According to Centers for Disease Control and Prevention, 11.6 million women used oral contraceptives (OC) in the United States in 2002, making it the most common method of birth control (17). Despite their pervasiveness, questions remain among scientists, health care providers, and users regarding possible adverse effects of OC.
Especially among young women, fear of weight gain (or body fat gain) has been cited as a primary reason for avoiding OC use (4,6,19,21,24). Many studies have investigated the OC-weight gain connection, and results have varied from a positive OC-weight gain association (18,23) to no association (1,5,12-14,25). In a review of 44 randomized controlled trials, Gallo et al. (9) concluded that the evidence was insufficient to rule out an association, but no large effect was evident; furthermore, no dose-response relationship was observed between the estrogen content of the various oral estrogen-progestin contraceptive formulations and weight gain. Gallo et al. described several mechanisms hypothesized to account for weight gain with OC, including fluid retention, increased subcutaneous fat, anabolic effects on appetite, and androgenic effects on muscle mass. Further investigation of this matter is warranted because it may affect the decision making of teenagers and young women in need of a reliable form of birth control. Also, any association between OC and body composition may be of special interest among the population of female runners because it may affect their running performance.
This article is a secondary analysis of data from a randomized trial that was designed primarily to assess the effect of a commonly used OC on bone mineral density and stress fracture incidence in female competitive runners (4). Here, we test the null hypotheses that OC use is not associated with weight, fat, or lean mass changes in these women, using intention-to-treat as well as treatment-received analyses.
Participants and recruitment.
Details of the study methodology have been described by Cobb et al. (4). One hundred and fifty competitive female runners were recruited between August 1998 and September 2003 from intercollegiate cross-country teams, postcollegiate running clubs, and road races mainly in the geographic areas of Palo Alto, CA, Los Angeles, CA, Ann Arbor, MI, West Haverstraw, NY, and Boston, MA. To be eligible, women had to be 18-26 yr old, run at least 40 miles·wk−1 during peak training times, and compete in running races. Exclusion criteria were medical contraindications to OC use, unwillingness to be randomized to OC use or nonuse, and any use of OC or other hormonal contraception within the past 6 months. All women were required to visit a study physician or student health service staff member before enrollment to rule out contraindications to OC use. Written informed consent was obtained from each subject after she received a detailed explanation of the study procedures. The protocol was approved by the institutional review boards of the US Army Medical Research and Materiel Command, the colleges from which participants were recruited, and the following clinical sites: Stanford University, the University of California Los Angeles, the University of Michigan, the Helen Hayes Hospital, and the Massachusetts General Hospital.
Randomization and intervention.
Eligible women were randomly assigned to receive an OC or no intervention for an intended 2 yr, stratified according to clinical site. An independent investigator who was not otherwise affiliated with the study performed the randomization using a random number table. The OC pill assigned in this study was Lo/Ovral (Wyeth Ayerst, 28-d pack), which contains 30 μg of ethinyl estradiol and 0.3 mg of norgestrel. For ethical reasons, neither the athletes nor the prescribing physicians were blinded to treatment assignment, and no placebo was used. Data collection personnel (i.e., those taking measurements) were not told the treatment assignment of the women, and attempts were made to hide treatment assignment from other research staff members, but we cannot rule out the possibility that the subjects revealed their treatment status to the staff.
Data collection and follow-up.
At baseline, participants visited one of the clinical sites for measurement of bone mineral density, body composition, height, and weight. Height and weight were measured using standard stadiometers and balance beam scales, respectively. Body composition (fat mass, percent body fat, and lean mass) was measured by dual-energy x-ray absorptiometry (Hologic QDR 4500A at four sites, QDR 2000W at one site). All women were asked to refrain from heavy physical activity 24 h before measurement to minimize fluctuations in hydration level.
Participants also filled out questionnaires on menstrual history, previous use of OC, injury and stress fracture history, training regimen, diet, eating attitudes, and eating behaviors, as previously described (3). Women were classified as amenorrheic, oligomenorrheic, or eumenorrheic on the basis of the number of menses they reported having in the previous 12 months. Amenorrhea was defined as 0-3 cycles in the past year; oligomenorrhea as 4-9 cycles in the past year; and eumenorrhea as 10 or more cycles in the past year. Participants were asked to return 1 and 2 yr later to repeat these measurements and questionnaires.
