Existence of multiple peaks in plasma ethinyl estradiol and norethindrone after oral administration of a contraceptive pill.

Previous measurements of plasma ethinyl estradiol (EE2) and norethindrone (NE) over 24 h after oral administration of a contraceptive pill have demonstrated a single steroid peak occurring 1-2 h after pill ingestion, with a gradual decline over the next 22 h. In the present study plasma concentrations of EE2 and NE were measured 0, 0.5, 0.75, 1, 2, 4, 12, and 24 h after oral ingestion of a contraceptive pill containing 35 micrograms EE2 and 1 mg NE at 0, 3, 6, and 9 months of use in 58 normal healthy women. Contrary to previous reports, analysis of the 464 steroid curves (58 subjects x 4 time periods x 2 steroids) revealed the presence of multiple hormone peaks. Two peaks of EE2 were identified in 44.8% of women during the first pill cycle and in 75.9%, 55.2%, and 67.2% of women after 3, 6, and 9 months of pill use. Two hormone peaks of NE were observed in 29.3% of women during the first cycle and in 36.2%, 50%, and 44.8% at 3, 6, and 9 months, respectively. Existence of these multiple peaks at the frequency observed has not previously been reported. Further quantification of the frequency and magnitude of these peaks could be helpful in explaining differences in biological responses associated with pill use.

The role of oral contraceptives in the treatment of hyperandrogenic disorders.

The clinical manifestations associated with hyperandrogenism, such as hirsutism and acne, are disturbing to most patients. In addition to correcting androgen-related problems, concerns such as contraception or other metabolic problems (for example, lipid/lipoprotein abnormalities, diabetes, hypertension) associated with these disorders and the effects of unopposed estrogen on the endometrium also need to be considered. Oral contraceptives are a therapeutic modality that may address these multiple problems. The potential mechanisms of action by which oral contraceptives correct excess androgen states include gonadotropin suppression, reduction of circulating androgens, increased androgen binding, suppression of adrenal androgen secretion and inhibition of 5 alpha-reductase, and androgen receptor binding. In normal women, there is good evidence that these actions occur with the use of oral contraceptives. Among women with anovulatory hyperandrogenic states, such as polycystic ovary syndrome, the response to oral contraceptives in each of these areas is somewhat more variable. However, oral contraceptive preparations that are more estrogen dominant appear to produce many of the desired effects. From a clinical standpoint, 60-100% of women with hirsutism improve on oral contraceptives; acne shows improvement in a high percentage of women as well. The use of oral contraceptives also reduces the risk of endometrial hyperplasia that may be associated with anovulatory states. Finally, current low-dose preparations containing the newer progestins (for example, norgestimate and desogestrel) appear to be either neutral, or perhaps beneficial, with respect to their metabolic impact.

Association between intrauterine device and pelvic inflammatory disease.

The Women's Health Study, a concurrent case-control study at 16 hospitals in 9 cities across the United States, examined the relationship between use of an intrauterine contraceptive device (IUD) and pelvic inflammatory disease (PID) severe enough to require hospitalization. After final classification, 1447 patients were eligible as PID subjects and 3453 patients were eligible as controls. For all current IUD users, the estimated relative risk of hospitalization with the diagnosis of PID was 1.6. However, the relative risk was about twofold when only individuals experiencing their first episode of the disorder were considered. The study also demonstrated an increased association for women aged 25 years or less and for nonblack women. Recent insertion or reinsertion of an IUD was associated with increased risk for PID, but total duration of use was not. Furthermore, the effect of IUD use on the development of PID persisted for several months after the IUD had been removed. The study also confirmed previous findings that the type of device does not markedly influence risk, and that the relative risk of PID is higher when IUD use is compared with use of other contraceptives or use of no contraceptives. Finally, the study suggests that users of nonpermanent forms of contraception other than the IUD have some form of protection against the development of PID.

A continuous quality improvement project to improve the quality of cervical Papanicolaou smears.

