Variations in Mortality and Morbidity by Gestational Age among Infants

Born at Term
XUN ZHANG, PHD,

AND

MICHAEL S. KRAMER, MD

Objective To examine the risks of infant death and neonatal morbidity by week of gestation at term.
Study design National U.S. birth cohort study on the basis of singleton live births in 1995-2001 at 37 to 41 completed weeks
gestational age (GA), with exclusion of congenital anomalies. Main outcomes included neonatal, postneonatal, and causespecific infant death; low-Apgar score at 5 minutes; receipt of neonatal mechanical ventilation >30 minutes; neonatal seizures;
birth injury; and meconium aspiration syndrome. To reduce confounding by indication, we carried out a secondary analysis
restricted to low-risk deliveries.
Results In non-Hispanic white women, the mortality rate decreased with increasing GA from 37 to 39 weeks, remained
stable from 39 to 40 weeks, and then (for neonatal death) increased at 41 weeks. Rates of low 5-minute Apgar score and
mechanical ventilation showed a U-shaped relation across term GAs, and rates of meconium aspiration syndrome and birth
injury rose with increasing GA. Results were similar among infants born to low-risk mothers and in non-Hispanic black women.
Conclusions Term infants show considerable heterogeneity across gestational age in neonatal and late infant outcomes, even
when born to mothers at low risk. Recent trends toward earlier labor induction may have adverse health impacts.
(J Pediatr 2009;154:358-62)

he World Health Organization defines preterm birth as birth at gestational age (GA) 37 completed weeks, a term
birth as birth at GA 37 to 41 completed weeks, and postterm birth as birth at GA 42 completed weeks.1 Most
attention has focused on the contribution of GA to adverse birth outcomes in the preterm and postterm periods,
including fetal and infant death, neonatal morbidity, and maternal complications.2-6 Preterm births account for most infant
deaths and neonatal morbidity,2,3 and most stillbirths also occur in the preterm period. Postterm births are associated with
increased risks of fetal and infant death and of neonatal morbidity.7,8 Postterm births occur in 8% or less of all pregnancies in
which GA is estimated by early ultrasound scanning, and the frequency of their occurrence has fallen with increased routine labor
induction after 40 or 41 weeks.9,10
The risks associated with preterm and postterm gestations have been compared with those in term births, which are usually
considered as a homogeneous group. Few investigators have studied variations in outcome by week of GA among term births.
This is an increasingly important area for study, given recent trends toward more frequent and earlier labor induction.11,12
Caughey et al13,14 reported that rates of meconium aspiration syndrome and macrosomia
increased beyond 38 completed weeks of gestation and that the risk of maternal peripartum complications increased beyond 40 weeks. Smith et al15 found that the risk of sudden
infant death syndrome (SIDS) declined with advancing weeks of gestation at term. To our
From the Departments of Pediatrics (X.Z.,
M.K.) and Epidemiology and Biostatistics
knowledge, however, variations in risks of other causes of infant death and of neonatal
(M.K.), McGill University Faculty of Medimorbidity have been less extensively studied among infants born at term. We studied a
cine, Montreal, Quebec, Canada.
large sample of U.S. births to explore heterogeneity in infant mortality and neonatal
Supported by a grant from the Canadian
Institutes of Health Research. The authors
morbidity among infants born at 37 to 41 completed weeks of gestation.

METHODS
We used United States birth cohorts for the years from 1995 to 2001 as our data
source. These linked live birthinfant death files, compiled by the U.S. National Center
for Health Statistics (NCHS), include information from death certificates linked to

GA
LMP

358

Gestational age
Last menstrual period

SIDS
OR

Sudden infant death syndrome

Odds ratio

declare no conflicts of interest.