Of 150 women randomized, 124 (83%) attended at least one follow-up appointment and 96 (64%) participants attended both, at an average of 14.4 months (median = 13.1 months) and 26.6 months (median = 25.4 months), respectively, after baseline. Loss to follow-up was somewhat greater in the group assigned to OC (22%) than in the control group (14%). Baseline characteristics of the participants with no follow-up data were similar to those with follow-up data, except that those with no follow-up were more likely to have a history of stress fracture before baseline (52% vs 32%, P = 0.05). Between clinic visits, participants filled out a monthly calendar on which they recorded menstrual bleeding and the use of OC pills.
Ascertainment of compliance.
Women in the treatment group were asked to report if and when they discontinued taking the study medication. Treatment compliance was also monitored through return of used pill packs, monthly calendars, and yearly questionnaires. If a woman reported having discontinued treatment, she was contacted by a research assistant to determine if and when the OC was discontinued and the reason why. Similarly, women in the control group were asked to contact us if they were planning to start an OC. If so, they were encouraged to take the study pill (Lo/Ovral) or a pill with a similar dose of estrogen. Compliance was also monitored on monthly calendars and yearly questionnaires. If a woman reported having started an OC, she was contacted by a research assistant to obtain the date of starting, as well as the formulation and the reason for starting the OC.
The study was powered to detect differences in bone mineral density and stress fracture occurrence between the OC group and the control group, the outcomes of the primary study, rather than to detect differences in rates of change in weight or body composition. Statistical analyses were performed using the SAS statistical package, version 8.2 (SAS Institute, Cary, NC). For preliminary descriptive analyses, means were compared between groups using t-tests for normally distributed variables and Wilcoxon's rank sum tests for nonnormally distributed variables. Proportions were compared using chi-square tests or Fisher's exact test in the case of small numbers in cells.
The intention-to-treat method was used for the primary analyses described here. The primary outcomes examined were rates of change in 1) weight, 2) body mass index (BMI), 3) fat mass, 4) body fat as percentage of total mass, and 5) lean mass. Linear mixed-effects models, taking into account data from all women with any follow-up measurements, were used to evaluate the effect of OC assignment on each outcome. Results for BMI change were identical to results for weight change because the women's heights were constant. BMI is not discussed further.
To reflect study design, clinical site was included as a covariate in all models. In addition, variables that differed (P < 0.10) between the OC and the control groups in baseline comparisons, along with two a priori variables of interest, menstrual status at baseline and eating disorder inventory score (10), were tested and included in the model if they had an interaction effect or confounding effect (>15% change in parameter) on the association of interest. In the end, the covariates included in each model were 1) for change in weight: clinical site, baseline menstrual status, and dieting frequency; 2) for change in fat mass and percent fat: clinical site and baseline menstrual status; 3) for change in lean mass: clinical site, with analysis stratified by baseline menstrual status because of an interaction effect between menstrual status and treatment. Because of the small number of women in the amenorrheic category of menstrual status, we combined amenorrheic and oligomenorrheic women into one category, irregular (0-9 menses in the past year), and contrasted them with eumenorrheic women, termed regular (10 or more menses in past year).
Because the proportion of women crossing over from their assigned protocol was greater than 25% in each group, secondary treatment-received analyses were performed. First, we evaluated the effect of the actual proportion of follow-up time that the women took any OC (including non-study-assigned OC) on weight and body composition. Each woman was assigned a proportion score ranging from 0 to 1, with 1 representing the use of any OC for 100% of the follow-up period, 0.5 representing 50%, and so forth. We also performed other treatment-received analyses, classifying women into the OC group if they used OC for 1) 3 months or more and 2) 6 months or more. Linear mixed-effects models, with the same covariates and stratification used in the intention-to-treat analyses, were used to evaluate the differences between outcomes for OC users and nonusers. We compared ≥6-month OC users versus <6-month users in baseline and follow-up characteristics to identify any potential confounders in the treatment-received analysis.
At baseline, the mean age for all 150 women was 22.1 yr, mean height was 165.6 cm, and mean weight was 58.2 kg, with a mean BMI of 21.1 kg·m−2. Mean caloric intake was 2278 kcal·d−1. The women ran for an average of 56.0 km·wk−1 and lifted weights 66.5 min·wk−1. Fifty participants (33%) reported irregular menses in the past year.
