National Academies Press: OpenBook

Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices (1987)

Chapter: 5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature

« Previous: 4 Teenage Pregnancy And Its Resolution
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 93
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 94
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 95
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 96
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 97
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 98
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 99
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 100
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 101
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 102
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 103
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 104
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 105
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 106
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 107
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 108
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 109
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 110
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 111
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 112
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 113
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 114
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 115
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 116
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 117
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 118
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 119
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 120
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 121
Suggested Citation:"5 The Health and Medical Consequences of Adolescent Sexuality and Pregnancy: A Review of Literature." National Research Council. 1987. Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. Washington, DC: The National Academies Press. doi: 10.17226/946.
×
Page 122

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Consequences of Early Sexual and Fertility Behavior CHAPTER 5 THE HEALTH AND MEDICAL CONSEQUENCE S OF ADOLESCENT SEXUALITY AND PREGNANCY: A REVIEW OF THE LITERATURE Donna M . St rob ino INTRODUCTION Dur ing the early 1970s, much of the concern about adolescent sexual behavior centered on the adverse social, economic and health consequences of early childbearing. As rates of sexual experience rose among adolescents throughout the decade, attention turned to parallel increases in use of induced abortion to terminate pregnancy and in rates of sexually transmitted diseases. This chapter discusses the effects of these increases on the health of the adolescent, her future reproduction and the health of her offspring, as well as the health consequences of adolescent pregnancy. This review of the literature is limited to those consequences of sexual behavior--sexually transmitted diseases, induced abortion, and birth--that are most likely to affect the physical well-being and future reproductive health of the adolescent population. Even still, it represents a major undertaking involving a vast literature, especially with regard to sexually transmitted diseases. The review of the literature on the health consequences of induced abortion and adolescent birth is a comprehensive compilation of recent studies con- ducted in the United States. For sexually transmitted diseases, a comprehensive review of the literature is presented only for studies of the prevalence of sexually transmitted diseases (STD) among adoles- cents and young adults. A complete review of the literature on the sequelae of STDs was beyond the scope of this chapter. In addition, because the focus of the review is reproductive health, the literature on STDs will include primarily studies of women. Several recommendations for future research are made here based on the review presented below. First, further research is needed to de- termine whether the elevated risk of STDs among adolescents is due to an increased biological susceptibility to these infections or to a pre- ponderance of other risk factors among young sexually active women. In particular, the role of patterns of sexual behavior among adolescents, such as unplanned sexual encounters, in increasing their risk needs elaboration. Secondly, age differences in the risk of sequelae of STDs, other than pelvic inflammatory disease, need further study as 93

94 well as age differences in the risk of some of the more recently publi- cized STDs, such as genital herpes. Thirdly, the risk of subsequent unfavorable outcomes of pregnancy following more than 2 induced abor- tions during adolescence remains an area for which few studies have been reported. Fourthly, the factors that place the adolescent at increased risk of unfavorable birth outcomes should be specified in future research. Fifthly, maternal age differences in the rate of fetal growth need further clarification. Finally, the effect of a young age of childbearing on the health of offspring during the early years of life has received little careful scientific scrutiny. In particular, the documented social disadvantages of adolescent child- bearing need to be studied in relation to the health of the children of adolescent parents. SEXUALLY TRANSMI TTED DI SEASES Several sexually transmitted diseases (STD) have received consider- able attention in recent literature. Their prevalence among teenagers and sequelae will be discussed below. They include genital infections caused by Neisseria gonorrhoeae, Chlamydia trachomatis and-herpes simplex virus. Increases in the rate of these STDs have generally paralleled the rise in rates of sexual experience among adolescents. A brief discussion of three other infections will also be included since they may affect the health of the adolescent or her offspring. Estimates of the prevalence of sexually transmitted diseases among adolescents vary depending on the source of data for the estimates. Sources generally include reported diseases (where the STD is report- able), surveys of visits to office-based practices, data on patients attending sexually transmitted disease clinics, and data on patients attending clinics or other health facilities. The data on reported diseases are limited by differences in the completeness of reporting of diseases for public and private health care sources and, thus, biases in the estimates of rates for individuals who are more or less likely to use public clinics. Survey data are constrained by lack of validation of diagnoses. Data from STD clinics or other health facili- ties provide information on isolation rates for sexually transmitted microorganisms and permit study of the risk of infection in relation to patient characteristics. Two major disadvantage of these sources of data are patient selection bias and differences between studies in isolation rates depending on the extent of symptoms among the patients studied. The review of the literature on STDs among adolescents is presented with these limitations in mind. Trends and Risk Factors for Gonorrhea Infections Among Teenage Women Starting in the mid-1960s, the number of reported cases of gonorrhea rose dramatically among teenagers, particularly among women.

95 It reached a peak at approximately 276,000 cases in 1979 and has de- clined since then (Mascola et al., 1983) about 243,000 cases in 1982 (CDC, 1983b). The rate of reported cases of gonorrhea in 1982 was 1425 cases per 100,000 women aged 15-19 years and 71 cases per 100,000 women aged 10-14 years. Among males, the respective rates were 980 and 23 per 100,000 population. Although in 1965 the rate of reported cases of gonorrhea was greater for males than females aged 15-19 years, the female rate surpassed the male rate in 1973 and has remained higher since then. The rate of reported cases of gonorrhea is especially high for non- white women aged 15-19 years, exceeding the white rate by almost ten- fold in 1979. Among women under 15 years, the nonwhite rate exceeds the white rate by more than tenfold. Although nonwhites are more likely to utilize public clinics than whites where reporting of gonorrhea cases is more complete (Barnes and Holmes, 1984), the large disparity in the rates between nonwhite and white teenagers is unlikely to be due solely to differences in the reporting of cases. Between 1967 and 1975, the rise in the rate of reported cases of gonorrhea was greatest among 15-19 year olds of all sex and race groups (Zaidi et al., 1983~. Since 1976, the overall rate of reported cases has declined in the United States (Par ra and Cates, 1985), while it has remained stable among teenage women. Indeed in 1983, the rate was highest among this age group (CDC, 1984b). Moreover, Bell and Holmes (1984) found that the rates of reported cases of gonorrhea among sexually experienced women declined steadily with age and were espe- cially high for very young teenagers. Their estimates of sexually experienced women were derived from data from the 1971 and 1976 surveys of Zelnik et al. The number of sexually experienced women over age 20 was overestimated by an unknown magnitude, and their rates of gonorrhea were accordingly underestimated by an unknown magnitude. The data on reported cases of gonorrhea may be compromised by age differentials in the reporting of cases, particularly since younger women are more likely to use public clinics where reporting is more complete (Bell and Holmes, 19847. Additional errors in reporting of gonorrhea may result from the lack of uniform criteria to diagnose gonorrhea, and from the difficulty in diagnosing gonococcal infections or invasive gonococcal disease in women (Barnes and Holmes, 19841. Apart from the limitations of the data, several explanations have been offered for the rise among teenagers in the number and rate of reported cases of gonorrhea between 1965 and 1975. The first is an increase in the numbers of teenagers and young adults in the population (the groups with highest rates of gonorrhea) as a result of the coming of sexual age of the baby boom population (Aral et al., 1983; Cates, 1984; Mascola et al., 1983; Zaidi et al., 1983~. This would influence the number of cases. A second explanation is the rise in sexual activity among teenagers, coupled with an earlier initiation of sexual activity and the use of contraceptives that are ineffective in pre- venting lower genital tract infections (Bell, 1983; Mascola et al.,

96 19831. Bell (1983) suggests that part of the rise in gonorrhea rates may be artifactual, related to improved diagnosis of Neisseria gonorrhoeae with the introduction of Thayer-Martin culture media in 1964 and to improved screening of cases, especially among women, through the National Gonorrhea Control Program begun in 1973. Cates (1984) notes that in the absence of efforts to control the spread of sexually transmitted disease (STD), the size of the population at risk will determine the magnitude of the problem. Apart from the obvious link with the size of the population at risk, i.e., sexually active, there is no direct scientific evidence to support these explanations. Whether the higher rates of gonorrhea among teenage women are a result of social and behavioral factors or an increased biological susceptibility remains unclear. There is recent indirect evidence to support the notion of greater biological susceptibility to Neisseria gonorrhoeae among adolescent women (Bell and Holmes, 1984; Washington et al., 1985; McGregor, 1985; Cates and Rash, 1985; Bell and Hein, 1984~. The columnar epithelium is more likely to be located at the porto vaginalis of the cervix among adolescents than among older women (Ostergard, 1977) and thus more exposed to the outside world (McGregor, 1985~. Neisseria gonorrhoeae as well as Chlamydia trachomatis appear to have a predilection for this columnar epithelial tissue (Washington et al., 1985; McGregor, 1985~. Another biologic factor is more specu- lative and is related to the unchallenged immune system of the adoles- cent who has not been previously exposed to sexually transmitted microorganisms "Washington et al., 1985; Bell and Holmes, 1984~. Within the teenage and young adult population, the prevalence of lower genital tract infections with Neisseria gonorrhoeae has been estimated recently in clinical studies of presumably sexually active women (Bowie et al., 1981; Chacko and Lovchik, 1984; Saltz et al., 1983; Anglin et al., 1981; Fraser et al., 1983; Shafer et al., 1984; Wiesmeier et al., 1983~. These studies provide estimates of recovery rates of Neisseria gonorrhoeae from 0 to 12 percent among women whose endocervix was cultured during a pelvic examination. The variation in the estimates of the prevalence of gonorrhea may arise from differences in the extent to which the women were currently sexually active and the extent to which the pelvic examination was performed because of symptoms of lower genital tract infection. Moreover, it may also result from differences in the socioeconomic characteristics of the samples studied. For example, in two of the three studies with a rate of recovery of N. gonorrhea exceeding 10 percent (Chacko and Lovahik, 1984; Fraser et al., 1983), the samples included predominantly indigent patients; in the third study, the sample was not described (Anglin et al., 1981~. In the four studies with a recovery rate of N. gonorrhoeae of less than 5 percent, part (Saltz et al., 1983; Shafer et al., 1983) or the majority (Bowie et al., 1981; Wiesmeier et al., 1983) of the study patients were from middle income or working class families. The prevalence of N. gonorrhoeae infections among adolescents and young adults may be influenced by demographic characteristics of the

