Young Maternal Age and Parity Influences on Pregnancy Theresa 0. Scholl, Francis E. Johnson,
Outcome
PhD, Mary L. Hediger, PhD, Jianping Huang, MA, PhD, Woollcott Smith, PhD, and Isadore G. Antes,
MD
The influence of very young maternal age and parity on pregnancy outcome was examined in a cohort of nearly 900 adolescents and mature women from Camden, New Jersey. Youngprimigravid primiparas (aged 12 to 15 years) were compared with mature prim&avid primiparas (18 to 29 years). Young multiparas (19 years or younger, with a first pregnancy at the age of J2 to J 5 years) were compared with mature, multiparas (19 to 29 years old, with a first pregnancy at 18 years or older). After controlling for confounding factors, young primiparas were found to have a modest increase in preterm delivery, which was not statistically significant. However, low gynecologic age contributed disproportionately to the risk ofpreterm delivery in this group, with risk decreasing with each year from menarche (Cox’s proportional hazard, 0.80; 95% confidence interval [CJ], 0.68 to 0.94). Among multipants, there were several statistical interactions associated with increased risk of small-for-gestational-age infants, including interactions between young age and low prepregnancy body mass (adjusted odds ratio [AOR], 5.74; 95% CJ, 2.18 to J5.08), young age and a prior low-birth-weight infant (AOR, J 0.58; 95% CJ, 3.89 to 28.77), and young age and a prior preterm delivery (AOR, 5.52; 95% CJ, 2.04 to 14.98). Thus, while chronologic age per se may not be a good predictor of pregnancy outcome, adolescents remain a high-risk group because of factors that are more common among them (e.g., biologic immaturity, inadequate prenatal care, poverty, minority status, low prepregnancy weight) and because factors associated with an early adolescent pregnancy, such as low gynecologic age, may continue to influence the outcome of subsequent pregnancies. Ann Epidemiol J 992;2:565-575. KEY WORDS:
gestation&age
Adolescent pregnancy, young maternal age, parity, preterm birth, small-forinfant, low birth weight.
INTRODUCTION There
has been ongoing
controversy
about what effect,
if any, young maternal age and
multiparity at a young age have on pregnancy outcome ( 1, 2). Several reports indicated that there is an increased risk of poor outcomes with young maternity and with repeat pregnancy during the teenage years (l-4)) and that such adverse outcomes may be due solely to a differential distribution of risk factors, such as cigarette smoking, black race, or poor nutritional status among the young (5-7). In other words, apart from these risk factors, the age and parity of the young mother may be of little consequence. On the other hand, even allowing that there may be effects of young maternal age and multiparity,
the nature of these effects is still debated. Some studies claimed that
poor outcomes among young gravidas may be at least partially attributable to physiologic parameters associated with their youth, such as biologic immaturity marked by low
From the Departments of Obstetrics and Gynecology, University of Medicine and Dentistry of New Jersey-School of Osteopathic Medicine and the Robert Wood Johnson Medical School, Camden, NJ (T.O.S., M.L.H., J.H., I.G.A.); the U mversity of Pennsylvania, Philadelphia (J.H., F.E.J.) ; and the Department of Statistics, Temple University, Philadelphia, PA (W.S.). Address reprmt requests to: Theresa 0. Scholl, PhD, Departments of Obstetrics and Gynecology, Unwersity of Medicine and Dentistry of New Jersey-SOM, 401 Haddon Avenue, Camden, NJ 08103. Received April 22, 1991; revised June 21, 1991. 0 1992 Elsev~erScience Publishing Co.,
Inc.
1047.2797192/$05.00
566
Scholl et al. YOUNG MATERNAL AGE, PARITY, AND FETAL OUTCOME
gynecologic (8-10).
age (years since menarche)
Others,
AEP Vol. 2, No. 5 September 1992: 565-575
or continued maternal growth during pregnancy
observing better development
and lower levels of mortality for infants
born to certain age groups of teenagers,
claimed that when the environment
the young actually have a reproductive
advantage over the more mature (11,
is poor, 12).
