Intrauterine and subdermal progestin administration postmenopausal hormone replacement therapy

FERTILITY AND STERILITY Vol. 63, No.2, February 1995 Copyright c 1995 American Society for Reproductive Medicine Printed on acid-free paper in U. S. A. Intrauterine and subdermal progestin administration postmenopausal hormone replacement therapy In Satu P. Suhonen, M.D.*t Teddy Holmstrom, M.D.* Hannu O. Allonen, M.D.:j: Pekka Liihteenmiiki, M.D.t Helsinki City Maternity Hospital, and University of Helsinki, Helsinki, Finland Objective: To compare the effects of intrauterine and subdermal administration of levonorgestrel on control of bleeding and on the endometrium in postmenopausal hormone replacement therapy. Interventions: Nineteen women started continuous oral E2 valerate therapy (2 mg daily) together with continuous parenteral progestin therapy. The subjects randomly received either a subdermallevonorgestrel-releasing implant (n = 9) or an intrauterine device (IUD) releasing levonorgestrel (n = 10). Main Outcome Measures: Serum concentrations of estrone, E2, FSH, sex hormone-binding globulin (SHBG) and levonorgestrel were followed. Endometrial biopsies and transvaginal ultrasonography were used to evaluate the endometrium. The subjects kept daily records of bleeding. The observation time was 1 year. Results: Serum concentrations of the hormones mentioned above and SHBG were similar in both groups during the observation time, but the patterns of bleeding differed. In the IUD group there were 0.9 days (mean, range 0 to 4 days) of spotting and no days of bleeding during the last month of the follow-up year. In the implant group there were 8 days (mean, range 0 to 25 days) of spotting and 3.4 days (mean, range 0 to 14 days) of bleeding. In histological examination there was uniform atrophy in the endometrial samples from the IUD group, and a weak or absent progestin effect in the implant group. Conclusions: In spite of similar serum levonorgestrel concentrations, local intrauterine administration of levonorgestrel resulted in better control of bleeding and in more effective endometrial suppression than subdermal administration. Fertil Steril 1995;63:336-42 Key Words: Continuous combined hormone replacement therapy, levonorgestrel, intrauterine levonorgestrel-releasing device, intrauterine progestin therapy, parenteral progestin therapy, endometrial suppression Received April 11, 1994; revised and accepted August 22, 1994. * Helsinki City Maternity Hospital. t Steroid Research Laboratory, Department of Medical Chern istry, University of Helsinki. :j: Leiras Oy, Turku, Finland. Reprint requests: Pekka Liihteenmiiki, M.D., Steroid Re search Laboratory, Department of Medical Chemistry, P.O. Box 8, SF-00014 University of Helsinki, Helsinki, Finland (FAX: 358-0-191-8276). The aim of continuous progestin administration in hormone replacement therapy (HRT) is suppression of the endometrium and the elimination of bleeding (1, 2). However, during the first months of treatment, irregular bleeding occurs in 30% to 50% of women (1-5), and even more frequently in perimenopausal women (2). Adjusting the dose of progestin has been suggested to overcome the problems associated with irregular bleeding (3). However, no significant differences in patterns of bleeding were observed when two doses of medroxyprogesterone acetate (4), norethisterone acetate and megestrol acetate (5) were compared. If postmenopausal women are using continuous combined HRT regi- 336 Suhonen et al. Parenteral use of levonorgestrel in HRT Fertility and Sterility