OBJECTIVE: To improve the quality of cervical Papanicolaou smears by continuous quality improvement techniques. METHODS: The study used a Papanicolaou smear data base of over 200,000 specimens collected between June 1988 and December 1992. A team approach employing techniques such as process flow-charting, cause and effect diagrams, run charts, and a randomized trial of collection methods was used to evaluate potential causes of Papanicolaou smear reports with the notation "inadequate" or "less than optimal" due to too few or absent endocervical cells. Once a key process variable (method of collection) was identified, the proportion of Papanicolaou smears with inadequate or absent endocervical cells was determined before and after employment of a collection technique using a spatula and Cytobrush. We measured the rate of less than optimal Papanicolaou smears due to too few or absent endocervical cells. RESULTS: Before implementing the new collection technique fully by June 1990, the overall rate of less than optimal cervical Papanicolaou smears ranged from 20-25%; by December 1993, it had stabilized at about 10%. CONCLUSION: Continuous quality improvement can be used successfully to study a clinical process and implement change that will lead to improvement.

Type of intrauterine device and the risk of pelvic inflammatory disease.

To study the association of pelvic inflammatory disease and various types of intrauterine devices (IUDs), data from the Women's Health Study were analyzed. The analysis included data from interviews of 622 women hospitalized with an initial episode of pelvic inflammatory disease and 2369 hospitalized control subjects reporting no history of pelvic inflammatory disease. Compared to the risk in women using no contraception, the relative risk of pelvic inflammatory disease in women currently using the Dalkon Shield was 8.3 (95% confidence limits 4.7 to 14.5). This represented a fivefold increase in risk compared to women currently using other types of IUDs. In this study, only 10% of women wearing an IUD were using the Dalkon Shield, yet they accounted for almost 20% of the excess risk of pelvic inflammatory disease occurring among all the IUD users. Most of the increased risk of pelvic inflammatory disease for women currently using other IUDs (excluding the Dalkon Shield) occurred in the first four months after insertion. These associations were not explained by differences between cases and controls in demographic variables, level of sexual activity, or medical history. The authors recommend that women still using a Dalkon Shield have it removed.

Intravascular spill of hyperosmolar urea during induced midtrimester abortion.

Intravascular spill of hypertonic sodium chloride and the resultant serious and occasionally fatal consquences have been reported in association with induced midtrimester abortion. This report details 3 cases of intravascular spill of hyperosmolar urea. Although urea may pose less concern, because of its ability to readily cross cell membranes and its action as an osmotic diuretic, inadvertent intravascular spill can produce symptoms that include nausea, headache, sensations of warmth, and intense uterine cramping. In addition, abnormal blood pigments may occasionally be noted in the urine. Treatment includes intravenous hydration, careful monitoring of fluid/electrolyte balance and renal function, and avoiding the use of oxytocic agents.

PIP: 3 cases of intravascular spill of hyperosmolar urea during induced midtrimester abortion are reported. In each case 80 gm of urea had been infused. Although urea may pose less concern than hypertonic sodium chloride because of its ability to readily cross cell membranes and its action as an osmotic diuretic, inadvertent intravascular spill can produce symptoms that include nausea, headache, sensations of warmth, and intense uterine cramping. Also, abnormal blood pigments may occasionally be noted in the urine. The 3 patients were 23, 20, and 24 years old, all para 0. The symptoms generally subsided after 30 minutes, and serum and urinary abnormalities returned to normal within 24 hours. Treatment included iv hydration, careful monitoring of fluid/electrolyte balance and renal function, and avoiding the use of oxytocic agents for at least 24 hours. The 3 patients eventually underwent successful abortions, 1 with intraamniotic prostaglandin F2 alpha (PGF2a), and 1 with urea-PGF2a abortion, without further complications.

Placental histopathology of midtrimester termination.

Placentas spontaneously passed after second-trimester terminations of pregnancy using either hypertonic sodium chloride or hyperosmolar urea plus prostaglandin F2alpha (PGF2alpha) were examined to determine histologic characteristics. The placentas of hypertonic sodium chloride terminations demonstrated a type of "coagulation necrosis" that has been described previously, while placentas of pregnancies terminated by hyperosmolar urea plus PGF2alpha showed a similar pattern in about one half the cases but a histologic pattern of less severe damage in the remaining cases. The 2 groups showed no significant differences when characteristics such as injection-abortion interval or estimated hypertonicity of the fluid were examined.