Submitted for publication Apr 3, 2008; last
revision received Jul 30, 2008; accepted
Sep 4, 2008.
Reprint requests: Michael S. Kramer, MD, The
Montreal Childrens Hospital, 2300 Tupper
Street (Les Tourelles), Montreal, Quebec
H3H 1P3 Canada. E-mail: michael.kramer@
mcgill.ca.
0022-3476/$ - see front matter
Copyright 2009 Mosby Inc. All rights
reserved.
10.1016/j.jpeds.2008.09.013

information from birth certificates for each infant born in the

United States who dies before his or her first birthday. The
birth data files also include information on several serious
newborn conditions.16
In the United States, GA is calculated from the first day
of the mothers last menstrual period (LMP) in completed
weeks (eg, 40 completed weeks of GA consists of 280 to 286
days, ie, 40 and 0/7 to 40 and 6/7 weeks). If the LMP date is
not reported, or if the length of gestation is not consistent
with birth weight (normal-weight births of apparently short
gestations or very low birth weight births reported to be
full-term), the clinical estimate of gestation is substituted. GA
is based on the clinical estimate of gestation for only about 5%
of births, most (about 97%) because of missing LMP.16 The
clinical estimate of gestation is recorded on U.S. birth certificates with no instructions (before the 2003 revision) to specify the basis of the estimate. California did not report the
clinical estimate. Recent evidence suggests that the clinical
estimate provides GA-specific rates and relative risks of adverse pregnancy outcomes that are more consistent with those
reported in other countries.17
We examined infant death (separating the neonatal
[0-27 days] and postneonatal [28-364 days] periods), major
causes of the death, and neonatal morbidity among term
births by week of gestation. We restricted our main study
sample to singleton non-Hispanic white live births to avoid
confounding by race/ethnicity. We also excluded all infants
with congenital anomalies diagnosed at birth or who died
from a congenital anomaly. Our study sample consisted of
12 762 098 births at 37 to 41 weeks of gestation. To assess the
generalizability of our results, however, we also repeated our
analysis in non-Hispanic blacks (n 2 902 922). Causes of
death were coded according to ICD-9 (1995-1998) and
ICD-10 (1999-2001) codes and classified according to the
recommendation of the International Collaborative Effort
(ICE) on Perinatal and Infant Mortality.18 Multiple causes
are converted to a single underlying cause of death by ACME
(Automated Classification of Medical Entities), a computer
software program developed by NCHS that uses World
Health Organization rules to select the underlying cause. The
major causes of infant death included asphyxia, infection, and
postneonatal SIDS. The ICD-9 and ICD-10 coding for the
causes of underlying death are given in the Appendix (available at www.jpeds.com). Neonatal morbidity outcomes included low (4) Apgar score at 5 minutes, receipt of mechanical ventilation 30 minutes, seizures, birth injury, and
meconium aspiration syndrome. Birth injury was defined as
any impairment of the infants body function or structure due
to adverse influences which occurred at birth.16 Nebraska and
Texas did not report birth injury and New York City did not
report assisted ventilation.16
Maternal demographic characteristics in the linked files
include age, marital status, and education. Medical risk factors
in the data include diabetes (including juvenile-onset, adultonset, and gestational diabetes), gestational hypertension,
chronic (pre-pregnancy) hypertension, eclampsia, and smok-