Studies typically have focused on teenagers of all ages and not necessarily those at highest risk of adverse outcomes, although they often do include a stratum in this age range (~15
years). Failure to control for well-known
a history of poor outcome
among multiparas,
confounding
is common.
variables,
such as
Since early pregnancy may
have sequelae at older ages (13), use of mature women, similarly exposed to pregnancy during adolescence, as a comparison may be inappropriate. In addition, much use has been made of heterogeneous outcomes, such as infants of low birth weight (LBW), which mixes the cause for preterm birth with that for intrauterine
growth retardation
(2, 4). Studies of pregnancy outcome in very young women are fairly uncommon (5, 14-16)) since pregnancy during the age range of 12 to 15 years is somewhat rare (1). As far we are aware, no study has ever examined the outcome of a subsequent pregnancy in young women who were 12 to 15 years old at their first pregnancy. things
in mind,
we examined
in a prospective
outcomes in very young gravidas, as well as the independent births following an early pregnancy.
MATERIALS
AND
With
these
study, the risk of poor pregnancy influence
of parity on
METHODS
Data were derived from a prospective
study of pregnancy outcome
Jersey, one of the poorest cities in continental
in Camden,
New
United States. From January 1985 until
August 1990, young and mature women from two prenatal clinics in the city of Camden were recruited based on age at first pregnancy, gravidity, and parity. Of the 1017 eligible registrants,
3 refused and 83 moved, terminated
the pregnancy,
or changed
providers before delivery. Of the remaining women, 40 had a fetal death and 4 a multiple pregnancy, leaving a base sample for analysis of 887 women with a live-born infant. Study groups included 256 primigravidas,
aged 12 to 15 years at the first pregnancy,
and 236 young multigravidas, aged 19 and under, with a first pregnancy between the ages of 12 to 15 years. The remainder were mature control subjects, aged 18 to 29 at the first pregnancy, and thus not exposed to an early adolescent pregnancy. Of these, 207 were primigravid and 188 were multigravid. In both adolescent and mature groups, we excluded multigravidas who did not have a prior pregnancy in the target age range that had extended beyond 20 weeks’ gestation. Women were interviewed, and anthropometric data, substance use information, and sociodemographic information were obtained at entry to prenatal care (mean, 19.2 -+ 7.0 weeks), and at 28 and 36 weeks’ gestation. Maternal weight was measured at each prenatal visit during pregnancy, maternal height was measured at entry to care, and prepregnancy weight was obtained by recall at entry to care. Information on current and past pregnancy outcomes and complications was abstracted from the prenatal record, the delivery record, delivery room logbooks, and the infants’ charts. Several variables were derived, and included gynecologic age, defined as the difference between maternal age at conception of the index pregnancy and age at menarche; low gynecologic age, defined as 2 or fewer completed years since menarche (8, 9); and body mass index (BMI), computed as prepregnancy weight divided by height squared
AEP Vol. 2, No. 5 September 1992: 565-575
YOUNG MATERNAL
Scholl et al. AGE, PARITY, AND FETAL OUTCOME
567
(kg/m2). Adequacy of weight gain was judged for the length of gestation within 2 weeks of delivery, and defined using published criteria (17). To distinguish between effects of age and parity, the outcomes of live-born singletons of young primiparas were compared only with those of mature primiparas, while young multiparas were compared with mature, multiparous control subjects. LBW infants (< 2500 g), preterm birth (< 37 weeks’ completed gestation), and small-for-gestational-age (SGA; < 10th percentile for standards) infants were examined as outcomes. SGA was defined using standards that adjust birth weight for length of gestation, ethnicity, maternal parity, and fetal sex (18). Gestation was reckoned from the mother’s last menstrual period (LMP), as well as from the obstetric estimate of length of gestation. The obstetric estimate is also based on the LMP, but is confirmed by an early ultasound scan (69.7%) or serial measurements of uterine fundal height. In the face of a size-for-dates discrepancy of greater than 2 completed weeks, or when the LMP was not known, the obstetric estimate was based solely on the ultrasound examination. In this, as well as in our other studies using this approach (8, 19, 20), comparable effects on outcome were obtained with both methods so that unless otherwise noted, the results based on the estimates of gestation from the LMP are reported. Cox’s proportional hazard model, with censoring at 37 completed weeks, that is, the point at which a delivery is no longer considered preterm, was used to study the influence of young maternal age and parity on time from the LMP to preterm delivery (21). Parallel plots of rates of preterm delivery by chronologic age and gynecologic age groups indicated that the underlying assumptions of the proportional hazards model were satisfied by the data. When independent variables are used as dummies, the antilog of the hazard coefficient yields a relative hazard interpretable as an odds ratio, which may be adjusted for confounding. Multiple logistic regression, with control for confounding, was used when LBW infants and SGA infants were the outcomes of interest. Adjusted odds ratios (AORs) and their 95% confidence intervals (CIs) were computed from the logistic regression coefficients and the corresponding covariance matrix. Confounding was assessed by comparing adjusted and crude odds ratios, since the problem of relying on tests of statistical significance alone is well recognized (22). Separate models were fitted for each outcome containing the risk factor, potential confounding variables (Table l), and interaction terms using forward inclusion and backwards deletion (23). To control more stringently for the effects of chronologic age, both age and age squared were included simultaneously in some models (24).
