Contraception in adolescence: a review. 2. Biomedical aspects.

Bnao,ses Bulletin ofthe WorldHealth Organization, 62 (2): 331-344 ( 1984) (3 World Health Organization 1984 Contraception in adolescence: a review 2. Biomedical aspects* ADELE D. HOFMANN 1 Although none of the currently available methods of contraception can be considered as idealfor use by adolescents, many come reasonably close. Thepresent article reviews the relative benefits and risks associated with each method. Oral contraceptives appear to bearfew specific risksfor adolescents. Pills containing 30-50 g of ethinylestradiol and 0.5-1.0 mg ofnorethisterone, or equivalents, offer a high degree ofprotection and are well tolerated. There is no evidence that the pill interferes with either pubertal growth or the achievement of regular ovulation. Risks of cardiovascular complications are lowerfor adolescents than for any other age group; however, changes in lipid profiles, which are thought to contribute to increased atherogenesis, may restrict total lifetime use of oral contraceptives. While the minimal compliance requirements of the intrauterine device (IUD) might suggest it as an appropriate method for adolescents, cramping and increased menstrual bloodflow often make it unacceptable. Expulsion rates are higher in nulligravidae than in parous females, although copper-wound or progestogen-impregnated devices are better retained. The increased risk ofpelvic inflammatory disease in unmarried adolescents using the IUD is ofsufficient magnitude to warrant the restriction ofthe device to those who have failed to use other methods. Depot medroxyprogesterone (injectable hormonal contraceptive) may be of use in particular groups of adolescents. Both the World Health Organization and the American Academy ofPediatrics recommend its use in selected groups ofadolescentsfor whom other methods are contraindicated. Oralprogestogen-containing "minipills" are not suitablefor teenagers, since compliance requirements are fairly rigid, and there is often excessive breakthrough bleeding, and a relatively low protection rate. All barrier methods have a significant placefor use by adolescents, although younger teenagers are often reluctant to insert the diaphragm. The condom, used alone or with spermicidal foam, has a particular advantage since it does not require a prescription, is widely available and suits the spontaneous, unplanned nature of coitus in this age group. Rhythm and otherforms of "natural" birth control are not suitablefor adolescents. They depend on regular ovulation and require extensive instruction, a high degree of motivation, and the ability to plan aheadfor coitus. None of thesefactors is characteristic of the teenage girl. An ideal contraceptive for adolescents would be lOOOo protective, wholly reversible, and medically safe with no health risks. It would also be easily * Part I of this review, covering psychosocial aspects, appeared in the Bulletin of the World Health Organization, 62 (1): 151-162 (1984). 1 Medical Director, Ambulatory Pediatrics, Children's Hospital of Orange County, PO Box 5700, Orange, CA 92667, USA. obtainable at low cost and be well within the young person's capacity to use properly; alternatively, a passive method with few compliance demands would be satisfactory. No currently available method meets all these requirements, but many come reasonably close and have a significant place in preventing teenage conceptions when judiciously prescribed with full appreciation of their relative benefits and risks as 4407 -331 A. D. HOFMANN measured against the hazards of adolescent preg- nancy. Maternal mortality rates for 15-19-year-old females in the United States of America, for example, are nearly ten times greater than mortality rates attri- butable to contraception (11.1 /100 000 births versus 1.2/100 000 users of oral contraceptives per year and 0.8/100 000 for users of the intrauterine device (1)). This disparity is even greater in developing couni ries (2). COMBINATION ORAL CONTRACEPTIVES Oral contraceptive pills containing various es- trogen-progestogen combinations constitute the pre- scription method most widely used by adolescents. Their advantages include a high degree of protection if taken regularly (Table 1), and dissociation from the coital act. Disadvantages include the need consciously to acknowledge and anticipate sexual intercourse (which requires a certain level of maturity), to obtain a prescription, and to take a pill every day. Adoles- cents with little motivation to use contraception tend to have low continuation rates. Even those who are highly motivated have difficulty with regular com- pliance if coitus is infrequent. Young people also tend to discontinue taking the pill on terminating a rela- tionship and then fail to resume it with a new partner. (All adolescents using the pill also need a back-up Table 1. Effectiveness of various contraceptive methods in theory and in practice' Theoretical Practical Method effectiveness effectiveness (%) (%)b Combination oral contraceptives 99.6-99.9 90.0-99.3 Progestogen-only oral contraceptives 97.5-97.7 90.0-100 Intrauterine device 97.0-99.0 94.4-99.0 Diaphragm 97.0-98.5 71.1-98.1 Condom 97.0-98.5 64.0-97.0 Spermicide 96.0-98.0 61.66-98.25 Condom plus 99.0+ 95.0 spermicide Periodic abstinence 90.0-95.0 53.0-99.7 Withdrawal 91.0 75.0-80.0 a Taken from: Tyrer, L. B. & Bradshaw, L. E. Barrier methods. Clinics in obstetrics and gynecology, 6: 39-55 (1979). b These rates primarily apply to adult women. The lower figures more accurately reflect effectiveness in adolescents. method such as the condom and/or spermicide to cover such exigencies.) Pills containing 30-50 Ag of ethinylestradiol and 0.5-1.0 mg of norethisterone (or equivalents) are most suitable for adolescents, the lower dosages being preferable. These are generally well tolerated with minimal side-effects and low risk rates while still affording high levels of protection provided that the pills are taken reasonably regularly (3). Frequent for- getting and "catch-up" doubling of low-dose pills more often results in breakthrough bleeding, and method failure will be more likely than with a 50 Ag estrogen dose. Variations in individual tolerance may also necessitate changes in formulation. Effect on growth It is well recognized that high doses of estrogen have the capacity to slow or arrest skeletal growth. This effect is thought to be mediated through the suppression of the bone-stimulating sulfation fac- tor, somatomedin, rather than through accelerated epiphyseal closure, an androgen-induced event in females as well as males (4). These observations have prompted concern about potential similar effects from the amount of estrogen in the pill. There is, however, no substantiated evidence for such effects. The clinical circumstances associated with suppression of growth by estrogen are significantly different from those encountered in persons using oral contraception. In the former case, the aim is solely to contain growth during the peri- menarcheal period in selected girls with excessive height prediction. A common regimen is daily oral administration of 0.25-0.5 mg of ethinylestradiol (or equivalent) together with a progestogen (0.5 mg of norethisterone) for the first 5 days of each month until a bone age of 15 years (990/0 of final height) is achieved. The total duration of treatment ranges between 1.5 and 2 years. The case histories of more than 950 children and adolescents treated in this manner have been de- scribed in the literature (5-12). Singularly high doses of estrogen (about 10 times the amount used for contraception) are required to achieve a significant reduction in final height. Lower doses, or similar doses given intermittently (i.e., for 3 out of every 4 weeks), have proportionately less effect (5). Treat- ment is also less effective if initiated after rather than before menarche (11). Thus the growth-suppressive effects of estrogen are directly dependent on dose and bone age at the time of initiation and would not be expected with the amount of estrogen in the pill administered postmenarcheally. Only one report (12) even suggests challenge to this conclusion. A mean reduction of 4.2 cm in predicted height was found among 18 girls taking oral contra- 332 BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE ceptives containing 0.01-0.05 mg of ethinylestradiol and 2 mg of norethisterone or norgestrel. However, they were all patients suffering from idiopathic scoliosis, and failure to achieve predicted height might have been due to residual spinal curvature. Furthermore, it is impossible to predict final height with great precision (10) and the estimated