ing during pregnancy.16 Maternal smoking was dichotomized

for our study as any or none. Parity was defined as the number
of live births including the index pregnancy and dichotomized
as primiparous versus multiparous. Obstetric procedures include induction and cesarean section. Method of delivery was
classified as noninstrumental vaginal, instrumental (vacuum
or forceps) vaginal, or cesarean.
Rates of infant mortality (neonatal and postneonatal),
cause-specific death, and neonatal morbidity were calculated and
compared by each GA (in completed weeks) at term. With the
gestational age of 40 completed weeks used as reference, multiple
logistic regression was used to estimate adjusted odds ratios and
their 95% confidence intervals, controlling for maternal demographic and clinical variables, including birth weight for GA
z-scores (1, 1 to 1, 1), based on an internal reference. Etiologic fractions (EFs) were calculated as EF (Pi[RRi
1])/(1 [RRi 1]) on the basis of the method of Miettinen,19 where Pi is the prevalence of the ith GA stratum, RRi is
the corresponding relative risk (based on the multiple logistic
regression analyses using 40 weeks as the reference), and
indicates the summation over the GA categories.
We also carried out an analysis among low-risk women
to reduce confounding by the reason for earlier birth, by
restricting the analyses to a subgroup with low-risk deliveries.
Low-risk delivery was defined as a spontaneous (noninduced)
vaginal delivery by a woman without any of the following
medical risk factors: anemia, heart disease, acute or chronic
lung disease, diabetes, genital herpes, hydramnios/oligohydramnios, hemoglobinopathy, chronic or pregnancy-associated hypertension (including preeclampsia), eclampsia, incompetent cervix, previous preterm or small infant, previous
infant 4000 g, kidney disease, Rh sensitization, or uterine
bleeding. Our analytic approach on the basis of low-risk
delivery has been used in previous studies, both in the United
States14 and Canada,20,21 although definitions have varied
slightly depending on available information.
Finally, to guard against potential misclassification of GA,
we also carried out a sensitivity analysis using the clinical estimate
of GA instead of the LMP estimate. All data were analyzed with
SAS version 9.1 (SAS Institute, Cary, North Carolina). Because
our analysis was based on public-use data files provided by
NCHS, no ethical approval was sought for this study.

RESULTS
Our main study sample included 1 116 817 (8.8%) nonHispanic white infants born at 37 weeks, 2 418 592 (19.0%)
at 38 weeks, 3 766 999 (29.5%) at 39 weeks, 3 546 328
(27.8%) at 40 weeks, and 1 913 362 (15.0%) at 41 weeks.
Maternal sociodemographic variables and clinical characteristics are summarized in Table I. Infants born at early term
GAs (37 to 38 weeks) were more likely to be boys and to be
delivered by cesarean but less likely to be induced than infants
born at later term GAs. Mothers of early term births were
more likely than those of later term births to be older (35
years), multiparous, and smokers and to have diabetes, hypertension, and eclampsia.

Variations in Mortality and Morbidity by Gestational Age among Infants Born at Term

359

Table I. Sociodemographic and clinical characteristics (%) by completed weeks of gestational age

N
Infant sex (male)
Primiparous
Mothers age
20
20-34
35
Mothers education
12 years
High school graduate
College graduate
Marital status (legally married)
Maternal smoking
Maternal diabetes
Chronic hypertension
Gestational hypertension
Eclampsia
Cesarean delivery
Induction

