JOLLY, M. C. MBBS, MRCOG; SEBIRE, N. MD; HARRIS, J.; ROBINSON, S. MD, FRCP; REGAN, L. MD, FRCOG
Imperial College of Science Technology and Medicine (St. Mary's Campus), London, United Kingdom.
Address reprint requests to: Lesley Regan, MD, Imperial College of Science Technology and Medicine, St. Mary's Campus, Department of Reproductive Science and Medicine (The Mint Wing), South Wharf Road, Paddington, London, W2 1NY, United Kingdom. E-mail: firstname.lastname@example.org
St. Mary's Hospital Joint Standing Research Committee funded MCJ's fellowship.
Received March 16, 2000. Received in revised form July 5, 2000. Accepted August 31, 2000.
Objective: To quantify the age-related risks of adverse outcome during pregnancy in women less than 18 years old.
Methods: We analyzed data from 341,708 completed singleton pregnancies in the North West Thames region between 1988 and 1997. Pregnancy outcomes were compared by age at delivery in women less than 18 years old (n = 5246) and 18–34 years old (n = 336,462); women 35 years old or older (n = 48,658) were excluded. Data are presented as percentages of women less than 18 and 18–34-year-old women, with adjusted odds ratios (OR) and 99% confidence intervals (CI).
Results: Pregnancy in women less than 18 years old was associated with increased risk of preterm labor before 32 weeks' gestation (OR 1.41, CI 1.02, 1.90), maternal anemia (OR 1.82, CI 1.63, 2.03), chest infection (OR 2.70, CI 1.21, 6.70), and urinary tract infection (OR 1.60, CI 1.11, 2.31), but less obstetric intervention. Operative vaginal delivery (OR 0.46, CI 0.41, 0.56), elective cesarean (OR 0.47, CI 0.35, 0.65), or emergency cesarean (OR 0.45, CI 0.38, 0.53) were all less likely in women aged less than 18 years. Women less than 18 years old were no more likely to have stillbirths (OR 0.75, CI 0.42, 1.34) or small-for-gestational-age infants (OR 0.95, CI 0.82, 1.09) than women aged 18–34 years.
Conclusion: Pregnant women less than 18 years old were more likely to deliver preterm than older women. In most other respects they have less maternal and perinatal morbidity and were more likely to have normal vaginal deliveries.
England and Wales have the highest teenage conception rate and the highest teenage motherhood rate in western Europe, four times higher than that of France and six times higher than that of The Netherlands. The teenage pregnancy rate in England and Wales continues to increase, and in 1997 more than 90,000 teenage girls conceived, of whom 8000 were under age 16 years and 2200 were under age 14 years.1
Although approximately half of pregnant teenage girls choose to terminate their pregnancies,2,3 many general practitioners, midwives, and obstetricians frequently are involved in providing antenatal care to teenagers. Those involved in maternity care should be aware of the particular obstetric risks associated with teenage pregnancy. The results from previous studies that investigated obstetric risks of teenage pregnancy are contradictory. Some reported statistically significant higher frequency of anemia,2,4,5 whereas others did not.6 Some studies reported association between low birth weight and young maternal age,7–9 but others did not.2 Many of the studies also lacked sufficient power to detect small but important differences in common outcomes or larger differences in outcomes with low prevalence, such as stillbirth. Pregnancy in young women is associated with many confounding factors and it is important to account for them. The aim of this study was to test the hypothesis that young maternal age is associated with altered obstetric risks and to quantify the risk after allowing for possible confounding factors using a validated database.10,11
Materials and Methods
Data were derived from the St. Mary's Maternity Information System database, a clinical database that records maternity information from National Health Service hospitals within the geographic boundaries of the North West Thames region. These state-funded hospitals serve northwest London and the surrounding area. The St. Mary's Maternity Information System contains data on more than 80% of all deliveries in the region, which has a population of more than 3.5 million. It has been validated for commonly recorded variables including fetal death.10,11 Postnatal mortality data were not analyzed in the present study because of incomplete ascertainment resulting from transfer of some neonates to tertiary referral units.
