OBSTETRICS

Obstetric management when normoglycemia is maintained in

diabetic pregnant women with vascular compromise

Radoslav Jovanovic, M.D., and Lois Jovanovic, M.D.

New York, New York

This study presents an obstetric protocol offering better management and prediction for normoglycemic

insulin-dependent patients (White Class 0 4 , F, R, or RF) who conceived after they were diagnosed as having vascular disease secondary to diabetes mellitus. Normoglycemia was accomplished during the

pregestational phase, and conception occurred only after the glycosylated hemoglobin level was

documented to be normal. Normoglycemia was maintained during pregnancy in the outpatient setting

through the use of blood glucose monitoring performed by the patient. The obstetric protocol emphasized three additional areas of attention: (1) assessment of fetal growth by serial uterine fundal measurement

and ultrasonography at gestational weeks 21 to 22; (2) assessment of fetal movement by patient-perceived

fetal movements for 1 hour a week starting at week 35, increasing to 2 hr/day at week 37, and increasing to 3 hr/day from week 38 onward; and (3) cervical assessment at week 37 and preparation for vaginal

delivery. Eight patients had a creatinine clearance of :s80 ml/min prior to conception (mean= 66 ± 6

ml/min). By 6 to 12 weeks' gestation all eight showed an increase in creatinine clearance

(mean = 91 ± 20, p < 0.01 ). There was no change in the third trimester, and postpartum creatinine

clearance was at antepartum levels. Proteinuria increased significantly by the end of the first trimester in all eight women and regressed postpartum. Proteinuria (>150 mg/24 hr) did not occur in the 14 women with normal antepartum creatinine clearance. Of 11 women with background retinopathy, six showed

improvement in retinal status by fundus stereophotography whereas five showed no change. Of 11 women with proliferative retinopathy, five improved, five required laser therapy, and one remained in stable

condition. Despite hemoglobin A, levels in the normal gestational range (3% to 7.5%), there was a

significant correlation of these levels with infant birth weights. None of the 22 infants died, and only one had any perinatal disease. Thus this protocol with its emphasis on fetal growth and size resulted in improvement in both maternal and infant outcome in pregnancies complicated by diabetes mellitus with

vascular compromise. (AM. J. OBSTET. GYNECOL. 149:617, 198<!.)

Although programs for maintaining normoglycemia

have improved the outcome of pregnancies compli­cated by diabetes mellitus, 1 the pregnancy complicated by vascular disease associated with diabetes mellitus is still reported to have an infant perinatal mortality of greater than 20% of cases.2 Improvement of the out-

From the Departments of Medicine and Obstetrics and Gynecology, Cornell University Medical College.

This work was funded in part iJy National Institutes of Health Gen­eral Clinical Research Center Grant No. RR00047 and iJy Na­tional Institute of Child Health and Human Development Diabetes in Early Pregnancy Study Grant No. NOJ-HD-0-2842. Dr. Lois Jovanovic is a recipient of the Andrew W. Mellon Teacher Scientist award.

come in this type of pregnancy can be possible only if two areas are specifically addressed. First, if normaliza­tion of the blood glucose level of the diabetic woman is accomplished before conception,3 the high incidence of malformations of the infants in this group can be prevented. Second, if normoglycemia is sustained throughout the pregnancy to decrease the possible potentiation effect of hyperglycemia on vascularly

compromised end-organs and thus prolong the gesta­tion, the morbidity associated with prematurity will

decrease.

Received for publication October 4, 1983; revised December 5, 1983; accepted December 15, 1983.

Reprint requests: Radoslav Jovanovic, M.D., Cornell University Medical College, Room 907, 515 East 71st St., New York, NY 10021.