37 weeks

38 weeks

39 weeks

40 weeks

41 weeks

Total

1 116 817
53.6
40.0

2 418 592
52.6
36.4

3 766 999
51.1
38.3

3 546 328
49.9
44.4

1 913 362
49.4
50.0

12 762 098
51.0
41.5

9.2
75.4
15.4

7.8
76.1
16.1

7.9
76.8
15.3

9.1
77.3
13.6

10.6
77.2
12.2

8.7
76.8
14.5

13.4
58.0
28.6
78.1
17.7
3.9
1.2
6.7
0.6
22.2
19.7

11.6
57.2
31.2
80.6
15.8
3.4
0.9
4.7
0.3
23.5
21.4

11.2
56.3
32.5
80.8
14.9
2.7
0.6
3.3
0.2
20.8
21.1

11.8
56.0
32.2
79.1
14.8
1.9
0.5
2.7
0.2
16.4
22.6

13.1
56.6
30.3
76.4
15.8
1.5
0.4
2.4
0.1
18.8
29.1

11.9
56.6
31.5
79.4
15.4
2.5
0.6
3.6
0.2
19.9
22.6

GA-specific infant mortality rates were highest at 37

completed weeks: 0.66 per 1000 in the neonatal period and
1.68 per 1000 in the postneonatal period (Table II). The rates
for neonatal death decreased with increasing GA from 37 to
39 weeks, remained stable from 39 to 40 weeks, and then
increased at 41 weeks, and the rates for postneonatal death
decreased from 37 to 40 weeks and then remained stable at 40
to 41 weeks. Table II also shows the adjusted odds ratios (very
similar to the crude odds ratios, which are not shown) for
infant death at each GA, with 40 weeks used as the reference.
The ORs for early neonatal (0-6 days) and late neonatal (7-27
days) mortality rate were very similar at all term GAs (data
not shown).
The major causes of infant death were asphyxia-related
(19.5% of neonatal deaths), infection (6.8% of neonatal
deaths, 5.8% of infant deaths), and SIDS (50.2% of postneonatal deaths) (Table III). The rates for infection and SIDS
decreased with increasing GA, with the highest rates (0.14
per 1000 for infection, 0.83 per 1000 for SIDS) at 37 weeks,
while the rates of neonatal death because of asphyxia showed
a slight U-shaped relation with GA: highest at 37 and 41
weeks (adjusted ORs 1.2 [95% CI, 0.9-1.5] and 1.2 [95% CI,
1.0-1.5]), respectively.
Patterns of GA-specific neonatal morbidity rates are
shown in Table IV. Neonatal seizure rates were stable across
GA. Rates of low 5-munite Apgar score and receipt of mechanical ventilation 30 minutes showed a U-shaped relation
with GA. Rates of birth injury and meconium aspiration
syndrome, however, increased with increasing GA.
Etiologic fractions (population attributable risks) for
individual completed weeks of GA were modest. For example, for infant death, the highest etiologic fractions were
observed at 37 weeks: 6.4% for neonatal death and 3.9% for
360

Zhang and Kramer

postneonatal death; for death caused by infection; they were

5.5% at 37 weeks and 6.8% at 38 weeks.
Our low-risk subgroup comprised 5 768 536 infants
(45.2% of the study sample). Among these low-risk deliveries,
absolute rates were similar to (for most outcomes) or lower
than (for some neonatal morbidities) those observed in the
total study sample shown in Tables II, III, and IV. The
pattern of odds ratios across GA was almost identical to that
observed in the total sample (data not shown).
Although absolute rates of mortality and morbidity
outcomes were higher among non-Hispanic blacks, contrasts
across GA for infant mortality, cause-specific infant mortality, and neonatal morbidity were similar to those observed
among non-Hispanic whites. For example, the rates of neonatal death decreased from 1.05 per 1000 at 37 weeks to 0.67
per 1000 at 39 weeks and then remained constant until 41
weeks. The corresponding adjusted ORs for 37 and 38 weeks
(vs 40 weeks) were 1.4 (95% CI, 1.2-1.7) and 1.1 (95% CI,
0.9-1.2), respectively.
In our sensitivity analysis on the basis of the clinical
estimate of gestational age, similar results were observed for
infant death, cause-specific death, and serious neonatal morbidities. Data and model information are available on request.

DISCUSSION
Even at late preterm (34 to 36 completed weeks) GA,
relative risks of death in the early neonatal, late neonatal, and
postneonatal periods are substantially higher than at term.2
Our results indicate that these increased risks persist at 37 and
38 weeks. Despite a low absolute risk of infant death at these
GAs, the risks were more than 50% higher at 37 weeks than
at 40 weeks. Older women (35 years), smokers, and nulliparas with medical risk factors are known to have elevated risks
The Journal of Pediatrics March 2009

Table II. Adjusted odds ratios for infant mortality by gestational age in completed weeks
Neonatal death
Gestational age
37
38
39
40
41

Postneonatal death*

Rate per 1000

Adjusted OR (95% CI)

Rate per 1000

Adjusted OR (95% CI)

0.66
0.42
0.33
0.34
0.40

1.8 (1.6-2.0)
1.2 (1.1-1.3)
0.9 (0.9-1.0)
reference
1.1 (1.0-1.3)

1.68
1.29
1.10
1.03
1.04

1.5 (1.4-1.6)
1.2 (1.1-1.3)
1.1 (1.0-1.1)
Reference
1.0 (0.9-1.0)

weeks
weeks
weeks
weeks
weeks

ORs were estimated from multiple logistic regression models adjusted for sex, parity, maternal age, education, marital status, smoking, diabetes, chronic hypertension, gestational
hypertension (including preeclampsia), eclampsia, mode of delivery, induction, and birth weight for gestational age z-score.
*On the basis of infants who survived the neonatal period.