On randomization, 69 women (46%) were assigned to the OC group and 81 (54%) to the control group (discrepancy due to chance). The two groups were similar at baseline in the four measurements of interest to this analysis: weight, fat mass, percent fat, and lean mass. The women were also comparable in mean age, height, caloric intake, total number of previous menstrual periods, and several other physiological and behavioral characteristics (Table 1). However, women assigned to the OC group had dieted more frequently and tended to score higher on the eating disorder inventory than control women.
Retention and adherence.
Twenty-six women (17%) withdrew from the study or provided no further body composition data after baseline (Fig. 1). This left 124 women (54 (78%) of the OC group and 70 (86%) of the control group) with at least one follow-up visit for the weight analysis. Lean mass, fat mass, and percent fat were not obtained from one additional control woman.
Within the treatment group, 14 (26%) stopped taking the OC after an average of 5.4 months, and 28 (40%) of the control group started taking them at an average of 11.3 months into the study. Four women in the control group and one woman in the treatment group switched groups twice. The reasons women gave for stopping the OC included the following (in decreasing order of frequency): fear of weight gain or perceived weight gain, adverse effects (irritability, abdominal symptoms, nausea, fatigue, or unspecified), and fear of detriment to athletic performance. The reasons control women gave for starting an OC included the following (in decreasing order of frequency): regulation of menstrual periods, alleviation of menstrual symptoms including cramps, prevention of pregnancy, and treatment of acne or allergies.
Women who stopped taking the OC had a lower percentage of body fat, fewer menstrual periods, and more disordered eating than those who continued to follow the OC protocol. Amenorrheic women were the least likely to comply with taking the OC. Of eight amenorrheic women who were assigned to the OC, only one took it through the entire study. Of the remaining seven, two were lost to follow-up, five discontinued use within 2 months, and one discontinued after 1.5 yr. However, these losses to the treatment group did not result in any significant baseline or follow-up differences between those assigned to OC and those who actually used OC for 6 months or more (results not shown).
Body composition changes by treatment randomization (intention-to-treat analysis).
- Change in Weight. The weight of women assigned to the OC group did not increase relative to controls. In fact, controlling for clinical site, baseline menstrual status, and baseline dieting frequency in linear mixed models analysis, OC assignment was associated with a trend toward lesser weight gain (P = 0.09; Table 2). This trend persisted when the above analysis was stratified by baseline menstrual status, eating disorder inventory score, or baseline weight (results not shown).
- Four women gained or lost more than 10 kg during the study. Removal of these women from the analysis resulted in a greater weight gain among controls relative to the OC group and thus only strengthened the observation that OC assignment did not contribute to weight gain.
- Change in Fat Mass and Body Fat Percentage. Women assigned to the OC group did not show an increase in fat mass or in percent body fat relative to controls. Mixed effects models controlling for clinical site and baseline menstrual status showed no statistically significant difference in the rate of change in fat mass between the two groups (Table 2). Stratifying the analysis by baseline menstrual category or eating disorder score did not change the direction or magnitude of this relationship (results not shown). For percent body fat, there was a trend toward relative fat loss in those assigned to OC use relative to controls (P = 0.08; Table 2). Absolute weight change and absolute fat change did not differ significantly from 0 in either group.
- Change in Lean Mass. The OC-lean mass association differed according to baseline menstrual status. Assignment to the OC group was associated with significantly greater lean mass gain relative to controls among women with regular menses (P < 0.0001; Table 2), but no association was seen among those with irregular menses (P = 0.01 for the statistical interaction with menstrual regularity). Eumenorrheic women assigned to OC use had an adjusted mean gain of 0.67 ± 0.15 kg·yr−1, whereas controls showed little change in lean mass (−0.10 ± 0.14 kg·yr−1). Women who had irregular menstrual cycles at baseline gained, on average, moderate amounts of lean mass regardless of treatment assignment. Omission of the four most extreme values for lean mass change did not affect the above results.
To preserve the benefits of random assignment, characteristics that differed between the two groups after randomization were not included as covariates in the primary intention-to-treat analyses. However, women assigned to OC use differed from controls by the second follow-up visit in three respects: a higher percentage of fat in the diet, more frequent weight lifting, and fewer kilometers run per week than control women (data not shown). When the measure of weight lifting was added to the mixed effects models above, the adjusted mean difference (AMD) in lean mass for eumenorrheic women fell by 15%, but the OC-lean mass association remained positive and statistically significant (P = 0.0004). Addition of other follow-up variables did not change the magnitude or significance of the OC-lean mass association in this group.
