Content uploaded by Terry R J Lappin
Author content
All content in this area was uploaded by Terry R J Lappin on Mar 27, 2014
Content may be subject to copyright.
n engl j med 358;19 www.nejm.org may 8, 2008
1991
The new england
journal of medicine
established in 1812
may 8, 2008
vol. 358 no. 19
Hyperglycemia and Adverse Pregnancy Outcomes
The HAPO Study Cooperative Research Group*
A B S T R AC T
The members of the Writing Group (Boyd
E. Metzger, M.D., Lynn P. Lowe, Ph.D.,
Alan R. Dyer, Ph.D., Northwestern Uni-
versity Feinberg School of Medicine, Chi-
cago; Elisabeth R. Trimble, M.D., Queen’s
University Belfast, Belfast, Northern Ire -
land; Udom Chaovarindr, M.D., Rajavithi
Hospital, Bangkok, Thailand; Donald R.
Coustan, M.D., Women and Infants’ Hos-
pital of Rhode Island–Brown University
Medical School, Providence, RI; David R.
Hadden, M.D., David R. McCance, M.D.,
Royal Jubilee Maternity Hospital, Belfast,
Northern Ireland; Moshe Hod, M.D.,
Helen Schneider Hospital for Women,
Rabin Medical Center–Sackler Faculty of
Medicine, Tel-Aviv University, Petah-
Tiqva, Israel; Harold David McIntyre, M.B.,
B.S., Jeremy J.N. Oats, M.D., Mater Mi-
sericordiae Mothers’ Hospital– University
of Queensland, Brisbane, Australia; Bengt
Persson, M.D., Ph.D., Karolinska Institute,
Stockholm, Sweden; Michael S. Rogers,
M.D., Prince of Wales Hospital–Chinese
University of Hong Kong, Hong Kong;
and David A. Sacks, M.D., Kaiser Foun-
dation Hospital, Bellflower, CA) of the
Hyperglycemia and Adverse Pregnancy
Outcome (HAPO) Study Cooperative Re-
search Group assume responsibility for the
overall content and integrity of the article.
Address reprint requests to Dr. Metzger
at the Northwestern University Feinberg
School of Medicine, Endocrinology, 645
N. Michigan Ave., Suite 530-22, Chicago,
IL 60611, or at bem@northwestern.edu.
*Members of the Hyperglycemia and Ad-
verse Pregnancy Outcome (HAPO) Study
Cooperative Research Group are listed in
the Appendix.
N Engl J Med 2008;358:1991-2002.
Copyright © 2008 Massachusetts Medical Society.
BACKGR OUND
It is controversial whether maternal hyperglycemia less severe than that in diabetes
mellitus is associated with increased risks of adverse pregnancy outcomes.
METHO DS
A total of 25,505 pregnant women at 15 centers in nine countries underwent 75-g
oral glucose-tolerance testing at 24 to 32 weeks of gestation. Data remained blinded
if the fasting plasma glucose level was 105 mg per deciliter (5.8 mmol per liter) or
less and the 2-hour plasma glucose level was 200 mg per deciliter (11.1 mmol per
liter) or less. Primary outcomes were birth weight above the 90th percentile for ges-
tational age, primary cesarean delivery, clinically diagnosed neonatal hypoglyce-
mia, and cord-blood serum C-peptide level above the 90th percentile. Secondary
outcomes were delivery before 37 weeks of gestation, shoulder dystocia or birth
injury, need for intensive neonatal care, hyperbilirubinemia, and preeclampsia.
RESU LTS
For the 23,316 participants with blinded data, we calculated adjusted odds ratios for
adverse pregnancy outcomes associated with an increase in the fasting plasma glu-
cose level of 1 SD (6.9 mg per deciliter [0.4 mmol per liter]), an increase in the 1-hour
plasma glucose level of 1 SD (30.9 mg per deciliter [1.7 mmol per liter]), and an in-
crease in the 2-hour plasma glucose level of 1 SD (23.5 mg per deciliter [1.3 mmol
per liter]). For birth weight above the 90th percentile, the odds ratios were 1.38
(95% confidence interval [CI], 1.32 to 1.44), 1.46 (1.39 to 1.53), and 1.38 (1.32 to 1.44),
respectively; for cord-blood serum C-peptide level above the 90th percentile, 1.55
(95% CI, 1.47 to 1.64), 1.46 (1.38 to 1.54), and 1.37 (1.30 to 1.44); for primary cesar-
ean delivery, 1.11 (95% CI, 1.06 to 1.15), 1.10 (1.06 to 1.15), and 1.08 (1.03 to 1.12);
and for neonatal hypoglycemia, 1.08 (95% CI, 0.98 to 1.19), 1.13 (1.03 to 1.26), and
1.10 (1.00 to 1.12). There were no obvious thresholds at which risks increased. Sig-
nificant associations were also observed for secondary outcomes, although these
tended to be weaker.
CONCL USIONS
Our results indicate strong, continuous associations of maternal glucose levels below
those diagnostic of diabetes with increased birth weight and increased cord-blood se-
rum C-peptide levels.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
T h e
ne w e ng l an d j our n a l
o f
me di ci ne
n engl j med 358;19 www.nejm.org may 8, 2008
1992
G
estational diabetes mellitus, de-
fined as “glucose intolerance with onset
or f irst recognition during pregnancy,”
1,2
has been the subject of considerable controversy.
Criteria for the diagnosis were initially established
more than 40 years ago
3
and, with minor modi-
fications, remain in use today. These criteria are
not designed to identify pregnant women who are
at increased risk for adverse perinatal outcomes
but rather women who are at high risk for the de-
velopment of diabetes after pregnancy,
3,4
or they
are the criteria used for the general population.
5
Overt diabetes mellitus during pregnancy is
associated with signif icantly increased risks of
adverse perinatal outcomes. Whereas some data
suggest that current diagnostic criteria for ges-
tational diabetes mellitus
1
are too restrictive and
that lesser degrees of hyperglycemia also increase
risk,
6-11
risks associated with hyperglycemia that
is less severe than that diagnostic of overt diabetes
mellitus are uncertain for a number of reasons.
First, there are no uniform international standards
for the ascertainment and diagnosis of gesta-
tional diabetes mellitus.
2
In addition, the extent
to which adverse outcomes associated with gesta-
tional diabetes mellitus may be explained by con-
founders (including obesity, advanced maternal
age, or associated medical complications) is un-
clear.
12-14
Caregiver bias (i.e., an expectation of
adverse outcomes due to gestational diabetes mel-
litus) may increase the likelihood of disorders or
problems due to increased intervention.
15
We conducted the Hyperglycemia and Adverse
Pregnancy Outcome (HAPO) study to clarify the
risks of adverse outcomes associated with various
degrees of maternal glucose intolerance less se-
vere than that in overt diabetes mellitus.
M e t h o d s
The protocol was approved by the institutional re-
view board at each field center. All participants
gave written informed consent. An external data
and safety monitoring committee provided over-
sight. The study methods have been published pre-
viously.
16,17
A brief overview is presented here.
Part icipant s
All pregnant women at a given center were eligi-
ble to participate unless they had one or more of
the following exclusion criteria
16
: age younger than
18 years, a plan to undergo delivery at another hos-
pital, an uncertain date of last menstrual period
and no ultrasonographic estimation between 6 and
24 weeks of gestational age, inability to complete
the oral glucose-tolerance test within 32 weeks
of gestation, multiple pregnancy, conception by
means of gonadotropin ovulation induction or in
vitro fertilization, glucose testing before recruit-
ment or a diagnosis of diabetes during the cur-
rent pregnancy, diagnosis of diabetes before the
current pregnancy and requiring treatment with
medication, participation in another study that
could interfere with the HAPO study, infection
with the human immunodeficiency virus or hep-
atitis B or C virus, previous participation in the
HAPO study, or inability to converse in the lan-
guages used on center forms without the aid of an
interpreter. If glucose measurements were made
outside the setting of the HAPO study after ini-
tial enrollment, participation was terminated. Age
and education level were recorded for women who
declined to participate.