mens, the patterns of bleeding become more acceptable along with time and amenorrhea is reached after 1 year in 77% to 95% of women (1-5). Most of the continuous combined regimens are oral. In sequential HRT, transdermal progestin is available (6) and other parenteral progestin alternatives are being investigated (7-9). In clinical practice in which oral continuous progestin therapy is used, the daily administration and the possibility of irregular bleeding can be regarded as disadvantages. Long-term, parenteral use ofprogestins with sustained release has been achieved successfully in contraception using polydimethylsiloxane (Silastic) as the carrier material. Silastic implants releasing levonorgestrel (N orplant; a registered trademark of The Population Council, Inc. for subdermal contraceptive implants) have been used with effective lifetimes of either 3 years (Norplant II, a set oftwo implants) or 5 years (Norplant, a set of 6 implants). The same polymer, polydimethylsiloxane, also has been used as a carrier for levonorgestrel in IUDs. It has been possible to extend the steroid release beyond 5 years (10) in intrauterine administration. In contraceptive use the levonorgestrel IUD has been shown to induce suppression of the endometrium (11), with a significant reduction of bleeding (12). Endometrial suppression with a high incidence of amenorrhea also has been observed when this IUD has been used in HRT (13). The purpose of this study was to compare two continuous, parenteral levonorgestrel-releasing alternatives, the primary interest being on the endometrium. Endometrial histology, patterns of bleeding and effects on endometrial thickness evaluated by ultrasound (US) were followed. MATERIALS AND METHODS The study was carried out at Helsinki City Maternity Hospital, Helsinki and the study protocol was approved by the ethics committee of the hospital. Nineteen volunteer women seeking treatment for their menopausal symptoms participated in the study, and they all gave written informed consent. Hysterectomized women or women with uterine anomaly were excluded. Otherwise the general contraindications for HRT were taken into account. Study Treatments To compare the patterns of bleeding during sequential oral and continuous parenteral use of the same progestin combined with oral E, oral sequen- tial HRT treatment (Cyclabil; Leiras Oy, Turku, Finland [license Schering AG Pharma, Berlin, Germany 1) was used for a minimum of 3 months. This regimen consisted of 2 mg E2 valerate daily for 21 days, together with 250,."g levonorgestrel daily from the 12th to the 21st day. A treatment-free month followed the last cycle on oral therapy. The proper study treatment was then started (baseline). According to the progestin component the women were allocated randomly into two groups using a randomization code. Nine women received one subdermal implant (Norplant II; Leiras Oy) containing 70 mg levonorgestrel. The initial daily release of levonorgestrel from this implant has been calculated to be 40,."g, with stabilization to 15,."g during the second half of the first year of use. The implant was inserted under local anesthesia in the upper arm. The other group (n = 10) received an IUD releasing 20,."g levonorgestrel daily (Leiras Oy). All subjects received oral continuous estrogen therapy: 2 mg E2 valerate daily (Progynova; Leiras Oy [license Schering AG Pharma]). The estrogen therapy was started at the same as the parenteral progestin therapy. Blood Samples and Laboratory Analyses Venous blood samples for determination of the serum concentrations of estrone (El), E2, FSH, sex hormone-binding globulin (SHBG), and levonorgestrel were taken during the last week of the last Cyclabil cycle, i.e., during the progestin phase (-1). The blood samples were taken within 8 to 12 hours of ingestion of the Cyclabil tablets because serum levonorgestrel concentrations have been shown to be rather stable after 6 hours from tablet intake (14). After the treatment-free month (baseline) serum El> E 2, FSH, and SHBG concentrations were measured. During the continuous combined therapy serum concentrations of E l, E2, FSH, SHBG, and levonorgestrel were determined after the 1st, 6th, and 12th months of treatment. The same timing for sampling was followed as during the sequential therapy, i.e., 8 to 12 hours after ingestion of E2 valerate tablets. Serum was separated by centrifugation and stored at -20 C until analyzed. All samples for one particular analysate were run in the same assay. Serum concentrations of El and E2 were measured by RIAs, using protocols from the World Health Organization (15), with diethyl ether extraction and no chromatography. Serum concentrations offsh and SHBG were measured by timeresolved fluoroimmunoassays (Delfia; Wallac Oy, Vol. 63, No.2, February 1995 Suhonen et a1. Parenteral use of levonorgestrel in HRT 337