These two endeavors necessitate a coordinated ap­proach to the care of the diabetic woman who has vas­

cular compromise. The obstetrician must practice watchful waiting while sustaining maternal normogly­cemia. This paper describes the obstetric protocol and

the outcome of 22 pregnancies complicated by diabetic

retinopathy and/or nephropathy in which the blood

617

618 Jovanovic and Jovanovic July 15, 1984 Am. J. Obstet. Gynecol.

Table I. Patient population in program for maintenance of normoglycemia

Duration of Case Age diabetes mellitus White Before No. (yr) (yr) Class gestation

I 25 21 D, 2 24 22 D, 3 27 21 D, 4 32 19 D, 5 31 24 D, 6 27 21 D, 7 26 17 D1 8 29 18 D, 9 31 20 D,

10 33 21 D, 11 29 17 D, 12 27 14 R 13 28 17 R 14 32 12 R 15 33 17 RF 16 31 13 RF 17 36 15 RF 18 36 19 RF 19 29 23 F 20 30 21 RF 21 29 22 RF 22 28 17 RF

glucose levels were normalized before conception and maintained at normal levels throughout pregnancy. Al­though serial antepartum electronic fetal monitoring was performed in the traditional fashion, a new pro­tocol emerged which proved to be a better guide for decisions of obstetric management.

Material and methods

Twenty-two diabetic women with renal and/or reti­nal disease (11, White Class D4 ; 3, R; 7, RF; l, F) (Table I) were included in a program designed to normalize blood glucose levels before conception and to maintain normoglycemia throughout pregnancy. This process was achieved by calculation of insulin need based on body weight (0.6 U/kg/24 hr = pregestational, 0.7 U /kg/24 hr = first trimester; 0.8 U /kg/24 hr = second trimester; 0.9 U/kg/24 hr =third trimester; 1.0 Ulkg/ 24 hr =term); the total amount of insulin required was divided into insulin amounts that provided the basal needs and the mealtime bolus needs. 2 An insulin infu­sion pump was used for six women, and 16 women received three to four injections a day of intermediate­and short-acting insulin. All patients monitored and charted the blood glucose levels five to 10 times a day with reflectance meters (Ames Eyetone Reflectometers or Dextrometers) and reagent strips (Dextrostix). The maternal age range was 24 to 36 years, and duration of diabetes in all of the women was 2:12 years (range, 12 to 24 years) (Table I).

90 100 110 140 115 120 100 90

120 110 140 105 110 105 70 74 80 60 70 70 60 43

Analysis of 24 hr urine specimen

Creatinine clearance (mllmin)

I I I First Third Post- Presence of

trimester trimester partum hypertension

120 150 80 120 140 90 130 160 100 140 120 100 130 150 105 130 140 110 120 120 100 110 150 100 150 140 110 120 150 100 140 150 120 120 160 100 120 180 90 130 120 110 76 70 72 + 79 80 72

110 120 110 80 110 80 +

120 120 90 110 120 110 90 96 80 + 61 23 143 +

After glycosylated hemoglobin levels in the patients were below 8.5%, which was the cutoff point for in­creased risk of malformations in the report by Miller et al.,4 the patients were given permission to conceive.

The pregnancy was documented by basal body tem­perature measurements, by human chorionic gonado­tropin levels by radioreceptor assay5 15 days from con­ception as indicated on the temperature graph, or by the human chorionic gonadotropin level 1 day after a missed menstrual period. All patients were immedi­ately hospitalized in a clinical research center for doc­umentation of glucose control, readjustment of insulin dosage, retinal photographs, and measurement of creatinine clearance and 24-hour protein excretion. Four of the patients were hypertensive at the time pregnancy was documented (blood pressure 2:140 mm Hg systolic and 2:90 mm Hg diastolic).

The antepartum surveillance program consisted of phone consultation after hospital discharge regarding questions of glucose control at a rate of three times per week and an office visit with the internist and obstetri­cian present every 2 weeks until 33 weeks' gestation, with weekly visits thereafter. Blood was measured for glycosylated hemoglobin level (total hemoglobin A1 ,

microcolumn, Helena Laboratories) every 2 weeks at the routine office visit. In addition, the uterine growth pattern received special attention. The uterine growth was measured serially, and whenever the uterine fun­dal growth measurement in centimeters was question-

Volume 149 Number 6

Before gestation

0 0 0 0 0.1 0 0 0.14 0 0 0 0 0.1 0 I 0.5 0.4 I 0.25 0.5 I 2

Analysis of 24 hr urine specimen

Proteinuria (gm/24 hr)