Gestational age and expected date of delivery
were determined from the date of the last men-
strual period, if the date was certain. If the date
was uncertain, the expected date of delivery was
estimated by means of ultrasonography performed
between 6 and 24 weeks of gestation. The final
expected date of delivery was also determined with
the use of ultrasonography if the gestational age
estimated on the basis of the date of the last men-
strual period differed by more than 5 days from
that based on ultrasonography performed between
6 and 13 weeks of gestation or by more than 10
days from that based on ultrasonography per-
formed between 14 and 24 weeks of gestation.
Ora l Gluco se-To leran ce Tes t
Participants underwent a standard oral glucose-
tolerance test, with the use of a 75-g dose of glu-
cose, between 24 and 32 weeks of gestation (tar-
get time of testing, 28 weeks). Height, weight, and
blood pressure were measured at the test visit. Data
concerning smoking and alcohol use, history of
diabetes and hypertension among first-degree fam-
ily members, and demographic characteristics were
collected by means of standardized questionnaires.
Race or ethnic group was self-reported by partici-
pants. A blood specimen was collected between
34 and 37 weeks of gestation for evaluation of the
random plasma glucose level, as a safety measure
to identify cases with hyperglycemia above a pre-
defined threshold.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
n engl j med 358;19 www.nejm.org may 8, 2008
1993
Gluco se Analys is
For purposes of clinical decision making, plasma
glucose levels were measured at center laborato-
ries by means of enzymatic methods, with exten-
sive quality control, as previously described.
17
To
avoid the confounding effects of analytic variation
among centers, aliquots of all oral glucose-toler-
ance test specimens were analyzed at the central
laboratory of the HAPO study, and the results were
used in the analyses reported here.
Unblind ing of Data
Fasting and 2-hour specimens from the oral glu-
cose-tolerance tests and the blood specimen taken
for determination of the random plasma glucose
level were analyzed at center laboratories. The data
were unblinded if the 2-hour plasma glucose lev-
el was diagnostic of diabetes (i.e., >200 mg per
deciliter [11.1 mmol per liter]) or, for ethical and
safety reasons, if the fasting plasma glucose level
exceeded 105 mg per deciliter (5.8 mmol per liter),
the random plasma glucose level was 160 mg per
deciliter (8.9 mmol per liter) or more, or any plas-
ma glucose level was less than 45 mg per deciliter
(2.5 mmol per liter). Otherwise, women, caregiv-
ers, and the staff of the HAPO study (except for
laboratory personnel) remained unaware of the
glucose values. Only women whose data were
blinded and who did not undergo any additional
glucose testing outside the HAPO study were in-
cluded in our analyses.
Cord -B lood P lasm a Gluco se and S erum
C-Pe ptid e Leve ls
Cord-blood specimens were collected at delivery
for the measurement of serum C-peptide and plas-
ma glucose levels. The specimens were analyzed
at the central laboratory with the use of an im-
munoassay (AutoDELFIA, PerkinElmer) for serum
C peptide and a chemical analyzer (Vitros 750,
Ortho-Clinical Diagnostics) for plasma glucose.
17
Because approximately 15% of cord-blood samples
have detectable hemolysis after serum or plasma
is separated out, because hemolysis is known to
increase insulin degradation but not to affect
C-peptide level,
17
and because C peptide and in-
sulin are secreted in equimolar levels, we used
cord-blood serum C-peptide level rather than in-
sulin level as our index of fetal β-cell function.
Prenatal Care , Del ivery, and Neonatal C are
Prenatal care, timing of delivery, and neonatal care
were determined by means of the standard prac-
tice at each center. No center arbitrarily induced
delivery before full term or routinely performed
cesarean delivery at a specified maternal or ges-
tational age. Medical records were abstracted to
obtain data regarding the prenatal course, labor
and delivery, the postpartum course, and the new-
born course.
Outc omes
Primary and Secondary Outcomes
The four primary outcomes were birth weight
above the 90th percentile for gestational age, pri-
mary cesarean delivery, clinical neonatal hypogly-
cemia, and cord-blood serum C-peptide level above
the 90th percentile (fetal hyperinsulinemia). Sec-
ondary outcomes were premature delivery (before
37 weeks of gestation), shoulder dystocia or birth
injury, need for intensive neonatal care, hyperbili-
rubinemia, and preeclampsia.
Possible Severe Adverse Outcomes
Additional data were abstracted at centers when-
ever a possible severe adverse event (e.g., death,
shoulder dystocia, birth injury, and major malfor-
mation) was identified. Data were reviewed by the
members of an outcome review committee, who
were unaware of the mother’s glycemic status, to
confirm whether the event was present. Perinatal
deaths were classified according to the Australian
and New Zealand Antecedent Classification of
Perinatal Mortality guidelines,
18
and major mal-
formations were classified according to codes of
the International Classification of Diseases, Tenth Revi-
sion.
19
The HAPO data and safety monitoring com-
mittee reviewed data regarding adverse outcomes
and deaths after having been made aware of the
results of the oral glucose-tolerance test and the
random plasma glucose levels.
Statis tic al Analys is
Mean and standard deviation are reported for con-
tinuous variables, and number and percentage are
reported for categorical variables. Pearson product-
moment correlations were used to assess associ-
ations among glucose measures. For associations
of glycemia with primary outcomes, as prespec-
ified in the HAPO study protocol, each glucose
measurement was considered as both a categori-
cal and a continuous variable in multiple logistic-
regression analyses. For secondary outcomes, only
results for continuous variables are presented. For
categorical analyses, each measure of glycemia
was divided into seven categories, such that the
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
T h e
ne w e ng l an d j our n a l
o f
me di ci ne
n engl j med 358;19 www.nejm.org may 8, 2008
1994
1- and 2-hour plasma glucose measures reflected
data for approximately the same number of wom-
en in each category as did the fasting plasma
glucose measure. Categories for the fasting plas-
ma glucose level were originally prespecified in
the HAPO study protocol as 100 mg per deciliter
(5.6 mmol per liter) or more, 95 to 99 (5.3 to 5.5),
90 to 94 (5.0 to 5.2), 85 to 89 (4.8 to 4.9), and less
than 85. Subsequently, the category of less than
85 mg per deciliter was further subdivided into
less than 75 mg per deciliter (4.2 mmol per liter),
75 to 79 (4.2 to 4.4), and 80 to 84 (4.5 to 4.7). The
highest and second-highest categories for each
glucose measure, accounting for 1% and 3% of
participants, respectively, were specifically chosen
to allow for assessment of whether there were
threshold effects.
For continuous-variable analyses, odds ratios
were calculated for a 1-SD increase in fasting,
1-hour, and 2-hour plasma glucose levels. As pre-
specified, to assess whether the log of the odds of
each outcome was linearly related to glucose level,
we added squared terms for glucose level for each
adverse pregnancy outcome to assess whether
there were significant quadratic associations. For
each outcome, two logistic models were fit.