Data from 341,708 completed singleton pregnancies between 1988 and 1997 were analyzed. The subjects were divided into groups according to maternal age at delivery, either less than 18 (n = 5246) or 18–34 (n = 336,462) years old. Women 35 years or older (n = 48,658) were excluded. Raw frequencies of various outcomes of pregnancy in the maternal age groups were calculated, and multiple logistic regression models were constructed to examine the magnitude and significance of the independent effects of age. All models included maternal age, body mass index (BMI), ethnic group, parity, and hypertension at booking. A medical classification was used for ethnic group, on the basis of the racial ancestry of each woman, to three generations, rather than nationality or place of birth. Parity was determined by the number of pregnancies that had been delivered after 24 weeks' gestation before the index pregnancy. Hypertension was defined as diastolic blood pressure (BP) greater than 90 mmHg. All models, except for the gestational diabetes mellitus model, included preexisting diabetes mellitus and gestational diabetes mellitus. Preexisting diabetes mellitus was defined as impaired glucose tolerance treated by restriction of carbohydrate intake or glucose-lowering medication while not pregnant. Because different diagnostic criteria were used within the region, gestational diabetes mellitus was defined as abnormal glucose tolerance diagnosed during pregnancy, which inevitably included a few women who had abnormal glucose tolerance outside pregnancy.
Other confounding factors of specific relevance to a particular outcome were included as appropriate. Proteinuria was defined as 2+ protein or more on dipstick testing in at least two consecutive midstream specimens without infection. Preeclampsia was defined as hypertension associated with proteinuria in women who were normotensive at booking.
Diagnoses of outcome measures such as abruption were based on clinical features. Urinary tract infection or genital tract infection was diagnosed when there was microbiologic evidence of infection from either a midstream urine specimen or high vaginal swab, respectively. Chest and wound infections were diagnosed clinically. Pyrexia of unknown origin was defined as a temperature of greater than 38C with no obvious source of infection. Diagnosis of postpartum hemorrhage was based on estimated blood loss of 500 mL or more at delivery and for up to 24 hours postpartum, excluding normal lochia. Operative vaginal delivery referred to the use of obstetric forceps or ventouse during vaginal delivery. Late booking was defined as booking with the hospital for obstetric care after 19 or more completed weeks' gestation. Anemia was defined as the lowest recorded antenatal hemoglobin concentration less than 10 g/dL.
Results are presented as frequencies by age group and odds ratios (OR) with 99% confidence intervals (CI). Birth weight was expressed as a delta value (the number of standard deviations by which the observed birth weight differed from the expected normal mean for males and females for each week of gestation). Delta birth weight was used to calculate birth weight centile to account for influence of gestational age and sex. All analyses were done with Statistical Analysis Software version 6.12 (SAS Institute, Cary, NC) using a UNIX server, SUN SPARC STATION 20 running the SOLARIS 2.6 operating system (Sun Microsystems, Inc., Palo Alto, CA).
Demographic characteristics of the groups are reported in Table 1. Table 2 shows antenatal characteristics and complications. The proportion of women under 18 years who were booked after 20 weeks' gestation was twice that of controls, and the young mothers were more likely to be anemic. The incidence of preeclampsia was higher in women under 18 years old, but after controlling for confounding variables, the difference was not statistically significant. Breech presentation was less common in younger women. Table 3 shows maternal complications. Chest infection, urinary tract infection, and pyrexia of unknown origin were more likely in women under 18 years old. The risk of post-partum hemorrhage was lower in younger women. Table 4 shows delivery complications. Induction of labor, operative vaginal delivery, and cesarean delivery were all less frequent in younger women. Table 5 shows fetal and neonatal complications. Preterm delivery was more common in the group under 18 years old, who were twice as likely as the older group to deliver before 32 weeks' gestation. The proportion of small for gestational age and large for gestational age infants was the same in both groups. There was no difference between the groups in the risk of stillbirth or a low Apgar score. Young mothers were much less likely to breastfeed at discharge.
This study demonstrated that the obstetric risks of pregnancy in women under 18 years old are generally low, except for preterm delivery, which is the most important determinant of perinatal morbidity.12 The large size of this study permitted inclusion of many confounding variables in the logistic regression models and enabled us to estimate specific age-related risks of pregnancy in women less than 18 years old. The notable confounding variable we were unable to include is social class. Sociodemographic factors are associated with teenage pregnancy and adverse pregnancy outcome.13,14 It is probable that inclusion of social class in the logistic regression models would have reduced the OR for adverse outcomes in the under 18 years old group.
The St. Mary's Maternity Information System has been validated, but some clinical diagnoses rely on a degree of subjective assessment. For example, postpar-tum blood loss is generally underestimated in routine clinical practice, but it is unlikely that those estimates are more or less accurate in one group compared with the other. Therefore, the effect on results will be small. The use of 99% CIs means that the results are likely to represent a true difference between groups.