97 population as well as by patterns of sexual activity. As noted above, rates of reported cases of gonorrhea are greater among black than white women. Shafer et al. (1983) also reported higher percentages of black women with positive cultures for N. gonorrhoeae than white. Kinghorn and his associates (1982) found a predominance of blacks among young adult British men and women with repeated cases of gonorrhea. Fraser et al. (1983) found no increased risk of gonorrhea among blacks. There is only limited evidence of a relationship between socioeconomic status and gonorrhea rates among adolescents (Ekstrom, 1970) or young adults (Kinghorn et al., 1982~. In a recent review of the epidemiology of gonorrhea, Barnes and Holmes (1984) note possible reasons for the high rates among black women. Persons with ABO blood group B appear to be more susceptible to gonococcal infections; this blood group is more frequently found among blacks than whites. Strains of N. gonorrhoeae that produce asymptomatic infections are found more frequently among whites than blacks. Sexual behavior, as measured by number of partners and age at first coitus, appears to be associated with gonorrhea! infections. In a study of patients attending STD clinics in Sacramento during the summer of 1971, Darrow (1975) found that the rate of gonococcal infections rose with increasing numbers of sexual partners, up to four, and then declined. However, 66 percent of the patients reported only one sexual partner in the past month. Fulford et al. (1983) found a direct rela- tionship between number of lifetime sexual partners and gonorrhea in- fections among male attenders of an STD clinic in England, while Ekstrom (1970) reported a relationship between a young age at first coitus and gonorrhea for 18-19 year old boys in Copenhagen. A related variable, use of oral contraceptives, does not appear to be associated with gonorrhea (Fraser et al., 1983; Darrow, 1975~. Darrow also found no relationship between use of the condom and gonorrhea. But many other studies have found a relationship with the use of contraceptives, especially barriers. Sequelae of Gonococcal Infections Pelvic inflammatory disease (PID) is the most severe complication of lower genital tract infections in women (Cates, 1984), resulting from an ascending spread of infection of the lower genital tract (Westrom, 19801. Although it has been held for a number of years that Neisseria gonorrhoeae is a common cause of PID, recent evidence sug- gests that it has a multifactorial microbial etiology (Thompson and Washington, 1983; Westrom, 1980~. Since PID can be caused by a number of microbial agents, but these agents have seldom been specifically detected in studies, it is difficult to determine the incidence of gonococcal PID or the factors exclusively associated with gonococcal PID. Barnes and Holmes (1984) indicate that PID may occur in 10 to 20 percent of women with gonorrhea.

98 PID is a vague term, frequently used to refer to salpingitis, in- flammation of the uterine tubes. Accurate diagnosis is problematic based on clinical criteria since they are not always confirmed by laproscopy (Westrom, 1980~. Many salpingitis patients have atypical signs and symptoms and some have none, especially those with infections caused by Chlamydia trachomatis (Westrom, 1980~. Moreover, while signs and symptoms are generally used to define PID in clinical studies, it is commonly defined in survey data by one or more International Classi- fication of Disease (ICDA) codes, some of which are noninfectious dis- orders (Westrom, 1980~. An analysis of data from the Hospital Discharge Survey (HDS) indi- cates that rates of hospitalization for PID rose between 1975 and 1981 among white women aged 15-24. In 1979-81, women aged 15-24 years had the highest rate, i.e., 100,000 women, of hospitalization for PID, sur- passing women aged 25-34 years for whom the rate was highest in 1975- 78. Among nonwhite women aged 15-24 years, rates of hospitalization remained stable between 1975 and 1981. Nonwhite women had higher rates of hospitalization during the entire period, but because of the rise in rates for young white women, the ratio of nonwhite to white rates de- clined from 1975 to 1981 (Washington et al., 1984~. The higher rates of PID among black women with gonorrhea may be explained by a high proportion of gonococcal infections caused by strains more likely to produce PID (Holmes et al., 1980~. While rates of hospitalization increased for young white women in the mid-1970s, the rate of visits for PID to office-based physicians declined, as reported from the National Disease Therapeutic Index (NDTI) and the National Ambulatory Medical Care (NAMC) Survey (CDC, 1980~. Most of the decline occurred among nonwhite women (Cates, 1984), but age-specific visit rates have not been reported. Washington et al. (1984) suggest that the rise in hospitalizations and the con- comitant decline in office visits for PID may be due to a greater con- cern for its consequences and, thus, a lower threshold of physicians for hospitalizing women with PID. When rates of hospitalization for PID are estimated for sexually experienced women, the rates decline exponentially with age (Bell and Holmes, 1984~. Similarly, West rom (1980) estimated that the risk of acquiring salpingitis in a sexually active girl was 1 in 8 for 15 year olds, 1 in 10 for 16 year olds but 1 in 80 for women aged 24. In con- trast, Chacko and Lovchik (1984) found the risk of PID, as measured by clinical signs, rose from 3 percent among 13-14 year olds to 13 per- cent for 17-18 year old sexually active women in a predominantly black, urban sample. Westrom (1980) argues that promiscuity may be the reason for the high risk of salpingitis among 15-16 year olds; among 15 year olds with coital experience, 50 percent reported four or more sexual partners in the past year compared with 20 percent among 18 year olds. A possible biological reason for the higher risk among young women is related to

99 the age differences in location of cervical columnar epithelium dis- cussed above. The Light transparent phenotype of N. gonorrhoeae is associated with infection of the fallopian tubes. It has been found to be more copious in the cervix of young women than older women, attaching itself to the columnar epithelium better than the less virulent opaque phenotype (Bell and Holmes, 1984; Bell and Hein, 1984~. The IUD has been implicated as a possible causative agent in PID. The risk of PID is about 2 to 4 times greater for IUD users than non- users, regardless of the study design or the study site (Senanayake and Kramer, 1980; Lee et al., 1983~. This risk is especially high for the Dalkon Shield (no longer available and recalled by FDA order) (Lee et al., 1983~. Few U.S. teenagers use IUDs and it is generally not recommended for women in this age group, especially if they have not had children. The risk of PID, on the other hand, is reduced with use of oral contraceptives {Senanayke and Kramer, 1980) and with use of barrier methods (Kelaghan et al., 1982~. Sterile inflammation in the endometrium and uterine tubes following IUD insertion and possible mechanical damage to the endometrial epithelium are two mechanisms hypothesized by which the IUD facilitates the ascent of microbes through the uterus to the fallopian tubes. Explanation for the pro- tective effect of oral contraceptives on PID are related to the pos- sible effect of steriods on the density of cervical mucus or uterine muscular activity (Senanayake and Kramer, 1980~. The rising rates of hospitalization for PID among young U.S. women are important because of their possible association with rising infer- tility and ectopic pregnancy rates. Between 1965 and 1976, the per- centage of infertile couples with the wife aged 25-29 rose in the United States. Among blacks, the increase was especially pronounced among couples with wives aged 20-24 for whom the percentage infertile rose from 3.3 percent in 1965 to 15 percent in 1976 (Aral and Holmes, 1984~. Between 1970 and 1978, the ectopic pregnancy rate rose by more than twofold. Again, the rise was greatest among black women. Maternal death-to-case rates from ectopic pregnancy are more than 3 times greater for black women than for white women (Rubin et al. 1983~. Tubal damage following PID is estimated to be associated with 30 to 40 percent of cases of female infertility and in 40 to 50 percent of ectopic pregnancies (Westrom, 1980~. However, determining a direct causal link between PID and its possible sequelae is difficult. Wes- trom (1980) found that infertility because of tubal occlusion occurred in 15.2 percent of women treated earlier for laproscopically verified salpingitis who later exposed themselves to the risk of pregnancy. The percentage with tubal occlusion infertility was twice as great for women aged 25-34 as for women aged 15-24, but age differences were found only for women with one infection. Infertility varied directly with the number of infections, reaching a peak of 54 percent among women with three or more infections. It also was directly associated with severity of infection among those with only one infection. The