Table 1. Characteristics of the Study Participants and Their Newborns and Frequency of Outcomes.*
Characteristic or Outcome
No. of Participants
(%) Mean ±SD
Range of Means
among Centers
Maternal characteristics
Age (yr) 23,316 (100) 29.2±5.8 25.4–33.6
Body-mass index 23,316 (100) 27.7±5.1 24.4–29.9
Mean arterial pressure (mm Hg) 23,316 (100) 80.9±8.3 75.9–84.1
Plasma glucose (mg/dl)
Fasting 23,316 (100) 80.9±6.9 78.2–83.7
1 hr 23,316 (100) 134.1±30.9 119.5–148.2
2 hr 23,316 (100) 111.0±23.5 99.6–120.9
Length of gestation at time of OGTT (wk) 23,316 (100) 27.8±1.8 25.9–29.5
Any prenatal smoking — % 1,581 (6.8) 0.2–23.7
Any prenatal alcohol use — % 1,612 (6.9) 0.1–26.5
Family history of diabetes — % 5,282 (22.7) 12.1–37.7
Parity (delivery ≥20 wk) at enrollment 12,233 (52.5) 34.8–68.9
Prenatal urinary tract infection — % 1,655 (7.1) 0.4–22.5
Hospitalization before delivery — % 3,271 (14.0) 2.3–33.2
Newborn characteristics
Gestational age at delivery (wk) 23,316 (100) 39.4±1.7 38.7–39.9
Birth weight (g) 23,217 (99.6) 3292±529 3109–3526
Cord-blood serum C peptide (μg/liter) 19,885 (85.3) 1.0±0.6 0.9–1.2
Cord-blood plasma glucose (mg/dl) 19,859 (85.2) 81.5±19.6 74.2–90.4
Male sex — % 12,003 (51.5) 49.3–54.0
Obstetrical outcomes
Cesarean delivery — %
Primary 3,731 (16.0) 8.6–23.5
Repeat 1,792 (7.7) 3.3–11.8
Hypertension — %†
Chronic hypertension 582 (2.5) 0.5–9.4
Gestational hypertension 1,370 (5.9) 0.7–17.7
Preeclampsia 1,116 (4.8) 1.4–11.4
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
n engl j med 358;19 www.nejm.org may 8, 2008
1995
Model I included adjustment for center or the
variables used in estimating the 90th percentile
for birth weight for gestational age (infant’s sex,
race or ethnic group, center, and parity). Model II
included adjustment for multiple potential pre-
specified confounders, including age, body-mass
index (BMI), smoking status, alcohol use, pres-
ence or absence of a family history of diabetes,
gestational age at the oral glucose-tolerance test,
sex of the infant, parity (0, 1, or ≥2, except for
primary cesarean deliveries), mean arterial pres-
sure and presence or absence of hospitalization
before delivery (except for preeclampsia), and pres-
ence or absence of a family history of hyperten-
sion and maternal urinary tract infection (for
analysis of preeclampsia only). Height was also
included as a potential confounder, on the basis
of post hoc findings of an association with birth
weight greater than the 90th percentile, and two
prespecified confounders (maternal urinary tract
infection and previous prenatal death) were exclud-
ed from primary and secondary outcome analy-
ses when neither was found to be related to any
primary outcome or to affect primary outcome–
glucose associations. Squared terms for age, BMI,
and mean arterial pressure were prescreened for
possible inclusion in model II adjustment when
only the center had been included (i.e., models
without glucose or other covariates); these terms
were included in model II when significant. Only
Table 1. (Continued.)
Characteristic or Outcome
No. of Participants
(%) Mean ±SD
Range of Means
among Centers
Newborn outcomes
Birth weight >90th percentile — %‡ 2,221 (9.5) 9.0–9.9
Clinical neonatal hypoglycemia — %§ 480 (2.1) 0.3–6.4
Cord-blood serum C peptide >90th percentile — %¶ 1,671 (8.4) 5.9–15.1
Premature delivery (before 37 wk) — % 1,608 (6.9) 3.9–9.1
Shoulder dystocia or birth injury 311 (1.3) 0.1–3.4
Intensive neonatal care — %‖ 1,855 (8.0) 3.0–28.8
Hyperbilirubinemia — %** 1,930 (8.3) 3.0–25.4
* The body-mass index (the weight in kilograms divided by the square of the height in meters), mean arterial pressure,
and glucose levels were obtained at the oral glucose-tolerance test (OGTT). To convert the values for glucose to milli-
moles per liter, multiply by 0.05551. For additional characteristics of the participants, see Table A in the Supplemen-
tary Appendix.
†
Chronic hypertension was defined as hypertension that was present before 20 weeks of gestation and that did not
progress to preeclampsia. Hypertension disorders occurring during pregnancy after 20 weeks of gestation were cate-
gorized according to the International Society for the Study of Hypertension guidelines.
20
Preeclampsia was defined
as systolic blood pressure of 140 mm Hg or more or diastolic blood pressure of 90 mm Hg or more on two or more
occasions a minimum of 6 hours apart and proteinuria of 1+ or more on a dipstick test or a protein level in the urine
of 300 mg or more for a 24-hour period. If the criteria for elevated blood pressure were met but those for proteinuria
were not, the hypertension was classified as gestational hypertension.
‡ For birth weight above the 90th percentile, the 90th percentiles for gestational age (30 to 44 weeks only) were deter-
mined with the use of quantile regression analyses for each of eight groups of newborns based on infant’s sex and
race or ethnic group (white or other, black, Hispanic, or Asian), with adjustment for gestational age, field center, and
parity (0, 1, or ≥2). A newborn was considered to have a birth weight above the 90th percentile if the birth weight was
greater than the estimated 90th percentile for the infant’s sex, gestational age, race or ethnic group, field center, and
maternal parity. Otherwise, the newborn was considered to have a birth weight at or under the 90th percentile. Birth
weights were available for 23,217 newborns.
§ Clinical neonatal hypoglycemia was defined as being present if there was a notation of neonatal hypoglycemia in the
medical record and there were symptoms or treatment with a glucose infusion or a local laboratory report of a glu-
cose value of 30.6 mg per deciliter (1.7 mmol per liter) or less in the first 24 hours after birth or 45.0 mg per deciliter
(2.5 mmol per liter) or less after the first 24 hours.
21
¶ A cord-blood serum C-peptide level above the 90th percentile was defined on the basis of the 19,885 newborns for
whom data were available.
‖ Intensive neonatal care was defined by admission to any type of unit for care more intensive than normal newborn
care and lasting more than 24 hours or by death of the baby or transfer to another hospital. Data were excluded for
admissions that were only for possible sepsis or sepsis, observation, or feeding problems.
** Hyperbilirubinemia was defined by treatment with phototherapy after birth, at least one laboratory report of a biliru-
bin level of 20 mg per deciliter (342 μmol per liter) or more, or readmission for hyperbilirubinemia.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
T h e
ne w e ng l an d j our n a l
o f
me di ci ne
n engl j med 358;19 www.nejm.org may 8, 2008
1996
the fully adjusted model results are presented in
this report. (Results for all models are given in
Tables B through F in the Supplementary Ap-
pendix, available with the full text of this article
at www.nejm.org.)
In post hoc analyses, we tested for interactions
of each glucose measure with center in models I
and II (except with regard to the outcomes of
clinical neonatal hypoglycemia and shoulder dys-
tocia or birth injury, owing to small numbers in
some centers [a total of 42 tests]). We also tested
for interactions of each glucose measure with BMI,
age, height, and mean arterial pressure in model
II (105 tests). P values less than 0.001 were con-
sidered to indicate statistical significance for
squared terms for glucose, age, BMI, and mean
arterial pressure and interaction terms for all out-
comes, except neonatal hypoglycemia and shoul-
der dystocia or birth injury, for which P values less
than 0.05 were considered to indicate statistical
signif icance, owing to the smaller numbers of
babies with these outcomes.
All analyses were conducted in SAS version 9.1
or Stata 10.0. All reported P values are two-sided
and were not adjusted for multiple testing.