Turku, Finland). The specificities, sensitivities, accuracies, and precisions of the analyses have been described recently (16). Serum levonorgestrel concentrations were determined by RIA as described by Weiner and Johansson (17). Tritiated levonorgestrel ([15,16-3 H]-levonorgestrel, specific activity 2,613.7 MBqfmg [22 Ci/mmol]) was obtained from Schering AG (Berlin, Germany). The detection limit of the standard curve in the levonorgestrel assay was 7 pg/ml (22 pmoljl). However, due to a small serum blank, the assay sensitivity was set at 50 pg/ml (160 pmoljl). The intraassay coefficient of variation (CV) was 3.6% to 5.4%. The interassay CV was 9.1 % to 13.1 %. In the levonorgestrel assay, the recovery was 96%. The Endometrium and Patterns of Bleeding To evaluate endometrial status, transvaginal ultrasonography and endometrial biopsies were used. These examinations were performed after the offtreatment month (baseline) and then 6 and 12 months after insertion of the levonorgestrel implant or levonorgestrel IUD. Transvaginal ultrasonography also was performed during the last week of the oral sequential therapy (-1). Ultrasound examination always preceeded the endometrial biopsy. The US examinations were carried out by a single investigator, and an Aloka Echo Camera SSD-550 scanner (Aloka Co., Ltd., Tokyo, Japan) and a 5-MHz vaginal probe was used. The thickness of the endometrium was evaluated from the longitudinal plane and both endometrial layers were ineluded (double-layer). In subjects with the levonorgestrel IUD, the measurement was performed near the fundal area as the shadow due to the vertical stem of the levonorgestrel IUD could thus be avoided. The endometrial samples were obtained with a curette designed for Vabra suction curettage (Ferrosan, Copenhagen, Denmark), but no suction was used. The samples were drawn around the uterine cavity to get a representative sample. The biopsies were placed in formolacetic acid immediately upon collection and after staining with hematoxylin and eosin, they were all examined by the same pathologist. The pathologist was not aware of the type of progestin therapy. In histological examination, the emphasis was on the intensity of the progestin effect on the endometrium. Abnormalities in endometrial morphology were reported separately. Transvaginal US examination, the endometrial biopsy, and all the laboratory tests also were carried out in cases of premature discontinuation. Each patient had a card on which eventual bleeding (menstrual-like bleeding) or spotting (bloody vaginal discharge not necessitating the use of pads) were recorded daily. Bleeding and spotting per cyele of oral treatment were recorded, as well as eventual bleeding during the off-treatment month. During the parenterallevonorgestrel therapy, the patterns of bleeding were analyzed as numbers of days of both bleeding and spotting in 30-day periods, starting from insertion of the IUD or implant. If the treatment was discontinued before completion of the 1-year follow-up time, the last whole 30-day period was taken into account. The definition for amenorrhea in this study was a period of 3 months without a single day of bleeding or spotting; this situation continuing thereafter for the remaining follow-up time. Statistics Statistical analyses were conducted using the SAS System (18) (SAS Institute Inc., Cary, NC). In this system, the data on bleeding were analyzed using Fisher's exact test, the Mann-Whitney U-test and Friedman's test. The histological results were analyzed using Fisher's exact test. The data from US examinations were analyzed using analysis of variance analysis for repeated measures, which also was used for analyses of serum concentrations of El> E 2, FSH, levonorgestrel, and