mean reduction in this group is close to the magnitude of prediction error. This study is of doubtful signifi- cance in the light of these two confounding factors and the absence of any corroborating reports. There is, then, no substantiated direct evidence that oral contraceptives containing 50 jAg or less of eth- inylestradiol or equivalent have any effect on growth in adolescents; indeed, there is considerable indirect evidence to the contrary. It may be presumed that this low dose, while sufficient to inhibit ovulation, is insufficient to suppress somatomedin production to below the level required for bone stimulation. It should also be noted that 99% of adult height has been achieved with a bone age of 15 years, at which time estrogen administration of any magnitude will have no significant effect. Most adolescents have reached this stage by two years after menarche and their coital debuts are unlikely to have taken place prior to this. Effect on maturation of the menstrual cycle A second issue of concern to adolescents relates to the potential inhibitory effects of oral contraception on the maturing hypothalamic-hypophyseal-gonadal axis and on achievement of regular ovulation (13). Vollman (14) demonstrated that cycle length is highly variable during the first gynaecological year (i.e., the first year after menarche), with a median length of 29 days and a 5th-95th percentile range of 17.8-76.5 days. Only in the seventh year do the menstrual cycles reach maximum stability with a median of 27.8 days and a 5th-95th percentile range of 19.2-42.5 days. Biphasic basal body temperature curves (presumed to indicate ovulation) occur in only 44% of cycles in the first gynaecological year with a linear increase to 93%o after 8 years. Apter (15) more precisely defined adolescent ovula- tory patterns by measuring serum progesterone on days 20-23 of the menstrual cycle; a rise of 2 ng/ml or more correlated well with the surge of luteinizing hormone accompanying ovulation. Of 138 cycles studied in adolescent girls of different gynaecological ages, 14% were ovulatory in the first year, 38%o in the second, 48%o and 50%o in years 3 and 4, and 64%o in the fifth year. This last figure compares favourably with the ovulatory frequency of 62%7o found by Met- calf (16) in women aged 20-24 years; only after the age of 30 years did the frequency exceed 90%. A number of reports have specifically examined the effects of estrogen administration on menstrual pat- terns in adolescents. Rey-Stocker et al. (17) studied hormonal profiles in 103 females of 5 gynaecological years or less at 3, 6, and 12 months after discontinuing oral contraceptives. Plasma estradiol and proges- terone levels, measured between days 20 and 25 of the menstrual cycle, rapidly and consistently approxi- mated levels found in gynaecological age-matched controls who had not used any contraception. Levels of follicle-stimulating hormone, luteinizing hor- mone, prolactin, and thyroid-stimulating hormone also rapidly returned to the control range, as did the frequency of ovulation, with rates of 50-70% in the study groups (divided according to the specific estrogen-progestogen formulation used) and 60%o among the controls. Results were similar for all types of contraceptive pill, although none contained more than 50 jig of ethinylestradiol or equivalent. Follow-up studies in "too tall" girls, treated with high doses of estrogen, found that nearly all subjects began menstruating within 2-3 months of stopping treatment. A few patients experienced somewhat ir- regular cycles for 7-13 months, but the number did not exceed what would be expected among females in their early postmenarcheal years under normal cir- cumstances. Only two patients failed to achieve regular menstruation within a year; one of these responded well to a short course of clomifene, leaving only a single patient who remained amenorrhoeic (6, 10). This case, however, cannot necessarily be ascribed to the effects of exogenous estrogen; one case of functional amenorrhoea among nearly 1000 pubertal girls is well within the general incidence of this condition in this age group. Both Bierich (8) and Hanker (9) carried these clinical observations one step further by conducting hormonal studies on a total of 18 "too tall" girls, aged 10-15 years, both during and after estrogen treatment. Luteinizing hormone, follicle-stimulating hormone and response to exogenous luteinizing hor- mone releasing factor (LH-RH test) were almost totally suppressed during and immediately after estrogen administration but returned to normal in all cases within 2-5 months. Indeed, one week after dis- continuation of treatment, endogenous estrogen had risen to the levels normally seen in menstruating adults in the midfollicular phase and previously elevated prolactin levels fell. The ultimate child-bearing capacity of these "too tall" young women is not yet known. Some have been lost to follow-up; others have not yet wished to become pregnant. At the time of reporting, however, 15 individuals are known either to be currently preg- nant or to have already delivered and at least 10 healthy offspring have been produced (5, 6, 10, 11). The weight of available evidence does not support the contention that estrogen administration of any 333 334 A. D. HOFMANN magnitude interferes with menstrual cycle maturation in normal adolescents. Post-pill amenorrhoea In adult women, menstruation normally resumes within 1-3 months of discontinuing oral contra- ceptives. In approximately 7007o of cases, the first cycle is ovulatory; 98% are ovulatory by the third cycle (18). Less than 1% of women remain amenor- rhoeic and half of these are subsequently found to have a clearly diagnosable cause unrelated to contra- ception, leaving only a small number of individuals with unexplained dysfunction (19). There is con- siderable controversy as to whether these cases are directly due to the contraceptive pill or result from a pre-existing problem; 8-55% of this group have a history of chronic oligomenorrhoea antedating use of oral contraceptives (20). Hull et al. (19) estimated a true relative risk for post-pill amenorrhoea of I per 1000 users, but others disagree with even this low incidence, finding little statistical support for any cause-and-effect relation- ship (21). There is no information on whether use of oral contraceptives during adolescence increases the prob- ability of post-pill amenorrhoea, but reported studies (18-21) have not included a disproportionate number of teenagers. Admittedly, this observation is com- promised by patient bias, since the main reason for seeking help for this problem is infertility, a concern which is unlikely to arise in adolescence. However, in the absence of data to the contrary, the observation of Ory et al. (22) - that post-pill infertility is at best an unfounded concern, at worst a modest problem. usually responsive to treatment - would appear to apply as much to adolescents as to adults. The primary concern associated with the use of oral contraceptives in the first few gynaecological years lies in the possibility of masking an underlying dys- function. This possibility, however, can be avoided in most instances if use of the pill is deferred until the normality or otherwise of the menstrual cycles can be assessed. This should not result in excessive delays because the onset of oligomenorrhoea usually occurs soon after menarche. Dramusic (23) found that in 71% of adolescents with prolonged secondary amen- orrhoea (unassociated with oral contraception) the onset had occurred during their first gynaecological year. Rodriguez-Rigau et al. (24) found that nearly 750o of adult women with chronic dysfunction dated the onset of this disorder back to their first two gy- naecological years. In contrast, only 10%o of normally menstruating controls had experienced significant oligomenorrhoea during the immediate postmen- archeal period. Cardiovascular complications It is now generally agreed that use of oral contra- ception carries an increased risk of myocardial infarc- tion, thrombotic stroke, haemorrhagic stroke, and venous thrombosis with and without pulmonary embolus (22, 25, 26). The major contributing factor is an estrogen-induced, reversible alteration in the blood clotting factors. Changes in blood lipid pro- files, particularly the suppression of high density lipo- proteins, may also contribute to increased risks for myocardial infarction by accelerating atherogenesis, although even here thrombosis appears to be the dominant cause (26). The actual degree of risk of cardiovascular disease directly associated with oral contraceptives varies according to the specific