ORIGINAL RESEARCH ARTICLE

Weight Change With OralContraceptive Use and Duringthe Menstrual CycleResults of Daily MeasurementsMichael Rosenberg

Although weight gain is among the most common com-plaints of women using oral contraceptives (OC) and afrequent reason for discontinuation, studies demonstratelittle basis for this perception. We explored this issue byanalyzing the daily weights of 128 women during fourcycles of triphasic OC use. The mean weight at the end ofthe fourth cycle of use was the same as baseline weight(average weight change, 0.0 pounds). The largest proportionof women, 52%, remained within 2 pounds (0.9 kg) of theirstarting weight, and 72% of women had either no weightchange or a loss. Over each menstrual cycle, regular butminor weight shifts were observed, with the mean weightrising by one-half pound (0.2 kg) during the first weeks ofeach cycle and falling by the same amount during the lastfew days. These results emphasize the lack of associationof OC use with weight gain but OC may be blamed at leastin part, based on cyclic fluctuations. Counseling shouldemphasize weight gain as a misperception and stress thefact that a highly effective and safe form of contraceptionshould not be ruled out or discontinued because of concernabout weight. CONTRACEPTION 1998;58:345–349 © 1998Elsevier Science Inc. All rights reserved.

KEY WORDS: oral contraceptive use, weight change, dailymeasurements

Introduction

Weight gain is one of the most frequently citedreasons for not using oral contraceptives(OC), complaints among women who use

OC, and reasons for discontinuation. A nationallyrepresentative sample of American women, for exam-ple, indicated that weight gain was the most common

single reason for discontinuing OC.1 At 11%, this wasa more frequent reason than nausea, headache, andmenstrual abnormalities. A survey of 6676 womenfrom several European countries similarly found thatweight gain was among the most common com-plaints and was associated with a 40% increasedlikelihood of discontinuation.2 A separate analysis ofthe same European data indicates that weight gain isthe most frequent reason for not initiating use ofOC.3 Other studies and surveys consistently indicatethat weight gain is among the most common reasonsgiven for OC discontinuation.

Despite the strong influence of the weight gainperception on a woman’s initial and ongoing choice ofcontraception, relatively few studies have examinedthe relationship between weight gain and OC use.What information is available indicates little or noweight change.4–6 A large longitudinal study found amodest but significant weight gain of 0.1–1 kg after 1year of use and 1.1–1.7 kg more after 2 years.7 How-ever, this study was not controlled so could notdifferentiate between weight gain as a normal part ofaging and that possibly resulting from the use of OC.A study comparing weight in women using differentmethods of contraception found a slight weight lossin the OC group and no significant differences be-tween any of the contraceptive user groups.8 Finally,a controlled study found that weight change wassimilar in groups of women who were randomized toplacebo and high-dose OC preparations.9 None of thesestudies measured daily weights, relying instead on pe-riodic clinic visits.

If the common perception that OC result in weightgain is incorrect, it means that a highly reliable andsafe contraceptive, which also has a number of im-portant noncontraceptive benefits, will not be used asfully as it might. To help clarify this issue, weanalyzed daily weight measurements among womenenrolled in a clinical trial.

Health Decisions, Inc., Chapel Hill, North CarolinaName and address for correspondence: Dr. M. Rosenberg, Health Decisions,

Inc., 1512 E. Franklin Street, Suite 200, Chapel Hill, NC 27514; Tel.: 919-967-1111; Fax: 919-967-1145; e-mail: mrosenberg@healthdec.com

Submitted for publication September 28, 1998Revised October 27, 1998Accepted for publication October 30, 1998

© 1998 Elsevier Science Inc. All rights reserved. ISSN 0010-7824/98/$19.00655 Avenue of the Americas, New York, NY 10010 PII S0010-7824(98)00127-9