The higher proportion of young mothers booked after 20 weeks' gestation confirms the results of earlier studies,3–5 which might be because teenagers are reluctant to seek medical advice. Pregnancy denial is also related to delayed booking in teenagers.15
Anemia was more common in younger women and might have been caused by poorer nutrition. Low-income teenage women are more likely to be in worse nutritional condition before conception than older women,16 and the effects of poor nutrition are exacerbated by increased use of essential substrates in adolescents.17 Anemia does not necessarily equate with disease. Hemodilution is an important physiologic adaptation to normal pregnancy, and women who do not hemodilute have higher incidence of preeclamp-sia.18,19 The frequency of preeclampsia in the younger group was nearly half that of controls, although after accounting for confounding variables it was not statistically significant.
Gestational diabetes mellitus was less common in younger women. Pancreatic B cell function and insulin sensitivity decrease with age. Older women with a predisposition to type II diabetes are more likely to have inadequate B cell responses to stimulation and be more insulin-resistant than teenage women. The incidence of breech presentation was lower in the younger group, possibly because of better abdominal and uterine tone, as it was independent of parity, which was included in the logistic regression model. Despite the lower incidence of breech presentation, there was no difference in incidence of vaginal breech delivery.
The women under 18 years old had fewer labors induced and operative vaginal deliveries and cesareans. Better myometrial function, greater connective tissue elasticity, and lower cervical compliance all might have contributed to higher normal vaginal delivery rates in younger women. It also might be that obstetricians and midwives strive harder to avoid delivering a teenage woman by cesarean. Uterine hypotonia is the major cause of postpartum hemorrhage, and better myome-trial function might be responsible for the lower proportion of women under 18 who had postpartum hemorrhages. More women under 18 had postpartum stays longer than 3 days, but once confounding variables were included in the logistic regression model, the OR (0.88, CI 0.79, 0.97) showed that women under 18 were less likely to have prolonged postpartum stays.
We confirmed the increased risk of preterm labor in young mothers.7,20,21 The mechanisms associated with it are not known. Smoking is associated with preterm labor and was included in the logistic regression model as a confounding variable.22,23 There is an association between bacterial vaginosis and preterm labor24; however, few of the participating hospitals would have screened for it. The OR for genital tract infection in women under 18 years old was not higher. The OR for urinary tract infection was higher in women under 18 years, and that infection is associated with preterm birth.22 A previous study showed that preterm delivery was more than twice as likely in women with iron-deficiency anemia.25 The same study showed that iron-deficiency anemia was associated with significantly lower energy. Iron intake early in pregnancy and the incidence of preterm labor were not increased with anemia from other causes. Iron has an important metabolic effect on function of coenzymes. The association between anemia and preterm labor might be caused by iron deficiency rather than anemia per se. In our study anemia was included as a confounding variable in the logistic regression model, but the OR for delivery before 32 weeks' gestation in anemic women was not statistically significant (OR 0.943, CI 0.814, 1.087).
Women under 18 years old were more likely than older women to have chest or urinary tract infections, despite controlling for delivery by cesarean and BMI. Those are clinical diagnoses and there is some heterogeneity between hospitals and doctors in their diagnostic criteria. Only speculation on the reasons why younger women had more chest and urinary tract infections is possible. One hypothesis is that pregnant women under 18 have reduced resistance to infection. Whether that is caused by immaturity of the immune system, more pronounced physiologic immunosup-pression of pregnancy, poorer nutritional status, or some other mechanism is not known.
Unlike other studies7–9 we found no association between young maternal age and low birth weight. Our findings might be different because we analyzed low birth weight in terms of small for gestational age infants and controlled for infant sex. We also included the additional confounding variables of smoking, pre-eclampsia, and diabetes in the logistic regression model. We found that women under age 18 years were at no greater risk of stillbirth than women 18–34 years old.
The ORs for many of the reported outcomes in this study were low. Those data showed that pregnancy in women less than 18 years old was generally associated with small changes in obstetric risk, with the exception of preterm labor. It is important to emphasize that because many confounding variables were included in the logistic regression models, the clinical significance of those results refers specifically to age-related risk of adverse obstetric outcomes, and that other factors such as smoking and ethnic origin were controlled for. Those data are equally applicable to pregnant white women 17 years old who smoke as they are to black-Caribbean 17-year-old women who do not smoke. It would be theoretically possible, using Bayesian analysis, to combine our data with data that described risks associated with smoking or ethnic origin to produce a more individualized risk score for particular women.