PIP: Second trimester placentas, passed spontaneously in 17 abortions induced by hypertonic sodium chloride and in 45 abortions induced by hyperosmolar urea plus prostaglandin F, were histologically compared; in 16 of the 17 sodium chloride cases, a pattern of severe tissue damage (Type A) was observed, and in the urea cases, approximately half of the cases exhibited Type A damage and the other half had a less severe form of damage (Type B). Type A damage is characterized by a zone of "coagulation necrosis" in which cell structure is lost, vessels are thrombosed, and there is severe inflammation. In Type B damage, cell structure is maintained, vessels are not thrombosed, and inflammation is diffuse. In previous studies of sodium chloride induced abortions, Type A damage had been noted and attributed to the hypertonicity of the abortifacient agent. Since half of the urea cases also had Type A damage, an attempt was made to discover clinical differences between the Type A and Type B urea cases. No significant differences were found; however, a scatter diagram suggests that patients with high concentrations of urea and with a longer interval between injection and subsequent abortion tended to exhibit Type A damage, while those with a shorter interval between injection and abortion manifested Type B damage. Tables include clinical characteristics of age, parity, duration of gestation, injection abortion interval, and amount of amniotic fluid removed for urea cases and a scatter diagram depicting interval time and estimated urea concentration. Photographs of tissue sections depicting Type and Type B damage are also included.

Cephalothin prophylaxis for midtrimester abortion.

A randomized prospective double-blind study was conducted to determine the efficacy of prophylactic antibiotics as compared with placebo in 198 women undergoing second-trimester intraamniotic injection abortions. Patients received either sodium cephalothin or placebo intravenously before the procedure and for up to 8 additional doses. In 11 patients postabortion endometritis developed; 2 had received the antibiotic and 9 had received a placebo (P less than .05). Prophylactic cephalothin decreased the incidence of endometritis in patients undergoing midtrimester injection abortion. An injection-abortion interval greater than 24 hours appears to identify patients at risk for the development of postabortion endometritis.

The management of midtrimester abortion failures by vaginal evacuation.

The management of 58 failed midtrimester primary abortion procedures by vaginal uterine evacuation is described. The indications, technic, and complications of this procedure are presented. This method was the only means of managing failed primary abortion over a 3-year period in which 2045 elective midtrimester abortions were performed. With the development of experience this technic offers significant advantages over other procedures, such as hysterotomy, in the management of such patients.

Lipid and lipoprotein changes associated with oral contraceptive use: a randomized clinical trial.

To determine the effects of oral contraceptives on lipids and lipoproteins over a six-month period, we randomized 266 women into four oral contraceptive groups: ethinyl estradiol 35 micrograms plus ethynodiol diacetate 1 mg, ethinyl estradiol 30 micrograms plus levonorgestrel 0.15 mg, ethinyl estradiol 35 micrograms plus norethindrone 1 mg, and ethinyl estradiol 35 micrograms plus norethindrone 0.5 and 1 mg (biphasic). For all groups, total cholesterol increased 5.9-9.1% from baseline values over the six months. Triglycerides increased with all preparations, with the ethynodiol diacetate group (37.6%) and the biphasic norethindrone group (45.3%) showing the greatest increase. Low-density lipoprotein cholesterol increased 10-15.6% among the groups; low-density lipoprotein-apolipoprotein B changed proportional to the low-density lipoprotein cholesterol increases. All groups except the ethynodiol diacetate group showed a decrease of high-density lipoprotein cholesterol, with the levonorgestrel group (8.7%) and biphasic norethindrone group (4.5%) showing the largest declines. Apolipoprotein A-1 increased in all groups, with the ethynodiol diacetate preparation (19.3%) showing the greatest increase and the levonorgestrel preparation (3.2%) showing the smallest increase from baseline values. The changes in apolipoprotein A-1 were out of proportion to the changes in high-density lipoprotein cholesterol, suggesting that the high-density lipoprotein particle may be undergoing some type of metabolic alteration.

Lipid and lipoprotein changes in relation to oral contraception and hormonal replacement therapy.

Estrogens and progestins for contraception or hormonal replacement therapy are widely used by practitioners. These steroids have substantial effects on lipids and lipoproteins that appear to be primarily related to chemical structure of the compound, dosage, and a patient's hormonal status. Although the mechanisms by which alterations in lipid and lipoproteins affect atherogenesis are not fully understood, epidemiologic studies clearly associate alterations with risk of coronary heart disease. Attention to these alterations by progestins and estrogens, as well as further research on how these steroids may exert other cardiovascular effects, is important because atherosclerotic heart disease is a major cause of morbidity and mortality for women as they age.