100 fertility prognosis for women under 25 was better with gonococcal- associated salpingitis than with nongonococcal salpingitis. Using the same data, Westrom (1980) studied the relationship of acute salpingitis with ectopic pregnancy. The ratio of ectopic to intrauterine pregnancy was 1:16 in the first pregnancy after salpingi- tis in 1968-74 and 1:24 in 1960-67. Among healthy control women (not described), the ratio was 1:147. Westrom estimated that women in the post-PID state accounted for about one-quarter of the rise in ectopic pregnancies in Lund, Sweden between 1960 and 1979 in women aged 20-29 years. Prevalence and Trends in Chlamydial Infections Chlamydial infections of the lower genital tract have surpassed gonococcal infections as the most prevalent STD among U.S. women and are a common STD among adolescents (Anglin et al., 1981; Cates, 1984; Fraser et al., 1983; Hare and Thin, 1983; Saltz et al., 1983; Schachter et al., 1975; Shafer et al., 1984; Thompson and Washington, 19831. Un- like gonorrhea, however, infections caused by Chlamydia trachomatis are not reportable conditions. Thus, most estimates of the prevalence of chlamydial infections are derived from clinical studies or reports of nonspecific lower genital tract infections in women (Hare and Thin, 1983) or of nongonococcal urethritis (NGU) in men (Cates, 1984~; both have been linked with C. trachomatis (Hare and Thin, 19831. The prevalence of chlamydial infections of the lower genital tract has been estimated to be between 7 and 23 percent among young women cultured during a pelvic examination. In all but one study the preva- lence of chlan~rdial infections was greater than the prevalence of gonococcal infections. Like gonorrhea, differences in recovery rates of C. trachomatis are likely due to variations in patient character- istics. For example, Shafer et al. (1984) noted that in adults the rate of chlamydial infection varies in adults from 4 to 8 percent among asymptomatic women to over 20 percent in women with symptoms of lower genital tract infection. There appears to be an increased risk of chlamydial infections among teenage women. Hobson et al. (1980) found that the degree of infection with C. trachomatis, as measured by the number of inclusions per coverslip of McCoy culture, was greater among women under 20 than older women. However, their findings did not suggest that this age effect was due to greater exposure of cervical columnar epithelial tissue (cervical ectopy) among the young women. They speculated that it may have an immunological basis. Harrison et al. (1983) reported a significant correlation between a positive culture for C. trachomatis and an age less than 24 years among pregnant low and middle income, white and Hispanic women. Within the teenage group, there does not appear to be an association of chlamydial infection with age (Anglin et al., 1981; Fraser et al., 1983; Wiesmeier et al., 1984) or gyne- cologic age (Shafer et al.~.

101 The evidence regarding an association between a positive culture for C. trachomatis and the demographic characteristics of the adoles- cent is conflicting. While Fraser et al. (1983) found no association with race, Shafer et al. (1984) reported a higher percentage of black females with a positive culture for C. trachomatis than whites or Hispanics. Shafer et al. found no relationship of socioeconomic status with a positive culture. On the other hand, Harrison et al. (1984) reported a significant correlation between a positive culture for C. trachomatis and maternal occupation and maternal education among pregnant women; their sample included older women as well as adoles- cents, although the reported relationship remained significant when adjustments were made for age differences in education and occupation. Sexual behavior, as measured by a young age at first intercourse (Shafer et al., 1984) and multiple recent sexual partners (Chacko and Lovahik, 1984; Wiesmeier et al., 1984), has been associated with in- creased rates of recovery of C. trachomatis, although Fraser et al. (1984) found no association between recovery of C. trachomatis and number of lifetime or recent sexual partners. There appears to be an increased risk of chlamydial infections among adolescents using oral contraceptives (Shafer et al., 1984; Fraser et al., 1983~. Moreover, Hob son et al. (1980) found an increased degree of infection with C. trachomatis among women using oral contraceptives, regardless of whether they had cervical ectopy; cervical ectopy was positively re- lated to the degree of infection. Shafer et al. (1984) and Fraser et al. (1983) have suggested that hormone-induced increases in cervical columnar epithelial tissue among oral contraceptive users may be the reason for their increased risk, but the work of Hobson et al. (1980) only partially supports their speculation. Fraser et al. (1983) also suggested that it may be due to more promiscuity among oral contracep- tive users but gave no data to support this speculation. Chacko and Lovahik (1984) did not find an association between oral contraceptive use and the prevalence of Chlamydial genital infections. Secuelae of Chlamydial Infections There is considerable evidence to suggest that infections with C. trachomatis are frequently asymptomatic in females (Fraser et al., 1983; Wiesmeier et al., 1984) or have nonspecific symptoms (Shafer et al., 1983~. Because of the asymptomatic nature of these infections, their sequalae become increasingly important in evaluating the extent to which they impact on the future reproductive health of the adoles- cent. In a review of studies of chlamydial infections of the pelvic region, Thompson and Washington (1983) reported that during the 1970s C. trachomatis was isolated from between 15 and 70 percent of cases of acute salpingitis in Scandinavian countries, but in only 0 to 10 per- cent of cases in the United States. The lower recovery rates in the U.S. may be due to an inability to utilize minute biopsies of tubal epithelium (Thompson and Washington, 1983~. Among women under 25 years in Lund, Sweden, Westrom (1980) found that over one-half of PID cases were caused by C. trachomatis.

102 The magnitude of infertility that can be directly attributed to C. trachomatis is unknown (Thompson and Washington, 1983~. Although re- sults of studies in which cultures were taken from the fallopian tubes during laproscopy are inconsistent, recent reports have shown a rela- tionship between elevated levels of antichlamydial antibodies and tubal factor infertility (Gibson et al., 1984; Thompson and Washington, 1983~. However, many of the infertile patients with elevated anti- chlamydial antibodies have reported no history of PID (Gibson et al., 1984; Thompson and Washington, 1983) , suggesting undetected infections in these women. The evidence of a link between chlamydial salpingitis or PID and ec topic pregnancy is indirect. For example, Thompson and Washington (1983) show a parallel rise in cases of PID and ectopic pregnancy rates between 1966 and 1973 as indirect evidence of the link between the two. Although Westrom (1980) shows more direct evidence of the association between salpingitis and ectopic pregnancy, the percentage of tubal pregnancies resulting from chlamydial induced tubal damage is unknown. C. trachomatis has also been implicated in conjunctivitis and an afebrile pneumonia syndrome in the newborn (Fraser et al., 1983; Schachter et al., 1979~. Estimates of the prevalence of C. trachomatis infections in pregnant women range from 7 to 27 percent in the U.S. (Thompson and Washington, 1983; Chacko and Lovahik, 1984~. The highest rates have been reported among young urban women (Cates, 1984; Harrison et al., 1983; Thompson and Washington, 1983~. Chacko and Lovahik (1984) reported a prevalence of 27 percent among pregnant teenagers aged 13-18. Schachter et al. (1979) estimated that given a cervical infection rate of 4 percent, there will be 14 cases of chlamydial conjunctivitis and 8 cases of chlamydial pneumonia per 1,000 live births. These estimates may be low for pregnant adolescents, given their high rates of chlamydial infections. There is considerable epidemiological evidence linking the number of sexual partners and a young age at first intercourse to an in- creased risk of cervical cancer (Hare and Thin, 1983; Rotkin, 1967; Schachter et al., 1982~. This evidence has led recent investigators to study the role of C. trachomatis in the development of cervical cancer (Hare and Thin, 1983; Schachter et al., 1982~. While it may play a role in a small proportion of cases of cervical cancer (Hare and Thin, 1983), Schachter et al. (1982) could not offer an explana- tion for its association with cervical cancer, particularly since the cell nucleus does not appear to be affected in chlamydial infections. Prevalence and Trends in Herpes Simplex Virus Infections of the Lower Genital Tract The widely held contention that herpes simplex virus (HSV) infec- tions of the lower genital tract increased during the 1970s is sup- ported by an analysis of data from the NDTI (Becker et al., 1985~.

103 The proportion of consultations for genital HSV infections in office- based fee-for-service practices rose from 30 per 100,000 consultations in 1966 to 336 in 1981 (Becker et al., 1985~. In contrast, the rate of consultations for oral herpes and ocular herpes infections rose by less than twofold between 1966 and 1979 (CDC, 1982~. In 1966, the number of visits to office-based practices for men outnumbered visits for women, but by 1980, visits by women for genital herpes surpassed those for men. The rise in the number of visits for genital herpes was very large for women aged 15-19, increasing from about 15,000 yearly visits in the beginning of the period to over 110,000 at the end (Becker et al., 19851. The NDTI data do not give estimates of the total occurrence of HSV infections in the United State since sources of health care such as HMOs, public clinics and hospital outpatient clinics are not included. Nevertheless, Sullivan-Bolyai et al. (1983) also reported a rise in the percentage of visits for genital herpes infections between 1976 and 1981 among patients attending STD clinics in King County, Wash- ington. An increased awareness and better diagnosis of HSV infections may explain same of the rise in genital HSV infections, but part of the rise occurred before intensive media campaigns about genital herpes (Becker et al., 1985~. Estimates of the prevalence of HSV genital infections are far more difficult to obtain. Depending on the sample population, they range from O to 6 (CDC, 1979; Josey et al., 1972; Schachter et al., 19751. Two studies have estimated the prevalence of genital HSV infections in a University sample. About 1 percent of the students who attended the Student Health Services (SHS) at UCLA in 1975 and 1976 had genital herpes infections, as defined by clinical criteria (Sumaya et al., 1980~. The mean age of the patients with a single first infection was similar to that for the general student population. The percentage with clinically diagnosed genital HSV infections was lower (0.6) among women who were treated for gynecological problems at the SHS at Penn State University in 1974 and 1975 (Kalinyak et al., 1977~. The per- centage of women with virologically confirmed infection was 0.3. There is little known about the risk factors associated with HSV genital infections, largely because descriptive studies of convenience samples of virologically confirmed cases of HSV infections are the primary source of this information (Nahmias et al., 1973~. Both cytological and virological evidence suggests that HSV genital infec- tions occur more frequently among lower socioeconomic groups (Josey, 1972; Baker and Amstey, 1983; Lancet, 1981; Rawls, 1971), although there is recent evidence that they also are common among higher socio- economic groups (Bierman, 1983; Knox, 1982~. The risk of experiencing recurrences has been reported to be greatest among higher socioeconomic groups (Lancet, 19811. Rawls et al. also found a younger mean age and a greater percentage of unmarried persons among patients with genital herpes infections than controls, but their control population was not well defined. Adler-Storthz et al. (1985) found that the frequency of