Materials and MethodsSubjects ranged in age from 18–35 years old, had nottaken hormonal contraception for at least 2 monthsbefore enrollment in the study, and had completedtwo normal menstrual cycles before enrollment. Sub-jects received a triphasic oral contraceptive (Tri-Nori-nyl) with 35 mg of ethinyl estradiol (EE) on each pillday and a variable amount of norethindrone (NET): 50mg on days 1–7, 100 mg on days 8–14, and 50 mg ondays 15–21. Inert pills were provided for the remain-ing 7 days of each cycle. Of the 177 women enrolled,133 (75%) completed all four cycles with dailyweights. Five women were excluded from analysis

because of insufficient data, leaving 128 women whocomprise the basis of this report.

During the four-cycle duration of the study, sub-jects noted weight in a daily record form, each day inthe morning, without clothes. Analyses are based ondifferences between baseline weight and weight at theend of the fourth cycle. We tested the possibility thatage, parity, and race may have influenced weight(effect modification) by stratifying according to eachand testing for heterogeneity, with p values of ,0.05interpreted as indicating a significantly different ef-fect among each of the different strata.

ResultsStudy participants were primarily young, nulliparous,and white (Table 1). The mean weight change be-tween study start and completion was 0.0 pounds(Figure 1). There was also no difference in meanweights over the four-cycle observation period ifweights are examined on any given cycle day. Forexample, when day 7 of each cycle is examined, nodifferences in mean weight are apparent.

Consideration of weight change for individuals alsoindicates that the majority of women experiencedlittle or no change (Figure 2). The largest single groupof women, 52%, experienced no weight change frombeginning to end of the study, defined as ,2 pounds

Table 1. Characteristics of study participants (n 5 128)

Characteristic (%)

Age (years)18–23 (54)24–29 (26)30–35 (20)

Parity0 (72)1 (13)21 (15)

RaceWhite (91)Black 1 Hispanic (9)

Figure 1. Mean weight change by day.

346 Rosenberg Contraception1998;58:345–349

(0.9 kg) difference. An additional 33% of women hada minimal change of ,5 pounds (2.3 kg), and theproportion who gained weight (28%) was similar tothat losing weight (20%). Thus, 85% of women in thestudy experienced weight change of #5 pounds (2.3kg). Women who had significant weight change wereapproximately equally distributed between those whogained and those who lost weight.

As a given weight change gain is more meaningfulfor a woman who weighs 90 pounds than for one whoweighs 190 pounds, percentage change in weight wasalso examined. As a percentage of body weight, for allwomen, there was a gain of 0.2%. This figure isslightly higher than the mean weight change becausethe greatest absolute weight change (loss) occurred inwomen who were heaviest, but this change was lessas a percentage of baseline body weight than forwomen who were lighter. The results of percentageweight change for individuals were also very similarto absolute changes. Seventy percent of subjects hadeither no weight change or a loss. Of the 31% ofsubjects who gained weight, more than four-fifthsgained ,5%. Thus, only 5% of subjects gained .5%in weight while using the OC.

The largest weight change was a loss of 31 poundsexperienced by a 29-year-old, para 4, white woman.Three nulliparous women, aged 18, 22, and 32 years

old, had weight losses of 11, 11, and 16 pounds,respectively. Two white women, a 22-year-old, para 2and 34-year-old, para 4, experienced weight gains of13 and 15 pounds. All other weight changes were inthe range of 210 to 110 pounds.

When women were stratified by age, race, andparity, minor differences in each group were foundthat did not achieve statistical significance (p ,0.05)and are considered chance findings. Among agegroups, the youngest women (18–23 years old) gaineda mean of 0.1 pounds, while women 22–29 years ofage had zero difference, and older women (30–35years of age) lost a mean of 0.3 pounds. For race, themean change was 0.0 for white and 0.3 for Hispanicand black women. Nulliparous women had weightchange of 0.0 pounds, while parity 1 women had again of 0.1 pounds and parity 21 women lost 0.2pounds, on average.