PIP: The use of estrogens and progestins for contraception for hormonal replacement therapy has substantial effects on lipids and lipoproteins. Elevations of lipoproteins such as total cholesterol, low density lipoprotein (LDL)-C, and apoprotein B, and reduced levels of others such as high density lipoprotein (HDL)-C and apoprotein A-1, are associated with increased risk of coronary heart disease, although the exact mechanism by which such changes contribute to atheroma formation is unknown. The effect of an estrogen or progestin on the lipid profile is dependent on the chemical structure, dose, and route of administration as well as the hormonal status of the patient. Moreover, when an estrogen and progestin are combined in a therapeutic regimen, additional interactions occur that further complicate understanding of this process. Some epidemiological studies have suggested that mechanisms other than those closely associated with the lipid/lipoprotein system may be operant in the development of coronary heart disease and that the protective effects of estrogen far outweigh any adverse effects of progestin. Other studies have found an increased risk of coronary heart disease among past oral contraceptive users. Since atherosclerotic heart disease is a major cause of morbidity and mortality for women as they age (a cumulative mortality for women 55-75 years of age of 10,500/100,000), further research on the effects of steroids on cardiovascular factors is urgently needed.

Effect of low-dose oral contraceptive on gonadotropins, androgens, and sex hormone binding globulin in nonhirsute women.

The effectiveness of a low-dose oral contraceptive (OC) in suppressing plasma levels of gonadotropins, ovarian, and adrenal androgens and stimulating sex hormone-binding globulin (SHBG) was evaluated prospectively in nonhirsute women. Thirty-three women ingested 35 micrograms of ethinyl estradiol and 1 mg of norethindrone beginning within day 1 to 5 of the menstrual cycle. Baseline levels of luteinizing hormone, follicle-stimulating hormone, total testosterone (T), androstenedione (A), dehydroepiandrosterone sulfate (DHEAS), and SHBG were obtained before ingestion of the OC and repeated after 3, 6 and, 9 months of OC use on day 1 to 5 of the OC "cycle". A significant suppression of gonadotropin levels is seen in nonhirsute women. Sex hormone binding globulin is consistently stimulated by the low-dose OC. A significant suppression of T and DHEAS is observed. No change was seen in levels of A. The demonstrated effects become evident at 3 months and are maintained at 6 and 9 months.

Intrauterine devices and pelvic inflammatory disease: evolving perspectives on the data.

PIP: In order to present an overview of current understanding about the relationship between pelvic inflammatory disease (PID) and IUD use, this paper reviews the evolution of pertinent studies and identifies study design limitations. The first section includes a tabulated outline of the initial epidemiological studies that indicated an association between PID and IUD use (1968-80) and describes the biases in these studies. Section 2 considers the contribution of the US Women's Health Study, the largest case-control study of the relationship, which addressed many of these concerns and indicated that the risk of IUD-associated PID was primarily related to insertion, that the Dalkon shield increased risk, and that monogamy reduced risk. A table illustrates relative risk found by this study for various scenarios. Next, the two analyses of data (1981 and 1990) from the Oxford Family Planning Association Study are reviewed and risk data from both publications are tabulated. Section 4 looks at the World Health Organization's 1992 study that used data collected in 12 randomized and one nonrandomized clinical trial from centers throughout the world. Data are tabulated to show rate of PID for various types of IUDs and two durations of use. This study confirmed that greatest risk is associated with insertion. The paper then provides an analysis of each of the following current conclusions regarding PID risk and IUD usage: 1) risk is primarily confined to time of insertion (except for the Dalkon shield); 2) in women with no risk factors, the IUD poses little risk of tubal infection; 3) screening for IUD use may reduce absolute risk; 4) copper IUDs (especially the copper T380A) pose lower risk; and 5) Actinomyces israelii is not thought to be a major pathogen in this clinical setting.^ieng

Ectopic pregnancy and prior induced abortion.

As part of the Women's Health Study, a case-control study conducted in nine cities in the United States, women hospitalized with an ectopic pregnancy and women hospitalized with non-gynecologic, medical or surgical diagnoses were interviewed concerning past reproductive history. There were 462 women meeting eligibility criteria in the ectopic pregnancy case group and 2326 women meeting the criteria for the control group. After adjustment for a number of possible confounders, the relative risk of ectopic pregnancy for women with a history of one induced abortion was 1.0 (95% confidence limits: 0.5 to 1.8) and was 0.9 (95% confidence limits: 0.8 to 1.1) for women with a history of two or more prior induced abortions. These results suggest that prior induced abortion does not significantly increase the risk of subsequent ectopic pregnancy.

