Norplant® implants and progesterone vaginal rings do not affect maternal bone turnover and density during lactation and after weaning

S. Díaz1,3, M.V. Reyes1, A. Zepeda1, G.B. González2, J.M. López2, C. Campino2 and H.B. Croxatto1

1 Instituto Chileno de Medicina Reproductiva, Consultorio de Planificación Familiar, J.V. Lastarria 29, Depto. 101 and 2 Departamento de Endocrinología, Facultad de Medicina, Universidad Católica de Chile, Santiago, Chile


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Bone density and turnover was assessed in a longitudinal study of healthy lactating women who initiated use of Norplant® implants (NOR, n = 29), progesterone vaginal rings (PVR, n = 28) or Copper T 380A intrauterine devices (T-Cu, n = 51, control group) around day 60 postpartum. Bone density, serum calcium, phosphorus, alkaline phosphatases, parathyroid hormone (PTH), follicle stimulating hormone (FSH), oestradiol and prolactin, and urinary hydroxyproline and creatinine were measured at postpartum months 1 (PM1), and 12 (PM12) and 6 or 12 months after weaning; at month 6 postpartum (PM6) serum and urine tests alone were performed. Baseline characteristics and lactation performance were similar between groups. Biochemical markers of bone turnover were higher at PM1, PM6 and PM12 than after weaning, with no differences between groups. Bone density in the lumbar spine (L2–L4) and femoral neck at PM1 and PM12 (~1.11 g/cm2) was similar in three groups. Lumbar spine values were found to be lower in lactating women than those present in non-lactating women, but increased after weaning to similar values. The two progestin-only contraceptives studied appear to have no deleterious effect upon bone density and metabolism in healthy lactating women.

Key words: bone density/bone metabolism/lactation/Norplant® implants/progesterone rings


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Lactation is associated with changes in maternal bone mass and metabolism (Sowers, 1996Go). Full breast-feeding is characterized by an increase in maternal bone turnover and a transient bone loss that is fully reversed after weaning (López et al., 1996Go). The mechanisms involved in these changes are not well understood but are independent of contemporary changes in ovarian function, parathyroid hormone (PTH) concentration and calcium intake (López et al., 1996Go; Sowers, 1996Go; Kalkwarf et al., 1997Go).

Although non-hormonal contraceptives are considered to be the first choice for lactating women, progestin only methods are a good option when non-hormonal methods are contra-indicated or are not acceptable (IPPF, 1989). They have no deleterious effect upon breast-feeding, infant growth or early development (Díaz et al., 1997Go) but, to avoid early transference of small amounts of steroid to the infant, it is recommended that their use be initiated not before 6 weeks postpartum (WHO, 1996).

However, there are only limited data about the impact of progestin only contraceptive methods upon bone mass and turnover during breast-feeding (Prior, 1990Go). A small study carried out during the first postpartum year suggested that progestin only pills may protect against bone loss (Caird et al., 1994Go). Further information is required to enable adequate counselling to be given to lactating women about potential risks of osteoporosis related to contraceptive choice.

The progestins selected for the present study were progesterone and levonorgestrel. Progesterone plays a role in bone turnover and may be a bone trophic factor (Prior, 1990Go), despite its weak anti-oestrogenic effect. Levonorgestrel has a strong progestogenic effect and a weak androgenic activity (Phillips et al., 1987Go), and may also be a bone trophic factor. However, its potent anti-oestrogenic effect may also have deleterious consequences with regard to bone metabolism.

The purpose of this study was to characterize the changes in bone density and turnover which occur during breast-feeding and after weaning in healthy lactating women who used Norplant® levonorgestrel implants or progesterone vaginal rings as their contraceptive method. The control group consisted of users of the Copper T 380A intrauterine device.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subjects
Volunteers were healthy lactating women, 18–35 years old, parity 1–3 who had a normal pregnancy ending at 38–40 weeks gestation in vaginal delivery of a single healthy child of normal weight (2500–3850 g). All women were regularly co-habiting and none had a chronic condition requiring continuous drug therapy or that could be considered a contraindication for the contraceptive methods used in the study. They belonged to an urban population of low or low middle socio-economic status. All women wanted to breast feed their child for as long as possible and had planned not to become pregnant for at least 2 years after admission to the study.