Table III. Adjusted odds ratios for major causes of death by gestational age in completed weeks
Death caused by asphyxia
Gestational age
37
38
39
40
41

weeks
weeks
weeks
weeks
weeks

Death caused by infection

Death caused by postneonatal

SIDS*

Rate per 1000

Adjusted OR
(95% CI)

Rate per 1000

Adjusted OR
(95% CI)

Rate per 1000

Adjusted OR
(95% CI)

0.10
0.08
0.07
0.08
0.10

1.2 (0.9-1.5)
0.8 (0.6-1.0)
0.7 (0.6-0.9)
Reference
1.2 (1.0-1.5)

0.14
0.11
0.09
0.08
0.06

1.7 (1.4-2.1)
1.4 (1.2-1.7)
1.1 (0.9-1.3)
Reference
0.7 (0.6-0.9)

0.83
0.66
0.56
0.51
0.50

1.5 (1.4-1.6)
1.2 (1.1-1.3)
1.1 (1.0-1.2)
Reference
1.0 (0.9-1.1)

ORs were estimated from multiple logistic regression models adjusted for sex, parity, maternal age, education, marital status, smoking, diabetes, chronic hypertension, gestational
hypertension (including preeclampsia), eclampsia, mode of delivery, induction, and birth weight for gestational age z-score.
*On the basis of infants who survived the neonatal period.

Table IV. Adjusted odds ratios for neonatal morbidity by gestational age in completed weeks
Gestational age
37 weeks
Rate/1000
OR (95% CI)
38 weeks
Rate/1000
OR (95% CI)
39 weeks
Rate/1000
OR (95% CI)
40 weeks
Rate/1000
OR (95% CI)
41 weeks
Rate/1000
OR (95% CI)

5-min Apgar score <4

Neonatal seizures

Ventilation >30 min

Birth injury

Meconium aspiration
syndrome

1.26
1.2 (1.1-1.3)

0.61
1.1 (1.0-1.2)

6.45
2.0 (1.9-2.1)

3.17
0.9 (0.9-1.0)

1.01
0.5 (0.4-0.5)

0.95
0.9 (0.9-1.1)

0.53
0.9 (0.9-1.0)

3.56
1.1 (1.1-1.2)

2.94
0.9 (0.9-0.9)

1.07
0.5 (0.4-0.5)

0.80
0.8 (0.8-0.9)

0.52
0.9 (0.9-1.0)

2.74
0.9 (0.9-1.0)

3.05
0.9 (0.9-1.0)

1.49
0.7 (0.6-0.7)

0.95
Reference

0.54
Reference

2.86
Reference

3.43
Reference

2.24
Reference

1.15
1.2 (1.1-1.3)

0.64
1.1 (1.0-1.2)

3.36
1.1 (1.1-1.2)

3.70
1.0 (1.0-1.1)

3.13
1.4 (1.3-1.5)

ORs were estimated from multiple logistic regression models adjusted for sex, parity, maternal age, education, marital status, smoking, diabetes, chronic hypertension, gestational
hypertension (including preeclampsia), eclampsia, mode of delivery, induction, and birth weight for gestational age z-score.

of infant death and neonatal morbidity. The adjusted odds

ratios shown in Tables II, III, and IV were obtained from
multivariate logistic regression models that included those
and other potential confounding variables and should there-

fore reflect the independent effects of gestational age. The

fact that these increased risks persisted even after restriction to
spontaneous-onset vaginal deliveries among women at low
risk adds strength to the inference that early term delivery per

Variations in Mortality and Morbidity by Gestational Age among Infants Born at Term