condition and the presence or absence of other contributing factors. For myocar- dial infarction, these include smoking and age. Non- smoking 15-19-year-olds enjoy the lowest risk of all. Tietze (1) estimated that the statistical risk of pill- related myocardial death among women under 20 years of age is approximately the same as that for women aged 20-29 years, i.e., 1.2 per 100 000 users per year for non-smokers and 1.4 per 100 000 for smokers. This is considerably less than the rates of 3.9 and 12.8 per 100 000, respectively, for non-smokers and smokers aged 35-39 years or 6.6 and 54.4 per 100 000 for those aged over 40 years. A similar pat- tern of risk exists for cerebral stroke. The excep- tionally low statistical risk of either of these con- ditions in adolescents is further supported by the present author's failure to discover even one reported case in the literature.' It is worth noting, however, that the idea of pill- associated risks for myocardial infarction in women of any age has recently been challenged. Studies of national vital statistics in a number of developed countries over the past few decades have shown a steady decline in deaths due to heart disease for both males and females. During the same interval, oral contraceptive use has become widespread. However, the decline in myocardial deaths in the 35-50-year age group is no less for females than for males, and this would scarcely have been expected if the pill were indeed responsible for an increasing incidence of such deaths (2, 27). Further research on the duration of pill use in adolescence and cumulative long-term cardio- vascular effects is required. Pill-associated venous thrombosis is less related to age, with a relative risk of 2 cases per 1000 users per a A number of reports of case-control studies on this topic have considered females "aged 15-44", the standard range of childbear- ing years employed in compiling fertility-related statistics, but it is never clear whether any cases under 20 years were actually found. While some data are broken down for the 15-25-year age group, none deal with 15-20-year-olds alone, presumably because the numbers are few and lack significance. BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE year (26). It does not appear that adolescents are at any greater risk and, considering the paucity of case reports in this age group, may be considerably less at risk. The only reported case concerned one adolescent receiving estrogen for excessive height prediction, and followed a trauma to the affected extremity (5). All cardiovascular risks associated with oral contraceptives appear to be dose-dependent to some degree. Risk rates are 50-80%7o less with pills contain- ing 30-50 Ag of ethinylestradiol and 1-2 mg of nor- ethisterone than for those containing 100-150 i4g of estrogen and 3-4 mg of progestogen (26). In add- ition, risks appear to be greatest during the period when the contraceptives are being used, and decrease with time after discontinuation. There is, however, some evidence for a cumulative effect of oral contra- ceptive use for five years or more at some time in the past for women aged 40-49 years (28). If borne out by further studies, this observation may be important for adolescents in terms of planning contraceptive use over their lifetime. The presumed mechanism for such a cumulative effect is accelerated atherogenesis secondary to estrogen-induced lipid changes. Plasma lipid changes in adolescents who smoke or use oral contraceptives are identical to those that occur in adults (29, 30). Conditions predisposing to cardiovascular disease are as much contraindications for oral contraceptive use in adolescents as they are in adults. The pill is absolutely contraindicated in individuals with hyper- lipidaemia or hypertension (although the pill itself has only minimal effects on blood pressure in normal individuals (26)). Its use in adolescent diabetics is controversial: some consider diabetes mellitus an absolute contraindication while others believe the high degree of protection afforded by this method warrants its use for a limited period of time (31, 32). Neoplasia Concern about neoplasia due to oral contraceptives extends to the breast, endometrium, cervix, and liver (33). The long latency period that commonly exists between the time of exposure to a carcinogen and appearance of disease, as well as the possibility of a cumulative increase in risk with long-term exposure, make the initiation of use of oral contraceptives at an early age a matter of particular concern. Data relative to breast cancer now indicate that there is no increased risk due to oral contraception. Moreover, the pill exerts a protective effect against benign breast disorders including fibrocystic disease and fibroadenoma (2, 22, 33-35). Only one, as yet un- confirmed, case-control study even suggests an increased incidence of malignancy under particular circumstances, but it is a study of some interest because of the young age group involved. Pike et al. (36) found a 3-fold greater risk of breast cancer in parous women under 32 years of age who had taken oral contraceptives for 97 or more months before their first full-term pregnancy. However, these findings are open to serious challenge in that only 163 among 245 eligible women (as determined by hospital records) were available for interview and it is not clear whether the case-control variables were adequately adjusted. Any apparent increased risk of breast cancer in younger women using oral contraceptives should be carefully distinguished from the well established risks associated with early menarche (before 12 years of age), obesity, a positive family history, and late menopause (33, 37). Endometrial cancer has been associated with the use of unopposed estrogen during the menopause (38), the sequential administration of estrogen and progestogen for contraception (33), and unopposed or sequential hormone replacement therapy in hypo- gonadal states (39). There is no such association with the use of combination oral contraceptives. The simultaneous use of estrogen and progestogen pro- tects against the precursor state of endometrial hyper- plasia seen with the administration of estrogen alone (2, 22). In relation to cervical carcinoma, aWHO Scientific Group stated in 1978 that the data "suggest the possi- bility of an increased risk among women with long- term [oral contraceptive] use and with other factors predisposing to [this] disorder" (33). More recent studies (40) found that the apparent association with the pill ceased to be of significance when corrected for sexual factors. In all probability, the recent notable increase in the neoplastic precursor states of cervical dysplasia and carcinoma in situ among sexually active adolescents and young adults (41, 42) is independent of the pill. Evidence for this comes not only from the study noted above (40) but also from the fact that few sexually active teenagers use any contraceptive for extended periods of time, if at all, giving such low rates of estrogen exposure as to make any statistical associ- ation with dysplasia improbable. There is significant correlation between cervical dysplasia and both an early age at initiation of coitus and exposure to multiple partners, either directly or indirectly through a single partner who himself has multiple exposures. Epidemiological evidence clearly points to the venereal transmission of factors that engender a dysplastic response. The precise nature of these factors is not yet fully understood, although it has been suggested that herpesvirus may play a con- tributory role (42). Hepatocellular adenoma has recently been associ- ated with oral contraceptive use (33). While no cases have occurred in adolescents, there have been a few cases in women aged 20-25 years. The greatest risk, 335 A. D. HOFMANN however, is to women over 27 years who have taken pills containing 50 Ag or more of estrogen for at least 7 years. The relative risk for all females is estimated to be 3.4 per 100 000 users of oral contraceptives per year (43). Hepatocellular adenoma is a benign condi- tion amenable to surgical excision, but I in 8 cases progresses to spontaneous rupture and haemorrhage with a high mortality rate. Early recognition and intervention greatly improves the prognosis. Since some cases have occurred in women in their early twenties, the possibility exists that adolescents may occasionally be afflicted. Hepatocellular adenoma should be considered in any teenager taking oral contraceptives presenting with a right upper quadrant mass, particularly if associated with pain and impend- ing shock. The relative risk to the teenage