Additional requirements were: body mass index (BMI) between 20 and 30 kg/m2; haemoglobin above 10 g/dl; first delivery after 20 years of age; smoking <20 cigarettes/week; alcohol ingestion <200 g/week; no chronic illness which could affect bone mass (kidney disease, hyper- or hypoparathyroidism, hyper- or hypothyroidism, etc.); and no consumption of drugs such as glucocorticoids, anticonvulsants, thiazides, calcium or thyroid hormones. All women had a sedentary lifestyle.

Admission and follow-up procedures
Volunteers were recruited in the first 2 days after delivery and follow-up visits were scheduled at days 7–10, 20 and 30 after delivery, at monthly intervals up to the end of the first year and at 3 month intervals thereafter. At each visit, the medical history of mother and infant and the bleeding and breast-feeding patterns were recorded. Pelvic and breast examinations were performed. Each child had a general physical examination, including determination of body weight. The clinical management of all subjects was similar, as described below.

Contraceptive management
During the second month postpartum, women were informed of the contraceptive methods available and of the purpose of the study, as well as of their freedom to participate or withdraw at any time, for any reason. Volunteers read and signed detailed informed consents, approved by the institution's Ethical Review Board. Methods were administered according to each woman's choice. Treatment was initiated at day 57 ± 3 after delivery while the women were fully breast feeding and the infant had adequate weight increase. Investigators from the Universidad Católica de Chile were not involved in contraceptive provision or services, which were the responsibility of the Instituto Chileno de Medicina Reproductiva.

Contraceptive devices
Types of contraceptives used were Norplant® levonorgestrel implants (NOR, n = 36), progesterone vaginal rings (PVR, n = 36) and Copper T 380 A intrauterine devices (T-Cu, n = 57). Women in the PVR group and in the NOR group were enrolled into the study in 1992–1993 and in 1994 respectively; each group had contemporary T-Cu controls. The sample size was estimated considering that, using analysis of variance (ANOVA), 28 subjects in each group would be needed to detect a difference of 0.05–0.1 g/m2 in bone density with a level of significance of 0.05 and a power of 90% (López et al., 1996Go).

Norplant® implants consist of six Silastic® capsules, inserted under the skin of the arm, that release ~85 µg/day of levonorgestrel initially, falling to ~35 µg/day in the second year of use (Fraser et al., 1998Go). The PVR is a homogeneous Silastic® ring that releases ~10 mg of progesterone per day for a 3 month period and provides progesterone plasma concentrations of ~25 nmol/l and 10 nmol/l during the first week and the third month of use respectively. The ring was replaced every 3 months, according to the study protocol (Sivin et al., 1997Go). Change to a more effective method at the time of weaning was recommended to women who selected the PVR because it is not considered effective in non-lactating women.

Breast-feeding management
Verbal and written instructions for breast-feeding on demand were given in the maternity ward and were reinforced at each visit. Mothers were instructed to use the breast as the only source of nutrients in the first 6 months. A midday meal was indicated at the seventh month of age and an evening meal at the ninth month. Each woman received powdered milk and rice as food supplements in the first postpartum year.

Infant weight increase was used as the main indicator of the adequacy of milk output. During the first 5 months of age, a weight increase of <20 g per day (mean since last visit) was considered an indication of inadequate milk output. Supplements were indicated if the growth rate was below this threshold, unless the child continued in the same weight percentile curve (within 10 percentile points) as before or had a normal weight for age and length. The Boston weight curve for boys (Nelson, 1964Go) was used as reference.