361

se may be the cause of the higher risks. Unfortunately, U.S.

birth certificates (which form the basis of the NCHS data
files) do not specify which labor inductions or cesarean deliveries are performed electively, that is, without clinical indication. Thus we cannot rule out the possibility that underlying
disease not captured by our low-risk criteria might be responsible both for the earlier delivery and the higher mortality and
morbidity rates we observed at early term GAs. A further
limitation of our study relates to the validity of the causes of
death noted on death certificates and the neonatal morbidity
diagnoses recorded on birth certificates. We observed that
among infant deaths caused by congenital anomalies, 43.4%
had not been diagnosed with congenital anomalies at birth.
Some degree of residual confounding may therefore be attributable to undiagnosed congenital anomalies that may have led
both to earlier birth and to subsequent morbidity or death.
Previous studies have also demonstrated underreporting of
medical risk factors, complications of labor and delivery, and
neonatal morbidity.22 Although such diagnostic misclassification is likely to have occurred in some cases, the degree of
misclassification should not differ by GA and thus would be
expected to attenuate, rather than accentuate, the associations
we observed.
Our results generally agree with those of Smith et al15
for SIDS and of Caughey et al13 for meconium aspiration
syndrome. Our study also reports several other important
findings. Although the risk of neonatal death at term decreases with advancing GA, the risk of overall neonatal death
because of birth asphyxia shows a U-shape relation, as do the
risks of several serious neonatal morbidities linked to asphyxia, including low 5-minute Apgar score and mechanical
ventilation. Although we restricted our main analysis to nonHispanic whites, we observed similar patterns among nonHispanic blacks, and thus our findings are likely to be generalizable.
Lindstrm et al23 recently reported higher risks of disability and lower disposable income in young adults who were
born at 37 to 38 weeks. These findings might be long-term
consequences of the higher risks of some serious neonatal
morbidities observed in our study. Although the absolute risks
are modest, the high (and increasing) frequency of early term
deliveries could represent a substantial public health burden.
The associations we and others have observed could be
confounded by the reasons for early delivery, but the possibility that earlier delivery per se could incur increased risks in
both the short and long term should be weighed whenever
elective labor induction or cesarean section are considered
before 39 weeks. Recent temporal increases in elective labor

362

Zhang and Kramer

induction and cesarean delivery suggest the need for a large

randomized trial to ensure that such interventions do more
good than harm.