population, however, is exceptionally low and insufficient to contraindicate pill use. Miscellaneous Specific instances of major morbidity associated with oral contraceptive use in persons under 20 years of age are singularly difficult to find. An extensive literature search revealed only two reports: Honore (44) noted a statistically significant but inter- dependent relationship between obesity, parity, and oral contraception in 9 out of 31 adolescents with cholelithiasis, compared with 16 out of 112 healthy controls; Wadlington et al. (45) recorded a single case of transient chorea in a 17-year-old girl who was taking the pill; this is a recognized but relatively rare idiosyncratic reaction. Summary There is no evidence to suggest a need for greater concern over the effects of oral contraception on adolescents than on women of any age, and in many respects, concern should be considerably less. There is no evidence for any adverse effect on growth or fer- tility and the relative risks for known serious com- plications are lower among teenagers than among other age groups. The only possible, but as yet unsub- stantiated, adverse effect, which might limit total life- time use of oral contraceptives, relates to estrogen- induced changes in lipid profiles and consequent possible accelerated atherogenesis. Additional studies are needed on this topic. We should be considerably more concerned about preventing smoking during adolescence a habit that may be difficult to abandon later. While the addi- tional risk of cardiovascular disease attributable to smoking is minimal in the teenage years, this risk increases with each decade and oral contraception becomes a considerably less viable option for the older woman who smokes than for the non-smoker. All in all, oral contraception for teenagers appears to have a wide margin of safety while offering the best available protection. As even the most modest of risks is even further reduced by lower-dose combinations, pills containing 30-50 Ag of ethinylestradiol and 1-2 mg of norethisterone, or their equivalents, are preferable. PROGESTOGENS The "minipill" Contraception with pills containing 0.35 mg of norethisterone or equivalent, taken daily without interruption, has been advocated by some as particu- larly suitable for adolescents. There are few serious risks and minimal effects on the hypothalamic-hypo- physeal-gonadal axis; ovulation is not regularly sup- pressed. Contraception results from the combined effects of increased viscosity of the cervical mucus, increased tubal motility, impaired corpus luteum function, and an altered endometrium (46). Experi- ence has not borne out the early optimism. Method failure rates are unacceptably high in adolescents- in the range of 2.5-3.7 per 100 wonran-years. In addition, menstruation is often irregular and un- predictable leading to a high level of method dissatis- faction and discontinuation. The "minipill" is, there- fore, not generally recommended for this age group (47). One possible exception is the adolescent with car- diac disease where a high degree of protection is re- quired but both combination oral contraceptives and the intrauterine device are contraindicated. Taurelle et al. (48) administered pills containing 500 Ag of lyn- estrenol to 40 adult cardiac patients for 6-30 months. There were no pregnancies and no evident alteration of haemodynamics, blood coagulation, blood lipids, liver metabolism, or glucose metabolism. Depot medroxyprogesterone acetate (DMPA) Considerable controversy has surrounded DMPA since the 1978 ban on its use in the United States as a consequence of the finding of an increased incidence of breast tumours in female beagle dogs given the drug (49). These findings have been challenged on the basis of the inapplicability of this particular animal model to humans and the absence of any similar find- ings in women given DMPA, despite an estimated 10 million past users (50). An expert panel, convened by the World Health Organization in 1981 to review all available data, con- cluded that "extensive clinical and epidemiological studies among women using... [DMPA and norethis- terone enanthate]... have thus far demonstrated no 336 BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE life-threatening side-effects, including any increase in the risk of neoplasia... [but] the potential long-term effects (more than 15 years) are not yet known" (51). This report goes on to support the use of DMPA in sexually active adolescents for whom other methods are contraindicated, either on medical grounds or because of poor compliance, observing that the social, medical and psychological consequences of unwanted pregnancy and abortion outweigh any reservations about the physiological effects of inject- able contraceptives. The American Academy of Pedi- atrics (52) has come to similar conclusions and sup- ports the use of DMPA in selected sexually active teenagers. Suggested candidates include mentally handicapped women at risk of sexual exploitation and unable to comply with other methods, as well as those with medical conditions for whom other contra- ceptives are either insufficiently effective or strongly contraindicated. DMPA is not recommended for use during the first two years after menarche because there is insufficient knowledge about its long-term effects on the matur- ing hypothalamic-hypophyseal-ovarian axis. Such caution, however, may not be entirely necessary. DMPA has been employed frequently in children with idiopathic isosexual precocious puberty to suppress endogenous gonadal hormone production and arrest pubertal progression. Dosages of about 150 mg per week, intramuscularly, have been used for as long as 5j years. On cessation of treatment, puberty again advances, picking up precisely where it had been stopped at the beginning of treatment. In Kauli's (53) series of 24 precociously pubertal girls, aged between 4 and 121 years at the beginning of DMPA adminis- tration, menarche occurred in all patients after dis- continuation of treatment at an appropriate time, as judged by pubertal stage. In most cases, the delay varied from a few days to a year, with only two cases taking up to 3 years. All children ultimately developed normal menstrual cycles. THE INTRAUTERINE DEVICE (IUD) In theory, the IUD would appear to be the ideal contraceptive for adolescents. Once in place, com- pliance requirements are few, refill prescriptions un- necessary, and there is no possibility of inadvertent discovery by others. Unfortunately, a number of problems seriously compromise these theoretical advantages. Expulsion and discontinuation The experience of Weiner et al. (55) led these inves- tigators to conclude that continuation rates with the IUD compared so poorly with those for oral contra- ceptives as to preclude this method from being a viable alternative. In their study of Swedish high school students using contraception, there were 4.6 pregnancies per 100 user-years and a 39lo discon- tinuation rate with the IUD compared with rates of 1 and 1507o respectively for pill users. Although Gold- man & Reichman in Israel (58) and Lane & Sobrero in the United States of America (61) reported somewhat more favourable outcomes, with one-year continua- tion rates of about 800Vo, such results appear to be the exception rather than the rule (Table 2). This discouraging experience must, however, be considered in relation to other factors. Adolescents who are not highly motivated are much more likely to discontinue the pill than the IUD. Thus, Jorgenson (62) found that only 47%7o of adolescents were still taking oral contraceptives one year after starting, compared with a continuation rate of 53%o for IUD users. Graves & Bradshaw (63) found that 30% of a group of adolescents given oral contraceptives became pregnant within one year, compared with only 9%o of those with an IUD. Success or failure with the IUD is both a relative and variable matter dependent on the interaction of a number of social, psychological, and medical factors. Significant psychosocial issues include the attitudes of the user and the provider towards the device, the ability of the adolescent to tolerate increased men- strual flow and cramping, and her willingness to insert her finger into her vagina to check on correct placement of the device. Significant medical factors include an adequate-sized uterus (sounding at least 5-6 cm), selection of a suitable device of the correct size, optimal timing of insertion (within 10 days of the onset of a menstrual period) (55), and skilful inser- tion to minimize pain and ensure proper placement (61). Both copper-wound and progesterone-impreg- nated devices have been used in adolescents with better success than other types (59). Progesterone- impregnated devices, however, have the disadvantage of requiring annual replacement. "T"-type copper- wound IUDs are somewhat better tolerated than "7"-types (60). The newer T-Cu-380A, with an 8-year life span and early indications of low preg- nancy and discontinuation rates (64), may prove to be the best choice. Any expulsions tend to be partial and to occur with- in six months of insertion; reinsertion results in reten- tion in about half of these cases (55-57). The overall experience with the IUD in nulliparous adolescents is similar to that in nulliparae of any age (60, 65). All have higher expulsion and removal rates than those who are parous, but these rates can be modified by proper insertion of a suitably sized device, use of analgesia for menstrual discomfort, and continuing psychological support. 