RESULTS Women were uniformly of low income, and primarily of a minority race (see Table 1). More of the mature gravidas smoked and smoked more heavily than did the teenagers. Reported use of marijuana differed little across age and parity groups, but alcohol and cocaine were used more often by the older women. Adolescent gravidas were more likely to be black and to be receiving Medicaid, and adolescent multiparas were more likely to have an inadequate weight gain for gestation than any of the other groups studied. Young primigravidas were more likely to have a lower prepregnancy BMI. Mature multiparas were more likely than adolescent multiparas to have had a prior spontaneous or induced abortion or a prior poor outcome (LBW infant or preterm delivery).
568
Scholl et al. YOUNG
MATERNAL
TABLE
1
AGE,
PARITY,
Background
AEP Vol. 2, No. 5 September 1992: 565-575
AND FETAL OUTCOME
characteristics Young
primigravidas
by maternal Young multiparas
age and parity groups’ Older primigravidas
Older multiparas
207 -
I88
-4~ (~1 256 0.4 8.2 33.6 57.8
n
12 13 14 15 16 17 18 19 20-24 25-29
236 0.4 2.6 8.1 22.1 33.2 28.9 4.7 -
46.6 23.8 26.7 2.9
3.7 54.3 42.0
Black race n %
256 67.6
236 63.1
207 54.6
188 59.0
Medicaid n %
256 80.2
236 86.0
207 67.6
188 72.9
Prepregnancy body mass (kg/m’) n 5 19.6 19.5-24.5 > 24.5
256 23.0 63.3 13.7
235 14.5 57.4 28.1
207 21.3 58.9 19.8
188 19.1 49.5 31.4
Inadequate weight gain for gestation n %
245 20.9
229 31.4
202 23.3
187 26.7
256 80.9 11.3
235 63.8 15.7
4.3 1.6 2.0
9.4 8.1 3.0
206 69.4 11.2 9.7 6.3 3.4
188 59.0 9.0 11.7 13.3 6.9
252 87.7 9.1 2.8 0.4
232 87.1 7.3 4.3 1.3
207 86.0 9.2 3.4 1.4
187 86.1 6.4 7.0 0.5
252 99.6 0.4
227 97.8 0.9 1.3
207 97.6 1.9 0.5
187 95.7 4.3
252 86.9 12.7 0.4
231 76.2 23.4 0.4
206 71.8 27.2 1.0
187 60.4 37.6 2.0
Cigarettes/d n None 54 5-9 10-19 2 20 Marijuana n None Monthly Weekly Daily Cocaine n None Monthly Weekly Daily Alcohol ;; one 19.5), there was no significant effect of young maternal age among multiparous women. However, when prepregnancy BMI was low (5 19.5), young multiparas had a fivefold
increase
in the risk of having
with older multiparas
with normal
a SGA
infant
while
or high BMI, but mature
underweight,
multipatas
compared
with a low BMI
did not (Table 5). In the absence of a history of having a LBW infant and/or pteterm delivery, multiparous women did not show an increased risk of having a SGA infant current
pregnancy.