Although age-related obstetric risks for women less than 18 years old are low, we believe that pregnancy is associated with considerable socioeconomic problems. It might lead to social exclusion and can reduce educational, career, and economic prospects.1 The consequences of teenage pregnancy can be detrimental to the health of women and children. Society also pays the cost of teenage pregnancy in welfare support for young mothers caught in a poverty trap.1
2. Konje JC, Palmer A, Watson A, Hay DM, Imrie A, Ewings P. Early teenage pregnancies in Hull. Br J Obstet Gynaecol 1992;99:969–73.
3. McGrew M, Shore W. The problem of teenage pregnancy. J Fam Prac 1991;32:17–21.
4. Goldberg GL, Craig CJ. Obstetric complications in adolescent pregnancies. S Afr Med J 1983;64:863–4.
5. Osbourne GK, Howat RC, Jordan MM. The obstetric outcome of teenage pregnancy. Br J Obstet Gynaecol 1981;88:215–21.
6. Bradford JA, Giles WB. Teenage pregnancy in western Sydney. Aust N Z J Obstet Gynaecol 1989;29:1–4.
7. Fraser AM, Brockert JE, Ward RH. Association of young maternal age with adverse reproductive outcomes. N Engl J Med 1995;332:1113–7.
8. Brown HL, Fan YD, Gonsoulin WJ. Obstetric complications in young teenagers. South Med J 1991;84:46–8,64.
9. Lao TT, Ho LF. Obstetric outcome of teenage pregnancies. Hum Reprod 1998;13:3228–32.
10. Cleary R, Beard RW, Coles J, Devlin HB, Hopkins A, Roberts S, et al. The quality of routinely collected maternity data. Br J Obstet Gynaecol 1994;101:1042–7.
11. Cleary R, Beard R, Coles J, Devlin B, Hopkins A, Schumacher D, et al. Comparative hospital databases: Value for management and quality. Qual Health Care 1994;3:3–10.
12. Draper ES, Manktelow B, Field DJ, James D. Prediction of survival for preterm births by weight and gestational age: Retrospective population based study. BMJ 1999;319:1093–7.
13. Hollingsworth DR, Felice M. Teenage pregnancy: A multiracial sociologic problem. Am J Obstet Gynecol 1986;155:741–6.
14. Reichman NE, Pagnini DL. Maternal age and birth outcomes: Data from New Jersey. Fam Plann Perspect 1997;29:268–72,295.
15. Bluestein D, Rutledge CM. Determinants of delayed pregnancy testing among adolescents. J Fam Pract 1992;35:406–10.
16. Stevens Simon C, McAnarney ER, Roghmann KJ, Forbes GB. Composition of gestational weight gain in adolescent pregnancy. J Matern Fetal Med 1997;6:79–86.
17. Frisancho AR, Matos J, Flegel P. Maternal nutritional status and adolescent pregnancy outcome. Am J Clin Nutr 1983;38:739–46.
18. Hays PM, Cruikshank DP, Dunn LJ. Plasma volume determination in normal and preeclamptic pregnancies. Am J Obstet Gynecol 1985;151:958–66.
19. Silver HM, Seebeck M, Carlson R. Comparison of total blood volume in normal, preeclamptic, and nonproteinuric gestational hypertensive pregnancy by simultaneous measurement of red blood cell and plasma volumes. Am J Obstet Gynecol 1998;179:87–93.
20. Zuckerman BS, Walker DK, Frank DA, Chase C, Hamburg B. Adolescent pregnancy: Biobehavioral determinants of outcome. J Pediatr 1984;105:857–63.
21. Olausson PM, Cnattingius S, Goldenberg RL. Determinants of poor pregnancy outcomes among teenagers in Sweden. Obstet Gynecol 1997;89:451–7.
22. Heffner LJ, Sherman CB, Speizer FE, Weiss ST. Clinical and environmental predictors of preterm labor. Obstet Gynecol 1993; 81:750–7.
23. Shah NR, Bracken MB. A systematic review and meta-analysis of prospective studies on the association between maternal cigarette smoking and preterm delivery. Am J Obstet Gynecol 2000;182:465–72.
24. Lamont R, Fisk N. The role of infection in the pathogenesis of of preterm labour. In: Studd JWW, ed. Progress in obstetrics and gynaecology. Edinburgh: Churchill Livingstone; 1993:135–58.
25. Scholl TO, Hediger ML, Fischer RL, Shearer JW. Anemia vs iron deficiency: Increased risk of preterm delivery in a prospective study. Am J Clin Nutr 1992;55:985–8.