SHBG. Logarithmic transformation was used if deviation from a normal distribution was found. A P value of <0.05 was considered statistically significant. RESULTS The groups did not differ in age, in body mass index (BMI), in time since last menstruation, or in serum FSH concentrations at baseline. In the levonorgestrel-iud group, the mean age was 53.8 ± 1.4 years (SE), the mean BMI 24.6 ± 1.0 (kg/m 2 ), the mean time from last menstruation 4.6 ± 1.4 years, and the mean serum FSH concentration at baseline 59.2 ± 8.5 miu/ml (conversion factor to SI unit, 1.00). The respective figures for the implant group were 50.9 ± 1.3 years, 25.6 ± 1.3 (kg/ m 2 ), 2.4 ± 1.0 years, and 46.3 ± 8.6 miu/ml. One subject in the levonorgestrel-iud group had heavy bleeding lasting for 25 days after insertion of the IUD. In the following controls her serum concentrations of levonorgestrel were undetectable «50 pg/ml [<160 pmoljl]), indicating unnoticed expulsion of the levonorgestrel IUD. This subject 338 Suhonen et a1. Parenteral use of levonorgestrel in HRT Fertility and Sterility

80 ::l -60 1 ~ 40 ~ rn 20 O+---~--~--r---r-~~-' o 2 4 6 8 10 12 Time (months) Figure 1 Serum SHBG concentrations (nmol/l) in the levonorgestrel-implant group (_) and the levonorgestrel-iud group (0) during the follow-up period. Data are presented as means ± SE. was excluded from the analysis. Hence, results from 18 women (i.e., 9 in the levonorgestrel-implant group and 9 in the levonorgestrel-iud group) are reported. Neither serum E 1, E 2, nor FSH concentrations differed between the groups during the follow-up time. Serum SHBG concentrations did not differ between the groups (Fig. 1) but rose significantly during the 12-month observation time. In the levonorgestrel-iud group, SHBG concentrations at baseline were 53.7 ± 7.4 nmol/l and at the 12- month control, 77.2 ± 10.8 nmol/l (P = 0.026). In this group, SHBG concentrations after the 3- month oral therapy were lower than after intrauterine use ofthe same progestin for 1 year (P = 0.011). In spite of different administration routes and different daily estimated releases, serum levonorgestrel concentrations were in the same range in both groups (P = 0.23; Fig. 2). During parenteral progestin therapy, serum levonorgestrel concentrations remained stable and were significantly lower than during the period on oral therapy at each time of evaluation (P < 0.0001). During the oral sequential therapy there were means of 2.7 days of bleeding (range 0 to 7 days) and 2.9 days of spotting (range 0 to 12 days) per cycle, when the recordings of three treatment cycles were combined. During the off-treatment month no bleeding was observed. In the implant group, during the follow-up year no significant changes in the numbers of days of bleeding and spotting were observed (Fig. 3). In the IUD group, during the observation time, the numbers of days of bleeding and spotting decreased significantly, compared with both the time on oral sequential therapy (P = 0.0008) and with the first 3-months use of the levonorgestrel IUD (P = 0.0005). When the I-year observation period was completed, five of the remaining eight women in the IUD group had had amenorrhea lasting for ~4 months (range 4 to 11 months). Only one of the seven women in the implant group had had the 2 last months without bleeding or spotting. The thickness of the endometrium as evaluated by US did not differ between the groups at any visit (Table 1). Compared with the ultrasonographic findings during oral sequential therapy, the endometrium was thinner at each evaluation during the use of parenteral progestin therapy. Two subjects, both from the implant group, had an endometrial thickness ~ 5 mm. One showed these findings at both the 6-month (endometrial thickness 5 mm) and 12-month visits (endometrial thickness 6 mm); the other only at the 12-month control (endometrial thickness 6 mm). Histological examination revealed cystic hyperplasia in one of the baseline samples from a subject in the implant group. Because the histological information came after