First Third Post-trimester trimester partum

0 0 0 0 0 0 0 0 0 0 0 0 0.15 0.12 0.1 0 0 0 0 0 0 0.157 0.15 0.1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 I I I 0.5 0.6 0.6 0.2 I 1.5 I 0.7 0.8 0.1 0.8 0.8 0.3 8 II 2 9 12 2

able for gestational age, a bimanual pelvic examination was performed. If a question remained after the bimanual pelvic examination, serial sonography was performed every 2 weeks. At 2I to 22 weeks' gestation all of the patients had an ultrasonic examination. In addition, they all had a routine pelvic examination at 26 to 28 weeks to rule out the threat of premature labor. At 33 weeks the patients were instructed on how to perceive the rhythm and frequency of movement of the fetus. At 35 weeks the patients were instructed to count fetal movement for 1 hour a day while resting on their left side.6 Less than four fetal movements per hour was considered as indication of an inactive fetus. Beginning at 36 weeks a nonstress test was performed weekly. From 37 weeks onward the patients were asked to count fetal movements two times a day. Also at 37 weeks pelvic examination was again performed to as­sess changes in the cervix; if cervical changes were min­imal or none, the cervix was stimulated to initiate cervi­cal changes. At 38 weeks the patients were told to count fetal movement three times a day. They also had non­stress testing two times a week, which was increased at 40 weeks to three times a week.

If the patients did not go into spontaneous labor by 40 to 4I weeks, labor was induced. The patients were not given insulin or breakfast the morning of labor. induction. Administration of intravenous fluid was started with one-half normal saline solution, and an­other bottle containing l 0 units of Pitocin in l 000 ml of

Obstetric management of diabetic pregnant women 619

N=B 200

-175 c:::

p < 0.001

~ 150 E -Cl) u 125 1: CIS I i~ ... CIS Cl) 100 0 ..... Cl)

::Et 1: 75 oot "E i::::tt

:0::: : ... CIS l!! :E.,;

50 <D 0

25

Pregestation First Third Postpartum Trimester

Fig. l. The creatmme clearance (milliliters per minute) throughout pregnancy of the eight women who at the start of pregnancy had a creatinine clearance of <80 ml/min. The four dots for each patient represent separate determinations at four discrete time points.

one-half normal saline solution was attached. The blood glucose level was determined every hour in order to maintain it in the range of 70 to 90 mg/dl in accord with a previously published protocol. 7

Results

A glycosylated hemoglobin level of <8.5% was achieved in all women before conception. Insulin re­quirement was 2: I SD below the published level of 0. 7 units/kg/24 hr for first-trimester diabetic women8 in all eight of the women with depressed kidney function. In addition, the insulin requirement rose to the same de­gree as in the nonrenal patients. Mean first-trimester blood glucose levels were 80 to 90 mg/dl, based on the five to IO blood glucose measurements per day.

Table II indicates that the minimal background ret­inopathy of II women (White Class D4 ) did not pro­gress when normoglycemia was maintained through­out pregnancy. The retinopathy of six of these II patients regressed, as documented by serial stereofun­dal photography. The IO women who began the preg­nancy with proliferative retinopathy had a variable course. Three of these women showed improvement throughout pregnancy, and two more improved post partum. Thus five had improved visual acuity after the pregnancy. Four women required laser therapy, but their postpartum visual acuity did not change from its status in the prepartum examination. These were the same four women who had hypertension complicating

620 Jovanovic and Jovanovic

Table II. Course of retinopathy in 21 patients

July 15, 1984 Am. J. Obstet. Gynecol.

Status of retinopathy during pregnancy 1--------r-------------i Postpartum visual acuity as compared

First trimester I Third trimester to pregestational status Pregestational retinal status

Background retinopathy (n = 11) 11, No change 6, Improved 6, Improved 5, No change 5, Improved 5, No change

Proliferative retinopathy (n = 10) 8, No change 3, Improved 2, Laser therapy 3, No change

4, Laser therapy (l repeat)

Table III. Obstetric maternal outcome

White Class

D4 (n = 11) R or RF (n = 6),

without hyper­tension

F (n = 1) RF (n = 4)

with hyper­tension

Gestational week at delivery (mean± SD)

39 + 1 40.5 ± 0.5

40 39 ± 0.5

(n = 3) 33 (n = 1)

*Total cesarean section rate, 54.5%.