TABLE
4
Explanatory Chronologic Chronologic Gynecologic Gynecologlc Gynecologic Gynecologic
However,
among
young mothers
Preterm birth: Influence of gynecologic variable
Proportional
a& ageb age‘ age” age’ age’
who had had prior poor outcomes,
and chronologic
hazard
0.97 1.00 0.80 0.72 0.76 0.79
young in the
age in primiparas 95%
Confidence
interval
0.89-1.05 0.83-1.22 0.68-0.94 0.58-0.89 0.60-0.94 0.61-1.01
i Model l--Cm’s proportional hazard adjusted for black race, adequacy of weight gam. prepregnancy body mass. and cigaretteaid. h Model I. excluding low-gynecologic-age (5 2 y) pnents.
LModel
I. unadjusted
a Model
1, adlusted
for chrwwlogic
age.
’ Model
1, adjuted
for chronologic
age and age’.
’Model
1, unadpted
for chronologic
for chronologlc
age.
age restncted,
to age I Z- I5 y.
AEP Vol. 2, No. 5 September 1992: 565-575
TABLE between
YOUNG
MATERNAL
5 Small-for-gestational-age effects of low prepregnancy
Outcome: Small-forgestational-age infant”
%
95% Confidence interval
AOR
7.25 9.86
Young multiparas Older multiparas
571
Scholl et al. AND FETAL OUTCOME
by maternal age group and parity: Comparison body mass in young and older multiparas
Unadjusted
Prepregnancy BMI > 19.5 Young multiparas Older multiparas Prepregnancy BMI S 19.5
AGE, PARITY,
32.35 5.71
0.71 1.00
0.44- 1.60 -
5.74 0.26
2.18-15.08 0.491.34
’ Adjusted odds ratio (AOR) adjusted for cigarettes/d and adequacy of weight gain for gestation, and prior low-birthweight infant. The results are from models contingent upon the significant interaction between young age and a low prepregnancy BMI (5 19.5).
the risk for a SGA
infant
was significantly
increased.
This
increase
in risk was not
observed among older mothers with prior poor outcomes (Table 6). Compared with all multiparas with no prior LBW infant or preterm
birth, the adjusted odds ratio for LBW was increased more than ninefold (AOR 9.10; 95% CI, 4.5 to 18.5) when there was history of a LBW infant (computed from Table 7) and the adjusted hazard for preterm birth was increased more than twofold (AOR, 2.25; 95% CI, 1.3 to 3.8) when preterm birth was reported for a prior pregnancy (model in Table 2). Neither of these adverse outcomes were specifically increased with young maternal age, as described for SGA infants. Finally, an examination of the more heterogeneous outcome, LBW infant, showed that young primiparas were somewhat less likely to be at risk than were older primiparas, while young multiparas had a twofold increase in risk compared with older multiparas (see Table 7). In neither case were there any statistically significant interactions when LBW infant was used as the outcome of interest.
TABLE 6 Increased risk associated with prior poor birth outcome’: young and older mulriparas with small-for*gestational-age infants Outcome: Small-for-gestational-age
infantb
low-birth-weight infant or preterm birth Young multiparas Older multiparas Prior low-birth-weight infant Young multiparas Older multiparas Prior preterm birth Young multiparas Older multiparas Prior low-birth-weight infant and/or preterm birth Young multiparas Older multiparas
Unadjusted
%
AOR
Comparison
between
95 % Confidence interval
No prior
6.80 9.45
0.71 1.00
0.63-
1.59
44.44 9.68
10.58 0.66
3.89-28.77 0.16- 2.69
31.03 8.00
5.52 0.56
2.04-14.98 O.ll2.82
33.33 7.69
6.39 0.54
2.49-16.42 0.14- 2.14
* Prior pour birth cutcome defined as preterm delivery and/or infant with low birth weight. b Adjusted odds ratio (AOR) adjusted for cigarettes/d, low prepregnancy BMI, and adequacy of weight gain for gestation from models contingent upon a significant interaction between maternal age and prior outcome.