Bone evaluation
Women were studied at month 1 (PM1), month 6 (PM6) and month 12 (PM12) postpartum and at month 6 (PVR group) or month 12 (NOR and T-Cu groups) after weaning. Month 6 after weaning was chosen in the PVR group because all women changed method when weaned and a longer interval may have reflected the influence of other contraceptives used upon bone turnover. Month 12 after weaning was chosen in the NOR group to allow a longer period for an eventual impact of levonorgestrel on bone density to be detected. At these four time points, blood and urine samples were obtained and an estimation of daily calcium intake in the preceding week (López et al., 1996Go) was performed. Bone densitometry (bone density) took place at PM1, PM12 and post-weaning.

Biochemical measurements
Blood
All volunteers were fasting at the time of blood sampling and at least 2 h had passed since the last episode of breast feeding. A total of 30 ml of blood was obtained. Samples were obtained between the fifth and tenth day of the menstrual cycle in those women who had reinitiated cycles. Serum was divided into separate aliquots for each of the following tests and stored at –20°C. Haemoglobin concentration (colorimetric method) and magnesium concentration (colorimetric method, Merck kit No. 3338) were measured and a biochemical profile was obtained (SMA-II autoanalyzer) of calcium, phosphorus, blood urea nitrogen, glucose, total proteins, albumin, cholesterol, uric acid, total alkaline phosphatases, bilirubin, serum glutamic-oxaloacetic transaminase (SGOT) and lactate dehydrogenase (LDH).

Parathyroid hormone (PTH) was determined by radioimmunoassay using TYR 43–68 labelled iodine-125 and a specific antibody for 44–68 fraction developed in sheep (donated by Dr D.Hesh, University of Hannover, Germany). The fragment 44–68 corresponds to the mid-region of PTH or M-PTH. The tracer used was this fragment, to which tyrosine (TYR) is attached in order to be able to label it with iodine. For this reason it is called TYR 44–68-labelled iodine-125. The fragment 44–68 TYR was obtained from Sigma. The I-labelling was carried out locally. A specific antibody for the 44–68 fraction was developed in sheep, also locally. The assay was set up and validated in the Endocrine Laboratory of the Faculty of Medicine in the Catholic University (Campino et al., 1987Go). The sensitivity of the assay was 20 ng/l and the interassay coefficients of variation for low concentration (35 ng/l) and high concentration (54.2 ng/l) pools were 20 and 11.1% respectively. Follicle stimulating hormone (FSH), prolactin and oestradiol determinations were performed using the reagents supplied by the WHO Programme for the Provision of Matched Assay Reagents for the Radioimmunoassay of Hormones in Reproductive Physiology (WHO, 1987). The intra-assay coefficients of variation for FSH, prolactin and oestradiol were 3.9, 6.4 and 5.0% for the high concentration pool and 18.5, 6.3 and 14.9% for the low concentration pool respectively. The inter-assay coefficients for the same hormones were 8.0, 16.6 and 12.5% for the high concentration pool and 24.9, 11.6 and 33.5% for the low concentration pool respectively. Assays were run simultaneously at the end of the study.

Urine
Morning fasting samples were collected to measure hydroxyproline and creatinine after 24 h on a diet free from hydroxyproline.

Bone
The mineral content and bone density (bone density) were measured in the total body, lumbar spine (L2–L4), right femoral neck and right trochanter by dual-energy X-ray absorptiometry using a Lunar Densitometer with a precision of 1–2%. A quality assurance programme devised at the Quality Assurance Center, Portland, Oregon, was incorporated into the study protocol.

Data analysis
Women were considered to be fully breast feeding when the breast was the infant's only source of nutrients up to 6 months of age and the only source of milk thereafter. Infant growth rate had to be adequate according to the criteria described above to classify a case in this category. Weaning was defined as less than one suckling episode per day during a week. The end of lactational amenorrhoea was marked by the occurrence of the first postpartum bleed, consisting of at least 1 day of normal bleeding or 3 consecutive days of spotting, followed by another bleed within 60 days.

Volunteers who weaned early (before 6 months postpartum), who had severe intercurrent systemic disease or who had their contraceptive device removed, were excluded from the study at the time of the event. Out of 129 women enrolled, 21 were excluded from the analysis because they participated only in the first evaluation (PM1).

