Paediatric and Perinatal Epidemiology 1989, 3, 357-366
Association between low gynaecological age and
Theresa 0. Scholl*, Mary L. Hediger*, Ruth W.
Salmoni-, DanieI H. Belsky* and Isadore G. Ancesl
"Department of Obstetrics and Gynecology, University of Medicine
and Dentistry of New Jersey, School of Osteopathic Medicine,
tour Lady of lourdes Medical Center, and $Cooper Hospital/
University Medical Center, Camden, New jersey
Summary. Low gynaecological age, defined as conception within 2
completed years of menarche, was examined for its association with
preterm birth, using data from a geographically based cohort of over
1700 young primigravidae aged 18 or younger at start of prenatal care.
After stratifying by chronological age and controlling for confounding
variables, low gynaecological age was associated with almost double the
risk of preterm delivery whether estimated from the mother's last
menstrual period (adjusted odds ratio (AOR)=l.77, 95% CI 1.19-2.64)
or using the obstetric estimate of gestation (AOR=2.10, 95y0 CI
1.36-3.25). Low gynaecological age was also associated with an increase
in risk of low birthweight (LBW) (AOR--1.70, 95% CI 1.01-2.88), but
not of small-for-gestational-age babies (AOR-0.94, 95% CI 0.49-1.81).
Thus low gynaecological age may be an important addition to assess-
ment systems to detect women at risk of pretem labour and delivery.
Young maternal age is associated with increased risk of low birthweight as well as
Address for correspondence: Dr Theresa 0. Scholl, Department of Obstetrics and Cynerol-
ogy, UMDNJ-SOM, 401 Haddon Avenue, Camden, New Jersey 08103, USA.
T.O. Scholl, M.1. Hediger, R.W. Salmon et al.
other poor pregnancy outcomes. Mechanisms which mediate these associations
are unclear. Low birthweight, in particular, is a heterogeneous category which
combines intrauterine growth retardation (IUGR) and preterm delivery.2 Presum-
ably, factors which shorten gestation and result in preterm birth differ from those
which give rise to IUGR. Such factors are poorly differentiated in general, and
during adolescent pregnancy in particular have yet to be identified.
Low gynaecological age, defined as conception within 2 completed years of
menarche, is sometime^'.^ but not always5-' linked to an increased incidence of
low birthweight. Studies showing a positive association do not appear to have
adequately controlled for background characteristics, such as maternal chrono-
logical age or ethnic group, which potentially confound effects of reproductive
immaturity on pregnancy outcome. Negative studies have either suffered from
inadequate statistical power to declare observed differences as statistically
significant or been carried out in developing countries in areas of the world where
malnutrition is endemic.
When chronic and clinical malnutrition are prevalent, menarche often is
delayed until the completion of breast and pubic hair stages at the end of puberty.8
In such instances, low gynaecological age may be neither a sensitive nor a specific
indicator of reproductive immaturity, but this does not appear to be the case in
developed countries such as the United States. In this report, using data from a
large, geographically based cohort of young pregnant women, we present
evidence suggesting that low gynaecological age is an independent risk factor for
low birthweight which is specific to a single underlying cause: preterm birth.
Material and methods
Between September 1983 and December 1986,2789 young pregnant women were
enrolled into the Camden County Adolescent Family Life Project (CCAFL), a
federally-funded demonstration project that provided prenatal care to pregnant
teenagers. The five hospitals and clinics in Camden County, New Jersey, that
provide prenatal care participated in the demonstration. Clinic patients were given
a comprehensive prenatal programme, which was standardised across sites, and
included prenatal and postpartum medical care, social services, health education,
and obstetric services.
Apart from age (I
18 years at entry to care), subjects were unselected and more
than 99% agreed to provide data. Of the 2789 registrants, 10.3% (n-288) moved,
terminated the pregnancy, or failed to return to the clinics before delivery. Of the
remaining 2501 for whom pregnancy outcomes were available, 4.0% (n= 101)
had a fetal demise, and 0.8% (n=20) had a livebom multiple birth. An additional
nine Asian patients were excluded, leaving a base sample for analysis of 2371
white, black and Hispanic (Puerto Rican) adolescents who had liveborn singleton
low gynaecological age
deliveries. Since only 17 multiparae were of low gynaecological age, this analysis
was confined to data from 1784 primigravidae (75.290 of the base sample).