R e s u lt s
Part icipant s
Among 53,295 eligible women (from 15 centers in
nine countries; see the Appendix), 28,562 (53.6%)
agreed to participate in this blinded study, between
July 2000 and April 2006. Among the women who
agreed to participate and those who refused to par-
ticipate, the mean age and years of education were
29.0 and 12.9 years and 28.5 and 12.5 years, re-
spectively. A total of 25,505 women completed an
oral glucose-tolerance test: 746 (2.9%) were exclud-
ed because their data were unblinded, 1412 (5.5%)
were excluded primarily because they had under-
gone glucose testing or delivery outside the con-
text of the HAPO study, and 31 (0.1%) were ex-
cluded owing to missing key data or an implausible
gestational age (>44 weeks); data from the remain-
ing 23,316 were available for analyses. Of these
23,316 patients, 48.3% were self-reported to be
white, 11.6% to be black, 8.5% to be Hispanic,
29.0% to be “Asian or Oriental,” and 2.6% to be
of other races or ethnic groups.
Characteristics of the mothers and newborns
and pregnancy outcomes are summarized in
Ta-
ble 1
. The mean age of participants was 29.2 years,
and the mean fasting, 1-hour, and 2-hour plasma
glucose levels were 80.9 mg per deciliter (4.5 mmol
per liter), 134.1 mg per deciliter (7.4 mmol per li-
ter), and 111.0 mg per deciliter (6.2 mmol per li-
ter), respectively. Correlations among these three
glucose measures were as follows: 0.38 for the
fasting plasma glucose level and the 1-hour plas-
ma glucose level, 0.30 for the fasting plasma glu-
cose level and the 2-hour plasma glucose level,
and 0.68 for the 1-hour plasma glucose level and
the 2-hour plasma glucose level (P<0.001 for all
three comparisons). There were 2 maternal deaths
(1 due to pulmonary embolism, the other due to
respirator y failure secondary to pneumonia), 14
cases of eclampsia, 321 cases of major malforma-
tion of the newborn, and 130 perinatal deaths (89
fetal and 41 neonatal or infant) (incidence, 5.6 per
1000) among the 23,316 deliveries.
Glycemia a nd Pregnan cy Ou tcomes
Categorical Analyses
The frequency of each primary outcome across the
seven glucose categories is shown in
Figure 1
. With
increasing maternal glucose levels, the frequency
of each primary outcome increased, although less
so for clinical neonatal hypoglycemia than for the
other outcomes. For example, for the fasting plas-
ma glucose level, frequencies in the lowest and
highest categories, respectively, were 5.3% and
26.3% for birth weight above the 90th percentile,
13.3% and 27.9% for primary cesarean section,
2.1% and 4.6% for clinical neonatal hypoglycemia,
and 3.7% and 32.4% for C-peptide level above the
90th percentile.
Table 2
shows the associations of maternal
glucose as a categorical variable with each primary
outcome, including odds ratios and 95% confi-
dence intervals for each category, as compared
with the lowest category, with adjustment for con-
founders. There were strong associations with
birth weight above the 90th percentile that in-
creased across the increasing glycemia categories.
Differences in mean birth weight between the low-
est and highest categories of the fasting, 1-hour,
and 2-hour plasma glucose levels were 242 to
305 g when birth weight was modeled as a con-
tinuous variable, with adjustment for multiple
confounders (data not shown). The odds ratio for
primary cesarean section increased across catego-
ries of maternal glycemia and was 1.86 in the
highest category of 1-hour plasma glucose; the
odds ratio in the highest category of 2-hour plasma
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
n engl j med 358;19 www.nejm.org may 8, 2008
1997
glucose did not differ significantly from 1.00. The
adjusted odds ratios for clinical neonatal hypogly-
cemia were substantially attenuated, and none of
the odds ratios for the highest glucose categories
were significantly different from 1.00. After ad-
justment for confounders, there was a strong as-
sociation between cord-blood serum C-peptide
level above the 90th percentile and maternal glyce-
mia, with the association increasing with increas-
ing glycemia category; the odds ratio was 7.65
(95% confidence interval [CI], 5.17 to 11.32) for the
highest category of the fasting plasma glucose.
Continuous Analyses
Results for analyses of glucose level as a continu-
ous variable, with model II adjustment for both
primary and secondary outcomes, are shown in
Table 3
. Among the primary outcomes, odds ra-
tios for an increase in the glucose level by 1 SD
were highest for birth weight greater than the 90th
percentile (range, 1.38 to 1.46) and cord-blood
serum C-peptide level above the 90th percentile
(range, 1.37 to 1.55). For primary cesarean sec-
tion and clinical neonatal hypoglycemia, the as-
sociations were weaker and the associations of
33p9
30
Frequency (%)
20
25
15
10
5
0
2 3 4 5 71 6
Glucose Category
CClinical Neonatal Hypoglycemia
ABirth Weight >90th Percentile
35
30
Frequency (%)
20
25
15
10
5
0
2 3 4 5 71 6
Glucose Category
BPrimary Cesarean Section
5
Frequency (%)
3
4
2
1
0
2 3 4 5 71 6
Glucose Category
35
30
Frequency (%)
20
25
15
10
5
0
2 3 4 5 71 6
Glucose Category
DCord-Blood Serum C Peptide >90th Percentile
AUTHOR:
FIGURE:
JOB:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Metzger
1 of 1
05-08-08
ARTIST: ts
35819 ISSUE:
Fasting glucose 1-Hr glucose 2-Hr glucose
Figure 1. Frequency of Primary Outcomes across the Glucose Categories.
Glucose categories are defined as follows: fasting plasma glucose level — category 1, less than 75 mg per deciliter
(4.2 mmol per liter); category 2, 75 to 79 mg per deciliter (4.2 to 4.4 mmol per liter); category 3, 80 to 84 mg per
deciliter (4.5 to 4.7 mmol per liter); category 4, 85 to 89 mg per deciliter (4.8 to 4.9 mmol per liter); category 5, 90 to
94 mg per deciliter (5.0 to 5.2 mmol per liter); category 6, 95 to 99 mg per deciliter (5.3 to 5.5 mmol per liter); and
category 7, 100 mg per deciliter (5.6 mmol per liter) or more; 1-hour plasma glucose level — category 1, 105 mg per
deciliter (5.8 mmol per liter) or less; category 2, 106 to 132 mg per deciliter (5.9 to 7.3 mmol per liter); category 3,
133 to 155 mg per deciliter (7.4 to 8.6 mmol per liter); category 4, 156 to 171 mg per deciliter (8.7 to 9.5 mmol per
liter); category 5, 172 to 193 mg per deciliter (9.6 to 10.7 mmol per liter); category 6, 194 to 211 mg per deciliter
(10.8 to 11.7 mmol per liter); and category 7, 212 mg per deciliter (11.8 mmol per liter) or more; and 2-hr plasma
glucose level — categor y 1, 90 mg per deciliter (5.0 mmol per liter) or less; category 2, 91 to 108 mg per deciliter
(5.1 to 6.0 mmol per liter); category 3, 109 to 125 mg per deciliter (6.1 to 6.9 mmol per liter); category 4, 126 to 139 mg
per deciliter (7.0 to 7.7 mmol per liter); category 5, 140 to 157 mg per deciliter (7.8 to 8.7 mmol per liter); category 6,
158 to 177 mg per deciliter (8.8 to 9.8 mmol per liter); and category 7, 178 mg per deciliter (9.9 mmol per liter) or more.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
T h e
ne w e ng l an d j our n a l
o f
me di ci ne
n engl j med 358;19 www.nejm.org may 8, 2008
1998
clinical neonatal hypoglycemia with the fasting
plasma glucose level and the 2-hour plasma glu-
cose level were not significant.