PIP: A case-control study conducted in 9 US cities as part of the Women's Health Study suggests that prior induced abortion is not associated with a subsequent markedly increased risk of ectopic pregnancy. Study subjects included 462 women hospitalized with an ectopic pregnancy at 16 hospitals in 1976-78. Controls included 2326 women hospitalized during the same period with nongynecologic, medical, or surgical diagnoses. After adjustment for several confounders that are known risk factors for ectopic pregnancy (i.e., race, a history of prior pelvic inflammatory disease, and prior ectopic pregnancy), the relative risk of ectopic pregnancy for women with a history of 1 induced abortion was 1.0 (95% confidence limits, 0.5-1.8). Among women with a history of 2 or more induced abortions, the relative risk of ectopic pregnancy was 0.9 (95% confidence limits, 0.8-1.1). Several earlier studies have suggested a potential increased risk of ectopic pregnancy among women with a history of abortion; however, these studies have generally failed to control for important risk factors, have involved populations from countries that are substantially different demographically from the US, and have included substantial numbers of women whose abortion procedures were illegal.

Morbidity risk among young adolescents undergoing elective abortion.

A cohort study comparing risks of morbidity at the time of elective abortion between 399 young adolescents aged 17 years or less matched to 399 women aged 20 to 29 years was completed. The study matched for duration of gestation, types of procedure, parity, race, month of procedure, and insurance status. Results indicated a relative risk of post-abortal endometritis for adolescents compared to older women of 2.5 (95% C.L. 1.48, 3.85). The reasons for this difference are unclear.

PIP: To determine whether younger teenagers (age 17 years or less) face differing morbidity than other women, a cohort study comparing risks of morbidity at the time of elective abortion between 399 young adolescents aged 17 years or less matched with 399 women aged 20-29 years was undertaken. Only women undergoing suction curettage up to 14 weeks duration of gestation or urea amnioinfusion in combination with intraamniotic prostaglandin F2alpha or intravenous oxytocin at 16-22 weeks duration of gestation were included. Cohert subjects were women who had undergone their abortion care in Fertility Control Center at Johns Hopkins Hospital (Baltimore, Maryland) between January 1, 1976 December 31, 1978. The occurence of endometritis was the only complication found more frequently among the adolescents to a significant degree. Among the endometritis patients, 7 adolescents and 2 control women were hospitalized for intravenous antibiotic therapy. The study indicates that adolescents undergoing abortion procedures are at increased risk for postabortal endometritis once other factors such as gestational age, type of abortion procedure, parity, race, level of operation traning, and socioeconomic status have been controlled. Cervical laceration has been shown previously to be about twice as common among women aged 17 years or less compared with older women. The current study demostrated a similar trend but not to a statistically significant degree. Reasons for the increased rated of postabortal endometritis are unclear. The study demonstrated increased rates of preexisting cervical gonorrhea and urinary tract infection among the adolescents compared with the older women, but all of these women underwent treatment prior to the procedure and none experienced endometritis. Careful monitoring after the procedure for the occurence of infection is of obvious importance. Whether routine use of prophylactic antibiotics in this group would be of substantial benefit may deserve further study.

Use of osmotic dilators to facilitate induced midtrimester abortion: clinical evaluations.

A number of studies evaluating an osmotic cervical dilator consisting of polyvinyl foam saturated with magnesium sulphate in women undergoing midtrimester abortion with intra-amniotic hyperosmolar urea plus prostaglandin F2a were completed. Comparisons with women receiving no pre-treatment with a laminaria tent or with one laminaria indicate that their use appears to shorten injection-abortion intervals, particularly in parous women, and reduce risk of endometritis and cervical laceration compared to women not receiving any type of device. The data suggests that two osmotic dilators may be more effective than one. Also, magnesium toxicity does not appear to be a substantial risk with their use.

The interaction of alcohol consumption and oral contraceptive use on lipids and lipoproteins.

Oral contraceptive (OC) use and alcohol consumption have been shown to alter the levels of lipids and lipoproteins in the blood. The effect of alcohol consumption on levels of total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, LDL-B, Apo-A1, the ratio of HDL cholesterol/total cholesterol, HDL cholesterol/LDL cholesterol, and the ratio of LDL cholesterol/LDL-B among normal healthy young women before initiation of oral contraceptives and after six months of oral contraceptive use are both described. Of primary interest is the mediating effect of alcohol consumption on the association between steroid usage and blood lipid values. At baseline, ethanol consumption was found to be positively associated with triglycerides, HDL-C, and Apo-A1 and negatively associated with LDL-C/LDL-B. After adjustment for several covariables, alcohol consumption was found to be positively associated with the increases in triglycerides and in Apo-A1 observed at 3 and 6 months after initiation of OCs. Since these two parameters are believed to have opposite relationships to cardiovascular disease, the effect of alcohol consumption remains uncertain.