Patients were interviewed either at entry to prenatal care or at the time of their
first physical examination, which was usually 1-2 weeks later. During the
interview, data on maternal behavioural, demographic and social characteristics,
reproductive history (including age at menarche), medical history, and usual
prepregnant weight, were obtained by trained clinical personnel following a
standard protocol. Information on pregnancy outcomes, complications, maternal
height and weight gain during pregnancy was abstracted from the prenatal record,
the delivery record, delivery room logbooks, and the infants' charts.
Several additional variables were derived from this information. Low
gynaecological age ( 5 2 completed years) was defined as the difference between
maternal age at conception and her age at menarche. Body mass index (BMI),
calculated as prepregnant weight-for-height2 (kg/m2), was derived to index
adiposity and prepregnant nutritional status. Adequacy of weight gain during
pregnancy was assessed for the entire pregnancy according to gestation at the
week of delivery using published charkg
Low birthweight ((2500 grams) and its two major underlying causes, preterm
delivery ( t 3 7 weeks completed gestation) and smallness-for-gestational-age
(< 10th percentile for standards), were examined as outcomes. Smallness-for-
gestational-age (SGA), usually taken to indicate intrauterine growth retardation,
was defined as a birthweight below the 10th percentile for gestational age, using
the birthweight standards compiled by Brenner et d . * O which adjust birthweight
for ethnicity (white vs black), maternal parity and fetal sex. Gestation was
reckoned using two sources: from the first day of the mother's last menstrual
period (LMP) and, from the obstetric estimate, based on the LMP confirmed either
by serial measurement of uterine fundal height and/or ultrasound (65%). In the
face of a size for dates discrepancy (>2 completed weeks) or when the LMP was
not known, the obstetric estimate was based on ultrasound examination of the
fetus alone. There was excellent agreement between the two estimates of
gestation. The LMP and the obstetric (OB) estimates agreed zk 1 week in 77.8% of
cases and k 2 weeks in 86.34b. In 4.2% the OB estimate exceeded the LMP
estimate by more than 2 weeks, and in another 9.5% the LMP estimate exceeded
the OB estimate by more than 2 weeks. Although the agreement was good, in
subsequent analyses both estimates of gestation were used separately to ensure
that the results were comparable.
The principal statistical technique used in analysis was multiple logistic
regression. In order to control more stringently for the effect of chronological age
on preterm birth, the logistic regression was stratified by narrow categories (age
12-14, 15, 16, 17, 18-19) of chronological age." This practice more effectively
controls for strong confounding, such as by chronological age, than does simple
linear adjustment for age. Adjusted odds ratios (AOR), adjusted for confounding,
were computed from the logistic regression coefficients; their 95 % confidence
intervals (CI) were computed from the covariance matrix. Multiple linear
regression,'* with adjustment for confounding, also was used to estimate linear
effects of low gynaecological age on birthweight and gestation.
Potential confounders were those variables associated with gynaecological age,
as well as with pregnancy outcome. Confounding was assessed by comparing
adjusted and crude odds ratiosI2 since the problem of relying on tests of statistical
significance alone to assess confounding is well re~ognised.'~
were fitted for each outcome and for both methods of estimating gestation. Each
model contained the risk factor, confounding variables, and interaction terms
Table 1. Selected background characteristics associated with low gynaecologi-
T.O. Scholl, M.L. Hediger, R.W. Salmon et al.
Low gynaecological age
High gynaecological age
Quetefet's Index (Body Mass Index)
Pregnancy weight gain for gestation
Clinic payment status
Low gynaecologicaf age
using forward inclusion and backward deletion.** The maternal characteristics
treated as potential confounding variables in preliminary models included
chronological age, ethnicity, prepregnant body mass index, cigarette and
marijuana smoking, adequacy of pregnancy weight gain for gestation, and clinic
payment status. No interaction terms were statistically significant (P>O.l).
The majority of young primigravidae with low gynaecological age were under the
age of 16 years. They were more likely to be black, to be non-smokers, and to have
prenatal care and delivery costs funded by Medicaid. Low gynaecological age was
also correlated with a lower prepregnant BMI and a slightly increased rate of
inadequate pregnancy weight gain for gestation (Table 1).