Twelve of the 15 analyses of secondary out-
comes showed significant positive associations
with maternal glycemia, after adjustment for con-
founders (
Table 3
). The strongest associations were
found for preeclampsia, for which the odds ratio
for each 1-SD increase in each glucose measure
ranged from 1.21 to 1.28; corresponding odds ra-
tios for shoulder dystocia or birth injury were ap-
proximately 1.20. Premature delivery, intensive
neonatal care, and hyperbilirubinemia were sig-
nificantly related to the 1-hour and 2-hour plasma
glucose levels but not to the fasting plasma glu-
cose level.
The only significant quadratic (nonlinear) as-
sociation found in these analyses was for the fast-
ing plasma glucose level with clinical neonatal
hypoglycemia (P = 0.01). Among the 147 tests for
interactions, 7 were significant: the fasting plas-
ma glucose level with field center, in relation to
primary cesarean delivery, in both models I and II
(P<0.001); age with the fasting, 1-hour, and 2-hour
plasma glucose levels, with regard to clinical neo-
natal hypoglycemia (P = 0.05, P = 0.03, and P = 0.001,
respectively); BMI and 1-hour plasma glucose level,
for clinical neonatal hypoglycemia (P<0.001); and
mean arterial pressure with fasting plasma glu-
cose level, for premature delivery (P<0.001).
We also examined associations of glucose mea-
sures with birth weight below the 10th percentile
for gestational age, using the same methods to
estimate the 10th percentiles as were used to es-
Table 2. Adjusted Odds Ratios for Associations between Maternal Glucose as a Categorical Variable and Primary Outcomes.*
Glucose
Category† Plasma Glucose Level
Fasting At 1 Hr At 2 Hr
total no.
(no. with outcome)
odds ratio
(95% CI)
total no.
(no. with outcome)
odds ratio
(95% CI)
total no.
(no. with outcome)
odds ratio
(95% CI)
Birth weight >90th percentile
1 4035 (213) 1.00 4177 (268) 1.00 4264 (297) 1.00
2 7501 (572) 1.37 (1.16–1.62) 7524 (584) 1.21 (1.04–1.41) 7422 (587) 1.11 (0.96–1.30)
3 6168 (622) 1.72 (1.46–2.03) 6003 (593) 1.65 (1.41–1.93) 5865 (580) 1.51 (1.30–1.75)
4 2741 (323) 1.95 (1.62–2.35) 2768 (352) 2.27 (1.91–2.71) 3024 (396) 2.15 (1.82–2.54)
5 1883 (310) 2.73 (2.25–3.31) 1858 (264) 2.66 (2.19–3.21) 1720 (210) 2.10 (1.73–2.56)
6 672 (124) 3.00 (2.34–3.86) 645 (111) 3.50 (2.72–4.50) 690 (101) 2.68 (2.08–3.45)
7 217 (57) 5.01 (3.54–7.09) 242 (49) 4.49 (3.16–6.39) 232 (50) 4.46 (3.15–6.33)
Primary cesarean section‡
1 3721 (495) 1.00 3826 (458) 1.00 3903 (535) 1.00
2 6806 (1151) 1.19 (1.06–1.34) 6792 (1113) 1.21 (1.07–1.36) 6664 (1032) 0.97 (0.86–1.09)
3 5483 (1014) 1.21 (1.07–1.37) 5311 (1032) 1.26 (1.11–1.42) 5201 (1017) 1.11 (0.99–1.26)
4 2378 (506) 1.33 (1.15–1.54) 2425 (522) 1.31 (1.13–1.52) 2650 (583) 1.15 (1.00–1.32)
5 1601 (380) 1.44 (1.23–1.69) 1623 (407) 1.48 (1.26–1.74) 1506 (350) 1.17 (0.99–1.37)
6 560 (134) 1.39 (1.11–1.75) 547 (132) 1.30 (1.04–1.64) 615 (162) 1.32 (1.08–1.63)
7 183 (51) 1.60 (1.12–2.27) 208 (67) 1.86 (1.35–2.57) 193 (52) 1.28 (0.91–1.81)
Clinical neonatal hypoglycemia
1 4043 (83) 1.00 4183 (72) 1.00 4266 (78) 1.00
2 7503 (144) 0.91 (0.69–1.21) 7523 (153) 1.12 (0.84–1.49) 7421 (134) 0.87 (0.66–1.17)
3 6164 (122) 0.92 (0.68–1.23) 6003 (131) 1.24 (0.92–1.68) 5868 (117) 0.96 (0.71–1.30)
4 2744 (59) 1.00 (0.70–1.43) 2772 (54) 1.11 (0.77–1.62) 3027 (80) 1.23 (0.88–1.71)
5 1884 (48) 1.19 (0.81–1.75) 1860 (45) 1.48 (0.99–2.22) 1720 (44) 1.13 (0.76–1.68)
6 672 (14) 1.01 (0.55–1.84) 643 (20) 2.17 (1.28–3.69) 693 (21) 1.36 (0.81–2.28)
7 217 (10) 1.98 (0.97–4.05) 243 (5) 1.29 (0.51–3.31) 232 (6) 1.12 (0.47–2.67)
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
n engl j med 358;19 www.nejm.org may 8, 2008
1999
timate the 90th percentiles (
Table 1
). In the con-
tinuous-variable models, odds ratios for each 1-SD
increase in glucose measures ranged from 0.77 to
0.80, with no evidence of nonlinear associations
(data not shown) and little difference from the
results from unadjusted models.
Although the HAPO study did not have the
statistical power to permit examination of perina-
tal death as a primary outcome, with only 130
deaths, unadjusted analyses showed no increase in
the risk of perinatal death with increasing glu-
cose levels. Unadjusted odds ratios for perinatal
death for each 1-SD increase in the fasting, 1-hour,
and 2-hour plasma glucose levels, respectively,
were 0.91 (95% CI, 0.76 to 1.08), 0.93 (95% CI, 0.78
to 1.11), and 0.99 (95% CI, 0.83 to 1.18), with no
evidence of nonlinear associations.
Di s c u s s io n
In 1952, Jorgen Pedersen
22
postulated that ma-
ternal hyperglycemia led to fetal hyperglycemia,
which evoked an exaggerated fetal response to in-
sulin. Since then, the Pedersen hypothesis has
formed the basis for understanding the pathophys-
iological consequences of diabetes during preg-
nancy. The objective of the HAPO study was to
clarify risks of adverse outcomes associated with
degrees of maternal glucose intolerance less se-
vere than overt diabetes mellitus. The data pre-
sented here show associations between increasing
levels of fasting, 1-hour, and 2-hour plasma glucose
obtained on oral glucose-tolerance testing and
birth weight above the 90th percentile and cord-
blood serum C-peptide level above the 90th per-
centile, with weaker associations between glucose
levels and primary cesarean delivery and clinical
neonatal hypoglycemia. We also found positive as-
sociations between increasing plasma glucose lev-
els and each of the f ive secondary outcomes ex-
amined: premature delivery, shoulder dystocia or
birth injury, intensive neonatal care, hyperbiliru-
binemia, and preeclampsia.
Associations between maternal glycemia and
Table 2. (Continued.)
Glucose
Category† Plasma Glucose Level
Fasting 1 Hr 2 Hr
total no.
(no. with outcome)
odds ratio
(95% CI)
total no.
(no. with outcome)
odds ratio
(95% CI)
total no.