337 A. D. HOFMANN Table 2. Reported experience with the intrauterine device in adolescents and adults Continu- ation Removal rate at Source Reference Sample Device Expulsions 1 year Pregnancy (%) Reason % (%)a rate (%)b 54 151 nulli- parous or very young parous adolescents aged 1 1 -20 years Lippes loop (91 subjects) Majzlin spring (60 subjects) 55 226 nulli- parous 13-20- year-olds; 3138 months of use Cu-T-200 or Cu-7 55 Infection 8 Bleeding/pain 7 Other 4 5 Infection 5 Bleeding/pain 2 Other 7 20 Bleeding/pain 1 5 Other medical 3 Personal 1 Huber (1979) 56 1 75 nulli- parous 14-20- year-olds; 4723 months of use 57 162 nulli- parous 13-13- year-olds; insertion at time of elective abortion; 1826 months of use 58 30 nulli- parous 14 18- year-olds; 554 months of use 59 94 nulli- parous and 22 parous 13-20- year-olds Mishell (1975) 60 Review of selected reports on adults Cu-T or Cu-7 11 Medical Personal Lippes loop (56 subjects) Cu-T or Cu-7 20 Medical Personal 7 Medical Personal Cu-T or Cu-7 7 Medical Personal Cu-7 17 (nulli- parous) 22 (parous) Lippes loop T-Cu-200 Cu-7 4 8 20 1 4 0 3 13 Bleeding/pain 21 Infection 3 Other medical 4 Personal 1 1 5 Bleeding/pain 10 Other medical 3 9 Bleeding/pain 9 Other medical 3 Personal 5 16 Bleeding/pain 11 Other medical 4 Personal 4 a Includes individuals with replacement and subsequent retention of a device following initial expulsion. b Includes only pregnancies occurring with the device in plaoe. 338 Johnson et al. (1971) Weiner et al. (1978) 26 2.2 81 1.6 61 4.6 77 4.0 Goldman et al. (1979) Goldman & Reichman (1980) Kulig et al. (1980) 3.7 77 3.3 50 3 79 2.4 74 1.6 65 2.9 BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE Ectopic pregnancy The IUD reduces the incidence of uterine implanta- tion to a greater extent than tubal implantation (99.50o versus 95%) and has no effect on ovarian implantation. This disparity results in an apparent increased risk of ectopic pregnancy (66). There is no evidence that ectopic pregnancies are more frequent in nulliparous than parous women, or in adolescents in particular in uncomplicated IUD use. Fertility Most studies on uncomplicated IUD use in adults demonstrate a prompt return of fertility on discon- tinuation, with rates equal to those of the population at large (66). Several reports, however, have contra- dicted this. Jain (67) found a higher incidence of infertility among Taiwanese women than could be accounted for by other causes of sterility. This was most marked in women over 30 years who had used an IUD for more than 3 years; women under 25 years were least affected. One study in the United Kingdom (68) demonstrated transient infertility for up to one year following IUD discontinuation, but by 31 years, the conception rate was similar to that for women who had previously used other methods. Here, too, the delays were greatest in the older age group. For practical purposes, there appears to be little risk of infertility among adolescents following uncompli- cated IUD use. Pelvic inflammatory disease (PID) Pelvic inflammatory disease is highly significant for adolescents and particularly compromises wide- spread IUD use among unmarried young people. Single, sexually active teenagers with an IUD in place are 3-10 times more likely to develop PID than those using other contraceptive methods or no method (69-71). Two confounding factors, however, must be considered in interpreting these data. The first is the possibility of a protective effect of oral contraception against PID, giving falsely high relative risks for the IUD (72). The second is the likelihood of substantial error in diagnosing salpingitis without laparoscopic confirmation, particularly in the absence of such indicators as a palpable salpingeal mass, fever, or elevated erythrocyte sedimentation rate (71). Even sophisticated examiners can be fooled, tending to overdiagnose when findings are limited to abdominal and pelvic tenderness and pain. After reviewing all available data, Senanayake & Kramer (72) concluded that the risk of PID in all women using an IUD is 1.5-4 times higher than for other groups of women. There is no greater risk with any particular device (with the exception of the Dalkon Shield, which has now been withdrawn from use). The bacteriostatic effects of copper-wound devices observed in vitro do not appear to operate in vivo (69, 73). Sexually active single adolescents are at particularly high risk for PID even in the absence of an IUD. Westrom (71) found that PID was more prevalent among sexually active 15 and 16-year-old girls in Sweden than among any other age group, witl' rates of 1 in 8 and 1 in 10, respectively, compared with 1 in 80 among sexually active women over 24 years of age. In the USA, adolescents aged 14-19 years accounted for 24% of all hospitalizations for acute salpingitis and PID, exceeded only by the 20-24-year-olds, who comprised 30% of the total (74). PID is rare in virginal adolescents and shows a strong correlation with the initiation of coitus at an early age, frequent coitus, and multiple partners (75). Westrom's study of Swedish girls (71) also found that half of all sexually active 15-16-year-olds (the group most likely to develop pelvic infections) had experi- enced intercourse with four or more partners in the previous year, compared with only 20/o of those aged 18 years. The chances of developing PID are increased by the presence of an IUD. In a case-control study of 1400 women admitted to hospital with this condition, Burkman (69) calculated that the overall risk was increased by a factor of 1.6 for women with an IUD. When age and number of partners were taken into account the relative risk increased to 1.9 for females under 25 years of age and 2.6 for women who had had two or more partners. Thus the greater probability of PID in sexually active adolescents derives from their pattern of coital behaviour, rather than from any particular physio- logical vulnerability. Repeated exposure to sexually transmitted diseases through multiple partners ap- pears to be the most significant factor leading to an increased risk of PID. The likelihood that exposure to a pathogen will progress to manifest infection is apparently significantly increased by the presence of an IUD. Precisely how this device facilitates infection is not clear. It has been suggested that the chronic sterile endometrial and tubal reaction noted with this method may well impair host defence mechanisms and decrease host resistance (66). The most fre- quently implicated organisms in PID include Neis- seria gonorrhoeae, Chiamydia trachomatis, and Mycoplasma hominis (75). A wide variety of other aerobes and anaerobes has also been recovered through culdocentesis or laparoscopy, implicating a polymicrobial etiology. Aside from the morbidity of the disease itself, the major consequences of PID are increased risks of ectopic pregnancy and infertility (71, 75). Women with a history of one or more episodes of salpingitis have a 1 in 16 probability that their first pregnancy will be ectopic, compared with 1 in 147 for those 339 A. D. HOFMANN without such a history; rates for infertility range between 18%o and 37%. The degree to which prompt recognition and early treatment of PID might alter these outcomes is unknown. While the IUD increases the probability of develop- ing PID in individuals at risk, there is as yet no indica- tion of a parallel increase in severity or likelihood of resultant infertility. However, no studies have been conducted on this specific topic. Most clinicians (57, 76) are agreed that, as a precautionary measure, an IUD should be removed in the case of serious infec- tion, although there is debate over whether this should be done at the time of diagnosis or after several days of antibiotic treatment. More controversy exists over