the macrosomic

newborn o a diabetic mother predisposes such neonates to

shoulder dystocia or cesarean delivery (Durnwald, 2004;

McFarland, 2000). However, one prospective study o etal

adipose measurements demonstrated no dierences between

measurements in 630 ospring o women with gestational

diabetes and 142 without diabetes (Buhling, 2012). Te

authors attributed this negative nding to successul treatment o gestational diabetes.

Insulin-like growth factors also infuence etal growth. Tese

proinsulin-like polypeptides are produced by virtually all etal

organs and are potent stimulators o cell dierentiation and

division. Although not a current clinical tool, one study ound

that maternal insulin-like growth actor 1 levels strongly correlated with birthweight (Luo, 2012). Te HAPO Study Cooperative Research Group (2008) also ound dramatic elevations

in cord serum C-peptide levels associated with rising maternal glucose levels ollowing a 75-g OG. C-peptide levels

>90th percentile were ound in almost a third o newborns

in the highest glucose categories. Other actors implicated in

macrosomia include epidermal growth actor, broblast growth

actor, platelet-derived growth actor, leptin, and adiponectin

(Grissa, 2010; Loukovaara, 2004; Mazaki-ovi, 2005).

Maternal body mass index (BMI) is an independent and

more substantial risk actor or etal macrosomia than is glucose

intolerance (Ehrenberg, 2004; Mission, 2013). In a secondary

analysis o women with either untreated mild gestational diabetes or normal glucose tolerance testing results, higher BMI

levels were associated with rising birthweight, regardless o glucose levels (Stuebe, 2012). Another analysis ound that gestational diabetes, compared with obesity or gestational weight

gain, contributed the least to the population-attributable raction o LGA neonates (Kim, 2014). Te highest raction o

LGA neonates was attributable to maternal obesity plus excessive gestational weight gain. Similarly, Egan and colleagues

(2014) ound that excessive gestational weight gain is common

in women with gestational diabetes and coners an additive risk

or etal macrosomia.

Weight distribution also seems to play a role because the risk

o gestational diabetes is greater with maternal truncal obesity.

One cohort study ound that increased maternal abdominal

subcutaneous at thickness at 18 to 22 weeks’ gestation correlated with BMI and was a better predictor o gestational diabetes (Suresh, 2012).

Neonatal Hypoglycemia

Hyperinsulinemia may provoke severe neonatal hypoglycemia

within minutes o birth, but only three ourths o these episodes occur in the rst 6 hours (Harris, 2012). Te denition

o neonatal hypoglycemia is debated, and recommended clinical

thresholds range rom 35 to 45 mg/dL. One NIH workshop

conerence on the topic supported a threshold o 35 mg/dL in

term newborns but cautioned that this practice is not strictly

evidence based (Hay, 2009). reating neonates with glucose levels <47 mg/dL did not improve outcomes compared with treating those with glucose levels <36 mg/dL (van Kempen, 2020).

Newborns described by the HAPO Study Cooperative

Research Group (2008) had an incidence o clinical neonatal

hypoglycemia that rose with increasing maternal OG result

values dened in Figure 60-4. Te requency varied rom 1 to 2

percent, but it was as high as 4.6 percent in women with asting

glucose levels ≥100 mg/dL. Similarly, an analysis o more than

3000 Korean women who underwent a 50-g OG ound that

neonates born to women with a screening result ≥200 mg/dL

were 84 times more likely to have hypoglycemia than those

born to women with a result <140 mg/dL (Cho, 2016). Te

risk o neonatal hypoglycemia correlates with umbilical cord

C-peptide levels. Te risk also rises with birthweight, independent o a maternal diabetes diagnosis (Mitanchez, 2014).

■ Management

Women with gestational diabetes can be divided into two

unctional classes using asting glucose levels. Pharmacological

methods are usually recommended i diet modication does

not consistently maintain the asting plasma glucose levels <95

mg/dL or the 2-hour postprandial plasma glucose <120 mg/dL

(American College o Obstetricians and Gynecologists, 2019a).

Whether pharmacological treatment should be used in women

with lesser degrees o asting hyperglycemia is unclear. No controlled trials have been done to identiy ideal glucose targets

or etal risk prevention. On the other hand, the HAPO Study

Cooperative Research Group (2008) did demonstrate increased

etal risk at glucose levels below the threshold used or diagnosis o diabetes. Te Fith International Workshop Conerence

recommended that asting capillary glucose levels be kept ≤95

mg/dL (Metzger, 2007).