572
AEP Vol. 2, No. 5 September 1992: 565-575
Scholl et al. YOUNG MATERNAL AGE, PARITY, AND FETAL OUTCOME
TABLE
7
Low birth
weight
Outcome: Lowbirth-weight infanP
(
Outcome
PhD, Mary L. Hediger, PhD, Jianping Huang, MA, PhD, Woollcott Smith, PhD, and Isadore G. Antes,
MD
The influence of very young maternal age and parity on pregnancy outcome was examined in a cohort of nearly 900 adolescents and mature women from Camden, New Jersey. Youngprimigravid primiparas (aged 12 to 15 years) were compared with mature prim&avid primiparas (18 to 29 years). Young multiparas (19 years or younger, with a first pregnancy at the age of J2 to J 5 years) were compared with mature, multiparas (19 to 29 years old, with a first pregnancy at 18 years or older). After controlling for confounding factors, young primiparas were found to have a modest increase in preterm delivery, which was not statistically significant. However, low gynecologic age contributed disproportionately to the risk ofpreterm delivery in this group, with risk decreasing with each year from menarche (Cox’s proportional hazard, 0.80; 95% confidence interval [CJ], 0.68 to 0.94). Among multipants, there were several statistical interactions associated with increased risk of small-for-gestational-age infants, including interactions between young age and low prepregnancy body mass (adjusted odds ratio [AOR], 5.74; 95% CJ, 2.18 to J5.08), young age and a prior low-birth-weight infant (AOR, J 0.58; 95% CJ, 3.89 to 28.77), and young age and a prior preterm delivery (AOR, 5.52; 95% CJ, 2.04 to 14.98). Thus, while chronologic age per se may not be a good predictor of pregnancy outcome, adolescents remain a high-risk group because of factors that are more common among them (e.g., biologic immaturity, inadequate prenatal care, poverty, minority status, low prepregnancy weight) and because factors associated with an early adolescent pregnancy, such as low gynecologic age, may continue to influence the outcome of subsequent pregnancies. Ann Epidemiol J 992;2:565-575. KEY WORDS:
gestation&age
Adolescent pregnancy, young maternal age, parity, preterm birth, small-forinfant, low birth weight.
INTRODUCTION There
has been ongoing
controversy
about what effect,
if any, young maternal age and
multiparity at a young age have on pregnancy outcome ( 1, 2). Several reports indicated that there is an increased risk of poor outcomes with young maternity and with repeat pregnancy during the teenage years (l-4)) and that such adverse outcomes may be due solely to a differential distribution of risk factors, such as cigarette smoking, black race, or poor nutritional status among the young (5-7). In other words, apart from these risk factors, the age and parity of the young mother may be of little consequence. On the other hand, even allowing that there may be effects of young maternal age and multiparity,
the nature of these effects is still debated. Some studies claimed that
poor outcomes among young gravidas may be at least partially attributable to physiologic parameters associated with their youth, such as biologic immaturity marked by low
From the Departments of Obstetrics and Gynecology, University of Medicine and Dentistry of New Jersey-School of Osteopathic Medicine and the Robert Wood Johnson Medical School, Camden, NJ (T.O.S., M.L.H., J.H., I.G.A.); the U mversity of Pennsylvania, Philadelphia (J.H., F.E.J.) ; and the Department of Statistics, Temple University, Philadelphia, PA (W.S.). Address reprmt requests to: Theresa 0. Scholl, PhD, Departments of Obstetrics and Gynecology, Unwersity of Medicine and Dentistry of New Jersey-SOM, 401 Haddon Avenue, Camden, NJ 08103. Received April 22, 1991; revised June 21, 1991. 0 1992 Elsev~erScience Publishing Co.,
Inc.
1047.2797192/$05.00
566
Scholl et al. YOUNG MATERNAL AGE, PARITY, AND FETAL OUTCOME
gynecologic (8-10).
age (years since menarche)
Others,
AEP Vol. 2, No. 5 September 1992: 565-575
or continued maternal growth during pregnancy
observing better development
and lower levels of mortality for infants
born to certain age groups of teenagers,
claimed that when the environment
the young actually have a reproductive
advantage over the more mature (11,
is poor, 12).