Values obtained in the T-Cu controls contemporary to the PVR (n = 38) and the NOR (n = 13) groups were pooled because there were no significant differences between them. In addition, values obtained in the control group at 6 (n = 34) and 12 (n = 9) months after weaning were pooled because there were no significant differences between these two sampling times; this pooled value was used for the comparisons with values obtained in the treated groups during lactation. Values obtained in non-lactating women in a previous study performed using the same methods (López et al., 1996Go) were used as reference values (0%) to establish percentage changes in hydroxyproline, alkaline phosphatases and bone density during lactation. Non-lactating controls were 26 women of similar admission characteristics to those involved in the present study, except that they were slightly older (29 years) than the lactating women (26–27 years). The women were sampled at three different times, providing 74 points for the basal values used. Baseline values used in Figure 1Go were 52.3 for alkaline phosphatase and 40.0 for hydroxyproline/creatinine ratio in urine. Baseline values used in Figure 2Go were 1.23 for bone density in L2–L4 and 0.989 for bone density in right femoral neck. The data from both studies were analysed in a similar way.



View larger version (21K):
[in this window]
[in a new window]
 
Figure 1. Variations with time in alkaline phosphatases (Alk Phos) and urinary hydroxyproline/creatinine ratio in urine (OH-P/Cr) according to the contraceptive method used. {square} Progesterone vaginal ring (PVR); {triangleup} Norplant® implants; {circ} Copper T IUD. Month 6 post-weaning included only women using PVR. The basal level (0) represents values observed in non-lactating women (Lopez et al., 1996Go).

 


View larger version (19K):
[in this window]
[in a new window]
 
Figure 2. Variations with time in bone density (BD) according to the contraceptive method used. L2–L4: lumbar spine; RFN: right femoral neck. {square} Progesterone vaginal ring (PVR); {triangleup} Norplant® implants; {circ} Copper T IUD. Month 6 post-weaning includes only women using PVR. The basal level (0) represents values observed in non-lactating women (Lopez et al., 1996Go).

 
The duration of lactation and of amenorrhoea was determined by life table analysis. Variance analysis was used for independent measurements and Tukey's method for multiple comparisons between each study group at admission and at each evaluation. Variance analysis for repeated measurements was performed for comparisons along time within groups. To meet the assumption of normality of prolactin, FSH and oestradiol values, comparisons were made using logarithmic transformations, since the values failed the test of normal distribution. Daily calcium intake, PTH and hydroxyproline/creatinine were analysed using Kruskal–Wallis and Dunn methods for multiple comparisons between each study group (Hollander and Wolfe, 1972Go). Parametric statistics were used for other comparisons between groups. A P value of <0.05 was considered significant. Statistical analysis was performed using SAS (SAS Institute, 1987, 1991). Results are presented as means ± SE, unless specified otherwise.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
There were no differences between the study groups in the general characteristics at admission (Table IGo). In the PVR group, 28 women reached PM 12 and 22 completed the study. In the NOR group, 24 and 22 women reached PM 12 and the post-weaning evaluation, respectively. In the control T-Cu group, 48 women reached PM 12 and 43 completed the study. Reasons for discontinuation from the study were pregnancy (PVR, n = 1); missed sampling (PVR, n = 5; NOR, n = 7; T-Cu, n = 6); breast feeding at the end of the study (NOR, n = 1; T-Cu, n = 1); and planning pregnancy (NOR, n = 1). The 22 women in the PVR group who completed the study were using other contraceptives at the time of the post-weaning sampling (combined pills, n = 9; T-Cu, n = 4; progestin-only pill, n = 3; NOR, n = 1; condoms, n = 5).


View this table:
[in this window]
[in a new window]
 
Table I. Baseline characteristics of the three groups of women using different contraceptive devices
 
There were no differences in lactation performance between the PVR and NOR groups and the control T-Cu group (Table IIGo). A significant difference in the duration of lactational amenorrhea was found between the PVR group (12 ± 0.7 months; mean ± SE) and the NOR group (12 ± 0.9 months) in comparison to the T-Cu group (7 ± 0.5 months).