REFERENCES
1.
International Statistical Classification of Diseases and Related Health Problems,
Tenth Revision, Volume 3. Alphabetical Index. Geneva, WHO, 1994.
2.
Kramer MS, Demissie K, Yang H, Platt RW, Sauv R, Liston R for the Fetal and
Infant Health Study Group of the Canadian Perinatal Surveillance System. The contribution of mild and moderate preterm birth to infant mortality. JAMA 2000;284:843-49.
3.
Rush RW, Keirse MJ, Howat P, Baum JD, Anderson AB, Turnbull AC. Contribution of preterm delivery to perinatal mortality. BMJ 1976;2:965-8.
4.
Divon MY, Haglund B, Nisell H, Otterblad PO, Westgren M. Fetal and neonatal
mortality in the postterm pregnancy: the impact of gestational age and fetal growth
restriction. Am J Obstet Gynecol 1998;178:726-31.
5.
Usher RH, Boyd ME, McLean FH, Kramer MS. Assessment of fetal risk in
postdate pregnancies. Am J Obstet Gynecol 1988;158:259-64.
6.
Hilder L, Costeloe K, Thilaganathan B. Prolonged pregnancy: evaluating gestation-specific risks of fetal and infant mortality. BJOG 1998;105:169-73.
7.
Clifford SH, Reid DE, Worcester J. Postmaturity. Am J Dis Child 1951;82:232.
8.
Savitz DA, Terry JW Jr, Dole N, Thorp JM Jr, Siega-Riz AM, Herring AH.
Comparison of pregnancy dating by last menstrual period, ultrasound scanning, and
their combination. Am J Obstet Gynecol 2002;187:1660-6.
9.
Wen SW, Joseph KS, Kramer MS, Demissie K, Oppenheimer L, Liston R, et al.
Recent trends in fetal and infant outcomes following post-term pregnancies. Chronic
Dis Can 2001;22:1-5.
10. Zeitlin J, Blondel B, Alexander S, Brart G, and the PERISTAT Group. Variation in rates of postterm birth in Europe: reality or artefact? BJOG 2007;114:1097-103.
11. Yuan H, Platt RW, Morin L, Joseph KS, Kramer MS. Fetal deaths in the United
States, 1997 vs 1991. Am J Obstet Gynecol 2005;193:489-95.
12. Kirby RS. Trends in labor induction in the United States: Is it true that what goes
up must come down? Birth 2004;31:148-51.
13. Caughey AB, Musci TJ. Complications of term pregnancies beyond 37 weeks of
gestation. Obstet Gynecol 2004;103:57-62.
14. Caughey AB, Stotland NE, Washington AE, Escobar GJ. Maternal and obstetric
complications of pregnancy are associated with increasing gestational age at term. AJOG
2007;196:155.e1-155.e6.
15. Smith GCS, Pell JP, Dobbie R. Risk of sudden infant death syndrome and week
of gestation of term birth. Pediatrics 2003;111:1367-71.
16. Vital Statistics of the United States: 2001 Natality. US Department of Health and
Human Services, Centers for Disease Control and Prevention, National Center for
Health Statistics. Hyattsville, Maryland: 2003.
17. Joseph KS, Huang L, Liu S, Ananth CV, Allen AC, Sauve R, et al. Reconciling
the high rates of preterm and postterm birth in the United States. Obstet Gynecol
2007;109:813-22.
18. Cole S, Hartford RB, Bergsj P, McCarthy B. International collaborative effort
(ICE) on birth weight, plurality, and infant mortality. Acta Obstet Gynecol Scand
1989;68:113-17.
19. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait
or intervention. Am J Epidemiol 1974;99:325-32.
20. Wen SW, Rusen ID, Walker M, Liston R, Kramer MS, Baskett T, et al. Comparison
of maternal mortality and morbidity between trial of labor and elective cesarean section
among women with previous cesarean delivery. AJOG 2004;191:1263-9.
21. Liu S, Liston RM, Joseph KS, Heaman M, Sauve R, Kramer MS. Maternal
mortality and severe morbidity associated with low-risk planned cesarean delivery versus
planned vaginal delivery at term. CMAJ 2007;176:455-60.
22. Piper JM, Mitchel EF Jr, Snowden M, Hall C, Adams M, Taylor P. Validation
of 1989 Tennessee birth certificates using maternal and newborn hospital records. Am J
Epidemiol 1993;137:758-68.
23. Lindstrm K, Winbladh B, Haglund B, Hjern A. Preterm infants as young adults:
a Swedish National Cohort Study. Pediatrics 2007;120:70-7.

The Journal of Pediatrics March 2009

APPENDIX
ICD-9 and ICD-10 Coding of Underlying Causes of
Infant Death
Asphyxia
ICD-9: 763, 766, 767, 768; 7616, 7617, 7620, 7621,
7622, 7624, 7625, 7626, 7701, 7722, 7790, 7792
ICD-10: P03, P10, P11, P12, P13, P14, P15, P20, P21,

P22, P24, P25, P26, P27, P28

Infection
ICD-9: 1-139; 320-326; 382; 420-422; 460-466; 475477; 480-491; 510, 511, 513, 540, 541, 566, 567, 570,
590, 591, 771, 790; 7700
ICD-10: A, B, P35, P36, P37, P38, P39
Postneonatal SIDS
ICD-9: 913; 7980, 7981, 7982, 7989
ICD-10: R95, R96, R99

Variations in Mortality and Morbidity by Gestational Age among Infants Born at Term

362.e1