the need for removal in the case of mild infec- tion; some clinicians advocate leaving the IUD in place. In any event, a past episode of confirmed PID generally constitutes an absolute contraindication to the initiation of contraception with this method. Summary The IUD, with its minimal demands for com- pliance, has much to commend it for adolescents. Un- fortunately, not only is there a high rate of expulsion and method dissatisfaction, but also a 2-3-fold increase in the risk of pelvic inflammatory disease, particularly in those exposed to multiple partners with a greater likelihood of acquiring sexually transmitted diseases. None the less, the IUD may be a viable alter- native for sexually active teenagers who are unable to use another method satisfactorily and have experi- enced one or more pregnancies. If used in this group, close supervision is required, with prompt attention to suspicious symptoms. These restrictions do not apply to married adoles- cents in a monogamous relationship, where recurrent exposure to sexually transmitted disease is unlikely. In this instance, the IUD would carry the same risk- benefit ratio as when used by older married women. BARRIER METHODS Diaphragm and spermicidal jelly The primary advantage of the diaphragm for adolescents is the almost total absence of risks and medical contraindications. The only significant con- sideration is the remote possibility of toxic shock syn- drome if the diaphragm is left in place for extended periods of time. One such instance in a 23-year-old woman unable to remove her diaphragm for two days has been reported in the literature (77). Other occasional problems include anatomic abnormalities precluding a proper fit and local allergic reactions to rubber or spermicides in either partner. The diaphragm does not enjoy great favour among adolescents. Reasons include the somewhat higher rates of method failure than with the pill, the need for trained personnel to ensure a proper fit, the require- ment for bulky and conspicuous supplies, the direct relationship with coitus, and patient bias, particularly among younger girls who are often reluctant to use any method requiring handling of the genital organs. Mature adolescents may be better users of this method. Lane & Sobrero (78) found that 83% of a group of 18-20-year-old women effectively used a diaphragm for more than one year with an accidental pregnancy rate of only 2.9 per 100 woman-years; younger teenagers had a continuation rate of only 70% (but with an incidence of pregnancy of only 1.9 per 100 woman-years, presumably due to the greater frequency of anovulatory cycles in this group). How- ever, even a 70%o continuation far exceeds the rates reported by others, and probably represents the maxi- mum degree of compliance that can be expected in adolescents. This particular study group consisted of highly motivated young women from the middle and upper middle classes and is not representative of the general population. Cervical cap The cervical cap is a relatively old device, which has recently been redesigned and reintroduced as an alternative to the diaphragm (79). The major advan- tage is that the cervical cap may be left in place for from 3-5 days to the entire intermenstrual period. Un- fortunately, initial trials have demonstrated a 6- month pregnancy rate of 8 per 100 woman-years and a high level of method dissatisfaction with a 39% dis- continuation rate (80). In addition, the risk of toxic shock is a more significant possibility than with the diaphragm because of the length of time that the device is left in place, although no cases have been reported to date. For all these reasons, the cervical cap cannot be recommended for adolescents. Spermicides While vaginal foams, creams, and suppositories have low reported rates of use among adolescents, their coital relationship and messiness placing them in some disfavour, they do offer significant advantages in that they are safe, inexpensive, do not require a clinic visit or prescription, and suit the spontaneous, unpredictable nature of teenage coital behaviour. A fair degree of protection against sexually transmitted diseases is an added advantage. Women using sper- micides have been found to have only one-quarter the incidence of gonorrhoea and four-fifths the incidence of Trichomonas or Haemophilus vaginitis compared with users of oral contraceptives or women who have 340 BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE been sterilized (81). There are no serious risks and complications are relatively few, with local allergic reactions being the most common. There has, however, been recent concern about possible teratogenic effects on fetuses. Jick (82) found a 1 Vo attributable increase in congenital anom- alies among infants born to women who had filled a prescription for a spermicide within 600 days of delivery. Further evaluation is necessary in light of the highly indirect nature of this study. At present, this conjectured risk is insufficient to contraindicate use of spermicides by adolescents. A very high degree of protection is afforded when spermicides are used in conjunction with the condom (Table 1). This may well be the method of choice for adolescents engaging in infrequent coitus. It also has a role as a back-up method in case of primary method problems or discontinuation. Condom A significant number of sexually active adolescents depend on the condom. It provides reasonably good protection with no medical risks apart from a rare rubber allergy, in which instance sheaths made of ani- mal products may be used. Substantial protection against the transmission of sexually transmitted dis- eases makes the condom a particularly important method for this population at risk. Male complaints of decreased sensation may be an advantage in re- ducing the tendency towards premature ejaculation commonly experienced by adolescent boys. In any event, insensitivity is less of a problem with newer products (83). The attitudes of many teenage boys and their un- willingness to accept responsibility for contraception are significant barriers to widespread use of the condom. However, a number of model programmes have demonstrated that these attitudes can be changed (89). For example, a free condom distri- bution programme in an urban community in the USA resulted in 75% of young men reporting use during their last sexual encounter, compared with 55%o of adolescents in a comparable community with no distribution programme (84). "NATURAL" BIRTH CONTROL The prevention of conception through abstinence during the periovulatory or "fertile" period depends on the ability accurately to predict this time by means of the calendar, basal body temperature, changes in cervical mucus, or all three in combination; this method is known as "fertility awareness", the symp- tothermal method, or the Billings method. Effective- ness depends on the menstrual cycles being of consis- tent length with ovulation occurring on the 14th day before the onset of each menses. The marked vari- ation in adolescent menstrual patterns makes such prediction uncertain at best for the first 5-6 gynaeco- logical years. In addition, basal body temperature is not wholly reliable; up to 2007o of regularly ovulating women (as determined by progesterone levels) have atypical temperature curves which are difficult for even skilled professionals to interpret (85). While advocates of the method claim good results among carefully instructed and highly motivated adult women (86), most investigators find unacceptably high pregnancy rates of from 4 to 40 per 100 woman- years and annual discontinuation rates of 38-44% (87, 88). Natural family planning is not a suitable method for adolescents. WITHDRAWAL There would be much to commend withdrawal for adolescents were it only more effective. It is always available, requires no advance planning or purchase, and is well suited to the spontaneous patterns of adol- escent coital behaviour. However, method failure rates range between 20% and 25%/6 (Table 1), which is only marginally better than no method at all. CONCLUSIONS Oral contraceptives, condoms, spermicides, and the diaphragm all have a significant place in family planning initiatives for adolescents. The intrauterine device may also be appropriate for married adol- escents and for selected unmarried teenagers who are unable to use other methods, have experienced one or more pregnancies, and can be treated promptly in the case of a pelvic infection. Depot medroxyproges- terone acetate also has a place in the management of adolescents unable to use other methods and for whom pregnancy carries greater physical, emotional, or social hazards than