104 antibodies to HSV-2 rose with the number of sexual partners. HSV infections have also been associated with other sexually transmitted infections (Josey, 1972; Rawls, 1971; Sumaya, 1980~. The two strains of HSV, HSV-1 and HSV-2, have been commonly thought to infect different sites in the body, HSV-1 with a predilection for sites above the waist and HSV-2 for the genital tract (Baker and Amstey, 1983; Josey et al., 1972; Lancet, 19811. Recent studies have questioned this distinction; suggesting that about one-third of genital herpes infections are HSV-1 infections (Kalinyak et al., 1977; Lancet, 19811. Changing sexual patterns are generally given as the reason for the increasing percentage of HSV-1 genital infections (Amstey and Baker, 1983; Lancet, 1981~. Seque~lae of Genital Herpes Infections The importance of recent trends in genital herpes infections lies in two features of these infections. First, the clinic manifestations of genital herpes infection tend to recur (Baker and Amstey, 1983; Cates, 1984;:CDC, 1982; Knox, 19827. Secondly, genital herpes infec- tions have been associated with cervical cancer in the woman and with life-threatening infections in the neonate (Cates, 1984; CDC, 1982~. Smith (1983) has recently reviewed the evidence regarding a rela- tionship between HSV-2 genital infection and cervical cancer. The literature is vast and has taken multiple approaches. Smith (1983) concluded that the greatest support for an association between the two This that, given the multiple study approaches, HSV-2 has yet to be scientifically discredited. Genital HSV infection, both asymptomatic and symptomatic, have been associated with HSV infection in the newborn (Sullivan-Bolyai et al., 1983, Whitley et al. 1980a; Whitley et al., 1980b). Infection in the newborn may be disseminated or localized. Disseminated and localized central nervous system (CNS) disease accounts for about 75 percent of the cases of neonatal herpetic infections; their respective mortality rates are about 80 and 30 percent (Whitley et al., 1980a). Severe CNS damage is common among survivors (Whitley et al., 1980a). Sullivan- Bolysi et al. (1983) reported a rise in the rate of neonatal herpes infections in Kings County, Washington between 1965 and 1982, mirror- ing the rise in consultations for genital herpes. Whitley et al. (1980b) reported mothers delivering infants with herpetic infections to be young and nulliparous. Sullivan-Bolyai et al. (1983) also found them to be young, with a mean age of 23 years and a range of 17 to 34 years. However, whether these results imply an increased risk of neonatal herpes infections among the newborn of teenager mothers is unclear.

105 Other Sexually Transmitted Infections Among Teenage Women Three STDs, syphilis, human papillomavirus, and group B strepto- coccal infections, warrant brief description here. Although rates of syphilis had been declining for several decades through the mid-1970s, there was a rise in the number of cases reported to CDC in 1982 (CDC, 1983~. Rates of syphilis are much lower than rates of gonorrhea, and they do not show the preponderance of females in cases among teenagers (Fichtner et al., 1983~. Nevertheless, estimated syphilis rates vary inversely with age among sexually experienced women (Bell and Holmes, 1984~. Moreover, the number of reported cases of congenital syphilis has risen in the United States in recent years (Mascola et al., 1984~. While most of these cases can be prevented by adequate prenatal care and appropriate diagnosis, some cases result from incubating infections at the time of delivery. They are more frequent among young mothers than older mothers and among minority women (MaScola et al., 1984~. The prevalence of genital warts among teenagers has not been re- ported, although they are generally at least three times more ire- quent than HSV infections (Parra and Cates, 1985~. Of major concern with human papillomavirus (HPV) is the association of subtypes 16 and 18 with cervical cancer (Bernstein et al., 1985; Crum et al., 1983; Crum et al., 1984~. Indeed, Crum et al. (1984} contend that HPV 16 is virtually specific for squamous-cell neoplasms. Aral and Holmes (1984) reported data indicating a rise in the rate of genital warts among men and women in England and Wales between 1970 and 1980. Sexual transmission of group B streptococcal infections remains unclear. (Aral and Holmes, 19841. Yet, they have been recently im- plicated as the major cause of neonatal sepsis in a number of insti- tutions (Aral and Holmes, 1984~. Because of the emergence of group B strep as a cause of early sepsis, their occurrence among teenage mothers is a concern. THE EFFECT OF INDUCED ABORTION ON THE HEALTH OF THE ADOLESCENT AND HER FUTURE OFF SPRING In 1981, the number of induced abortions occurring to teenagers was close to 450,000. They represented 28 percent of all induced abor- tions in the United States in that year (Henshaw et al., 1985~. In sheer numbers alone, complications associated with induced abortion could have a major impact on the health of the adolescent and her sub- sequent offspring. This review will first discuss the literature com- paring complications following induced abortion between adolescent and older women. It will be followed by a review of the literature on the impact of induced abortion on subsequent fertility and pregnancy out- comes. Only recent U.S. studies or studies in other countries with induced abortion laws and practices similar to the U.S. will be in- cluded. Since very few studies compare outcomes by maternal age, the review must focus on studies of women of all ages. -

106 Variations in abortion practices by age warrant comment here as they may influence complications following induced abortion and subse- quent pregnancy outcomes. Teenagers are twice as likely as older women to obtain an abortion after 12 weeks gestation and the youngest teen- agers are most likely to delay seeking an abortion (Cates, 1980; Lang et al., 19831. Length of gestation is the most important determinant of abortion complications (Cates, 1980~. It largely determines the abortion method, and methods used later in pregnancy have elevated risks of complications (Cates et al., 1977~. Moreover within methods, complication rates vary directly with gestation (Cates et al., 1977~. In 1981, suction curettage was virtually the only method used in the United States for pregnancies under 13 weeks gestation. At 13-15 weeks and 16 weeks, dilatation and evacuation was the most common method used tHenshaw et al., 19851. In 1980, it replaced saline infusion as the most frequently used method at 16-20 weeks (Lang et al., 1983~. Teenagers are less likely to be represented among women with repeat abortions, the most probable reason because of less time exposed to the risk of pregnancy and, thus, abortion (Burnham, 1983~. Nevertheless, because they experience an induced abortion at a young age, they are at greater risk of a repeated abortion than older women by virtue of their greater length of time remaining for exposure to an unwanted pregnancy. Number of previous abortions has been associated with subsequent pregnancy outcomes (Maine, 1979), although the evidence regarding increased risk of subsequent pregnancy problems is not con- firmatory (Hogue et al., 19821. Complications Following Induced Abortions Among Teenagers The most comprehensive analysis of complications of abortions among teenagers has been reported by Cates and his colleagues (1983) using data from the Joint Program for the Study of Abortion (JPSA) and the CDC's surveillance of abortion-related mortality. The JPSA included data on morbidity for over 80,000 legal abortions in 32 institutions in 1971-1975, and for over 84,000 legal abortions in 13 institutions in 1975-1978. Over 50,000 abortions to teenagers were included. The data for the two periods were analyzed separately because of slightly different terms used in the studies. An overall complication rate, including fifteen major complications, as well rates for fever Item perature of 38 degrees C or above for 3 or more days), hemorrhage requiring transfusion, and unintended abdominal surgery were studied by abortion method. These latter complications accounted for 80 per- cent of the major complications between 1971 and 1975. The overall risk of major complications following abortions by suction curettage at 12 weeks of gestation or earlier was similar for teenage and older women, 3 per 1000 procedures in 1971-75 and between 1 and 2 per 1000 in 1975-78. Complication rates for fever and for uterine hemorrhage did not vary with age, but rates of transfusion for hemorrhage were lower among teenagers. Teenagers also had lower rates

107 of uterine perforation, particularly if they had no previous term pregnancy. On the other hand, teenagers aged 17 or less had higher rates of cervical injury, even when adjustment was made for other factors related to cervical trauma such as method of dilatation, gravidity, level of medical training, use of general anesthesia, and gestation. Complication rates associated with abortions by dilatation and evacuation or saline administration were lower for teenagers than any other age group; rates generally varied directly with age and were greater following saline administration than following dilatation and evacuation. The rate of febrile morbidity was lower among teenagers than older women in 1971-75. Transfusion rates were lower among teen- agers for all years. Following abortions from saline administration, teenagers had lower rates of hemorrhage requiring transfusion, unin- tended major surgery and retained products of conception. They had lower failure rates and shorter administration to abortion times as well (Cates et al., 1983~. The risk of dying from legally induced abortion was also lower among teenagers than older women between 1972 and 1978; death rates varied directly with age. Within gestation, teenagers had the lowest death rates. They had slightly higher death rates from infectious com- plications than older women but slightly lower death rates from embolic events or from reactions to anesthesia (Cates et al., 1983~. Grimes et al. (1981), however, reported that the death-to-case rate from sepsis after legal abortion was almost six times greater among teen- agers than among women over 20 years. In contrast, death rates from hemorrhage were considerably higher among women over 20 than among teenagers (Cates et al., 1983; Grimes et al., 1983~. Two potential sources of bias in the analysis of JPSA data were possible selection of more skilled physicians by teenagers and less frequent reporting of complications by teenagers (Cates et al., 1983~. Cates and his associates concluded that like any surgical procedure, induced abortion was not without risks, but these risks were not greater for the teenager than for the older women and in some circuit stances, they were less. The one exception, also noted by Tyler (1983), was the greater risk of cervical trauma among teenagers. Cervical trauma following induced abortion may increase the risk of unfavorable outcomes in subsequent pregnancies because of the presumed association between cervical trauma and the occurrence of incompetent cervix (Tyler, 1983; Hogue et al., 1982~. Despite their thorough analysis of the risk of induced abortion Cates et al. (1983) did not report the overall morbidity rate fol- lowing abortion for teenagers. The extent to which the greater delay in obtaining abortion among teenagers increases their overall risk of morbidity, regardless of their lower rates of morbidity following second trimester induced abortions, cannot be evaluated from their analysis. Harman (1981), using data on abortions performed at a large