View this table:
[in this window]
[in a new window]
 
Table II. Lactation performance and infant growth in the three groups of women using different contraceptive devices
 
A few minor differences were found between the three contraceptive groups in the nutritional parameters examined (Table IIIGo). The average daily calcium intake was lower in the T-Cu group than the NOR group at PM6. Daily calcium intake decreased in all women over time, but remained at approximately the recommended dietary allowance during lactation (Riggs et al., 1991Go). Haemoglobin concentrations were higher in the PVR group at PM12 and in the NOR group from PM6 to post-weaning, in comparison with the control T-Cu group. Haemoglobin concentrations remained within the normal range in all groups, although minor changes occurred with time (Table IIIGo). No differences were found between PVR and NOR groups and the control T-Cu group in the biochemical mean values.


View this table:
[in this window]
[in a new window]
 
Table III. Post-partum nutritional parameters in the three groups of women using different contraceptive devices
 
No differences in prolactin and FSH concentrations were found between groups. Oestradiol mean values were significantly lower in the NOR group than in the PVR and T-Cu groups at PM1, PM6 and PM 12 (Table IVGo). A borderline but significant difference (P = 0.047) in PTH concentrations was found between the NOR group and the PVR and T-Cu groups at PM12.


View this table:
[in this window]
[in a new window]
 
Table IV. Post-partum endocrine parameters in the three groups of women using different contraceptive devices
 
Measurements of bone metabolism and turnover showed increased bone remodelling during lactation in all groups (Table VGo). There were no differences between the groups with the exception of serum alkaline phosphatases which were lower in the PVR group than in the control T-Cu group at PM6. Values of hydroxyproline and alkaline phosphatases changed over time in a similar way in the three groups (Figure 1Go) and reached, after weaning, the values observed in non-lactating women (see above and Lopez et al., 1996).


View this table:
[in this window]
[in a new window]
 
Table V. Biochemical markers of bone metabolism and turnover according to contraceptive method and study stage
 
No differences were found between groups in any of the bone density measurements, at any point after contraceptive initiation (Table VIGo). The bone density in the lumbar spine (L2–L4) was significantly less in the first month after delivery in all groups, in comparison to non-breast-feeding women (P < 0.05), but no difference was found after weaning. Either during breast feeding or after weaning, the bone density in the right femoral neck did not differ in any of the study groups from the values obtained in non-breast-feeding women (Figure 2Go).


View this table:
[in this window]
[in a new window]
 
Table VI. Bone density according to contraceptive method and study stage
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
This study suggests that there is no difference in lactation performance, bone turnover or bone density between postpartum women who used Norplant® implants, progesterone vaginal rings or Copper T IUDs. This finding suggests that these two progestin-only methods do not interfere with the changes in bone turnover and in bone mass that occur during lactation. The transient bone loss and the changes in metabolic markers of bone turnover that occurred during breast feeding, as well as the recovery of bone density after weaning, were similar in the three groups. The lack of impact of the progestins on bone turnover during lactation was observed in the presence of both progesterone, a weak anti-oestrogen, and levonorgestrel, a strong anti-oestrogen.

Our study also confirms that lactation is associated with a transient decrease in bone mass, that is reversed after weaning, and with changes in bone turnover. The lack of deleterious effects ascribable to the contraceptive methods studied has to be considered against this particular physiological background. The devices may have a different influence on bone metabolism at other stages of life. However, it is reassuring that a limited study carried with Norplant® implants showed no negative impact in adolescents (Cromer et al., 1996Go) or in women aged 19–42 years (Intaraprasetrt et al., 1997Go), while a positive effect was found in perimenopausal women (Naessen et al., 1995Go).