In their systematic review, Hartling and associates (2013)

concluded that treating gestational diabetes resulted in signi-

cantly lower incidences o preeclampsia, shoulder dystocia, and

macrosomia. For example, the calculated risk ratio was 0.50 or

delivering a newborn >4000 g ater treatment. Tese investigators caution that the attributed risk or these outcomes is low,

especially when glucose values are only moderately elevated.

Tey were unable to demonstrate an eect on neonatal hypoglycemia or on uture metabolic outcomes in the ospring.

Diabetic Diet

Nutritional instructions generally include a carbohydrate-controlled diet sucient to maintain normoglycemia and avoid ketosis. On average, this includes a caloric intake o 30 to 35 kcal/

kg/d. In one study, 152 women with gestational diabetes were

randomly assigned to a 40- or a 55-percent daily carbohydrateCHAPTER 60

Diabetes Mellitus 1083

diet, and no dierences in insulin levels and pregnancy outcomes

were ound (Moreno-Castilla, 2013). Te American College o

Obstetricians and Gynecologists (2019a) suggests that carbohydrate intake be limited to 33 to 40 percent o the total daily calories. Te remaining calories are apportioned to give 20 percent as

protein and 40 percent as at.

Te most appropriate dietary approach or women with gestational diabetes has not been established. One metaanalysis

o trials o low-glycemic-index diets ound that diets higher in

complex carbohydrates and dietary ber reduced the risk o

macrosomia and likelihood o insulin use in women with gestational diabetes (Wei, 2016). However, there clearly are limits

to what can be accomplished with various dietary approaches

alone. Most and Langer (2012) ound that insulin was eective in reducing the risk o excessive birthweight in ospring o

obese women with gestational diabetes. Casey and colleagues

(2015b) also ound that dietary treatment alone or morbidly

obese women with mild gestational diabetes did not reduce

neonatal at mass or LGA birthweights.

Exercise

Few trials have evaluated exercise specically or women with

gestational diabetes. Te American College o Obstetricians

and Gynecologists (2020b) recommends regular physical activity that incorporates aerobic and strength-conditioning exercise

during pregnancy and extends this to women with gestational

diabetes. wo recent metaanalyses demonstrate that structured exercise programs during pregnancy diminish weight

gain during pregnancy and even reduce the risk o developing

gestational diabetes (Russo, 2015; Sanabria-Martinez, 2015).

Exercise during pregnancy in women with gestational diabetes

also lowers glucose levels (Jovanovic-Peterson, 1989).

Glucose Monitoring

Hawkins and associates (2008) compared outcomes in 315

women with diet-treated gestational diabetes who used personal

glucose monitors against those o 615 gestational diabetics who

were also diet-treated but who underwent intermittent asting

glucose evaluation during weekly obstetrical visits. Women

using daily blood-glucose sel-monitoring had signicantly

ewer macrosomic newborns. Tey also gained less weight ater

diagnosis than women evaluated during clinic visits only. Tese

ndings support the common practice o blood-glucose sel-

monitoring or women with diet-treated gestational diabetes.

Postprandial surveillance or gestational diabetes is superior

to preprandial surveillance (DeVeciana, 1995). Te American

College o Obstetricians and Gynecologists (2019a) and the

American Diabetes Association (2019) recommend glucose

assessment our times daily. Te rst check is perormed asting, and the remainder are done 1 or 2 hours ater each meal.

At Parkland Hospital in women with diet-treated gestational

diabetes, changing to postprandial monitoring signicantly

reduced weekly maternal weight gain (0.45 lb) compared with

preprandial monitoring (0.63 lb).

Insulin Treatment

Historically, insulin has been considered standard therapy in

women with gestational diabetes when target glucose levels

cannot be consistently achieved through nutrition and exercise.

Insulin does not cross the placenta, and tight glycemic control

can usually be achieved. Insulin therapy is typically added i

asting levels persist above 95 mg/dL in women with gestational

diabetes. Te American College o Obstetricians and Gynecologists (2019a) also recommends that insulin be considered in

women with 1-hour postprandial levels that persistently exceed

140 mg/dL or those with 2-hour levels >120 mg/dL. Importantly, all o these thresholds are extrapolated rom recommendations or managing women with overt diabetes.

Te starting insulin dose is typically 0.7 to 1.0 U/kg/d and is

given in divided doses (American College o Obstetricians and

Gynecologists, 2019a). A combination o intermediate-acting

and short-acting insulin may be used, and dose adjustments are

based on glucose levels at particular times o the day.

At Parkland Hospital, the starting daily dose is divided so

that two thirds is given in the morning beore breakast and

one third in the evening beore dinner. In the morning dose,

one third is regular insulin and two thirds are NPH (neutral

protamine Hagedorn) insulin. For the evening dose, one hal is

regular insulin and the other hal is NPH. Insulin instruction

or these women is accomplished either in a specialized outpatient clinic or during a short hospital stay.