(no. with outcome)
odds ratio
(95% CI)
Cord-blood serum C peptide >90th percentile
1 3546 (131) 1.00 3593 (176) 1.00 3599 (193) 1.00
2 6453 (378) 1.41 (1.15–1.74) 6372 (366) 1.07 (0.88–1.29) 6353 (401) 1.06 (0.88–1.27)
3 5255 (429) 1.75 (1.42–2.15) 5132 (458) 1.62 (1.34–1.95) 5039 (440) 1.44 (1.20–1.73)
4 2308 (266) 2.36 (1.88–2.97) 2424 (274) 1.95 (1.58–2.41) 2609 (286) 1.72 (1.40–2.11)
5 1592 (181) 3.62 (2.87–4.58) 1607 (251) 2.76 (2.21–3.43) 1495 (202) 2.21 (1.77–2.76)
6 561 (131) 4.46 (3.36–5.93) 549 (95) 2.91 (2.18–3.89) 596 (109) 2.86 (2.18–3.77)
7 170 (55) 7.65 (5.17–11.32) 208 (51) 4.65 (3.19–6.79) 194 (40) 3.48 (2.33–5.21)
* Associations were adjusted for the following variables: field center, age, body-mass index, height, smoking status, alcohol use, presence or
absence of family history of diabetes, gestational age at oral glucose-tolerance test, infant’s sex, presence or absence of hospitalization be-
fore delivery, mean arterial pressure (in all models), parity (0, 1, or ≥2; not included in the model for primary cesarean delivery), cord-blood
plasma glucose level (included in the model for cord-blood serum C-peptide level >90th percentile only). Additional details are shown in
Tables B, C, D, and E in the Supplementary Appendix.
† Glucose categories are defined as follows: fasting plasma glucose level — category 1, less than 75 mg per deciliter (4.2 mmol per liter); cat-
egory 2, 75 to 79 mg per deciliter (4.2 to 4.4 mmol per liter); category 3, 80 to 84 mg per deciliter (4.5 to 4.7 mmol per liter); category 4, 85
to 89 mg per deciliter (4.8 to 4.9 mmol per liter); category 5, 90 to 94 mg per deciliter (5.0 to 5.2 mmol per liter); category 6, 95 to 99 mg
per deciliter (5.3 to 5.5 mmol per liter); and category 7, 100 mg per deciliter (5.6 mmol per liter) or more; 1-hour plasma glucose level —
category 1, 105 mg per deciliter (5.8 mmol per liter) or less; category 2, 106 to 132 mg per deciliter (5.9 to 7.3 mmol per liter); category 3,
133 to 155 mg per deciliter (7.4 to 8.6 mmol per liter); category 4, 156 to 171 mg per deciliter (8.7 to 9.5 mmol per liter); category 5, 172 to
193 mg per deciliter (9.6 to 10.7 mmol per liter); category 6, 194 to 211 mg per deciliter (10.8 to 11.7 mmol per liter); and category 7, 212
mg per deciliter (11.8 mmol per liter) or more; and 2-hr plasma glucose level — category 1, 90 mg per deciliter (5.0 mmol per liter) or less;
category 2, 91 to 108 mg per deciliter (5.1 to 6.0 mmol per liter); category 3, 109 to 125 mg per deciliter (6.1 to 6.9 mmol per liter); category
4, 126 to 139 mg per deciliter (7.0 to 7.7 mmol per liter); category 5, 140 to 157 mg per deciliter (7.8 to 8.7 mmol per liter); category 6, 158
to 177 mg per deciliter (8.8 to 9.8 mmol per liter); and category 7, 178 mg per deciliter (9.9 mmol per liter) or more.
‡ Data for women who had had a previous cesarean section were excluded.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
T h e
ne w e ng l an d j our n a l
o f
me di ci ne
n engl j med 358;19 www.nejm.org may 8, 2008
2000
adverse outcomes generally remained significant
after adjustment for multiple potential confound-
ers — 10 of 12 associations for primary outcomes
and 12 of 15 for secondary outcomes — and were
generally consistent across centers, except the as-
sociation for the fasting plasma glucose level and
primary cesarean delivery. Furthermore, findings
with respect to cord-blood serum C-peptide levels,
and hence fetal insulin levels, support Pedersen’s
proposed mechanism to explain the propensity
for excessive growth in fetuses of mothers with
hyperglycemia. When associations between ma-
ternal glucose level and birth weight were esti-
mated with the use of birth weight as a continu-
ous variable, the difference in mean birth weight
between the lowest and highest glucose categories
was in the range of 240 to 300 g, even after full
adjustment for potential confounders.
Two primary outcomes — birth weight above
the 90th percentile and cord-blood serum C-pep-
tide level above the 90th percentile — though
strongly associated with maternal glycemia, could
be viewed as physiological consequences of ma-
ternal glycemia rather than as true disorders or
problems. However, the other two primary out-
comes (primary cesarean delivery and clinical
neonatal hypoglycemia) and the five secondary
outcomes reported (premature delivery, shoulder
dystocia or birth injury, intensive neonatal care,
hyperbilirubinemia, and preeclampsia) also showed
continuous linear associations with the 1-hour
plasma glucose level (seven analyses), the 2-hour
plasma glucose level (six analyses), and the fast-
ing plasma glucose level (three analyses). These are
well-recognized complications of pregnancies in
mothers with preexisting or gestational diabetes,
as currently defined.
Questions have been raised regarding the
benefits of treating “mild” gestational diabetes
mellitus.
13,23,24
However, one recently published
randomized clinical trial, the Australian Carbo-
hydrate Intolerance Study in Pregnant Women
(ACHOIS), found reduced perinatal morbidity and
mortality when standard contemporary treatment
of gestational diabetes mellitus was compared
with no intervention
25
; another study on treatment
of “mild” gestational diabetes mellitus is ongo-
ing.
26
Taken together, the current results and re-
sults of the ACHOIS trial
25
indicate that mater-
nal hyperglycemia less severe than that used to
define overt diabetes is related to clinically impor-
tant perinatal disorders or problems and that their
Table 3. Adjusted Odds Ratios for Associations between Maternal Glycemia as a Continuous Variable and Primary
and Secondary Perinatal Outcomes.*
Outcome Plasma Glucose Level
Fasting At 1 Hr At 2 Hr
odds ratio (95% CI)
Primary outcome
Birth weight >90th percentile 1.38 (1.32–1.44) 1.46 (1.39–1.53) 1.38 (1.32–1.44)
Primary cesarean section† 1.11 (1.06–1.15) 1.10 (1.06–1.15) 1.08 (1.03–1.12)
Clinical neonatal hypoglycemia 1.08 (0.98–1.19)‡ 1.13 (1.03–1.26) 1.10 (1.00–1.12)
Cord-blood serum C peptide >90th percentile 1.55 (1.47–1.64) 1.46 (1.38–1.54) 1.37 (1.30–1.44)
Secondary outcome
Premature delivery (before 37 wk) 1.05 (0.99–1.11) 1.18 (1.12–1.25) 1.16 (1.10–1.23)
Shoulder dystocia or birth injury 1.18 (1.04–1.33) 1.23 (1.09–1.38) 1.22 (1.09–1.37)
Intensive neonatal care 0.99 (0.94–1.05) 1.07 (1.02–1.13) 1.09 (1.03–1.14)
Hyperbilirubinemia 1.00 (0.95–1.05) 1.11 (1.05–1.17) 1.08 (1.02–1.13)
Preeclampsia 1.21 (1.13–1.29) 1.28 (1.20–1.37) 1.28 (1.20–1.37)
* Odds ratios were for an increase in the glucose level of 1 SD (6.9 mg per deciliter [0.4 mmol per liter] for the fasting plas-
ma glucose level, 30.9 mg per deciliter [1.7 mmol per liter] for the 1-hr plasma glucose level, and 23.5 mg per deciliter
[1.3 mmol per liter] for the 2-hr plasma glucose level). The model for preeclampsia did not include adjustment for hos-
pitalization or mean arterial pressure, and presence or absence of family history of hypertension or prenatal urinary tract
infection was included in the model for preeclampsia only. See Table 2
for other details about adjustments in each model.