Actual use of OC.
Among the 124 women with follow-up, 75 took OC for 3 months or more and 59 used them for 6 months or more, ignoring treatment assignment. Among women in the control group who took OC, the majority took Lo/Ovral or Ortho Tri-Cyclen (Ortho-McNeil Pharmaceutical, Inc; 35 μg of ethinyl estradiol). We combined all OC users into a single OC group for all secondary analyses.
Baseline and follow-up characteristics.
Women who used OC for 6 months or more differed from those who did not in a few baseline characteristics (data not shown). Those who used OC for at least 6 months had lower mean baseline bone mineral content. They less frequently reported a history of menstrual irregularity and menstrual irregularity at baseline. However, the two groups did not differ in number of total lifetime menstrual periods from menarche to baseline.
Follow-up measures of menstrual cycles, diet, and exercise characteristics revealed that the women who took OC for at least 6 months had predictably regular periods (having a greater number of menses per year than non-OC users); the only other follow-up difference between the groups was that OC users had a higher percentage of dietary fat (21.3% vs 18.5%; P = 0.04) at the second follow-up.
Body composition changes by actual OC use (treatment-received analysis).
- Change in Weight. When treatment assignment was ignored and the analysis was performed only on the basis of actual use of OC, the results were similar to those of the intention-to-treat analysis: there was little association between the proportion of follow-up time on OC and the rate of change in weight (Table 3). Adding covariates from the set of variables that differed between OC users and nonusers did not change the results, so we did not include these covariates in the models presented here. When women were classified by duration of OC use, namely, 1) 3 months or more and 2) 6 months or more, comparisons to the non-OC group again revealed little difference in weight change (Table 3). In all analyses, the OC users gained, on average, slightly less weight than nonusers.
- Change in Fat Mass and Body Fat Percentage. There was little association between body fat change and actual OC use in the proportionate, 3-month-use, or 6-month-use analysis (Table 3).
- Change in Lean Mass. As in the intention-to-treat analysis, women with regular menses at baseline showed a significant relative increase in lean mass associated with the proportion of time that they were actually taking OC (P < 0.0001; Table 3). The association with OC use was also seen in the analyses of 3-month users (P = 0.005) and 6-month users (P = 0.0007). No association between actual OC use and lean mass change was found among women with irregular menses (Table 3).
Because weight lifting affected lean mass, we explored adding it to the models above, although the two treatment-received groups did not differ significantly in their frequencies of weight lifting (P = 0.14). The AMD fell by 15-32%, but the OC-lean mass association remained positive and statistically significant in the proportionate treatment-received analysis (P = 0.002) and in the 6-month treatment-received analysis (P = 0.01). The association was of borderline significance in the 3-month-use analysis (P = 0.06).
Our finding of little difference in weight gain or fat mass gain between women assigned to 30 μg of ethinyl estradiol plus 0.3 mg of norgestrel versus no treatment further supports the conclusion reached in the review by Gallo et al. (9) that a large effect of OC use on weight is unlikely. That we found a (nonsignificant) trend toward less weight and fat mass gain among OC users compared with nonusers should reassure athletes that use of this OC will not cause performance-impairing weight changes. The results of our treatment-received analyses bolster our confidence in this conclusion. Because of this study's focus on young female runners, we cannot rule out that the OC could cause weight gain among inactive women or those outside the 18- to 26-yr age range. Also, our findings may not apply to injectible hormonal contraceptives or other formulations or dosages of OC, which may affect body composition differently than the medications tested in this study (2).
We confirm others' findings (21) that weight gain is commonly believed to be an adverse effect of OC use despite evidence to the contrary. Fear of or perceived weight gain was the most common reason given for discontinuing OC use in our study (cited by 6 of 14 switchers). Although we do not have weight measurements from women at the point when OC use ceased, when we compared weight gain of switchers in the first year to that of control women, the difference was insignificant (0.4 kg·yr−1 for switchers vs 0.1 kg·yr−1 for controls, P = 0.56).
The unanticipated positive association between OC use and lean mass gain observed among eumenorrheic women merits further investigation. The association was strongest in the intention-to-treat analysis, and was also present, although slightly weaker, in the treatment-received analysis. There are few other reports with which to compare our findings. Machado et al. (15) found that women randomized to estradiol-gestodone for one cycle only showed a significant increase in fat-free mass relative to controls. Women assigned to estradiol-drospirenone did not.