Type of delivery

Normal sponta-neous

delivery (n)

6 2

Cesarean section*

n I 5 4

0 3

%

45 66

0 75

the pregnancy. One patient with proliferative retinop­

athy who did not require laser therapy also did not

experience change in visual acuity throughout preg­

nancy. Table I also summarizes the status of the kidneys

before and at each trimester of the pregnancy. In 14 women with a creatinine clearance >80 ml!min before pregnancy, creatinine clearance improved appropri­

ately throughout gestation. Of the eight women with creatinine clearances ~80 ml/min, four began their pregnancies with hypertension (160/90, 150/90, 150/ 100, and 145/96 mm Hg). All four were managed without antihypertensive medications during the preg­

nancy by means of bed rest at home with daily home blood pressure determinations until the blood pressure rose >I 0 mm Hg above the baseline systolic or diastolic

levels. Two patients required hospitalization for blood

pressure rise at 26 and 30 weeks, respectively. These

patients were delivered at 33 and 38 weeks because of

the superimposed preeclampsia. Fig. 1 illustrates that in the patients with Class F dia­

betes the creatinine clearances improved (mean preges­

tational creatinine clearance of 65.5 ± 4 to a postpar­tum level of 94.6 ± 12 ml/min). Three of the eight patients did not improve throughout the pregnancy,

but after delivery the creatinine clearance of one of

these women had a remarkable rise from a third­

trimester level of 23 to 140 ml/min postpartum.

Fig. 2 illustrates the increased proteinuria through­

out pregnancy in all eight of these patients. Only the

four patients with hypertension developed massive proteinuria by the third trimester (2, 7, 11, and 12

gm/24 hr). The proteinuria of all eight patients re­gressed to pregestational values at 6 weeks post partum.

Delivery results

Based on perinatal outcome, conclusions could be

made as to the necessity of the various antepartum sur­

veillance techniques used. The serial nonstress testing

after gestational week 35 did not play a significant role

in the decision-making for early delivery. All nonstress

tests performed were reactive. Instead, the mean

glycosylated hemoglobin level measured twice monthly emerged as the best predictor of fetal growth. Normal third-trimester glycosylated hemoglobin levels in 10

nondiabetic pregnant women were <7%, and there­

fore a glycosylated hemoglobin level >7.0% was an in­dication for initiating preterm delivery after documen­

tation of positive lung maturity. Two of the 22 women had a glycosylated hemoglobin level > 7% (7 .1% and

7.2%). These women delivered after amniocentesis was performed for assessment of lung maturity at 37.5 and

38 weeks' gestation. The body weights of their infants were above the ninetieth percentile. Two other women had premature deliveries as a consequence of pre­eclampsia superimposed on chronic hypertension (they were the two women with proteinuria of :2:11 gm in the third trimester). The other 18 women with a glycosy­

lated hemoglobin level <7.0% and with no evidence of

hypertension were allowed to go into spontaneous

labor or labor was induced or a scheduled repeat cesar­

ean section was performed on their calculated date of confinement. There were 10 spontaneous deliveries

and 12 cesarean sections (Table III).

Infant outcome

Fig. 3 shows the relationship of infant birth weight to

the mean third-trimester glycosylated hemoglobin level

Volume 149 Number 6

10

9

8 -.c ..,. 7

~ E Dl - 6

111 ·;:: 5 ::::J

1: ·a; 4 0 ...

Q. 3

2

N=S

N.S.

First · Third Trimester Trimester

Fig. 2. The 24-hour protein excretion into the urine of the eight patients who had creatinine clearances of <80 ml/min. Each bar represents separate determinations at four discrete time points. The dots represent mean ± SD.

(measured every 2 weeks from weeks 30 to 40, so each

patient's data point equals a mean of six determina­

tions). The five smallest infants (mean weight of 2450 gm) had the lowest glycosylated hemoglobin levels

(mean hemoglobin A1 of 5.6% ). The largest infants

(mean weight of 3820 gm) had the highest glycosylated

hemoglobin levels (mean hemoglobin A1 of 6.8%). Table IV summarizes the metabolic status of the in­

fants at birth. There were no incidences of blood glu­

cose levels <40 mg/dl; bilirubin levels >12 mg/dl; he­

matocrit values >65%; or calcium levels <7.5 mg/dl.

There was one mild episode of respiratory distress syndrome in the infant born of the mother who deliv­

ered at 33 weeks because of severe preeclampsia.

There were no malformations.