Studies typically have focused on teenagers of all ages and not necessarily those at highest risk of adverse outcomes, although they often do include a stratum in this age range (~15
years). Failure to control for well-known
a history of poor outcome
among multiparas,
confounding
is common.
variables,
such as
Since early pregnancy may
have sequelae at older ages (13), use of mature women, similarly exposed to pregnancy during adolescence, as a comparison may be inappropriate. In addition, much use has been made of heterogeneous outcomes, such as infants of low birth weight (LBW), which mixes the cause for preterm birth with that for intrauterine
growth retardation
(2, 4). Studies of pregnancy outcome in very young women are fairly uncommon (5, 14-16)) since pregnancy during the age range of 12 to 15 years is somewhat rare (1). As far we are aware, no study has ever examined the outcome of a subsequent pregnancy in young women who were 12 to 15 years old at their first pregnancy. things
in mind,
we examined
in a prospective
outcomes in very young gravidas, as well as the independent births following an early pregnancy.
MATERIALS
AND
With
these
study, the risk of poor pregnancy influence
of parity on
METHODS
Data were derived from a prospective
study of pregnancy outcome
Jersey, one of the poorest cities in continental
in Camden,
New
United States. From January 1985 until
August 1990, young and mature women from two prenatal clinics in the city of Camden were recruited based on age at first pregnancy, gravidity, and parity. Of the 1017 eligible registrants,
3 refused and 83 moved, terminated
the pregnancy,
or changed
providers before delivery. Of the remaining women, 40 had a fetal death and 4 a multiple pregnancy, leaving a base sample for analysis of 887 women with a live-born infant. Study groups included 256 primigravidas,
aged 12 to 15 years at the first pregnancy,
and 236 young multigravidas, aged 19 and under, with a first pregnancy between the ages of 12 to 15 years. The remainder were mature control subjects, aged 18 to 29 at the first pregnancy, and thus not exposed to an early adolescent pregnancy. Of these, 207 were primigravid and 188 were multigravid. In both adolescent and mature groups, we excluded multigravidas who did not have a prior pregnancy in the target age range that had extended beyond 20 weeks’ gestation. Women were interviewed, and anthropometric data, substance use information, and sociodemographic information were obtained at entry to prenatal care (mean, 19.2 -+ 7.0 weeks), and at 28 and 36 weeks’ gestation. Maternal weight was measured at each prenatal visit during pregnancy, maternal height was measured at entry to care, and prepregnancy weight was obtained by recall at entry to care. Information on current and past pregnancy outcomes and complications was abstracted from the prenatal record, the delivery record, delivery room logbooks, and the infants’ charts. Several variables were derived, and included gynecologic age, defined as the difference between maternal age at conception of the index pregnancy and age at menarche; low gynecologic age, defined as 2 or fewer completed years since menarche (8, 9); and body mass index (BMI), computed as prepregnancy weight divided by height squared
AEP Vol. 2, No. 5 September 1992: 565-575
YOUNG MATERNAL
Scholl et al. AGE, PARITY, AND FETAL OUTCOME
567
(kg/m2). Adequacy of weight gain was judged for the length of gestation within 2 weeks of delivery, and defined using published criteria (17). To distinguish between effects of age and parity, the outcomes of live-born singletons of young primiparas were compared only with those of mature primiparas, while young multiparas were compared with mature, multiparous control subjects. LBW infants (< 2500 g), preterm birth (< 37 weeks’ completed gestation), and small-for-gestational-age (SGA; < 10th percentile for standards) infants were examined as outcomes. SGA was defined using standards that adjust birth weight for length of gestation, ethnicity, maternal parity, and fetal sex (18). Gestation was reckoned from the mother’s last menstrual period (LMP), as well as from the obstetric estimate of length of gestation. The obstetric estimate is also based on the LMP, but is confirmed by an early ultasound scan (69.7%) or serial measurements of uterine fundal height. In the face of a size-for-dates discrepancy of greater than 2 completed weeks, or when the LMP was not known, the obstetric estimate was based solely on the ultrasound examination. In this, as well as in our other studies using this approach (8, 19, 20), comparable effects on outcome were obtained with both methods so that unless otherwise noted, the results based on the estimates of gestation from the LMP are reported. Cox’s proportional hazard model, with censoring at 37 completed weeks, that is, the point at which a delivery is no longer considered preterm, was used to study the influence of young maternal age and parity on time from the LMP to preterm delivery (21). Parallel plots of rates of preterm delivery by chronologic age and gynecologic age groups indicated that the underlying assumptions of the proportional hazards model were satisfied by the data. When independent variables are used as dummies, the antilog of the hazard coefficient yields a relative hazard interpretable as an odds ratio, which may be adjusted for confounding. Multiple logistic regression, with control for confounding, was used when LBW infants and SGA infants were the outcomes of interest. Adjusted odds ratios (AORs) and their 95% confidence intervals (CIs) were computed from the logistic regression coefficients and the corresponding covariance matrix. Confounding was assessed by comparing adjusted and crude odds ratios, since the problem of relying on tests of statistical significance alone is well recognized (22). Separate models were fitted for each outcome containing the risk factor, potential confounding variables (Table l), and interaction terms using forward inclusion and backwards deletion (23). To control more stringently for the effects of chronologic age, both age and age squared were included simultaneously in some models (24).