usual. With the exception of an increased risk of pelvic in- fection in IUD users frequently exposed to sexually transmitted diseases, contraception in adolescence appears to be remarkably safe. There are no greater complications associated with oral contraceptives in this age group than in any other, and in all instances risks are equal to or less than those experienced by women aged 20-25 years. In particular, there is no confirmed evidence that estrogens interfere with pubertal growth or the achievement of regular ovu- lation; indeed, there is considerable evidence to the contrary. The only possible concern relates to the possibility of accelerated atherosclerosis, which may 341 342 A. D. HOFMANN limit the total lifetime use of the pill. When measured against the costs of an unwanted pregnancy in a sexually active adolescent, the benefits of contraception by any means far outweigh the risks. Childbearing in the teenage years not only places the adolescent's health and that of her infant in jeopardy, but also forecloses her personal options for the future, too often consigning her to a life of poverty and hardship. RESUME LA CONTRACEPTION DANS L'ADOLESCENCE: UNE REVUE 2. ASPECTS BIOMEDICAUX Aucune des methodes courantes de contraception ne peut etre consideree comme ideale pour les adolescentes, mais beaucoup le sont presque. En fait, lorsqu'on les compare avec les risques de grossesse non desiree, les avantages de la contraception, quel qu'en soit pratiquement le rmoyen, l'em- portent de loin sur les risques. Les contraceptifs oraux paraissent presenter peu de risques specifiques pour l'adolescente et ils sont aussi surs pour ce groupe d'age que pour les adultes. Les pilules come- nant 30 a 50 jig d'ethinyl-oestradiol et 0,5 a 1 mg de nore- thisterone ou des equivalents offrent une bonne efficacite et sont bien tolerees. Par ailleurs, rien n'indique que la pilule compromette le developpement sexuel ou une ovulation reguliere. Les risques de complication cardio-vasculaire sont plus faibles pour les adolescentes que pour tout autre groupe d'age. Toutefois, la modification du profil lipidique, qui est consideree comme contribuant a augmenter l'atherogenese, peut empecher la prise de contraceptifs oraux indefini- ment. Quant aux dispositifs intra-uterins, bien qu'ils n'en- trainent qu'un minimum de contraintes, ce qui en fait une methode appropriee pour les adolescentes, ils sont souvent inacceptables parce qu'ils provoquent des crampes et aug- mentent le flux menstruel. Le taux d'expulsion est plus eleve chez les femmes nullipares, mais les dispositifs en cuivre ou. liberant des progestogenes sont mieux toleres. Le risque accru d'inflammation pelvienne chez l'adolescente non mariee est assez grand pour que le dispositif intra-uterin soit limite aux sujets qui ne peuvent pas utiliser d'autres methodes. La medroxyprogesterone-retard (contraceptif hormonal injectable) peut etre utilisee chez l'adolescente comme chez l'adulte. L'Organisation mondiale de la Sante et l'Academie de Pediatrie des Etats-Unis en recommandent l'usage pour certaines categories d'adolescentes chez lesquelles d'autres methodes seraient contre-indiquees. Par contre, la <mini- pilule>> contenant des progestogenes ne convient pas aux adolescentes, car elle s'accompagne de contraintes assez rigides, elle entraine souvent un saignement excessif et offre un taux de protection assez faible. Toutes les methodes mecaniques sont assez repandues chez les adolescentes, encore que les plus jeunes repugnent souvent a mettre le diaphragme en place. Les preservatifs, utilises seuls ou en association avec un gel spermicide, pre- sentent l'avantage de ne pas etre delivres sur ordonnance, d'etre faciles a se procurer et de convenir au caractere spontane et non premedite du coit dans ce groupe d'age. La methode de la continence periodique et les autres formes de contraception <<naturelle>, ne conviennent pas pour les adolescentes. Ces methodes sont tributaires d'une ovulation reguliere et exigent une information poussee, une forte motivation et l'aptitude A prevoir le coit, facteurs dont aucun n'est habituel chez l'adolescente. REFERENCES 1. TIETZE, C. New estimates of mortality associated with fertility control. Family planning perspectives, 9: 74-76 (1977). 2. RINEHART, W. & PIOTROW, P. T. OC's - update on usage, safety, and side effects. Population reports, Series A, (5): 134-186 (1979). 3. LAWSON, J. S. ET AL. Optimum dosage of an oral con- traceptive. A report from the study of seven combina- tions of norgestimate and ethinyl estradiol. American journal of obstetrics and gynecology, 135: 315-320 (1979). 4. PHILLIPS, L. S. & VALLILOPOULOV-SELLIN, R. Somato- medins. New England journal of medicine, 302: 371-380, 438-444 (1980). 5. CRAWFORD, J. D. Treatment of tall girls with estrogen. Pediatrics, 62 (Suppl.): 1189-1195 (1978). 6. ANDERSEN, H. ET AL. Treatment of girls with excessive height prediction. Acta paediatrica Scandinavica, 69' 293-297 (1980). 7. CONTE, F. A. & GRUMBACH, M. M. Estrogen use in children and adolescents: a survey. Pediatrics, 62: 1091-1097 (1978). 8. BIERICH, J. R. Estrogen treatment of girls with con- stitutional tall stature. Pediatrics, 62 (Suppl.): 1196-1201 (1978). 9. HANKER, J. P. ET AL. Regulation of the menstrual cycle following high-dose estrogen treatment in excessively tall girls (abstract). Fertility and sterility, 28: 363 (1977). 10. PRADER, A. & ZACHMANN, M. Treatment of excessively tall girls and boys with sex hormones. Pediatrics, 62 (Suppl.): 1202-1210 (1978). 11. WETTENHALL, H. N. B. ET AL. Tall girls: a survey of 15 years of management and treatment. Journal of pediatrics, 86: 602-610 (1975). 12. ZACHMANN, M. ETAL. Estrogen treatment of excessively tall girls. Helvetica paediatrica acta, 30: 11-30 (1975). BIOMEDICAL ASPECTS OF CONTRACEPTION IN ADOLESCENCE 343 13. LUNENFELD, B. ET AL. The ovary learns to ovulate. Journal of biosocial science, 5 (Suppl.): 27-42 (1978). 14. VOLLMANN, R. F. The menstrual cycle. Philadelphia, W. B. Saunders, 1977. 15. APTER, D. Serum steroids and pituitary hormones in female puberty: a partly longitudinal study. Clinical endocrinology, 12: 107-120 (1980). 16. METCALF, M. G. & MACKENZIE, J. A. Incidence of ovulation in young women. Journal of biosocial science, 12: 345-352 (1980). 17. REY-STOCKER, I. ET AL. Der Sensibilitat der Hypophyse, der Gonaden und der Schilddruse biem jungen Madchen vor und nach hormoneller Kontrazeption. Pharmako- therapia, 3: 42-45 (1980). 18. HUGGINS, G. R. Contraceptive use and subsequent fer- tility. Fertility and sterility, 28: 608-612 (1977). 19. HULL, M. G. R. ET AL. Post-pill amenorrhea: a causal study. Fertility and sterility, 36: 472-479 (1981). 20. SMILJANIC, N. ET AL. Amenorrhea following oral contraception. Acta obstetricia et gynecologica Scandi- navica, 59 (B): 261-264 (1980). 21. TOLIS, G. ET AL. Prolonged amenorrhoea and oral con- traceptives. Fertility and sterility, 32: 265-268 (1979). 22. ORY, H. W. ET AL. The pill at 20: an assessment. Inter- national family planning perspectives, 6: 125-130 (1980). 23. DRAMUSIC, V. The importance of prompt treatment and long follow-up of cases of menstrual disorders in adolescence concerning their fertility. Proceedings of the VI Simposio Mundial de Ginecologia Pediatricay de la Adolescencia, Punta del Este, Uruguay, 2-5 December 1981, p. 101. 24. RODRIGUEZ-RIGAU, L. J. Ovarian dysfunction in the adolescent. Proceedings of the VI Simposio de Ginecologia Pediatrica y de la Adolescencia, Punta del Este, Uruguay, 2-5 December 1981, p. 25-51. 25. DALEN, J. E. & HICKLER, R. B. Oral contraceptives and cardiovascular disease. American heart journal, 101: 626-639 (1981). 26. STRADEL, B. V. Oral contraceptives and cardiovascular disease. New England journal of medicine, 305: 612-618, 672-677 (1981). 27. WISEMAN, R. A. & MACRAE, K. D. Oral contraceptives and the decline in mortality from circulatory disease. Fertility and sterility, 35: 277-283 (1981). 28. SLOAN, D. ET AL. Risk of myocardial infarction in relation to current and discontinued use of oral contra- ceptives. New England journal of medicine, 305: 420-424 (1981). 29. WALLACE, R. B. ET AL. Plasma lipids, lipoproteins and blood pressure in female adolescents using oral contra- ceptives. Journal ofpediatrics, 95: 1055-1059 (1979). 30. WEBBER, L. S. ET AL. The interaction of cigarette smoking, oral contraceptive use, and cardiovascular risk factor variables in children: the Bogalusa heart study. American journal ofpublic health, 72: 266-274 (1982). 31. ANDREWS, W. C. Oral contraceptives. Clinics in obstetrics and gynecology, 6: 3-26 (1979). 32. GREYDANUS, D. E. & MCANARNEY, E. R. Contra- ception in the adolescent: current concepts for the pediatrician. Pediatrics, 65: 1-12 (1980). 