At the University o Alabama at Birmingham, a basal-bolus

approach using insulin glargine with rapid-acting insulin at

each meal is preerred. Ater calculating an initial insulin dose,

hal is given as long-acting glargine at bedtime, and the other

hal is administered as rapid-acting insulin split into three doses

given beore breakast, lunch, and dinner. As shown in able

60-8, when using insulin analogues such as insulin aspart and

insulin lispro, the more rapid onset o action must be considered during postprandial glucose management.

Oral Hypoglycemic Agents

Insulin is the preerred rst-line agent or persistent hyperglycemia in women with gestational diabetes (American College

o Obstetricians and Gynecologists, 2019a; American Diabetes

Association, 2017b). Both organizations acknowledge that several studies support the saety and ecacy o either metormin

(Glucophage) or glyburide (Micronase), which is also called

glibenclamide (Feig, 2020; Langer, 2000; Nicholson, 2009).

Balsells and coworkers (2015) completed a metaanalyses o trials that compared both agents with insulin or with each other.

In the seven trials comparing glyburide with insulin, glyburide

was associated with higher birthweight, more macrosomia, and

more requent neonatal hypoglycemia. In the six trials comparing metormin with insulin, metormin was associated with

less maternal weight gain, more preterm birth, and less severe

neonatal hypoglycemia. Women requiring insulin initiation or

insulin addition were considered to have ailed treatment. On

average, such ailures occurred in 6 percent o women treated

with glyburide and 34 percent o those treated with metormin.

In the two studies directly comparing oral hypoglycemic agents,

however, treatment ailure rates o both agents were equivalent.

Moreover, reminiscent o ndings rom the trial o adjunctive metormin therapy in women with type 2 diabetes previously described, metormin treatment was associated with less

maternal weight gain, lower birthweight, and less macrosomia.1084

Section 12

Medical and Surgical Complications

Conversely, in a randomized trial o glyburide treatment as an

adjunct to diet therapy in 395 women with mild gestational diabetes, Casey and colleagues (2015a) did not identiy any signi-

cant improvements in pregnancy outcomes in women treated

with adjunctive glyburide.

Concerns have also emerged regarding potential adverse

outcomes among women treated with glyburide. First, glyburide crosses the placenta and reaches concentrations in the

etus that are more than two thirds o maternal levels (Caritis,

2013). In a study o more than 9000 women with gestational

diabetes treated with either insulin or glyburide, a signicant

rise in rates o neonatal intensive care unit admission, respiratory distress, and neonatal hypoglycemia was associated with

glyburide use (Castillo, 2015).

Similarly, metormin reaches etal serum concentrations

nearly equal to maternal levels. However, in one study o 751

women with gestational diabetes who were randomly assigned

to metormin or insulin treatment, short-term perinatal adverse

events such as neonatal hypoglycemia, respiratory distress syndrome, phototherapy, or birth trauma did not dier between

groups (Rowan, 2008; 2011). Te at distribution in children

exposed to metormin showed a tendency toward a more avorable pattern. From a smaller randomized metormin trial, at 18

months, ospring exposed as etuses to metormin were slightly

heavier. However, markers o early motor or language development did not dier compared with those in ospring exposed

as etuses to insulin (Ijäs, 2015).

Te Food and Drug Administration has not approved glyburide or metormin use or treatment o gestational diabetes.

Te American College o Obstetricians and Gynecologists

(2019a) recognizes both as reasonable choices or second-line

glycemic control in women with gestational diabetes. Because

long-term outcomes have not been ully studied, the College

recommends disclosure o the limitations in current saety data.

■ Obstetrical Management

In general, or women with gestational diabetes who do not require

insulin, early delivery or other interventions are seldom required.

Tere is no consensus regarding the value or timing o antepartum

etal testing. It is typically reserved or women with pregestational

diabetes because o the greater stillbirth risk. Te American College

o Obstetricians and Gynecologists (2019a, 2020c) endorses etal

surveillance in women with gestational diabetes and poor glycemic

control. At Parkland Hospital, women with gestational diabetes

are routinely instructed to perorm daily etal kick counts in the

third trimester. As previously discussed, insulin-treated women are

oered inpatient admission ater 34 weeks’ gestation. Approximately hal o these women accept admission, and antepartum

etal monitoring is perormed three times each week.