RESULTS Women were uniformly of low income, and primarily of a minority race (see Table 1). More of the mature gravidas smoked and smoked more heavily than did the teenagers. Reported use of marijuana differed little across age and parity groups, but alcohol and cocaine were used more often by the older women. Adolescent gravidas were more likely to be black and to be receiving Medicaid, and adolescent multiparas were more likely to have an inadequate weight gain for gestation than any of the other groups studied. Young primigravidas were more likely to have a lower prepregnancy BMI. Mature multiparas were more likely than adolescent multiparas to have had a prior spontaneous or induced abortion or a prior poor outcome (LBW infant or preterm delivery).
568
Scholl et al. YOUNG
MATERNAL
TABLE
1
AGE,
PARITY,
Background
AEP Vol. 2, No. 5 September 1992: 565-575
AND FETAL OUTCOME
characteristics Young
primigravidas
by maternal Young multiparas
age and parity groups’ Older primigravidas
Older multiparas
207 -
I88
-4~ (~1 256 0.4 8.2 33.6 57.8
n
12 13 14 15 16 17 18 19 20-24 25-29
236 0.4 2.6 8.1 22.1 33.2 28.9 4.7 -
46.6 23.8 26.7 2.9
3.7 54.3 42.0
Black race n %
256 67.6
236 63.1
207 54.6
188 59.0
Medicaid n %
256 80.2
236 86.0
207 67.6
188 72.9
Prepregnancy body mass (kg/m’) n 5 19.6 19.5-24.5 > 24.5
256 23.0 63.3 13.7
235 14.5 57.4 28.1
207 21.3 58.9 19.8
188 19.1 49.5 31.4
Inadequate weight gain for gestation n %
245 20.9
229 31.4
202 23.3
187 26.7
256 80.9 11.3
235 63.8 15.7
4.3 1.6 2.0
9.4 8.1 3.0
206 69.4 11.2 9.7 6.3 3.4
188 59.0 9.0 11.7 13.3 6.9
252 87.7 9.1 2.8 0.4
232 87.1 7.3 4.3 1.3
207 86.0 9.2 3.4 1.4
187 86.1 6.4 7.0 0.5
252 99.6 0.4
227 97.8 0.9 1.3
207 97.6 1.9 0.5
187 95.7 4.3
252 86.9 12.7 0.4
231 76.2 23.4 0.4
206 71.8 27.2 1.0
187 60.4 37.6 2.0
Cigarettes/d n None 54 5-9 10-19 2 20 Marijuana n None Monthly Weekly Daily Cocaine n None Monthly Weekly Daily Alcohol ;; one 19.5), there was no significant effect of young maternal age among multiparous women. However, when prepregnancy BMI was low (5 19.5), young multiparas had a fivefold
increase
in the risk of having
with older multiparas
with normal
a SGA
infant
while
or high BMI, but mature
underweight,
multipatas
compared
with a low BMI
did not (Table 5). In the absence of a history of having a LBW infant and/or pteterm delivery, multiparous women did not show an increased risk of having a SGA infant current
pregnancy.