If there is an anabolic effect of combination OC formulations on lean mass, the mechanism remains unclear. Friend et al. (7) proposed that estrogen may have an anabolic action by increasing serum growth hormone concentrations. Skeletal muscle contains some estrogen receptors and thus may be affected directly as well. Phillips et al. (20) demonstrated measurable increases in muscular force of the adductor pollicis when estrogen levels rose in the follicular phase of young women's menstrual cycles. In contrast, multiple studies have found that OC use is associated with decreased free testosterone and dehydroepiandrosterone in healthy women of reproductive age (16), potentially causing a decrease in lean mass relative to controls.
It is impossible to tell whether the estrogen and/or progestin component of the OC used in this trial might be associated with the lean mass increase observed. The progestin component of Lo/Ovral, Norgestrel, has the highest androgenic activity of the four first-generation progestins investigated in a 2006 US Pharmacist report (22) and so could have a muscle-building effect.
We observed a stimulatory effect of the OC on lean mass in eumenorrheic women but not in women with irregular menstruation. The reason for this subgroup difference is not apparent. Irregular periods could be a marker for caloric deprivation. Menstrual irregularity was strongly related to disordered eating in this study, although self-reported calories did not differ between menstrual groups at baseline (3). Ihle and Loucks (11) found that caloric deprivation promoted bone resorption and interfered with bone formation, and Frost (8) proposed that the "mechanostat" that adjusts the bone to the mechanical stresses put on it might be reset by energy availability or hormone levels. Similar mechanisms could apply to muscle mass. Participants who menstruated fewer than 10 times per year at baseline may have had a dietary deficit that prevented a gain in lean mass.
Our study illustrates the difficulty of conducting a randomized controlled trial with OC, which greatly affects the personal lives of participants. Noncompliance was high in both treatment and control groups. We found women in the age range 18-26 yr to be a difficult group to observe in a study of this length because of their high mobility and multiple interests. We were unable to obtain any follow-up weight and body composition measurements on 17% of the subjects, and one third were lost from the study by the end of the follow-up period. To address the possible impact of loss to follow-up on our results, we performed an intention-to-treat sensitivity analysis, imputing a change of 0 for all measures among women lost to follow-up. Although this decreased the magnitude of the AMD slightly, eumenorrheic women assigned to OC still gained lean mass relative to controls (AMD = 0.70 + 0.14 kg·yr−1, P < 0.0001), and the direction of the differences in weight and fat was still negative.
For ethical reasons, women were not blinded regarding their use or nonuse of the OC, and this could have introduced bias into the study, especially given the widespread belief among young women that OC use leads to weight gain. Possibly, the women assigned to treatment compensated for that assignment in some way, for example, by cutting back on calories or increasing their activity. As noted above, by the second follow-up visit, the OC group had a higher mean frequency of weight training, but they had a lower mean number of kilometers run per week, and their self-reported calorie consumption was indistinguishable from that of controls. Although self-reported dietary questionnaires are notoriously inexact, there is little reason to believe that errors in reporting would be differentially erroneous between the two groups in this study. It is also possible that research assistants taking the weight and body composition measurements were made aware of treatment assignment by a few of the subjects themselves. However, these measurements were not the primary focus of the research, and given the prevalent belief that OC cause weight gain, if anything, any bias would be expected to occur in the direction opposite to our results.
Because this project was initially designed to assess bone mineral density and stress fracture as primary outcomes (not weight change), standardized conditions at the weigh-in regarding clothing worn, time of day, or day in menstrual cycle were not imposed, and this could have introduced some nondifferential inexactness of measurement.
Despite these weaknesses, the similarity of results between the randomized, intention-to-treat analysis and the actual-use analysis, the assessment of a great number of lifestyle variables, and the duration of this trial should lend confidence to our conclusions that young female distance runners can use this OC without fear of resultant weight or fat gain. Evaluation in other studies of the effect of OC on lean mass may clarify whether OC can enhance muscle strength.
This study was funded by a grant from the US Army Medical Research and Materiel Command, award number DAMD17-98-1-8518. Wyeth-Ayerst provided the oral contraceptives. The authors thank Dr. Robert Marcus and Dr. Rita Popat for their help in the trial, and Dr. Wayne Gray for statistical consultation.