Comment

Previous obstetric protocols recommended early de­

livery in the case of pregnancies complicated by dia­betic vascular disease. These protocols did not differ

with regard to maternal glucose levels, and therefore

all infants were delivered as soon as the amniotic fluid

indicated lung maturity. The 22 patients in this study

had blood glucose levels between 60 and 140 mg/dl

before and during the entire pregnancy. Therefore

this study not only offered the possibility of observing

the course of retinopathy and nephropathy during

pregnancy but also allowed an obstetric protocol to be developed specifically for normoglycemic pregnancies.

Maintaining blood glucose levels in the correct physio-

Obstetric management of diabetic pregnant women 621

p<O.OI

7

0 <;t 6

I

0 r0

<f) ~ 5

<1.> <1.> 3

";:§ ~ 4

<! .0 I c 3 0 <1.>

E 0

2 c .... <1.> -0 ~

0 2000- 2501- 3001- 3501-2500 3000 3500 4000

Infant birth weight (gm)

Fig. 3. Relationship of hemoglobin A1 (HbA1 [%]) to birth weight of infant. The hemoglobin A1c values for each patient represent the mean of at least six determinations between gestational weeks 30 and 40. Each bar represents separate de­terminations at four discrete time points. The dots represent mean± SD.

logic range can only be achieved through the collabor­ative effort of the obstetrician and the most important

member of the team, the patient, through patient­

monitoring of blood glucose. The main function of the

physician and diabetes educator then becomes that of consultant and supporter.

This study revealed that rigorous antepartum sur­

veillance cannot substitute for maintenance of norma­glycemia. The main function of antepartum obstetric

care was observing fetal growth, which reflects glucose

control, while also watching glycemic control as indi­

cated in the patient's self-monitored glucose diaries

and the frequent glycosylated hemoglobin determina­

tions. Sonography at 21 to 22 weeks proved to be useful

for confirmation of date of confinement, for a baseline

value of fetal size, and for detection of congenital mal­

formations. Fetal movement records as recorded by the patients

proved to be helpful in management decisions. For

622 Jovanovic and Jovanovic

Table IV. Neonatal outcome

White Class

D, R RF F

Gestational week at delivery

(mean+ SD)

39.0 ± I 40.5

39.0 ± 1.5 40

No. of cases

II 3 7 7

Birth weight (gm-mean ± SD)

2998 ± 400 3050 ± 450 3197 ± 150

3100

Glucose <40 mgldl

0 0 0 0

these 22 patients and for an additional 60 well-con­

trolled Class A, B, and C diabetic patients studied,

counting fetal movements was a simple, practical, and reliable sign of fetal well-being. All 82 mothers with

active babies had a good outcome. These data confirm

the study of Rayburn et al.6• The prescribed amount of

bed rest, which was increased to 3 hours a day, may also

have had a beneficial effect on fetal growth and well­

being. Nonstress testing did not add to the information ob­

tained from the fetal movement records. During the

observation period for fetal movement an inactive in­

fant occasionally became active during a repeat count a few hours later. Four of the 82 patients reported the

fetus was inactive after the second count. These pa­tients were in prodromal labor or in early dysfunctional

labor. Further investigation by nonstress testing was

not necessary, because active labor ensued, and all four

patients were delivered of healthy infants. In order to rely so heavily on the fetal movement

records of the patients, it is important that they learn to detect rhythm and frequency of fetal movement early

in the third trimester. This enables the women to keep

track of fetal activity later in the pregnancy when im­

portant decisions must be made. From our experience

nonstress testing is no longer the mainstay in obstetric care of well-controlled diabetic patients. It did not play

a significant role in the management of any of our 82 diabetic patients. It appears from this study that elimi­

nation of maternal hyperglycemia eliminates the need

for nonstress testing.

Other investigators agree that nonstress testing is not

as reliable as it was once hoped. These investigators

have reported false positive fetal activity test rates in

excess of 50% when fetal heart rate acceleration was

insufficient by their standards.9 It has been shown by

Leveno et al. 10 that antepartum acceleration of the fetal

heart rate in response to movement does not rule out

serious uteroplacental insufficiency, nor does it guar­

antee that the fetus is safe for a predictable length of

time. This is particularly true in the pregnancies of

diabetic patients, who can have rapid metabolic changes that may lead to fetal acidemia. 11 The results

of this study would suggest that nonstress testing

should be performed only when a patient reports a

Bilirubin >12

mgldl

0 0 0 0

Incidence of

Hema-tocrit

>65%

0 0 0 0

Calcium <7.5 mgldl

0 0 0 0

July 15, 1984 Am. J. Obstet. Gynecol.