108 university hospital in Winnipeg, Manitoba between 1974 and 1979, re- ported that teenagers had a 35 percent greater chance of experiencing at least one complication than older women. This higher risk was due to the predominance of teenagers among women obtaining second trimester abortions. Similarly, Wadhera (1982) reported almost a twofold greater complication rate following induced abortion during 1975-1980 for Canadian teenage women than older women. Again, one important reason for their high complication rate was the later gestations at which teenagers obtained abortions. While the analysis of Canadian data suggests a higher complication rate among teenagers because of delay in seeking abortion, the Canadian experience with abortion is not di- rectly comparable with that in the U.S. The Canadian policies with regard to access to abortion are far more restrictive than those in the United States, and the complication rates reported by Wadhera (1982) are considerably higher than those reported by Cates et al. (1983~. The Effect of Induced Abortion on Outcomes of Subsequent Pregnancies The method of pregnancy termination (WHO, 1979) and whether or not abortion is legal are two important factors that may influence the impact of induced abortion on subsequent pregnancies (Hogue et al., 1982~. Moreover, differences between women who seek induced abortion and those with no abortion experience may also influence the risk of unfavorable pregnancy outcomes. In this review, only studies that in- cluded women who obtained an induced abortion after 1970 will be dis- cussed; the degree and extent of cervical dilation during evacuation procedures has been reduced since 1970 (Hogue et al., 19821. In addi- tion, the vast majority of studies reported here were conducted in the United States. Two studies from other countries (Meirik and Bergstrom, 1983; Van Der Slikke and Treffers, 1978) will also be discussed since they include women whose abortions were known to have been performed by vacuum aspiration with minimal dilatation. Several studies from other countries have been excluded because induced abortion was illegal in the study country (Daring and Emanuel, 1975; Papevangelou et al., 1973; Pantelakis et al., 19737; the method of abortion was not reported (Ratter and Beischer, 1979; Richardson and Dixon, 1976~; or data on subsequent pregnancy outcomes were obtained from respondent reports to a mailed questionnaire and no subsequent follow-up or validation was performed (Dalaker et al., 19791. One additional methodological issue is important in evaluating the studies reported here. There is considerable debate regarding what constitutes an appropriate comparison group for women with a prior in- duced abortion, particularly with regard to gravidity and parity. Both cannot be controlled in a comparison group unless the group experienced a prior spontaneous abortion tHogue et al., 1982~; the risk of subse- quent unfavorable pregnancy outcomes is increased for women experi- encing a prior spontaneous abortion. On the other hand, unfavorable pregnancy outcomes, such as preterm or low birth weight infants, occur

109 more frequently among primiparous women than among multiparas. The conclusions derived from a given study depend on whether the comparison group was experiencing a first or second pregnancy and whether, if a second pregnancy, the first resulted in a spontaneous abortion or live birth (Hogue et al., 1982~. There appears to be no association between a history of one (Daring and Emanuel, 1977; Madore et al., 1981; Schoenbaum et al., 1980) or more (Daring and Emanuel, 1977) induced abortion and the birth weight of infants born in subsequent pregnancies. Linn et al. (1983) initial- ly found a direct relationship between LBW rates and number of prior induced abortions, but it was not signifioant when adjustment was made for differences by abortion status in age, race, welfare status, parity and cigarette smoking. Daling and Emanuel (1977) found no differences in birth weights of infants of primiparous women under age 20 with and without a prior induced abortion, but their numbers were quite small (65 women with a prior induced abortion and 65 controls similar on religion, marital status and payment for medical expenses). There also appears to be no association between a history of in- duced abortion (Daring and Emanuel, 1977; Madore et al., 1983) or number of induced abortions (Daring and Emanuel, 1977; Meirik and Bergstrom, 1983) and preterm deliveries, before 37 weeks of gestation. Linn et al. (1983) found similar results for preterm births similar to those for LBW infants. Daling and Emanuel (1977) reported no associa- tion between a history of induced abortion and subsequent preterm de- liveries in women under 20 years. Van Der Slikke and Treffers (1978) found an increased percentage of deliveries between 17 and 31 weeks among primiparous women with at least one induced abortion. However, they argued that their findings were applicable only to the Nether- lands. If there is an effect of induced abortion on subsequent pregnancy outcomes, the evidence appears to be most convincing regarding second trimester spontaneous abortions among nulliparous women. Harlap et al. (1979) found a twofold increased risk of second trimester spontaneous abortions among nulliparous women with a prior induced abortion com- pared with nulliparous women with no prior induced abortion. Among women with two or more prior induced abortions the risk was increased by threefold. There was no increased risk of second trimester spon- taneous abortions among multiparous women with a history of induced abortion. Madore et al. (1981) also reported more second trimester losses among women in their induced abortion group (nulliparous and Primiparous). On the other hand, Chung et al. (1982b) found no association between history or number of induced abortions and subse- quent second or third trimester spontaneous abortions. A major limi- tation of their study is the use of fetal death records to measure outcomes of subsequent pregnancies; these records are notoriously underreported, underreporting varying inversely with length of ges- tation. Underreporting is a greater problem in studying first trimester spontaneous abortions, and because of it, they are not discussed here.

110 Two case-control studies show conflicting results regarding the risk of spontaneous abortions for women with a history of induced abor- tions. Kline et al. {1978) found no association between second tri- mester spontaneous abortions and an induced abortion for any prior pregnancy or for the pregnancy immediately preceding the study preg- nancy. Levin et. al. (1980) also reported no significant association between second trimester abortions and one prior induced abortion. However, the odds of a second trimester abortion increased with in- creasing number of prior induced abortions. Selection bias in the choice of controls cannot be ruled out as an explanation for the findings in either of these studies. There is mixed evidence regarding whether the number of induced abortions is related to subsequent ectopic pregnancies. Madore et al. (1981) and Chung et al. (1982a) found no association between history or number of induced abortions and subsequent ectopic pregnancy. Chung et al. reported higher ectopic pregnancy rates for women with infection complications and retained products of conception following induced abortion, but their numbers were too small to make valid inferences. A more efficient design for studying this rare event is the case-control study of Levin et al. (1982~. They found no significant association between a subsequent ectopic pregnancy and one induced abortion, but the odds of an ec topic pregnancy was 2.6 for women with two or more prior induced abortions. Even in this study, the number of women with ec topic pregnancies was small (851. In a review of the literature on abortions (Hogue et al. (1982) concluded that even if there is an ele- vated risk of ectopic pregnancy following an induced abortion, it is not large enough to yield consistent results in relatively large case- control studies. There is no evidence in prospective studies of a relationship between a history of induced abortion and subsequent offspring with congenital anomalies (Linn et al., 1983; Madore et al., 1983~. In a case-control study, Bracken and Holford (1979) reported an increased risk of congenital malformations following induced abortion, but only among infants born to black women under 30 years of age. Congenital malformations were not adequately defined in any of these studies, particularly with regard to the ages at which the diagnoses were made. Linn et al. (1983) also studied a variety of complications of preg- nancy in relation to a history of induced abortion. First trimester bleeding was the only complication of pregnancy significantly associa- ted with only one prior induced abortion. Abnormal presentations and premature rupture of the membrane in addition to first trimester bleeding were significantly associated with two or more prior induced abortions. Finally, induced abortion does not appear to affect subsequent fer- tility. In a prospective study with repeated follow-up for 30 months, Stubblefield et al. (1984) found no significant difference in pregnancy rates for women with a prior induced abortion and control women with a