TABLE
4
Explanatory Chronologic Chronologic Gynecologic Gynecologlc Gynecologic Gynecologic
However,
among
young mothers
Preterm birth: Influence of gynecologic variable
Proportional
a& ageb age‘ age” age’ age’
who had had prior poor outcomes,
and chronologic
hazard
0.97 1.00 0.80 0.72 0.76 0.79
young in the
age in primiparas 95%
Confidence
interval
0.89-1.05 0.83-1.22 0.68-0.94 0.58-0.89 0.60-0.94 0.61-1.01
i Model l--Cm’s proportional hazard adjusted for black race, adequacy of weight gam. prepregnancy body mass. and cigaretteaid. h Model I. excluding low-gynecologic-age (5 2 y) pnents.
LModel
I. unadjusted
a Model
1, adlusted
for chrwwlogic
age.
’ Model
1, adjuted
for chronologic
age and age’.
’Model
1, unadpted
for chronologic
for chronologlc
age.
age restncted,
to age I Z- I5 y.
AEP Vol. 2, No. 5 September 1992: 565-575
TABLE between
YOUNG
MATERNAL
5 Small-for-gestational-age effects of low prepregnancy
Outcome: Small-forgestational-age infant”
%
95% Confidence interval
AOR
7.25 9.86
Young multiparas Older multiparas
571
Scholl et al. AND FETAL OUTCOME
by maternal age group and parity: Comparison body mass in young and older multiparas
Unadjusted
Prepregnancy BMI > 19.5 Young multiparas Older multiparas Prepregnancy BMI S 19.5
AGE, PARITY,
32.35 5.71
0.71 1.00
0.44- 1.60 -
5.74 0.26
2.18-15.08 0.491.34
’ Adjusted odds ratio (AOR) adjusted for cigarettes/d and adequacy of weight gain for gestation, and prior low-birthweight infant. The results are from models contingent upon the significant interaction between young age and a low prepregnancy BMI (5 19.5).
the risk for a SGA
infant
was significantly
increased.
This
increase
in risk was not
observed among older mothers with prior poor outcomes (Table 6). Compared with all multiparas with no prior LBW infant or preterm
birth, the adjusted odds ratio for LBW was increased more than ninefold (AOR 9.10; 95% CI, 4.5 to 18.5) when there was history of a LBW infant (computed from Table 7) and the adjusted hazard for preterm birth was increased more than twofold (AOR, 2.25; 95% CI, 1.3 to 3.8) when preterm birth was reported for a prior pregnancy (model in Table 2). Neither of these adverse outcomes were specifically increased with young maternal age, as described for SGA infants. Finally, an examination of the more heterogeneous outcome, LBW infant, showed that young primiparas were somewhat less likely to be at risk than were older primiparas, while young multiparas had a twofold increase in risk compared with older multiparas (see Table 7). In neither case were there any statistically significant interactions when LBW infant was used as the outcome of interest.
TABLE 6 Increased risk associated with prior poor birth outcome’: young and older mulriparas with small-for*gestational-age infants Outcome: Small-for-gestational-age
infantb
low-birth-weight infant or preterm birth Young multiparas Older multiparas Prior low-birth-weight infant Young multiparas Older multiparas Prior preterm birth Young multiparas Older multiparas Prior low-birth-weight infant and/or preterm birth Young multiparas Older multiparas
Unadjusted
%
AOR
Comparison
between
95 % Confidence interval
No prior
6.80 9.45
0.71 1.00
0.63-
1.59
44.44 9.68
10.58 0.66
3.89-28.77 0.16- 2.69
31.03 8.00
5.52 0.56
2.04-14.98 O.ll2.82
33.33 7.69
6.39 0.54
2.49-16.42 0.14- 2.14
* Prior pour birth cutcome defined as preterm delivery and/or infant with low birth weight. b Adjusted odds ratio (AOR) adjusted for cigarettes/d, low prepregnancy BMI, and adequacy of weight gain for gestation from models contingent upon a significant interaction between maternal age and prior outcome.
572
AEP Vol. 2, No. 5 September 1992: 565-575
Scholl et al. YOUNG MATERNAL AGE, PARITY, AND FETAL OUTCOME
TABLE
7
Low birth
weight
Outcome: Lowbirth-weight infanP
(