Respiratory distress

syndrome

0 0 I 0

Malfor-motions

0 0

decrease in fetal activity. A better use of energy would

be to concentrate all effort on keeping the maternal

blood glucose level within the physiologic range and on

using only patient-monitored fetal activity as a confir­

mation of fetal well-being.

A secondary goal would be to achieve as high a vagi­

nal delivery rate as possible. The protocol to achieve

this goal included a pelvic examination at 37 weeks to

assess changes in the cervix. If changes were minimal or nonexistent, the cervix was stimulated, which led to

rapid cervical changes in most of the patients because

of increasing release of prostaglandin F, as demon­

strated by Mitchell et al. 12 Before cervical stimulation

was performed, a cervical culture was obtained to rule

out the presence of pathogens. According to Stys

et al., 1 ~ prostaglandin E2 produced by the cervix is ca­

pable of stimulating the changes in cervical collagens and/or ground substances which are thought to permit

the dramatic alteration in cervical consistency necessary

for the dilation process. Also, the main process that

allows cervical dilation by promoting tissue softening is collagenolysis caused by collagenase from neutrophils.

Thus cervical examination stimulates intense infiltra­

tion by neutrophils. 14 If labor was not spontaneous in

our well-controlled patients, it was induced at 40 to 41

weeks. Ten vaginal deliveries resulted. It is interesting to note that in our study of 82 well­

documented euglycemic obstetric patients none of the infants developed hypoglycemia or hypocalcemia.

Soler and Maling 15 are in agreement with us that main­tenance of normoglycemia in diabetic mothers would

eliminate neonatal complications in their infants. Our

study of 22 infants delivered of mothers with the most

severe diabetic complications and of the additional 60

infants we cared for substantiates this view.

The only infant from this group who developed mild

respiratory distress syndrome was born at a gestational

age of 33 weeks. The pregnancy was interrupted be­

cause of severe maternal chronic hypertension and the

sudden onset of preeclampsia.

From our study of 82 diabetic euglycemic patients, it

appeared that the process of lung maturation in their infants was the same as it was in the infants of our

nondiabetic patients. We did not observe lung matura­

tion delay in the infants of the diabetic mothers. The

Volume 149 Number 6

evidence was convincing that lung maturation of in­fants of diabetic mothers is indirectly related to mater­nal blood glucose levels and resulting fetal insulin levels. When maternal normoglycemia is maintained, it

appears that lung maturation of infants of diabetic mothers is equal to that of infants of nondiabetic

mothers. No infant in our study group had any malformation,

which was contrary to the results in a study by Pedersen and Molsted-Pedersen 16 in which six of 22 infants de­

livered of similarly compromised mothers had severe congenital anomalies. However, they concluded that

the congenital malformations that occurred in this study were related more to the poor control of blood glucose levels during the first weeks of pregnancy rather than to the maternal vascular status.

Weights of the infants in our series are of interest because of the correlation between increases in infant size and worsening of the mother's diabetic classifica­

tion (Table IV). The largest infants were those of mothers with White Class RF diabetes. The infants of

diabetic mothers with vascular compromise were not small as expected, but they were not large either. Their

weights do not seem to substantiate the belief that in­fants of mothers with White Class RF diabetes always have retarded growth, but their weights do seem to be

directly related to maternal blood glucose levels during

pregnancy. The relationship between the infant's birth weight and the third-trimester glycemic levels is illus­

trated in Fig. 3. It is clear from this figure that the

mothers of the smaller infants had lower hemoglobin A1 levels than the mothers of the larger infants, as determined by the mean of six glycosylated hemoglo­

bin levels in each patient from weeks 30 to 40. Our observations suggest that uteroplacental blood

flow remains normal in the normoglycemic diabetic patient with retinopathy and nephropathy until the blood pressure reaches 160/ 100 mm Hg. The infant is not in jeopardy nor is there deterioration of the mater­nal retinopathy or nephropathy.