111 previous pregnancy. The one exception was at 9 months postabortion, when the abortion group reported greater use of contraceptives. At the time of enrollment, sample women were aged 18-20 years. In two case control studies, Daling and Emanuel (1981) and Daling et al. (1985) found no significant association between secondary infertility and a history of one or more induced abortions. In both studies, the cases of secondary infertility were limited to women who sought medical care for their problem. The major limitation of the studies reported here are that with the exception of the work of Daling and Emanuel (1977), where the numbers are small, and the research of Bracken and Holford (1979) and Stubble- field et al. (1984), the effect of a prior induced abortion on subse- quent pregnancy outcomes has not been evaluated for young age groups. Given the increased risk of cervical injury as a complication of abor- tion for teenagers and its possible association with subsequent un- favorable pregnancy outcomes, more research on the delayed effects of induced abortion among teenagers is needed. The sheer numbers of abortions performed on teenagers annually in the U.S. also justify continued research on the subject. THE MEDICAL AND HEALTH CONSEQUENCES OF ADOLE SCENT BI RTH The literature concerned with the medical and health consequences of adolescent childbearing presents mixed conclusions ranging from statements of the biological advantages of youthful birth to conclu- sions of an increased incidence of a variety of complications of chilo- bearing and unfavorable birth outcomes (Morris, 1981~. Most recently, there is increasing evidence to support the conclusion that adoles- cence, per se, may not be a risk factor for poor health outcomes of the mother or her offspring, but rather, that the preponderance of other risk factors such as low socioeconomic status, poor prenatal care and primiparity is the reason for their poor outcomes. This review will compile the evidence published since 1960 to evaluate this general conclusion. The literature abounds with descriptive studies of a series of adolescent patients, but few of these studies will be reviewed here. Rather, this review focuses on studies that included an adolescent sample and a comparison sample of older pregnant women. There are a few investigations reported here that explored explanations for the increased incidence of complications and unfavorable outcomes among adolescent mothers only. The following complications of pregnancy: pregnancy induced hypertension; anemia; cesarean deliveries; and con- tracted pelvis, and unfavorable pregnancy outcomes: low birth weight; prematurity; perinatal mortality; neonatal mortality; infant mortality; neonatal morbidity; and morbidity during the first year of life, are compared for adolescents and older mothers. Maternal mortality was not valuated in the vast majority of the studies because it occurred very infrequently among the women studied.

112 Pregnancy Induced Hypertension Pregnancy induced hypertension (PIH)is one of the most frequently and consistently reported complication of adolescent childbearing. An increased incidence of PIH among adolescents appears, in part, to be explained by a preponderance of women who are black, experiencing their first pregnancy and who receive inadequate prenatal care among adoles- cents, when compared with older women. Yet, understanding of the etiology of PIH among adolescents as well as older women remains limited. There is no uniform definition of PIH or preeclampsia used in studies of adolescent pregnancy and frequently no definitions are given. Terminology varies, including pregnancy induced hypertension and preeclampsia in more recent studies, and toxemia in older studies. When a definition of preeclampsia is not given in a study, it is assumed that it refers to a clinical diagnosis noted in the medical record with the inherent problems of differences in the degree to which providers diagnose or record the condition. The number of cases of eclampsia were too few to evaluate in the adolescent studies. In those studies where PIH was investigated as a complication of pregnancy, the percentage of adolescents with PIH ranged from a low of 2.5 percent among whites studied by Israel and Wouteraz (1963) to a high of 34.0 percent among the predominantly black adolescents studied by Duenholter et al. (1975~. In the majority of the studies, the per- centage of adolescents with PIH ranged from 7 to 17 percent. Moreover, the percentage of women with PIH was generally higher among adolescent women than the older control group. Differences in the definition of PIH may influence the magnitude of the differences in the percentage of women with PIH between adolescent and control women, especially if underlying hypertension is included. Yet, it is more likely that they are influenced by differences in con- founding variables between the adolescent and older control group, particularly race and parity. Among the studies where there was no attempt to adjust for differences between adolescent and older women on race or parity, the adolescents had higher percentages of PIH (Claman and Bell, 1964; Semmens, 1965; Utian, 1967; Jovanovic, 1972~. The one exception was Briggs et al. (1962), a study including a very unusual comparison sample. When rates of PIH were compared in studies by race, differences by age were markedly reduced (Battaglia et al., 1963; Haskin, 1963; Israel and Wouterez, 1963; Spellacy et al., 1978~. Indeed, Haskin (1963) showed greater between race differences within age than between age differences within race. Parity also appears to be related to age differences in rates of PIH in that most of the studies in which primigravidous adolescents and primigravidous older women were compared showed only small dif- ferences between the two age groups (Bochner, 1962; Hassans and Falls, 1964; Poma, 1981; Osbourne et al., 1981; Graham, 1981; Lee and Walters,

113 1983~. The results of Coates (1970), Hulka and Schaaf (1964), and Duenholter et al. (1975) are exceptions to this general relationship. Coates' control sample included white women while the adolescent sample did not. Despite similar study designs, Hulka and Schaaf (1964) found a higher percentage of older mothers with PIH, while Duenholter et al. (1975) found a higher percentage among adolescent mothers. Both studies were limited by their definitions of PIH, and this limitation may account for their opposite results. One reason for the particularly high rates of PIH in the study of Duenholter et al. (1975) may be their very young sample, adolescents aged 14 years or less. Battaglia et al. (1963), similarly, showed higher percentages with PIH in this age group compared with nonwhite primigravidas aged 15-19 years. The reasons for the elevated rates of PIH among the youngest adolescents are unclear, but they do not appear to be due to physical immaturity. Erkan et al. (1971) reported that the percentage of adolescents with preeclampsia (undefined) was greater among adolescents with a postmenarcheal age (PMA) less than 24 months, as measured by the difference between the given age at menarche and the date of last menstrual period (LMP), than among those with a PMA of 24 months or more, but the difference was not statistically significant. Hollingsworth et al. (1981) reported that hypertension (undefined) was related to race, but it was not associated with gynecologic age; they gave no definition for gynecologic age. In two additional studies, inadequate prenatal care was investi- gated as a reason for the reported high rates of pregnancy complica- tions among adolescents. McAnarney and her associates (1978) studied the Rochester Adolescent Maternity Program (RAMP), a program providing more prenatal visits and psychosocial services than two other study sites. They found no significant differences in the frequency of hypertension or preeclampsia among adolescents receiving services in the three sites, but their sample size was very small. Jorgensen (1972), on the other hand, found a marked drop in the percentage of adolescents with preeclampsia at Pennsylvania Hospital after introduc- tion of an adolescent clinic; the clinic was developed on the premise that intensive prenatal care and health education would lower the risk of adolescent pregnancy. It is impossible to evaluate this before/ after comparison since the characteristics of adolescents using the hospital before the clinic was introduced were not described and the availability of the clinic to all adolescents was not discussed. Based on the studies reported here, it appears that much of the risk of preeclampsia or PIH associated with adolescent pregnancy may be due to a predominance of primigravidas and blacks among pregnant adolescents. There may be a somewhat elevated risk of PIH among the youngest adolescents, but this elevated risk does not appear to be a result of their physical immaturity.

114 Anemia During Pregnancy Anemia is a frequently cited complication of adolescent child- bearing. Yet, the evidence supporting this conclusion is very limited. Like PIH, anemia, is not uniformly defined in the adolescent pregnancy literature, although most investigators measure it by low hemoglobin or hematocrit levels. In many studies, these levels are not taken at a uniform time during pregnancy and for convenience are frequently mea- sured at the time of registration for prenatal care. When comparing adolescents and older women, hematocrit or hemoglobin levels may vary simply because adolescents are more likely to begin prenatal care in the second trimester and hemoglobin levels drop in the normal pregnancy during this trimester. The percentage of adolescents with anemia ranged from 0.8 to 19.7 percent in the studies reviewed. Even when the studies in which no definition of anemia was given (Bochner, 1962; Utian, 1967; and Poma, 1981) or a conservative definition was used (Spellacy et al., 1978) are excluded from comparisons, the range in estimates of adolescents with anemia during pregnancy is still large (3.6 to 19.7 percent). In most of the studies, there was no difference in the percentage of women with anemia between adolescent and control women, even when no adjustment was made for difference by age in race or other poten- tial confounding variables. Israel and Woutersz (1963) and Osbourne et al. (1981) were the only two investigators to report a signifi- cantly higher percentage of adolescent than older women with anemia. In contrast, although Hulka and Schaaf (1964) found no differences by age in antepartum anemia, the older control mothers more frequently had postpartum anemia than the adolescent mothers. In a study of the incidence of folacin and iron deficiency among predominantly black, low income women, Bailey et al. (1980) found no age differences in mean hematocrit levels and serum and red blood cell levels of folacin taken at first prenatal visit. Serum iron concentra- tion was significantly lower among the older women than among the adolescents and there was a trend for transferrin saturation to also be lower in older women. The authors suggested that iron depletion may be less in the adolescents because of their comparably shorter menstrual histories. Israel and Wouteraz (1963) noted a greater percentage of women with anemia among both nonwhite teenagers and controls than among all women delivering in 10 collaborative study institutions. Similarly, McGanity et al. (1969) found signif icantly lower mean hemoglobin, hematocrit and corpuscular hemoglobin concentration among black adoles- cents than among whites. Their mean plasma iron levels, however, were similar. McGanity et al. did not report the trimester during which these measures were taken.