The incidence of preterm delivery, reckoned from the LMP as well as the
obstetric estimate, along with rates of low birthweight (LBW) and smallness-for-
gestational-age (SGA) were computed within chronological age strata according to
gynaecological age (Table 2). At every chronological age, primigravidae with low
gynaecological age had higher rates of preterm delivery regardless of whether the
LMP or the OB estimate was used to estimate gestation. In addition, low
gynaecological age mothers had a tendency toward more low birthweight infants,
particularly at low chronological ages ((16 years) but this was not as marked as
Table 2. Rates of low birthweight (LBW), small-for-gestational-age
(SGA) and preterm birth by gynaecological age and chronological age
Low gyn. High gyn. Low gyn. High gyn.
YO Preterm LMP
O/O Pretenn OB
Low gyn. High gyn. Low gyn. High gyn.
T.O. Scholl, M.L. Hediger, R.W. Salmon et al.
Table 3. Adjusted odds ratios; LBW, SGA and preterm birth by gynaeco-
Outcome Unadjusted 70
1. Low birthweight" Low gyn. 13.4
High gyn. 8.5
2. SGAb Low gyn. 6.3
High gyn. 7.0
Low gyn. 29.1
High gyn. 13.6
3. Preterm LMP'
4. Preterm OBd Low gyn. 19.8
High gyn. 8.7
Stratified by chronological age (12-14, 15, 16,17,18-19) and adjusted
for ethnicity (black, Puerto Rican vs white), BMI ((19.5 vs other), adequacy
of weight gain (yes/no), smoking (yeslno).
Less than 10% of Brenner standards which adjust for maternal
ethnicity, panty and fetal sex, stratified by chronological age and adjusted
Computed from LMP. Regression is stratified by chronological age and
adjusted for ethnicity, adequacy of weight gain during pregnancy, smoking
Obstetric estimate of gestation. Regression is stratified by chronologi-
cal age and adjusted for ethnicity, smoking, BMI and adequacy of weight
gain during pregnancy.
the association with preterm delivery. There was little association between low
gynaecological age and SGA after stratification by chronological age.
These associations with low gynaecological age were confirmed in Iogistic
regression analyses stratified by chronological age and controlled for additional
confounding variables (Table 3). In the logistic models, low gynaecological age
was associated with approximately double the adjusted odds ratio for low
birthweight (AOR- 1.70, 9546 CI 1.01-2.88), for preterm delivery as assessed
from the LMP (AOR- 1.77, 959/0 CI 1.19-2.64), and for preterm delivery assessed
from the obstetric estimate of gestation (AOR-2.10,
Consistent with the univariate analysis, low gynaecological age was not
associated with an increased odds ratio for SGA births (AOR=0.94, 95% CI
95% CI 1.36-3.25).
Logistic regressions for preterm birth were recomputed after confining the
analysis to categories of patients where gestation was more secure. These
included: patients with entry to care at 24 weeks or earlier, patients with a known
LMP, patients with both earlier entry to care and a known LMP, and patients with
estimates of gestation which agreed within 2 completed weeks. There was a
modest increase in the adjusted odds ratio in the 1218 patients (68.3%) who
Low gynaecological age
entered care earlier than 24 weeks (AOR= 1.88, 95% CI 1.19-3.00 from the LMP,
AOR=2.51, 9596 CI 1.53-3.11 from the OB estimate). In the other categories
examined, the adjusted odds ratio relating low gynaecological age to preterm
births stiIl showed an approximate doubling.
Linear models, estimating the effect of gynaecological age on gestation and
birthweight also indicated that low gynaecological age was associated with
decreased gestation (Table 4). After control for maternal age and other
confounding variables, low gynaecological age was associated with a decrease in
gestation by 0.69 or 0.59 weeks according to whether gestation was estimated
from the LMP or the OB. After controlling for chronological age and other
confounding variables but not gestation, infants of low gynaecological age
mothers weighed 87.6 grams less. However, after adjusting for gestation from the
OB estimate, low gynaecological age was associated with a reduction in
birthweight of only 19.1 grams, suggesting shortened gestation as the main reason
for the lower birthweights in infants of primigravidae of low gynaecological age.
In 1977, Zlamik & BurrneisteF' coined the term 'low gynaecological age' to
distinguish young gravidae who were not physiologically mature from those who
were. Their rationale may have originated in the well-known differences in the
onset and timing of adolescent development. l4 At similar chronological ages some
girls have experienced a growth spurt, menarche and, according to indicators such
as skeletal age, breast and pubic hair development, are physiologically mature,
while others are not and may not yet be mature for several months or even
Table 4. Linear models for effects of low gynaecological age on birthweight and gestation
Birthweight not adjusted for
Birthweight adjusted for gestation
- 19.1 grams'
Gestation by LMP
Gestation by OB
a Adjusted for chronological age, ethnicity, smoking, low and high BMI (vs average),
adequacy of weight gain during pregnancy.