starting the study treatment and because the endometrial thickness at baseline and this subject's previous bleedings were not suspicious, she was allowed to continue in the study. The next samples showed a weak estrogen effect and an absent progestin effect, but no hyperplastic changes. Only in 3 of 18 biopsy samples during the 2500 2000 ~ 1500 -bo c.. '-' 3 1000 500 0-1 baseline 1 6 12 Time (months) Figure 2 Serum levonorgestrel concentrations (pg/ml) during oral sequential therapy (-1) and during the follow-up year in the levonorgestrel-implant group (_) and in the levonorgestrel IUD group (!!!il). Data are presented as means ± SE (conversion factor to SI unit, 3.195). Vol. 63, No.2, February 1995 Suhonen et al. Parenteral use af levanargestrel in HRT 339

15... 10 ~ c:l 5 A irregular bleeding: one after 8 months and the other after 10 months. The former woman was the subject with a finding of cystic hyperplasia at baseline biopsy, the latter, the one mentioned above with a finding of a weak progestin effect on the endometrium at termination. Suction curettage was carried out in these two cases at termination. No endometrial pathology was noticed. In ultrasonography at termination, the first subject had an endometrial thickness of 5 mm; the other, 2 mm. 0 15 8 10 5 o 1 B 2 3 4 5 6 7 8 9 10 11 12 2 3 4 5 6 7 8 9 10 11 12 Time (months) Figure 3 Numbers of days of spotting (~) and bleeding (_) per periods of 30 days in the levonorgestrel-implant group (A) and in the levonorgestrel-iud group (B). Data are presented as means ± SE. whole follow-up period was a progestin effect observed in the levonorgestrel-implant group. Two of these observations were at the 12-month control and one at a premature termination because of irregular bleeding. In the endometrial biopsies of the two subjects mentioned above who had an endometrial thickness of 5 to 6 mm, neither progestin effect nor hyperplastic signs were observed. In contrast, the levonorgestrel IUD induced a strong progestin effect and atrophy ofthe endometrium uniformly in all samples as soon as after the first half of the study. No clear association between the endometrial thickness (US) and proliferation of the endometrium (histological examination) was observed. There were three discontinuations during the year. One woman in the IUD group decided to end the study treatment after 3.5 months because of vertigo and headache. These symptoms, however, continued even after termination. Two women in the implant group terminated the study because of DISCUSSION Progestins are recommended generally to nonhysterectomized women receiving estrogen replacement therapy (ERT), to protect the endometrium (19). Thus, a sufficient local effect on the endometrium with minimal systemic effects is the most important characteristic of the progestin component in HRT. During the use of levonorgestrel-releasing IUDs in premenopausal women, high locallevonorgestrel concentrations in the endometrium have been reported (20). In this study by Nilsson et al. (20), the endometrial concentrations during the use of 250 Ilg levonorgestrel orally were <1 % of the tissue concentrations during intrauterine levonorgestrel administration. When levonorgestrel is used in sequential regimens, 250 Ilg/d is the most commonly used dose, even though an oral dose of 75 Ilg has been found to counteract the estrogen stimulus on the endometrium (21). In continuous combined regimens, lower progestin doses than in sequential regimens are sufficient to suppress the endometrium. When the implant, with an approximate daily release of 15 Ilg levonorgestrel was used, the mean serum levonorgestrel concentrations were 17% of the concentrations in same women during oral therapy with 250 Ilg levonorgestrel daily. Based on patterns of bleeding and endometrial histology, our results suggest that the tissue concentrations Table 1 Thickness of the Endometrium in Transvaginal Ultrasonography in Both Groups* Levonorgestrel Levonorgestrel Time implant IUD mm -1 mo 6.2 ± 0.7 5.2 ± 1.1 Baseline 2.1 ± 0.4 2.2 ± 0.7 +6mo 3.1 ± 0.5 3.3 ± 0.2 +12mo 2.4 ± 0.6 3.0 ± 0.4 * Values are means ± SE. mm 340 Suhonen et ai. Parenteral use of levonorgestrel in HRT Fertility and Sterility