115 In two final studies, the role of a special program for adoles- cents in reducing complications of pregnancy was evaluated. Neeson and her associates (1983) reported that on admission for labor, adoles- cents attending the Young Womens Clinic at University of California had hematocrits similar to older controls using the regular obstetrics clinic, but that teenagers using the regular clinic had significantly lower hematocrits. Similarly, Jorgensen (1972), reported that adoles- cents attending an adolescent clinic at the University of Pennsylvania were less likely to have anemia, as measured by a hemoglobin under 10.5 at delivery, than adolescents receiving care at the hospital before the clinic's inception. The possibility of self-selection of the adoles- cents into the special program cannot be ruled out as an explanation for the results of this latter study. On balance, the literature does not suggest that anemia is a major complication of adolescent pregnancy. A high frequency of anemia among black adolescents may be cause for some concern by health care pro- viders. Most studies are limited by small sample sizes, failure to measure anemia uniformly during pregnancy and almost complete absence of adjustment for confounding variables. With few exceptions (Bailey et al., 1980), these studies were not designed to specifically investi- gate differences in anemia between adolescent and older mothers and, accordingly, do not address normal physiologic changes in pregnancy. Cesarean Sections and Cephalopelvic Disproportion Cesarean delivery rates have been studied as a possible complica- tion of adolescent childbearing, largely because of their indication as a method of delivery for women with cephalopelvic disproportion or contracted pelvis. Because of their young age, adolescents have been presumed to be more likely than older women to have cephalopelvic disproportion. Among the studies reviewed, the majority show no difference in cesarean delivery rates between the adolescent and older control mothers. Nevertheless, several studies have shown the advantage to the adolescent (Briggs et al., 1962; Israel and Woutersz, 1963; Poma, 1981; Osborne et al., 1981; Nee son et al., 1983; Lee and Walters, 1983), but many of these studies did not exclude repeat cesareans or multiparous women from the control sample. Even among the more recent studies in which only primiparous women were studied (Duenholter et al., 1975; Poma, 1981; Osbourne et al., 1981; Graham, 1981; Lee and Walters, 1983), there were no differences in cesarean delivery rates for the two age groups or lower rates were reported for the adoles- cent. Excluding the study by Poma because it included only women with x-ray pelvimetry, estimates of cesarean delivery rates in the most recent U.S. studies (Duenholter et al., 1975; Spellacy et al., 1978; Neeson et al., 1983; Graham, 1981) range from 9.2 to 14.7 percent.

116 The incidence of contracted pelvis among adolescents and older con- trol women was evaluated in some studies as a complication of preg- nancy. With the exception of the work of Hulka and Schaaf (1964), these studies indicate an increased incidence of contracted pelvis among the adolescent group, particularly when x-ray pelvimetry was per- formed to diagnose the contracted pelvis. Duenholter et al. (1975) found significant differences only for contraction of the pelvic inlet of less than 85 percent of the normal area. Similarly, Poma (1982) re- ported that on admission for labor and delivery, there were no signi- ficant differences in the average diameter of the pelvic midplane and outlet between adolescent and older mothers, but the average pelvic inlet antereoposterior and transverse diameters were significantly lower for the adolescent. He argued that these differences were mini- mal and unlikely to be clinically meaningful. Cephalopelvic dispropor- tion was the most frequent indication for cesarean births in his study sample, and it was given with similar frequency in both age groups (Poma, 1981~. Birth Weight and Gestational Age of Infants Born to Adolescent Mothers Birth weight has received the most careful scientific scrutiny of all outcomes of adolescent childbearing. Many studies have reported a higher proportion of low birth weight (LBW) and lower mean birth weights of infants born to adolescents than those born to an older control sample. Maternal age differences in LBW rates and in mean birth weights may in part be explained by racial differences by age; black infants in general weigh less than white infants and pregnant adolescents are more likely to be black than older mothers. More re- cent studies suggest that not only race but other risk factors such as low socioeconomic status, lower prepregnancy weights and later initia- tion of prenatal care may also explain maternal age differences in birth weight. Many of the early studies of adolescent pregnancy interchanged the terms prematurity and LBW. Currently, a premature infant refers to an infant born before week 37 of gestation and a LBW infant to one weigh- ing 2500 grams or less at birth. Table 10 provides estimates of the percentage of LBW infants in 15 studies of pregnant adolescents; in the early studies it is assumed that prematurity referred to LBW. The percentage ranges from 6.3 to 23.3, with the majority of the studies reporting between 10 and 20 percent. The percentage of LBW infants is almost two times higher among nonwhite than white adoles- cents. Many of the studies of obstetrical populations in the 1950s and 1960s showed significantly greater percentages of LBW infants among adolescents than older mothers (Hulka and Schaaf, 1964; Battaglia et al., 1963; Utian, 1967; and Jovanovic, 1972~. Hulka and Schaaf and Battaglia et al. reported significant differences in spite of simi

118 In three recent studies, Horon et al. (1983), Rothenberg and Varga (1981) and Zuckerman et al. (1983) found no significant difference in mean birth weights between infants born to adolescent mothers and in- fants born to older mothers, when adjustment was made for age differ- ences in a number of confounding variables. Horon et al. found similar birth weights in spite of a higher percentage of clinic patients, un- married women and women living in less affluent neighborhoods, and lower pregnancy weights, shorter statures, shorter gestations and later initiation of prenatal care among the adolescents than among the older controls. In the study by Zuckerman et al. (1983), the adolescent mothers began prenatal care later, had lower prepregnancy weights, had more gonorrhea! infections during pregnancy, were more likely to be black, had fewer religious affiliations, were less likely to smoke, use psychoactive drugs or drink alcohol during pregnancy and were less likely to have had an x-ray during pregnancy than the older mothers. Rothenberg and Varga (1981) did not describe the differences between their adolescent and control sample. All investigators concluded that factors other than a young age were associated with low birth weight among pregnant adolescents. The conclusion that a young age, per se, is not the reason for poor outcomes of adolescent pregnancies is supported only in part by studies of the relationship of gynecologic age and measures of fetal growth. Hollingsworth and her associates (1981) found no association of gynecologic age with birth weight, length of the infant, head cir- cumference of the infant, or gestation. On the other hand, Erkan (1971) and Zlatnik and Burmeister (1977) found higher LBW rates among offspring of adolescents with a low gynecologic age than among off- spring of more mature adolescents. The measure of gynecologic age used by Zlatnik and Burmeister was confounded with gestation since it was defined as the length of time between age at menarche and age at delivery. The possible role of adequate or intensive prenatal care in re- ducing LBW rates among adolescents has received considerable attention. Neeson and her colleagues (1983), Jorgenson (1972), Zackler et al. (1969) and Felice et al. (1981) reported lower percentages of LBW in- fants born to adolescents attending an intensive prenatal care program than infants born to adolescents receiving routine prenatal care. Zackler et al. (1969) found the advantage of the program to be greatest for black adolescents. Neeson et al. also found birth weights to be similar for the adolescents in the Special Program and older women attending the regular obstetrics clinic. Although Felice and her associates (1981) raised the possibility of self-selection of teens into the special care program as a reason for their better outcomes, they dismissed it because of their study design. It cannot be ruled out as a possible explanation in their study or those of Jorgenson (1972) and Zackler et al. (1969~. McAnarney and her associates (1978) did not find differences in mean birth weights or length of gestation of offspring of adolescents served in three different sites.

rarities between the adolescent and control group on race and parity. On the other hand, Briggs (1962), Bochner (1962), Coates (1970) and Semmens (1965) reported no significant differences in the proportion of LBW infants between adolescents and their controls. In more recent studies (Duenholter et al., 1975; Spellacy et al., 1978; Poma, 1981; Osbourne et al., 1981; Graham, 1981; Lee and Walters, 1983), only Spellacy and his colleagues found a significantly greater percentage of LBW infants born to adolescent mothers than to older mothers. The elevated percentage for infants of adolescent mothers held when only black women were compared. Most other studies have shown greater variation in LBW rates between whites and nonwhites within age groups than between adolescent and older mothers within racial groups (Graham, 1981; Haskin, 1963; Israel and Wouterez, 1963~. The risk of LBW among adolescents appears to increase with the number of previous births of the mother (Graham, 1981; Israel and Deutschberger, 1964; Jekel et al., 1975~. Graham (1981) also re- ported a higher percentage of preterm births among multiparous adolescents than older multiparas. A number of investigators (Hulka and Schaaf, 1964; Israel and Deutschberger, 1964; Utian, 1967) have also reported a greater percentage of pregnancies ending before 37 weeks of gestation among adolescents than older controls. On the other hand, Poma (1981) did not find any significant differences by age in length of pregnancy. Both Poma (1981) and Hulka and Schaff (1964) only studied primigravidas. Garn and Petzold (1983) found no relationship between maternal age and mean length of gestation using data from the National Collaborative Perinatal Project, but the per- centage of pregnancies ending before 38 weeks varied inversely with age, as did the percentage of LBW infants. While these results sup- ported previous studies of age variations in birth weight and mea- sures of maturity of infants at birth, they provided little explana- tion for these variations. Naeye {1981) also analyzed the data from the Collaborative Peri- natal Study to determine if young teenage mothers have slower rates of fetal growth than older mothers. He included black singleton infants whose mothers were between 10 and 32 years of age and whose infants were born between 38 and 44 weeks of pregnancy. Among under weight and normal weight mothers, mean birth weights were significantly lower for the adolescent than for the older mothers across most weight gain groups. Although Naeye concluded that fetuses of most 10-16 year olds grow more slowly than those of older mothers, this conclusion was pre- sumptous in that births before 38 weeks gestation were not included. Horon et al. (1983) have recently shown that birth weights of pre- mature infants born to young adolescents were significantly greater than birth weights of premature infants born to older mothers. Be- cause their premature infants weighed more than those of the older mother, the infants of the adolescents in their sample, on average, had similar birth weights to the infants of older mothers despite a greater frequency of premature births among the adolescents.