† Data for women who had had a previous cesarean section were excluded.
‡ The P value for the quadratic (nonlinear) association was 0.013.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
n engl j med 358;19 www.nejm.org may 8, 2008
2001
effects can be reduced by means of treatment, al-
though a threshold for the need for treatment is
not established.
The individual measures from the oral glucose-
tolerance tests were not highly correlated, and no
single measure was clearly superior in predicting
the primary outcomes. When adjusted for poten-
tial confounders, relative increases in each glucose
measure were similarly predictive of birth weight
above the 90th percentile. When the glucose mea-
sures were analyzed as continuous variables, each
was a signif icant predictor of primary cesarean
delivery, with 1-SD increases in glucose level be-
ing associated with an increase of 8 to 11% in the
odds of delivery by cesarean section. Clinical neo-
natal hypoglycemia was infrequent (overall inci-
dence, 2.1%), and when adjusted for confounders,
only the 1-hour plasma glucose level remained a
signif icant predictor of this outcome. All three
measures of plasma glucose were highly predictive
of cord-blood serum C-peptide values, with the
fasting plasma glucose level being the strongest
predictor.
Our study had some limitations. The nutritional
status and gestational weight gain of the partici-
pants could affect fetal growth and other perinatal
outcomes; we do not have data on these variables.
Some confounders, such as previous gestational
diabetes mellitus, maternal BMI, or previous mac-
rosomia, may have inf luenced clinical decisions
such as the choice of route of delivery. Because of
the observational design of our study, we cannot
conclude that maternal glycemia is causally related
to the adverse outcomes observed; however, such a
relationship is plausible. Although the rate of par-
ticipation in this blinded study was 54%, we be-
lieve this is unlikely to materially affect our es-
timates of associations. Differences in age and
education level were small between those who
agreed to participate and those who did not.
The broad inclusion criteria, the large number
and the geographic distribution of centers in-
volved, and the similarity across centers in the as-
sociations we found between maternal glycemia
and outcomes provide support that our results can
be generalized to develop outcome-based criteria
for classifying glucose metabolism in pregnancy
that can be applied worldwide. Lack of clear
thresholds for risk and the fact that the four pri-
mary outcomes are not necessarily of equal clini-
cal importance make direct translation of our re-
sults into clinical practice challenging. However,
our findings of significant associations between
adverse outcomes and higher levels of maternal
glucose within what is currently considered a non-
diabetic range indicate the need to reconsider cur-
rent criteria for diagnosing and treating hyper-
glycemia during pregnancy.
Supported by grants from the Eunice Kennedy Shriver Na-
tional Institute of Child Health and Human Development and
the National Institute of Diabetes and Digestive and Kidney
Diseases (R01-HD34242 and R01-HD34243); the National Center
for Research Resources (M01-RR00048 and M01-RR00080); and
the American Diabetes Association; and g rants to local field
centers from Diabetes UK (R D04/0002756), Kaiser Permanente
Medical Center, KK Women’s and Children’s Hospital, Mater
Mother’s Hospital, Novo Nordisk, the Myre Sim Fund of the
Royal College of Physicians of Edinburgh, and the Howard and
Carol Bernick Family Foundation.
Dr. Metzger reports receiving an educational grant from Novo
Nordisk; Dr. Hadden, an honorarium from Novo Nordisk; Dr.
McCance, an honorarium from Takeda; and Dr. Persson, hono-
raria from Novo Nordisk. No other potential conflict of interest
relevant to this article was reported.
APPENDIX
The members of the HAPO Study Cooperative Research Group were as follows: North American Field Centers — Kaiser Foundation Hospi-
tal, Bellflower, CA: M. Contreras, D.A. Sacks, W. Watson (deceased); Prentice Women’s Hospital of Northwestern Memorial Hospital–Northwestern
University Feinberg School of Medicine, Chicago: S.L. Dooley, M. Foderaro, C. Niznik; MetroHealth Medical Center–Case Western Reserve University,
Cleveland: J. Bjaloncik, P.M. Catalano, L. Dierker, S. Fox, L. Gullion, C. Johnson, C.A. Lindsay, H. Makovos, F. Saker; Women and Infants’
Hospital of Rhode Island–Brown University Medical School, Providence: M.W. Carpenter, J. Hunt, M.H. Somers; Sunnybrook and Women’s College
Health Sciences Centre–University of Toronto, Toronto: K.S. Amankwah, P.C. Chan, B. Gherson, E. Herer, B. Kapur, A. Kenshole, G. Lawrence,
K. Matheson, L. Mayes, K. McLean, H. Owen; European–Caribbean Field Centers — Queen Elizabeth Hospital–School of Clinical Medicine and
Research, University of the West Indies, Barbados: C. Cave, G. Fenty, E. Gibson, A. Hennis, G. McIntyre, Y.E. Rotchell, C. Spooner, H.A.R.
Thomas; Royal Jubilee Maternity Hospital, Belfast, Northern Ireland: J. Gluck, D.R. Hadden, H. Halliday, J. Irwin, O. Kearney, J. McAnee, D.R.
McCance, M. Mousavi, A.I. Traub; St. Mary’s Hospital–Manchester University, Manchester, United Kingdom: J.K. Cruickshank, N. Derbyshire, J.
Dry, A.C. Holt, F. Khan, C. Lambert, M. Maresh, F. Prichard, C. Townson; University Hospital–University Medical Center Utrecht, Utrecht, the
Netherlands: T.W. van Haeften, A.M.R. van de Hengel, G.H.A. Visser, A. Zwart; Middle Eastern–Asian Field Centers — Rajavithi Hospital,
Bangkok, Thailand: U. Chaovarindr, U. Chotigeat, C. Deerochanawong, I. Panyasiri, P. Sanguanpong; Soroka Medical Center–Ben-Gurion
University, Beersheba, Israel: D. Amichay, A. Golan, K. Marks, M. Mazor, J. Ronen, A. Wiznitzer; Helen Schneider Hospital for Women, Rabin
Medical Center–Sackler Faculty of Medicine, Tel-Aviv University, Petah-Tiqva, Israel: R. Chen, D. Harel, N. Hoter, N. Melamed, J. Pardo, M. Wit-
shner, Y. Yogev; Australasian Field Centers — Mater Misericordiae Mothers’ Hospital–University of Queensland, Brisbane, Australia: F. Bowling, D.
Cowley, P. Devenish-Meares, H.G. Liley, A. McArdle, H.D. McIntyre, B. Morrison, A. Peacock, A. Tremellen, D. Tudehope; Prince of Wales
Hospital–Chinese University of Hong Kong, Hong Kong: K.Y. Chan, N.Y. Chan, L.W. Ip, S.L. Kong, Y.L. Lee, C.Y. Li, K.F. Ng, P.C. Ng, M.S.
Rogers, K.W. Wong; John Hunter Hospital, Newcastle, Australia: M. Edgar, W. Giles, A. Gill, R. Glover, J. Lowe, F. Mackenzie, K. Siech, J.