Adjusted for gestation (obstetric estimate), and other variables in above model.
Adjusted for chronological age, maternal ethnicity and adequacy of weight gain
Adjusted for chronological age, maternal ethnicity and smoking.
T.O. Scholl, M.L. Hediger, R.W. Salmon et al.
years.I4 Thus, it was suggested, gynaecological age might be used to further refine
the risk of low birthweight associated with young maternal age.3 But effects of low
gynaecological age were not found in subsequent studies so that its utility was
doubted.Is These studies, however, focused on size at birth rather than on any
association between a gynaecological age and shortened gestation.
In the present study, low gynaecological age appeared to have had an effect on
pregnancy outcome which was detectable even after stringent control for maternal
chronological age and other confounding variables. Low gynaecological age was
associated with almost double the risk of low birthweight. Furthermore, for the
first time, we have demonstrated that the association was specific to length of
gestation. Low gynaecological age was associated with a decreased average
duration of gestation by approximately two-thirds of a week. However gestation
was estimated the risk of preterm delivery was approximately doubled when
gynaecological age was low. There was no association between gynaecological
age and smallness-for-gestational-age or birthweight adjusted for gestation, again
demonstrating specificity for length of gestation.
The sensitivity of these findings to misclassification was tested by recomputing
adjusted odds ratios, after excluding certain cases where gestation was less secure.
In each sub-group the association between low gynaecological age and preterm
birth was confirmed. Adjusted odds ratios were modestly increased by confining
the analysis to those entering prenatal care at 24 weeks or earlier. With early entry
to prenatal care, the LMP is more likely to be recalled accurately. Furthermore,
fetal age estimated using early ultrasonography is less dependent on fetal size
than later in pregnancy when fetal size is more variable.I6 Confining the analysis
to cases where differences in estimates of gestation were 2 2 weeks did little to
increase or diminish adjusted odds ratios for preterm birth.
Use of the LMP as opposed to the obstetric estimate increased the absolute risk
of preterm birth from 10.1% to 15.54b. Rates of preterm birth based on the LMP
alone are in agreement with US data for minorities" and with information from
the Collaborative Perinatal Project.'* A lower incidence of preterm birth with the
obstetric estimate accords with published data from large-scale studies which
included ethnic minoritie~.'~ Poor memory or vaginal bleeding early in pregnancy
would be the most likely causes of higher rates of preterm birth reckoned from the
LMP. Despite differences in absolute level of risk, adjusted odds ratios for preterm
birth computed from each estimate of gestation were consistent, making
inferences about the influence of low gynaecological age on preterm birth more
secure. Thus, the addition of low gynaecological age to screening protocols to
detect increased risk of preterm labour and delivery may improve predictive value
of these instruments. Improved detection is especially important among primi-
gravidae, where sensitivity of risk assessment systems is low.
There are several possible explanations for the association between low
gynaecological age and preterm birth. For example, low gynaecological age may
Low gynaecological age
be a marker for social disadvantage and concomitant risk factors such as exposure
to alcohol or drugs which may affect gestation. Alternatively, low gynaecological
age may be indicative of poor nutritional status associated with continued
maternal growth during pregnancy.20 Another possibility is that low gynaecologi-
cal age is associated with maturational differences in the hypothalamic-pituitary
axis or in hormonal milieu of pregnancy which theoretically would shorten
ge~tation.~ Finally, low gynaecological age may reflect immaturity of the uterus,
cervix or their blood supply which may predispose young gravidae to infection
and secondarily increase the incidence of preterm birth.*'
Preterm birth is a leading cause of perinatal morbidity and mortality in the
United States. It is also an important underlying cause of low birthweighk2 Yet
few studies have examined specifically the causes of preterm birth, instead mixing
preterm and small-for-gestational-age infants into the composite category of low
birthweight. To improve utility of risk assessment for preterm birth, risk factors
specific to pretenn delivery need to be defined. Based upon the current study, one
specific risk factor appears to be low gynaecological age. A prior history of preterm
delivery is one of the strongest risk factors for subsequent preterm birth.'