119 The elevated risk of delivering a LBW infant among adolescents appears to be due to a preponderance of other risk factors among the adolescent. Exactly which other risk factors remain in question, how- ever, as does possible differences in the growth rate of the fetuses of adolescents and older mothers. Both areas warrant future research. The LBW infant, in addition to experiencing an increased risk of neo- natal death, is more likely than the normal weight infant to succumb to a variety of problems (Shapiro et al., 1980~. Prevention of LBW infants among adolescents and at all maternal ages is a major public health priority. Perinatal, Neonatal and Infant Mortality Among the Offspring of Adolescent Mothers Maternal age differences in perinatal, neonatal and infant mortal- ity have been investigated in a number of the studies including cross- sectional studies of births occurring in a geographic area. These latter studies have generally used linked birth and death certificates as the source of data. These data have the major advantage of being available for all births in a geographic area but the major disadvan- tage of containing limited information. Israel and Woutersz (1963), Osbourne et al. (1981~; Graham et al. (1981) and Duenholter et al. (1975) found no significant differences in perinatal mortality rates between infants born to adolescent mothers and infants born to older mothers. Duenholter et al. (1975) also found no significant maternal age differences in neonatal mortality rates. On the other hand, Battaglia et al. (1963) reported a higher perinatal mortality rate for nonwhite infants born to mothers under 15 than for all nonwhite infants born in Baltimore City, but the higher rate was due almost entirely to the greater frequency of infants weighing under 1000 grams among the young mothers. Similarly, Hulka and Schaaf (1964) noted an elevated neonatal mortality rate among in- fants born to adolescents that was due to LBW infants. Israel and Wouterez (1963) also found greater neonatal mortality rates among off- spring of teenage mothers, but differences in rates were less by age than by race. In several additional studies, the number of perinatal or neonatal deaths was too small to evaluate maternal age differences (Utian, 1967; Coates, 1970; Spellacy et al., 1978; Poma, 1981; Lee and Walters, 1983~. Dott and Fort (1975; 1976) found a J-shaped relationship between maternal age and perinatal mortality among 1972 Louisiana births. The risk of a neonate' death was greater for the offspring of the very young and elderly mothers than for the offspring of women in their twenties. Varva and Querec (1973) reported a U-shaped relationship between age of mother and mortality rates for deaths occurring in the first day of life, the first week of life and from day 7 to 27, using data from the 1960 U.S. birth cohort study. Postneonatal death rates were highest for infants of teenagers. In Louisiana, they were high

120 est for infants born to mothers aged 10-14 years, followed by infants of 15-19 year old mothers (Dots and Fort, 1975~. Within the teenage births, infant mortality rates were higher for infants of married women than for infants of unmarried women. Varva and Querec (1973) found differences in the relationship of maternal age to infant mortality by birth order. As with LBW rates, infant mortality rates were elevated for second and higher order births to teenage mothers. Jekel et al. (1975) also reported a higher peri- natal mortality rate for higher pregnancy order births among adoles- cents. In contrast to several of the above studies, the classic study of Shah and Abbey (1971) indicated no maternal age effect on neonatal mortality when adjustment was made for birth weight differences by age. However, next to birth weight, maternal age and parity showed the strongest relationship with postneonatal mortality. Moreover, adjust- ment for parity increased the risk of postneonatal mortality for in- fants born to mothers less than 20. Thus, the results of Shah and Abbey showed that maternal age was an important predictor of infant death only in the postneonatal period. It appears that much of the risk of elevated neonatal mortality rates among infants born to adolescent mothers is due to the increased proportion of LBW infants born to these mothers. Neonatal and First-Year Morbidity Among Offspring of Adolescent Mothers The relationship of maternal age with morbidity in the newborn or the infant during the first year of life has been much less exten- sively studied than birth weight or mortality. There is some evidence of an elevated risk of morbidity among infants of young mothers, but the variation by maternal age is less than reported for LBW rates. Garn and Petzold (1983), Osbourne et al. (1981) and Poma (1981) found no consistent relationship between maternal age and the percent- age of infants with low Apgar scores. Zuckerman et al. (1983) and Finkelstein et al. (1982) noted lower one minute Apgar scores among in- fants of adolescent mothers than among infants of older mothers, but five minute scores were similar. Neeson et al. (1983) reported similar results for infants of teenagers receiving care in the regular obstet- rics clinic compared with infants of older mothers receiving care there or infants of teenagers in a special program. Only Rothenberg and Varga (1981) found lower five minute Appear scores for infants of adolescent mothers, but the association of Apgar scores with maternal age was weak. There appears to be little association between a young age of childbearing and other measures of neonatal morbidity, including conjunctivitis, omphalitis, septicemia, skin infections, convulsive disorders, intracranial hemorrhage, hyaline membrane disease, excessive weight loss, duration of nursery stay (Duenholter et al., 1975) and jaundice (Osbourne et al., 19811.

121 Morbidity during the first year of life has been investigated in relation to maternal age in a number of recent studies (Hendershot, 1979; Shapiro et al., 1980; Smolen et al., 1984; Rothenberg and Varga, 1981; Finkelstein et al., 1982~. In a study of 4,327 surviving infants born in 8 geographic regions in the United States in 1975, Shapiro et al. (1980) reported that infants of women under 18 years at the time of their birth had more significant illnesses during the first year and greater rates of postneonatal death, especially if they were LBW. Mor- bidity data were collected through a household interview. Finkelstein et al. (1981) also reported significantly more acute episodes of ill- ness among the offspring of white adolescents than the offspring of older white mothers. Hendershot (1979) reported more hospitalizations among infants of mothers under 20 than infants of mothers 20 or older among ever-married, primiparous U.S. women included in Cycle I of the 1973 National Survey of Family Growth. Rothenberg and Varga (1981) and Smolen et al. (1984), on the other hand, found no differences in reports of hospitalization or the need to see a physician regularly for a medical problem in their children for adolescent and older mothers. However, Rothenberg and Varga re- ported more burns and injurious conditions for infants of adolescent mothers than for infants of older mothers. Wicklund et al. (1984) also have recently reported higher rates of fatal infant accidents for the offspring of mothers under 20 than for those of mothers over twenty in North Carolina and Washington State. Maternal age differences in rates remained within race and education groups when standardized for parity. Smolen et al. (1984) noted a greater percentage of the infants of adolescent mothers below the~~fifth percentile for weight gain, while the infants of older mothers had significantly more clinic visits for medical problems. In contrast, Finkelstein et al. (1982) found no maternal age difference in mean heights or weights of children during the first two years of life. There appears to be only limited evidence of an increased risk of morbidity for infants of young mothers. This evidence suggests that accidental injuries may be more frequent among offspring of adolescent mothers. The differences in morbidity by maternal age are small even when reported in large cross-sectional studies (Shapiro et al., 1980; Hendershot, 1979~. Most other studies are limited by small sample sizes or sampling methods that may have resulted in a biased adoles- cent sample or an inappropriate control sample. CONCLUSI ONS This review of the literature indicates that rates of sexually transmitted diseases rose from adolescents during the 1970s, and for some infections, such as genital herpes or chlamydial infections, they could continue to climb. The risk of gonorrhea, syphilis, and

122 chlamydial infections is highest among the teenager, particularly when rates are estimated for sexually experienced women. Although the pos- sibility of an increased biological susceptibility to sexually trans- mitted infections has been related to age differences in the location of cervical columnar epithelial tissue, the role of age differences in patterns of sexual behavior in explaining elevated rates among adoles- cents remains to be explored. The risk of genital herpes and genital warts among adolescents has not been described and is an area for future research, especially given the devastating effects of herpetic infections in the newborn and the clear role of human papillomavirus in cervical cancer. N. gonorrhoeae and C. trachomatis have been impli- cated in pelvic inflammatory disease and its consequences of infertil- ity and ec topic pregnancy; both have been increasing recently among young black women. Complications following induced abortion are generally lower among adolescents than older women, regardless of the gestation at which the abortion was performed or the method used. Two exceptions are cervical injury and death-to-case rates from sepsis which are more frequent among teenagers. If there is an increased risk of unfavorable outcomes in pregnancies following an induced abortion, the risk is small. Even for second trimester abortions, where the evidence to date is most convinc ing, the increased r isk following induced abortion is at most twofold, if it exists at all. The r isk of second trimester abortion rises somewhat with increasing numbers of prior induced abortions. Much of the risk of subsequent unfavorable outcomes of pregnancy following induced abortion appears to be associated with differing characteristics of women with and without a history of induced abor- tion. Whether a young age at the time of the abortion confers an in- creased risk of subsequent unfavorable pregnancy outcomes has not been studied. Most recent research indicates that the elevated risk of poor pregnancy outcomes among adolescents is most likely explained by a preponderance of risk factors among young mothers. Although race, primiparity and poor prenatal care have been suggested as possible risk factors, research is still needed to specify the factors that are most likely to explain their increased risk of poor outcomes. High rates of perinatal and neonatal deaths have been associated with high LBW rates among adolescents. However, an elevated risk of postneonatal deaths among the offspring of adolescent mothers appears to be indepen- dent of age differences in birth weight. The research on early child- hood morbidity is too limited to determine whether this increased risk extends to morbidity as well.

Next: 6 Social and Economic Consequences of Teenage Childbearing »
Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices Get This Book
×
Buy Paperback | $130.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

More than 1 million teenage girls in the United States become pregnant each year; nearly half give birth. Why do these young people, who are hardly more than children themselves, become parents? The statistical appendices and working papers for the report Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing provide additional insight into the trends in and consequences of teenage sexual behavior.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!