achieved with this continuous subdermal progestin dose were insufficient to counteract the estrogen stimulus. In the levonorgestrel-iud group, the numbers of days of bleeding and spotting decreased through the observation year. The patterns of bleeding were quite similar to the oral continuous combined regimens: bleeding and spotting observed in the beginning became less frequent with time and amenorrhea could be reached in majority of women. However, the effective endometrial suppression, as judged by histological examination, and the possibility of long-term use of the levonorgestrel IUD once it has been inserted can be regarded as the major advantages of this intrauterine progestin administration. In contraceptive use, the recommended time of use for a single levonorgestrel IUD is 5 years (10) and the typical histological characteristics in the endometrium induced by the levonorgestrel IUD are maintained at least for 7 years (11). After insertion of a levonorgestrel-releasing IUD, in fertile women serum levonorgestrel concentrations will stabilize in a few weeks to a level of 94 to 188 pg/ml (300 to 575 pmol/l) (10). In the present study, with postmenopausal women, serum levonorgestrel concentrations were markedly higher than during the use of the same IUD in fertile women. The profile of serum concentrations during the I-year follow-up period was also different. In our study, serum levonorgestrel concentrations remained stable and were still elevated at the last control (317 ± 37 pg/ml [1,012 ± 118 pmol/l]). Levonorgestrel is known to induce a decline in serum SHBG concentrations (22) and on the other hand, oral ERT is known to induce an increase (23). No decline in serum SHBG concentrations was observed; rather there was an elevation. This elevation can be explained by the effect of oral estrogen treatment in inducing an elevation in serum SHBG concentrations through the first-pass mechanism in the liver. The higher serum levonorgestrel concentrations in women participating in this study compared with previous data on women of fertile age may thus reflect the effects of oral estrogen administration. Transvaginal ultrasonography did not reveal differences in endometrial thickness between the groups, nor did it predict the patterns of bleeding. The almost uniformly absent progestin effect in the implant group did not correlate with endometrial thickness. An endometrial thickness of 4 to 5 mm has been regarded as being the limit under which hyperplastic or malignant changes seldom exist (24, 25). Two women, both from the levonorgestrel-implant group showed findings exceeding this limit, combined with an absent progestin effect. With the levonorgestrel IUD, the uniform histology observed during intrauterine administration suggests that a sufficient endometrial response as regards HRT possibly could be achieved with a lower daily release and thus lower serum concentrations of levonorgestrel. The intrauterine administration ofprogestins in HRT seems to be a logical route where safety and acceptability, and thus better compliance, can be combined. Acknowledgments. We thank the staff of the laboratory, especially Ms. Eeva Harju, Ms. Sirpa Ranta, and Ms. Marjatta Tevilin, for their technical expertise and help. REFERENCES 1. Staland B. Continuous treatment with natural estrogens and progestogens. A method to avoid endometrial stimulation. Maturitas 1981;3:145-56. 2. Mattsson L-A, Cullberg G, Samsioe G. Evaluation of a continuous oestrogen-progestogen regimen for climacteric complaints. Maturitas 1982;4:95-102. 3. Magos AL, Brincat M, Studd JWW, Wardle P, Schlesinger P, O'Dowd T. Amenorrhea and endometrial atrophy with continuous oral estrogen and progestogen therapy in postmenopausal women. Obstet Gynecol 1985;65:496-9. 4. Weinstein L, Bewtra C, Gallagher JC. Evaluation of a continuous combined low-dose regimen of estrogen-progestin for treatment of the menopausal patient. Am J Obstet Gynecol 1990;162:1534-42. 