Consequently, factors associated with adolescent pregnancy, such as low
gynaecological age, may continue to influence the outcome of subsequent
pregnancies in the mature years.
Supported in part by grants HD 18269 from NICHD and APH 000522 from the
Office of Population Affairs, Department of Health and Human Services.
Washington: Institute of Medicine, National Academy Press, 1985; pp. 46-93.
Kessel, S.S., Villar, J., Berendes, H.W.
in the United States. Journal of the Anrerican Medical Association 1984; 251:1978-1982.
Zlatnik, F.J., Burmeister, L.F. Low 'gynecologic age': an obstetric risk factor. American
Journal of Obstetrics and Gynecology 1977; 128:183-186.
Ma yland State Medical Journal 1971; 2050-52.
Hollingsworth, D.R., Kotchen, J.M. Gynecologic age and its relation to neonatal
outcome. In Birth Defects: Original Article Series Vol XVZI. New York: Alan R. Liss, 1981; pp.
Committee to Study the Prevention of Low Birthweight. Preventing low birthweight.
et al. The changing patterns of low birthweight
Erkon, K.A., Rimer, B.A., Stine, O.C. Juvenile pregnancy - role of physiologic maturity.
gynecologic age in pregnant adolescents. Yale Journal of Biology and Medicine 1984;
Felice, M.E., James, M., Shragg, P. et al. Observations related to chronologic and
7 Frisancho, A.R., Matos, J., Bolletino, L.A. Role of gynecological age and growth
T.O. Scholi, M.L. Hediger, R.W. Salmon et al.
maturity status in fetal maturation and prenatal growth of infants born to young still-
growing adolescent mothers. Human Biology 1984; 56:583-593.
Cameron, N., Mitchell, J., Meyer, D. et al. Secondary sexual development of 'Cape
Coloured' girls following kwashiorkor. Annals of Human Biology 1988; 1565-76.
Butman, M. Prenatal Nutrition: A Clinical Manual, 2nd edition. Boston: Massachusetts
Department of Health, 1983; pp. 23-32.
10 Brenner, W.E., Edelman, D.A., Hendricks, C.H. A standard for fetal growth for the
United States of America. American Journal of Obstetrics and Gynecology 1976; 126:555-564.
1 1 Breslow, N.E., Day, N.E. Statistical Methods in Cancer Research. Volume 1. The Analysis
of Case-Control Studies. Lyon: IARC Scientific Publications, 1980; pp. 192-247.
12 Kleinbaum, D.G., Kupper, L.L., Morgenstern, H. Epidemiologic Research: Principies and
Quantitatiae Methods. Belmont, CA: Lifetime Learning, 1982; pp. 419-491,
13 Dales, L.G., Ury, H.K. An improper use of statistical significance testing in studying
covariables. lnternational Journal of Epidemiology 1978; 7:373-375.
14 Tanner, J.M. Growth at Adolescence. Oxford Blackwell Scientific, 1962; pp. 94-155.
15 Stevens-Simons, C., Forbes, G.F., Kreipe, R.E. et al. A comparison of chronologic age
and gynecological age as indices of biologic maturity. American journal of Diseases of
Children 1986; 140702-705.
16 Manning, F.A. Ultrasound in perinatal medicine. In Maternal-Fetal Medicine, Editors:
Robert K. Creasy & Robert Resnick. Philadelphia: W.B. Sanders, 1984; pp. 210-211.
17 National Center for Health Statistics. Advance report of final natality statistics, 1985.
Monthly vital statistics report, Volume 36, No. 4, Suppl. DHHS Publication No (PHS) 87-
1120. Hyattsville, MD: Public Health Service, 1985.
18 Hardy, J.B., Drage, J.S., Jackson, E.C. The First Year of Life. The Collaborative Perinatal
Project of the National Institute of Neurological and Communicative Disorders aird Stroke.
Baltimore: The John Hopkins University Press, 1979; pp. 40-41.
19 Shiono, P.H., Klebanoff, M.A. Ethnic differences in preterm and very preterm
delivery. American Journal of Public Health 1986; 76:1317-1321.
20 Scholl, T.O., Hediger, M.L., Ances, I.G. et al. Growth during early teenage
pregnancies. Lancet 1988; i:701-702.
21 MeGregor, J.A. Adolescent misadventures with urethritis and cervicitis. journal of
Adolescent Health Care 1985; 6:286-297.