A similar study of eight women showed a slight positive effect of progestin only pills upon bone turnover during lactation (Caird et al., 1994Go). Our data do not support this beneficial effect, since bone recovery after weaning was similar in the Norplant® and the control T-Cu groups. Whether or not the various progestins may have different effects attributable to their pharmacological profile or the route of administration remains to be established.

This study covered only a relatively short period after weaning. The long term impact of progestin-only methods also needs to be assessed, particularly because the information obtained from non-lactating women is contradictory. Negative effects on bone metabolism have been reported following long term use of Depo-Provera in some (Cundy et al., 1991Go) but not all (Virutamasen et al., 1994Go; Naessen et al., 1995Go) the studies. If such a deleterious effect on bone mass occurs during long term administration of medroxyprogesterone acetate, it may be explained by the ability of this progestin to exert a stronger inhibition of the endogenous secretion of oestrogens in comparison to that of progesterone rings or levonorgestrel implants.

Oestrogens protect bone mass in other situations and it has been suggested that progesterone may have a trophic effect on bone mass (Prior, 1990Go). We found no differences in bone parameters between the three contraceptive groups although endogenous concentrations of steroids differed significantly between them. Norplant® users had lower oestrogen concentrations than PVR and IUD users. We assume that all women in the Norplant® group remained anovulatory, since they had prolonged amenorrhoea and implants inhibit ovulation during the first year of use in almost all cycling women (Croxatto et al., 1982Go). Progesterone ring users were exposed continuously to relatively high plasma progesterone concentrations, ranging from ~25 nmol/l to 10 nmol/l between the first week and the third month of use (Sivin et al., 1997Go). These women also remained in prolonged amenorrhoea and are unlikely to have ovulated (Díaz et al., 1991Go). On the other hand, women using a Copper T IUD were exposed to variable oestrogen and progesterone concentrations, since menstruation resumed in half of this group within the first 7 months postpartum. These three different patterns of plasma oestrogen and progesterone concentrations were not translated into differences in bone turnover or density.

The findings of this study have implications for the counselling of lactating women on contraceptive choices. Progestin-only methods initiated after 6 weeks postpartum are safe regarding lactation and infant growth (WHO, 1996). Our results suggest that the two progestin-only methods studied do not interfere with changes in bone turnover and in bone mass that occur during lactation or with the recovery of bone mass after weaning.


    Acknowledgments
 
This work was undertaken as part of the contraceptive development program sponsored and co-ordinated by the International Committee for Contraceptive Research of The Population Council, Inc., New York, New York. The content of this report does not necessarily reflect the policy of the funding sources.


    Notes
 
3 To whom correspondence should be addressed at: Instituto Chileno de Medicina Reproductiva, José Ramón Gutiérrez 295, Depto. 3, Santiago, Chile Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Caird, L.E., Reid-Thomas, V., Hannan, W.J. et al. (1994) Oral progestogen-only contraception may protect against loss of bone mass in breast-feeding women. Clin. Endocrinol., 41, 39–45.

Campino, C., Bruggendieck, B., Pumarino, H. et al. (1987) Desarrollo de un radioimmunoanalisis para la fraccion media de hormona paratiroidea (m-PTH): comparacion con dos kits comerciales para carboxilo terminal. Revista Medica de Chile, 115, 763–767.[ISI][Medline]

Cromer, B.A., McArdle Blair, J., Mahan, J.D. et al. (1996) A prospective comparison of bone density in adolescent girls receiving depot medroxyprogesterone acetate (Depo-Provera), levonorgestrel (Norplant), or oral contraceptives. J. Pediatr., 129, 671–676.[ISI][Medline]

Croxatto, H.B., Díaz, S., Pavez, M. et al. (1982) Plasma progesterone levels during long-term treatment with levonorgestrel silastic implants. Acta Endocrinol., 10, 307–311.

Cundy, T., Evans, M., Roberts, H. et al. (1991) Bone density in women receiving depot medroxyprogesterone acetate for contraception. Br. Med. J., 303, 13–16.[ISI][Medline]

Díaz, S., Miranda, P., Brandeis, A. et al. (1991) Mechanism of action of progesterone as contraceptive for lactating women. In Seppala, M. and Hamberger, L. (eds) Frontiers in Human Reproduction. The New York Academy of Sciences, New York, NY, USA, pp. 11–21.