33. WHO Technical Report Series, No. 619, 1978 (Steroid contraception and the risk of neoplasia: report of a WHO Scientific Group). 34. JICK, H. ET AL. Oral contraceptives and breast cancer. American journal of epidemiology, 112: 577-585 (1980). 35. ROYAL COLLEGE OF GENERAL PRACTITIONERS. Breast cancer and oral contraceptives: findings in Royal College of General Practitioners' study. British medical journal, 282: 2089-2093 (1981). 36. PIKE, M. E. ET AL. Oral contraceptive use and early abortion as risk factors for breast cancer in young women. British journal of cancer, 43: 72-76 (1981). 37. SHERMAN, B. ET AL. Relationship of body weight to menarcheal and menopausal age: implications for breast cancer risk. Journal ofclinical endocrinology and metabolism, 52: 488-493 (1981). 38. OBRINK, A. ET AL. Estrogen regimen of women with endometrial carcinoma. a retrospective case-control study at Radiumhemmet. Acta obstetricia et gyneco- logica Scandinavica, 60: 191-197 (1981). 39. LEVINE, L. A. Treatment of Turner's syndrome with estrogen. Pediatrics, 62 (Suppl.): 1178-1188 (1978). 40. SWAN, S. H. & BROWN, W. L. Oral contraceptive use, sexual activity, and cervical carcinoma. American journal of obstetrics and gynecology, 139: 52-57 (1981). 41. DELKE, I. M. ET AL. Abnormal cervical cytology in adolescents. Journal ofpediatrics, 89: 985-987 (1981). 42. SCHYDLOWER, M. ET AL. Adolescents with abnormal cervical cytology. Clinical pediatrics, 20: 723-726 (1981). 43. ROOKS, J. B. ET AL. Epidemiology of hepatocellular adenoma; the role of oral contraceptive use. Journal of the American Medical Association, 242: 644-648 (1979). 44. HONORE, L. H. Cholesterol cholelithiasis in adolescent females: its connection with obesity, parity, and oral contraceptive use-a retrospective study of 31 cases. Archives of surgery, 115: 62-64 (1980). 45. WADLINGTON, W. B. ET AL. Chorea sssociated with the use of oral contraceptives; report of a case and review of the literature. Clinical pediatrics, 20: 804-806 (1981). 46. FREUTNER, A. K. & HOLLINGSWORTH, D. R. Adoles- cent obstetrics and gynecology, Chicago, Illinois, Year Book Medical Publishers, 1978. 47. MISHELL, D. R. Contraception. American journal of diseases in children, 32: 912-920 (1978). 48. TAURELLE, R. ET AL. [Contraception using a proges- togen-only minipill in cardiac patients.] Archives des maladies du coeur et des vaisseaux, 72: 98-106 (1979). 49. Approval of Depo-Provera for contraception denied. Federal Drug Administration drug bulletin, 8: 10 (1978). 50. GOLD, R. B. & WILLSON, P. D. Depo-Provera: new developments in a decade-old controversy. Inter- national family planning perspectives, 6: 156-160 (1980). 51. Injectable hormonal contraceptives: technical and safety aspects. Geneva, World Health Organization, 1982 (Offset Publication, No. 65). 344 A. D. HOFMANN 52. AMERICAN ACADEMY OF PEDIATRICS, COMMITTEE ON DRUGS. Medroxyprogesterone acetate (Depo-provera). Pediatrics, 65: 648 (1980). 53. KAULI, R. & LARON, Z. Pubertal development in girls with sexual precocity after discontinuation of treatment with cyproterone acetate. Helvetica paediatrica acta, 35: 149-154 (1980). 54. JOHNSON, L. B. ET AL. Problems with contraception in adolescents: the successful use of an intrauterine device. Clinical pediatrics, 10: 315-319 (1971). 55. WEINER, E. ET AL. Contraception for teenage girls: combination pill or IUD? Acta obstetricia et gyneco- logica Scandinavica, 88 (Suppl.): 65-69 (1979). 56. HUMBER, A. IUDs for adolescent girls. In: Hafez, E.S.E. & Van Os, W.A.A., ed., Medicated IUDs and polymeric delivery systems. Proceedings of an inter- national symposium. Amsterdam, North Holland Pub- lishing Co., 1979. 57. GOLDMAN, J. A. ET AL. Immediate postabortion intra- uterine contraception in nulliparous adolescents. Israeli journal of medical science, 15: 522-525 (1979). 58. GOLDMAN, J. A. & REICHMAN, J. Contraception in the teenager: a comparison of four methods of contracep- tion in adolescent girls. Israeli journal of medical science, 16: 510-513 (1980). 59. KULIG, J. W. ET AL. Experience with the copper-7 intra- uterine device in an adolescent population. Journal of pediatrics, 96: 746-750 (1980). 60. MISHELL, D. R. Assessing the intrauterine device. Family planning perspectives, 7: 103-111 (1975). 61. LANE, M. E. & SOBRERO, A. J. Experience with intra- uterine contraception by adolescent women. Mount Sinai journal of medicine, 42: 337-344 (1975). 62. JORGENSEN, V. One year contraceptive follow-up of adolescent patients. American journal of obstetrics and gynecology, 115: 484-486 (1973). 63. GRAVES, W. L. & BRADSHAW, E. R. Early reconception and contraceptive use among black teenage girls after an illegitimate birth. American journal of public health, 65: 738-740 (1975). 64. Roy, S. ET AL. Comparison of three different models of the copper-T intrauterine contraceptive device. American journal of obstetrics and gynecology, 134: 568-574 (1979). 65. BERGQVIST, A. & RYBO, G. IUD compared with oral contraception in nulliparae. Contraception, 20: 407-415 (1979). 66. PIOTROW, P. T. ET AL. IUDs - update on safety, effec- tiveness and research. Population reports Series B, 7(3): 49-98 (1979). 67. JAIN, A. K. & MOOTS, B. Fecundity following the dis- continuation of IUD use among Taiwanese women. Journal of biosocial science, 9: 137-151 (1977). 68. VESSEY, M. P. ET AL. Fertility after stopping different methods of contraception. British medical journal, 1: 265-267 (1978). 69. BURKMAN, R. T. Woman's health study: association between intrauterine device and pelvic irflammatory disease. Obstetrics and gynecology, 57: 269-276 (1981). 70. VESSEY, M. P. ET AL. Pelvic inflammatory disease and the intrauterine device: findings in a large cohort study. British medical journal, 282: 855-857 (1981). 71. WESTROM, L. Incidence, prevalence and trends of acute pelvic inflammatory disease and its consequences in industrialized countries. American journal of obstetrics and gynecology, 138: 880-892 (1980). 72. SENANAYAKE, P. & KRAMER, D. G. Contraception and the etiology of pelvic inflammatory disease: new per- spectives. American journal of obstetrics and gynec- ology, 138: 852-860 (1980). 73. ESCHENBACH, D. A. ET AL. Pathogenesis of acute pelvic inflammatory disease: role of contraception and other visible factors. American journal of obstetrics and gyn- ecology, 128: 838-850 (1977). 74. WEISNER, P. J. Magnitude of the problem of sexually transmitted diseases in the United States. In: NIAID Study Group, Sexually transmitted diseases: 1980 status report. Washington, DC, US Government Printing Office, 1961, pp. 21-32 (NIH publication, No. 81- 2213). 75. ESCHENBACH, D. A. & HOLMES, K. K. Acute pelvic inflammatory disease: current concepts of patho- genesis, etiology and management. Clinical obstetrics and gynecology, 18: 35-56 (1975). 76. MISHELL, D. R. Intrauterine devices. Clinics in ob- stetrics and gynecology, 6: 27-37 (1979). 77. LooMis, L. & FEDER, H. M., JR. Toxic-shock syndrome associated with diaphragm use. New Englandjournal of medicine, 305: 1585 (1981). 78. LANE, M. & SOBRERO, A. Successful use of the dia- phragm and jelly by a young population: report of a clinical study. Family planning perspectives, 8: 81-86 (1976). 79. WORTMAN, J. The diaphragm and other intravaginal barriers - a review. Population reports, Series H, (4): 57-75 (1976). 80. High rates of pregnancy and dissatisfaction mark first cervical cap trials. Family planning perspectives, 13: 48 (1981). 81. JICK, H. ET AL. Vaginal spermicides and gonorrhea. Journal of the American Medical Association, 248: 1619-1621 (1982). 82. JICK, H. ET AL. Vaginal spermicides and congenital dis- orders. Journal of the American Medical Association, 245: 1329-1332 (1981). 83. DUMM, J. ET AL. The modern condom - a quality product for effective contraception. Population report, Series H, (2): 21-36 (1974). 84. ARNOLD, C. B. The sexual behavior of inner city adolescent condom users. Journal of sex research, 8: 298-309 (1972). 85. LENTON, E. A. ET AL. Problems in using basal body tem- perature recordings in an infertility clinic. British med- ical journal, 1: 803-805 (1977). 86. LANCOT, C. A. Natural family planning. Clinics in obstetrics and gynecology, 6: 109-127 (1979). 87. KAMBIC, R. ET AL. A thirty-month clinical experience in natural family planning. American journal of public health, 71: 1255-1258 (1981). 88. MASTROIANNI, L. JR. Rhythm: systematized chance- taking. Family planning perspectives, 6: 209-212 (1974). 89. HOFMANN, A. Contraception in adolescence: a review. I. Psychosocial aspects. Bulletin of the World Health Organization, 62: 151-162 (1984).