Verma, A. Wright; KK Women’s and Children’s Hospital, Singapore City, Singapore: Y.H. Cao, J.J. Chee, A. Koh, E. Tan, V.J. Rajadurai, H.Y. Wee,
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
n engl j med 358;19 www.nejm.org may 8, 2008
2002
Hy pergl ycem ia a nd A dver se Pre gnanc y O ut come s
G.S.H. Yeo; Regional Centers — Providence, RI: D. Coustan, B. Haydon; Belfast, Northern Ireland: A. Alexander, D.R. Hadden; Petah-Tiqva,
Israel: O. Attias-Raved, M. Hod; Brisbane, Australia: J.J.N. Oats, A.F. Parry; Clinical Coordinating Center — Northwestern University Feinberg
School of Medicine, Chicago: A. Collard, A.S. Frank, L.P. Lowe, B.E. Metzger, A. Thomas; Data Coordinating Center — Northwestern Univer-
sity Feinberg School of Medicine, Chicago: T. Case, P. Cholod, A.R. Dyer, L. Engelman, M. Xiao, L. Yang; Central Laboratory — Queen’s Univer-
sity Belfast, Belfast, Northern Ireland: C.I. Burgess, T.R.J. Lappin, G.S. Nesbitt, B. Sheridan, M. Smye, E.R. Trimble; Steering Committee
— Providence, RI: D. Coustan; Chicago: A.R. Dyer; Belfast, Northern Ireland: D.R. Hadden; Petah-Tiqva, Israel: M. Hod; Chicago: B.E. Metzger,
L.P. Lowe (ex officio); Brisbane, Australia: J.J.N. Oats; Stockholm: B. Persson; Belfast, Northern Ireland: E.R. Trimble; Data and Safety Monitor-
ing Committee — G.R. Cutter, S.G. Gabbe, J.W. Hare, L.E. Wagenknecht; Consultants — Y. Chen, J. Claman, J. King.
Re fer enc es
journal
editorial
fell ow
The Journal’s editorial office invites applications for a one-year
research fellowship beginning in July 2009 from individuals at any
stage of training. The editorial fellow will work on Journal projects
and will participate in the day-to-day editorial activities of the Journal
but is expected in addition to have his or her own independent
projects. Please send curriculum vitae and research interests
to the Editor-in-Chief, 10 Shattuck St., Boston, MA 02115
(fax, 617-739-9864), by September 30, 2008.
American Diabetes Association. Clini-
cal practice recommendat ions 2001: ges-
tationa l diabetes mellitus. Diabetes Care
2001;24:Suppl 1:S77-S79.
Metzger BE, Coustan DR. Summary
and recommendations of the Fourt h In-
ternational Workshop-Conference on Ges-
tationa l Diabetes Mellitus. Diabetes Care
1998;21:Suppl 2:B161-B167.
O’Sullivan JB, Mahan C. Criteria for
oral glucose tolerance test in pregnancy.
Diabetes 1964;13:278-85.
Metzger BE, Buchanan TA, Coustan
DR, et al. Summary and recommenda-
tions of the Fifth International Work-
shop-Conference on Gestational Diabetes
Mellitus. Diabetes Care 2007;30:Suppl 2:
S251-S260. [Erratum, Diabetes Care 2007;
30:3154.]
WHO Expert Committee on Diabetes
Mellitus: second report. World Health Or-
gan Tech Rep Ser 1980;646:1-80.
Jensen DM, Damm P, Sorensen B, et
al. Clinical impact of mild carbohydrate
intolerance in pregnancy: a study of 2904
nondiabetic Danish women with risk fac-
tors for gestational diabetes. Am J Obstet
Gynecol 2001;185:413-9.
Yang X, Hsu-Hage B, Zhang H, Zhang
C, Zhang Y, Zhang C. Women with im-
paired glucose tolerance during pregnan-
cy have significantly poor pregnancy out-
comes. Diabetes Care 2002;25:1619-24.
Vambergue A, Nuttens MC, Verier-
Mine O, Dognin C, Cappoen JP, Fontaine
P. Is mild gestational hyperglycemia as-
sociated with maternal and neonatal com-
plications? The Diagest Study. Diabet Med
2000;17:203-8.
Langer O, Brustman L, Anyaegbunam
A, Mazze R. The significance of one ab-
normal glucose tolerance test value on
adverse outcome in pregnancy. Am J Ob-
stet Gynecol 1987;157:758-63.
1.
2.
3.
4.
5.
6.
7.
8.
9.
Sacks DA, Greenspoon JS, Abu-Fadil
S, Henr y HM, Wolde-Tsadik G, Yao JFF.
Toward universal criteria for gestational
diabetes: the 75-gram glucose tolerance
test in pregnancy. Am J Obstet Gynecol
1995;172:607-14.
Sermer M, Naylor CD, Gare DJ, et al.
Impact of increasing carbohydrate intoler-
ance on maternal-fetal outcomes in 3637
women without gestational diabetes. Am J
Obstet Gynecol 1995;173:146-56.
Jarrett R J. Reflections on gestational
diabetes mellitus. Lancet 1981;2:1220-1.
Hunter DJS, Keirse MJNC. Gestational
diabetes. In: Chalmers I, Enkin M, Kierse
M, eds. Effective care in pregnancy and
childbirt h. Oxford, England: Oxford Uni-
versity Press, 1989:403-10.
Spellacy WN, Miller S, Winegar A, Pe-
terson PQ. Macrosomia: maternal charac-
teristics and infant complications. Obstet
Gynecol 1985;66:158-61.
Coustan DR. Management of gesta-
tional diabetes: a self-fulfilling prophecy?
JAMA 1996;275:1199-200.
HAPO Study Cooperative Research
Group. The Hyperglycemia and Adverse
Pregnancy Outcome (HAPO) Study. Int J
Gynecol Obstet 2002;78:69-77.
HAPO Study Cooperative Research
Group, Nesbitt GS, Smye M, Sheridan B,
Lappin TR, Trimble ER. Integration of lo-
cal and cent ral laborator y functions i n a
worldwide multicent re study: experience
from the Hyperglycemia and Adverse
Pregnancy Outcome (HAPO) Study. Clin
Trials 2006;3:397-407.
Chan A, King JF, Flenady V, Haslam R,
Tudehope D. Classif ication of perinatal
deaths: development of the Australian
and New Zealand classifications. J Paedi-
atr Child Health 2004;40:340-7.
International statistical classification
of diseases and related health problems,
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
10th revision: ICD-10. Geneva: World
Health Organi zation, 1992.
Brown MA, Lindheimer MD, de Swiet
M, Van Assche A, Moutquin JM. The clas-
sification and diagnosis of the hyperten-
sive disorders of pregnancy: statement
from the International Society for the
Study of Hy pertension i n P regnancy
(ISSHP). Hy pertens Preg nancy 2001;20:
ix-xiv.
Alkalay AL, Sarnat HB, Flores-Sarnat
L, Elashoff JD, Farber SJ, Simmons CF.
Population meta-analysis of low plasma
glucose thresholds in full-term normal
newborns. Am J Perinatol 2006;23:115-
9.
Pedersen J. Diabetes and pregnancy:
blood sugar of newborn infants. (Ph.D.
thesis. Copenhagen: Danish Science Press,
1952:230.)
Brody SC, Harris RH, Whitener BL, et
al. Screening for gestational diabetes: sys-
tematic evidence review. Rock ville, MD:
Agency for Healthcare Research and Qual-
ity, 2003.
Tuffnell DJ, West J, Walkinshaw SA.
Treatments for gestational diabetes and
impaired glucose tolerance in pregnancy.
Cochra ne Database Syst Rev 20 03;3:
CD003395.
Crowther CA, Hiller JE, Moss JR, et al.
Effect of treatment of gestational diabetes
on pregnancy outcomes. N Engl J Med
2005;352:2477-86.
Landon MB, Thom E, Spong CY, et al.
The Nationa l Instit ute of Child Health
and Human Development Maternal-Fetal
Medicine Unit Network randomized clini-
cal trial in progress: standard therapy
versus no therapy for mild gestat ional
diabetes. Diabetes Care 2007;30:Suppl 2:
S194-S199.
Copyright © 2008 Massachusetts Medical Society.
20.
21.
22.
23.
24.
25.
26.
The New England Journal of Medicine
Downloaded from nejm.org at QUEENS UNIV MED LIB on March 27, 2014. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.