1. Bonny AE, Ziegler J, Harvey R, Debanne SM, Secic M, Cromer BA. Weight gain in obese and nonobese adolescent girls initiating depot medroxyprogesterone, oral contraceptive pills, or no hormonal contraceptive method. Arch Pediatr Adolesc Med
2. Clark MK, Dillon J, Sowers MF, Nichols S. Weight, fat mass and central distribution of fat increase when women use depot-medroxyprogesterone for contraception. Int J Obesity
3. Cobb KL, Bachrach LK, Greendale G, et al. Disordered eating, menstrual irregularity, and bone mineral density in female runners. Med Sci Sports Exerc
4. Cobb KL, Bachrach LK, Sowers MF, et al. The effect of oral contraceptives on bone mass and stress fractures in female runners. Med Sci Sports Exerc
5. de Melo NR, Aldrighi JM, Faggion D Jr, et al. A prospective open-label study to evaluate the effects of the oral contraceptive Harmonet (gestodene75/EE20) on body fat
6. Emans SJ, Grace E, Woods ER, Smith DE, Klein K, Merola J. Adolescents' compliance with the use of oral contraceptives. J Am Med Assoc
7. Friend KE, Hartman ML, Pezzoli SS, Clasey JL, Thorner MO. Both oral and transdermal estrogen increase growth hormone release in postmenopausal women-a clinical research center study. J Clin Endocrinol Metab
8. Frost HM. Bone's mechanostat: a 2003 update. Anat Rec A Discov Mol Cell Evol Biol
9. Gallo MF, Lopez LM, Grimes DA, Schulz KF, Helmerhorst FM. Combination contraceptives: effects on weight. Cochrane Database Syst Rev
10. Garner DM, Olmsted MP. Manual for Eating Disorders Inventory
. Odessa (FL): Psychological Assessment Resources, Inc; 1984. pp. 1-100.
11. Ihle R, Loucks AB. Dose-response relationships between energy availability and bone turnover in young exercising women. J Bone Miner Res
12. Lech MM, Ostrowska L. Oral contraceptives use and weight gain in women with a Central European lifestyle. Eur J Contracept Reprod Health Care
13. Lloyd T, Lin HM, Matthews AE, Bentley CM, Legro RS. Oral contraceptive use by teenage women does not affect body composition. Obstet Gynecol
14. Lloyd T, Taylor DS, Lin HM, Matthews AE, Eggli DF, Legro RS. Oral contraceptive use by teenage women does not affect peak bone mass: a longitudinal study. Fertil Steril
15. Machado RB, Tachotti F, Cavenaque G, Maia E. Effects of two different oral contraceptives on body water: a randomized study. Contraception
16. Miller KK, Lawson EA, Mathur V, Wexler TL, et al. Androgens in women with anorexia nervosa and normal-weight women with hypothalamic amenorrhea
. J Clin Endocrinol Metab
17. Mosher WD, Martinez GM, Chandra A, Abma JC, Wilson SJ. Use of contraception and use of family planning services in the United States: 1982-2002. Adv Data
18. Neel EU, Litt IF, Jay MS. Side effects and compliance with low- and conventional-dose oral contraceptives among adolescents. J Adolesc Health Care
19. Oddens BJ. Women's satisfaction with birth control: a population survey of physical and psychological effects of oral contraceptives, intrauterine devices, condoms, natural family planning, and sterilization among 1466 women. Contraception
20. Phillips SK, Sanderson AG, Birch K, Bruce SA, Woledge RC. Changes in maximal voluntary force of human adductor pollicis muscle during the menstrual cycle. J Physiol
21. Picardo CM, Nichols M, Edelman A, Jensen JT. Women's knowledge and sources of information on the risks and benefits of oral contraception. J Am Med Womens Assoc
22. Rice C, Thompson J. Selecting and monitoring hormonal contraceptives: an overview of available products. US Pharm
23. Rickenlund A, Carlstrom K, Ekblom B, Brismar TB, Von Schoultz B, Hirschberg AL. Effects of oral contraceptives on body composition and physical performance in female athletes. J Clin Endocrinol Metab
24. Rosenberg M. Weight change with oral contraceptive use and during the menstrual cycle: results of daily measurements. Contraception
25. Sule S, Shittu O. Weight changes in clients on hormonal contraceptives in Zaria, Nigeria. Afr J Reprod Health