In summary the results of this study show that nor­moglycemia was maintained in the outpatient setting, which resulted in a dramatic cost reduction for each pregnancy. In a well-controlled diabetic pregnancy, gestation was allowed to continue to 40 or 41 weeks to

avoid (1) iatrogenic premature delivery, (2) the risk of amniocentesis, (3) failed induction of labor because of

an unfavorable cervix, and (4) unnecessary cesarean

section. The obstetric protocol that best correlated with outcome in these normoglycemic diabetic pregnancies

proved to be (1) intensive diabetes control as docu­mented by five to 10 glucose levels a day and bimonthly glycosylated hemoglobin levels; (2) patient-monitored

fetal movement records; (3) uterine growth as assessed by serial bimanual pelvic examinations, fundal mea-

Obstetric management of diabetic pregnant women 623

surement, and ultrasonic examination; (4) cervical as­

sessment and preparation to increase the chances of vaginal delivery.

This study reveals that pregnancy in normoglycemic

patients does not affect the course of nephropathy and retinopathy or vice versa. The infant's size is directly

related to maternal blood glucose levels, as docu­

mented by serial glycosylated hemoglobin levels. Thus newborn weight could be predetermined by hemoglo­

bin A1 and optimal delivery management can be pre­scribed.

We wish to thank Ms. Felicia Anzel for her assistance in editing the manuscript.

REFERENCES

I. Jovanovic L, Druzin M, and Peterson CM. Effect of eu­glycemia on the outcome of pregnancy in insulin-depen­dent diabetic women as compared with normal control subjects. Amj Med 1981;71:921.

2. White P. Pregnancies in diabetes complicated by vascular disease. Diabetes 1977;26:953.

3. Mills JL, Baker L, Goldman AS. Malformations in infants of diabetic mothers occur before the seventh gestational week: implications for treatment. Diabetes 1979;28:292.

4. Miller E, Hare JW, Clogerty JP, Dunn PJ, Gleason RE, Soeldner JS, Kitzmiller JL. Elevated maternal hemoglo­bin A1c in early pregnancy and major congenital anoma­lies in infants of diabetic mothers. N Eng! J Med 1981; 304:1331.

5. Saxena BB, Hasan SH, Haour F, Schmidt-Gollwitzer M. Early diagnosis of pregnancy. Science 1974;184:793.

6. Rayburn W, Zuspan F, Motley ME, Donaldson M. An alternative to antepartum fetal heart rate testing. AM J 0BSTET CYNECOL 1980; 138:223.

7. Jovanovic L, Peterson CM. Insulin and glucose require­ments during the first stage of labor in insulin­independent diabetic women. Am J Med 1983;75:607.

8. Jovanovic L, Peterson CM, Saxena BB, Dawood MY, Saudek CD. Feasibility of maintaining normal glucose profiles in insulin-dependent pregnant diabetic women. Am] Med 1980;68:105.

9. Weingold AB, Yonekura ML, O'Kieffe J. Nonstress test­ing. AMJ OBSTET CYNECOL 1980;138:195.

10. Leveno KJ, Williams ML, DePalma RT, Whalley PJ. Peri­natal outcome in the absence of antepartum fetal heart rate acceleration. Obstet Gynecol 1983;61:347.

11. Philipps AF, Dubin JW, Matty PJ, Raye JR. Effects of chronic hyperglycemia in the fetal lamb. Pediatr Res 1981; 15:638.

12. Mitchell MD, Keirse MJ, Anderson AB. Evidence for a local control of prostaglandins within the pregnant human uterus. Br J Obstet Gynaecol 1977;84:35.

13. Stys SJ, Dresser BL, Otte, TE, Clark KE. Effect of pros­taglandin E2 on cervical compliance in pregnant ewes. AM j 0BSTET CYNECOL 1981;140:415.

14. Junqueria LCU, Zugaib M, Montes GS, Toledo OMS, Krisztan RM, Shigihara KM. Morphologic and histo­chemical evidence for the occurrence of collagenolysis and for the role of neutrophilic polymorphonuclear leu­kocytes during cervical dilation. AM J 0BSTET CYNECOL 1980;138:273.

15. Soler G, Maling JM. Diabetic pregnancy: management of diabetes on the day of delivery. Diabetologia 1978; 15:441.

16. Pedersen JF, Molsted-Pedersen L. Early fetal growth delay detected by ultrasound marks increased risk of congenital malformation in diabetic pregnancy. Br Medj 1981;283:269.