Women with gestational diabetes and adequate glycemic

control are managed expectantly. Elective labor induction to

prevent shoulder dystocia compared with spontaneous labor

remains controversial. In the truncated GINEXMAL randomized trial o 425 women with gestational diabetes, outcomes o

labor induction between 38 and 39 weeks’ were compared with

expectant management until 41 weeks’ gestation (Alberico,

2017). Although underpowered, this trial demonstrated no

clinically meaningul dierence in the cesarean delivery rate

between the induction and expectantly managed groups—12.6

versus 11.8 percent. However, with early labor induction, neonatal hyperbilirubinemia rates were signicantly higher, and

there was a nonsignicant threeold greater shoulder dystocia

rate. In a cohort study o 8392 women with gestational diabetes, routine delivery at 38 or 39 weeks’ gestation was associated

with a lower rate o cesarean delivery but with an elevated rate

o neonatal intensive care unit admission (Melamed, 2016).

Te American College o Obstetricians and Gynecologists

(2019a) recommends that routine labor induction in women

with diet-treated gestational diabetes should not occur beore

39 weeks’ gestation. As mentioned previously, at Parkland

Hospital, those treated with insulin are delivered at 38 weeks’

gestation. At the University o Alabama at Birmingham, delivery is carried out ater 39 weeks.

Elective cesarean delivery to avoid brachial plexus injuries

in overgrown etuses is another issue. Te American College o

Obstetricians and Gynecologists (2019b) concludes that data are

insucient to determine whether women with gestational diabetes

whose etuses have a sonographically estimated weight ≥4500 g

should undergo cesarean delivery to avoid risk o birth trauma.

In one systematic review, Garabedian and associates (2010) estimated that as many as 588 cesarean deliveries in women with

gestational diabetes and an estimated etal weight o ≥4500 g

would be necessary to avoid one case o permanent brachial

plexus palsy. In one analysis o 903 women with gestational diabetes, sonographic estimates o etal weight within 1 month o

delivery typically overdiagnosed etuses as being LGA. Only 22

percent o women estimated to have an LGA etus actually delivered an overgrown newborn (Scires, 2015). Still, the American

College o Obstetricians and Gynecologists (2020b) acknowledges that prophylactic cesarean delivery may be considered in

diabetic women with an estimated etal weight ≥4500 g.

■ Postpartum Evaluation

Recommendations or postpartum evaluation are based on the

50- to 75-percent likelihood that women with gestational diabetes will develop overt diabetes within 15 to 25 years (American

Diabetes Association, 2019). Te Fith International Workshop

Conerence on Gestational Diabetes recommended that women

diagnosed with gestational diabetes undergo postpartum evaluation with a 75-g OG (Metzger, 2007). Tese recommendations and the classication scheme o the American Diabetes

Association are shown in Table 60-14. However, one study o

insurance claim data rom 2000 to 2013 showed that only 24

percent o women with a pregnancy complicated by gestational

diabetes underwent postpartum screening within a year. Less than

hal o those underwent a 75-g OG (Eggleston, 2016). Te

American College o Obstetricians and Gynecologists (2019a)

recommends either a asting glucose assessment or a 75-g, 2-hour

OG at 4 to 12 weeks postpartum or the diagnosis o overt

diabetes. Te American Diabetes Association (2019) recommends

testing every 1 to 3 years in women with a history o gestational

diabetes but normal postpartum glucose screening.

Women with a history o gestational diabetes are also at risk

or cardiovascular complications associated with dyslipidemia,CHAPTER 60

Diabetes Mellitus 1085

hypertension, and abdominal obesity—the metabolic syndrome

(Chap. 51, p. 903). In a study o 47,909 parous women, the

nearly 5000 women with prior gestational diabetes were 2.6

times more likely to be hospitalized or cardiovascular morbidity

(Kessous, 2013). Another study evaluated 483 women between 5

and 10 years ater being diagnosed with mild gestational diabetes

(Varner, 2017). Investigators ound no increased risk or developing metabolic syndrome associated with additional pregnancies.

However, risk or subsequent diabetes rose almost ourold i gestational diabetes complicated at least one subsequent pregnancy.

■ Recurrent Gestational Diabetes

In one large metaanalysis, the pooled gestational diabetes recurrence rate was 48 percent. Te same investigative group identi-

ed elevated maternal BMI, insulin use, etal macrosomia, and

weight gain between pregnancies as additional risk actors or

gestational diabetes recurrence (Schwartz, 2015, 2016).

Liestyle behavioral changes that include weight control and

exercise between pregnancies would seem likely to prevent gestational diabetes recurrence. However, women randomized to

an exercise program that started beore 14 weeks’ gestation in

a subsequent pregnancy did not have a lower recurrence rate

(Guel, 2016). Conversely, Ehrlich and coworkers (2011)

ound that prepregnancy loss o at least two BMI units was

associated with a lower subsequent risk o gestational diabetes

in women who were overweight or obese in the rst pregnancy.

Nhận xét