5. Sporrong T, Hellgren M, Samsioe G, Mattsson L-A. Comparison of four continuously administered progestogen plus oestradiol combinations for climacteric complaints. Br J Obstet Gynecol. 1988;95:1042-8. 6. Whitehead MI, Fraser D, Schenkel L, Crook D, Stevenson JC. Transdermal administration of oestrogen/progestagen hormone replacement therapy. Lancet 1990;335:310-2. 7. Suhonen S, Haukkamaa M, Liihteenmiiki P, Holmstrom T. Endometrial effect of transdermal estradiol and progestin ST-1435 in postmenopausal women. Fertil Steril 1992;57:1211-5. 8. Cicinelli E, Cignarelli M, Resta L, Scorcia P, Petruzzi D, Santoro G. Effects of the repetitive administration of progesterone by nasal spray in postmenopausal women. Fertil Steril 1993;60:1020-4. 9. Critchley HOD, Buckley CH, Anderson DC. Experience with a 'physiological' steroid replacement regimen for the establishment of a receptive endometrium in women with premature ovarian failure. Br J Obstet Gynecol 1990; 97:804-10. 10. Nilsson CG, Liihteenmiiki PLA, Luukkainen T, Robertson DN. Sustained intrauterine release of levonorgestrel over five years. Fertil SterilI986;45:805-7. 11. Silverberg SG, Haukkamaa M, Arko H, Nilsson CG, Luukkainen T. Endometrial morphology during long-term use of Vol. 63, No.2, February 1995 Suhonen et al. Parenteral use of levonorgestrel in HRT 341

levonorgestrel-releasing intrauterine devices. Int J Gynecol Pathol 1986;5:235-41. 12. Milsom I, Andersson K, Andersch B, Rybo G. A comparison of fiurbiprofen, tranexamic acid and a levonorgestrel-releasing intrauterine contraceptive device in the treatment of idiopathic menorrhagia. Am J Obstet Gynecol 1991; 164:879-83. 13. Andersson K, Mattsson L-A, Rybo G, Stadberg E. Intrauterine release of levonorgestrel - a new way of adding progestogen in hormone replacement therapy. Obstet Gynecol 1992;79:963-7. 14. Victor A, Edqvist L-E, Lindberg P, Elamsson E, Johansson EDB. Peripheral levels of d-norgestrel in women after oral administration of d-norgestrel and when using intravaginal rings impregnated with dl-norgestrel. Contraception 1975;12:261-78. 15. Sufi SB, Donaldson A, Jeffcoate SL. Programme for the provision of matched assay reagents for the radioimmunoassay of hormones in reproductive physiology. Method manual. London: World Health Organization, 1993. 16. Suhonen SP, Allonen HO, Liihteenmiiki P. Sustained-release subdermal implants: a new alternative in estrogen replacement therapy. Am J Obstet GynecoI1993;169:1248-54. 17. Weiner E, Johansson EDB. Plasma levels of d-norgestrel, estradiol and progesterone during treatment with Silastic implants containing d-norgestrel. Contraception 1976; 14:81-92. 18. SAS Institute Inc., SAS Technical Report P-229, SASj STAT Software: changes and enhancements, release 6.07. Cary (NC): SAS Institute Inc., 1992:620. 19. Whitehead M, Lobo RA. Progestagen use in postmenopausal women. Consensus Conference. Lancet 1988;2: 1243-4. 20. Nilsson CG, Haukkamaa M, Vierola H, Luukkainen T. Tissue concentrations of levonorgestrel in women using a levonorgestrel-releasing IUD. Clinical Endocrinol 1982; 17:529-36. 21. King RJB, Whitehead MI. Assessment of the potency of orally administered progestins in women. Fertil Steril 1986;46:1062-6. 22. Victor A, Johansson EDB. Effects of d-norgestrel induced decreases in sex hormone binding globulin capacity on the d-norgestrel levels in plasma. Contraception 1977;16:115-23. 23. Mashchak CA, Lobo RA, Dozoko-Takano R et al. Comparison of pharmacodynamic properties of various estrogen formulations. Am J Obstet GynecoI1982;144:511-8. 24. Granberg S, Wikland M, Karlsson B, Norstrom A, Friberg L-G. Endometrial thickness as measured by endovaginal ultrasonography for identifying endometrial abnormality. Am J Obstet GynecoI1991;164:47-52. 25. Varner RE, Sparks JM, Cameron CD, Roberts LL, Soong S-J. Transvaginal sonography of the endometrium in postmenopausal women. Obstet Gynecol 1991;78:195-9. 342 Suhonen et ai. Parenteral use of levonorgestrel in HRT Fertility and Sterility