PIP: The effects of alcohol consumption and oral contraceptive (OC) use on lipid and lipoprotein parameters were analyzed among 267 healthy women. Levels of cholesterol, low density lipoprotein (LDL)-B, triglycerides, LDL-cholesterol, and apolipoprotein A1 increased over the 6 months following OC initiation, while levels of high density lipoprotein (HDL)-cholesterol and the ratios HDL-cholesterol/LDL-cholesterol, HDL-cholesterol/cholesterol, and LDL-cholesterol/LDL-B tended to decrease with time. The greatest proportion of overall OC-related changes occurred between the baseline and 3-month measurement. Also observed was an association between baseline ethanol consumption and lipid parameters. Alcohol consumption was positively associated with triglycerides, HDL-cholesterol, and apolipoprotein A1 and negatively associated with LDL-cholesterol/LDL-B. After adjustment for several covariables, alcohol consumption was found to be positively associated with the increases in triglycerides and in apolipoprotein A1 observed at 3 and 6 months after initiation of OC use. The fact that apoliproprotein A1 levels increased with high levels of ethanol intake without a concurrent increase in HDL-cholesterol probably reflects substantial compositional changes in HDL. Overall, these findings suggest that women who initiate OC use can expect that the resulting increases in triglycerides and apolipoprotein A1 will be enhanced by baseline alcohol consumption. Since very high levels of triglycerides and low levels of apolipoprotein A1 are believed to be risk factors for cardiovascular disease, the net effect of alcohol consumption remains uncertain.

Bacteriologic culture results obtained before and after elective midtrimester urea abortion.

Twenty patients undergoing elective midtrimester abortion by urea amnioinfusion had bacteriologic culture data obtained before and after the procedure. Despite the presence of pathogens in the vagina and cervix prior to the procedure, only 20% of patients exhibited endometrial growth following abortion. Factors that seem to favor such endometrial growth include increased injection-abortion intervals and the requirement for curettage to manage incomplete abortion.

Oral contraceptives and lipids and lipoproteins: Part I--Variations in mean levels by oral contraceptive type.

A randomized clinical trial of oral contraceptives evaluated 67 women on 50 micrograms ethinyl estradiol (EE) and 1.0 mg norethindrone (NE), 61 women on 35 micrograms EE and 1.0 mg NE, and 64 women on 35 micrograms EE and 0.5 mg NE. Fasting lipids and lipoproteins were measured at baseline, three, six and nine months. All groups showed an increase in plasma total cholesterol, triglycerides, low-density lipoprotein/apolipoprotein B and apolipoprotein A-1. The group taking the preparation with 0.5 mg of NE was the only one to result in an elevation of high-density lipoprotein cholesterol; the other two groups showed declines in the mean levels of this lipoprotein over the study time period. Mean changes in lipid/lipoprotein levels associated with oral contraceptive use appear to be at least partially related to the doses of the contraceptive steroids.

Oral contraceptives and lipids and lipoproteins: Part II--Relationship to plasma steroid levels and outlier status.

A randomized clinical trial of oral contraceptives evaluated 67 women on 50 ug ethinyl estradiol (EE) and 1.0 mg norethindrone (NE), 61 women on 35 ug EE and 1.0 mg NE, and 64 women on 35 ug EE and 0.5 mg NE. At baseline, three, six and nine months, lipids and lipoproteins were measured as well as selected plasma determinations of the contraceptive steroids. Data was related to change in outlier status for lipids/lipoproteins (less than 10th percentile for high-density lipoprotein cholesterol and apolipoprotein A-1, greater than 90th percentile for all others). Women on the lowest dose preparation had the smallest trend towards outlier status for high-density lipoprotein cholesterol and apolipoprotein A-1. An increase over the time period of the study in the initial slope (one hour level minus zero hour level over time) of NE and a decrease in the initial slope of EE was associated with a shift to outlier status for low-density lipoprotein cholesterol. A number of other relationships were also shown. Use of the techniques described might assist in identifying sub-groups of women at risk for adverse cardiovascular sequelae associated with oral contraceptive use.