Díaz, S., Zepeda, A., Maturana, X. et al. (1997) Fertility regulation in nursing women: IX. Contraceptive performance, duration of lactation, infant growth and bleeding patterns during use of progesterone vaginal rings, progestin-only pills, Norplant® implants and Copper T 380-A intrauterine devices. Contraception, 56, 223–232.[ISI][Medline]

Fraser, I.S., Tiitinen, A., Affandi, B. et al. (1998) Norplant® consensus statement and background review. Contraception, 57, 1–9.[ISI][Medline]

Hollander, M. and Wolfe, D. (eds) (1972) Nonparametric Statistical Methods. John Wiley & Sons, New York, pp. 115–125, 131.

Intaraprasetrt, S., Taneepanichskul, S., Theppisai, U. and Chaturachinda, K. (1997) Bone density in women receiving Norplant® implants for contraception. J. Med. Assoc. Thai., 80, 738–741.[Medline]

International Planned Parenthood Federation (1989) Statement on Breast Feeding, Fertility and Post-partum Contraception. Statement by the International Medical Advisory Panel, November, 1–5.

Kalkwarf, H.J., Specker, B.L., Bianchi, D.C. et al. (1997) The effect of calcium supplementation on bone density during lactation and after weaning. N. Engl. J. Med., 337, 523–558.[Abstract/Free Full Text]

López, J.M., González, G., Reyes, M.V. et al. (1996) Bone turnover and density in healthy women during breast-feeding and after weaning. Osteoporosis Int., 6, 153–159.[ISI][Medline]

Naessen, T., Olsson, S.-V. and Gudmundson, J. (1995) Differential effects on bone density of progestogen-only methods for contraception in premenopausal women. Contraception, 52, 35–39.[ISI][Medline]

Nelson, W.E. (ed) (1964) Textbook of Pediatrics, 8th edn. W.B. Saunders Company, Philadelphia, 42 pp.

Phillips, A., Hahn, D.W., Klimek, S. et al. (1987) A comparison of the potencies and activities of progestogens used in contraceptives. Contraception, 36, 181–192.[ISI][Medline]

Prior, J.C. (1990) Progesterone as a bone-trophic hormone. Endocrine Rev., 1, 386–398.

Riggs, B.L., Peck, W.A. and Bell, N.H. (eds) (1991) Physician's Resources Manual on Osteoporosis. 2nd edn. Current Medical Directions, New York, 19 pp.

Sivin, I., Díaz, S., Croxatto, H.B. et al. (1997) Contraceptives for lactating women: a comparative trial of a progesterone-releasing vaginal ring and the copper T 380A IUD. Contraception, 55, 225–232.[ISI][Medline]

SAS Institute Inc. (1987) SAS/STATTM Guide for Personal Computers, 6th edn. SAS Institute Inc., Cary, NC.

SAS/STAT Software (1991) The PHREG Procedure, Preliminary Documentation. SAS Institute Inc., Cary, NC.

Sowers, M. (1996) Pregnancy and lactation as risk factors for subsequent bone loss and osteoporosis. J. Bone Miner. Res., 11, 1052–1060.[ISI][Medline]

Virutamasen, P., Wangsuphachart, S., Reinprayoon, D. et al. (1994) Trabecular bone in long-term depot-medroxyprogesterone acetate users. Asia–Oceania, J. Obstet. Gynaecol., 20, 269–274.

World Health Organization (1987) Special programme of research development and research training in human reproduction. Programme for the provision of matched assay reagents for the RIA of hormones in reproductive physiology. Method Manual, 11th edn. WHO, Geneva.

World Health Organization (1996) Improving access to quality care in family planning. Medical Eligibility Criteria for Contraceptive Use. WHO/FRH/FPP Geneva.

Submitted on February 5, 1999; accepted on June 21, 1999.