Weight Fluctuation Throughout theMenstrual CycleDaily weights indicate that women tended to gainweight during the first few weeks of each treatmentcycle and lose approximately the same amount ofweight during menstruation. These fluctuationswere, however, relatively small: the mean weight

Figure 2. Mean weight change from baseline.

347Contraception Oral Contraceptives and Weight Gain1998;58:345–349

fluctuated between a gain of one-half pound (0.2 kg)early in the cycle to a loss of one-half pound towardsthe end. There was also a suggestion that weightlosses tended to be slightly greater during the firstcycle than during the second or later cycles.

DiscussionMost women starting OC experienced no change inweight, and those who did were approximately equallydistributed between gain and loss. These results didnot differ when considered based on percentagechange or between different groupings of age, race,and parity.

Our findings are consistent with those from otherstudies that find little or no weight change associatedwith OC use. Few of these, however, measured otherfactors that might have affected weight, such as dietand physical activity. One study that did measurethese factors found no weight gain and a slight butsignificant increase in basal metabolic rate among OCusers.10 A second study compared 50 women usingOC with the same number of users of Norplant andDepo-Provera, finding a slight drop of 2 pounds inweight among OC users and no significant differencein women using other methods.8 The largest studywas a 2-year longitudinal evaluation of 4342 women.This study involved weights every 3 months andfound significant weight increases of 0.1–1.0 kg overbaseline after the first year and 1.1–1.7 kg the follow-ing year. Because this was uncontrolled, however, itcould not differentiate between weight gain associ-ated with OC or other activities such as physicalactivity and diet. A second uncontrolled study of OCusers found a nonsignificant weight gain of 0.2 kgover a six-cycle duration.5 Three studies that in-cluded a comparison group all found nonsignificantdifferences between women using OC and those us-ing other methods. One was a retrospective chartreview in women 12–25 years of age,6 one a prospec-tive study in similarly young women,4 and the last acomparison of 138 OC users and a group of 35 womenusing intrauterine devices (IUD), foam, diaphragms,or condoms over a year’s time.8

The present study is the first to measure dailyweights throughout each cycle, a subject on which wefound no reference in gynecology textbooks. Thefinding of slight but consistent weight fluctuationsover each cycle indicates that when in the cycleweight is measured is important for accurate weightdetermination. Use of OC may in some cases beblamed for weight gain, and possibly be discontinuedas a result, although at least some change may beattributable to normal cyclic fluctuations. This study

also involved women broadly representative of OCusers in the United States.

The present study also has a number of weaknesses.First, it was a relatively brief investigation, of onlyfour cycles. Use of OC could conceivably be associ-ated with weight gain over a longer interval, in whichcase it may not have been apparent during this briefexamination. Second, we relied on measurementsfrom patients themselves, using uncalibrated scales.However, our concern was in relative weights, andsubstantial drift over 4 months seems unlikely. Inaddition, women were carefully instructed aboutcompleting the daily record form, including a stan-dardized time of weighing themselves, and specifyingthat weighing be done without clothes. A third weak-ness was the lack of a control group or informationabout food intake, physical activities, or other activ-ities that might affect caloric balance. Women alsomay have altered their behavior because they wereaware that weight was being recorded. A controlgroup would be important if we found increases inweight, to separate the effect of OC from othercauses. Our finding of an absence of a difference inoverall weight makes a control group less important.Similarly, lack of information about other factorsaffecting weight are of lesser import because, if OCreally did cause weight gain, one would have tohypothesize that women increased caloric expendi-ture or decreased intake to account for our findings ofno change. This seems unlikely. Finally, because westudied only a single OC preparation, our findings,although consistent with studies of other formula-tions, may be limited to this one.

Clinical ImplicationsThis and other studies consistently indicate thatweight change is unrelated to OC use, but there areminor fluctuations during each cycle. These findings,however, remain of scant comfort in the shadow ofthe common perception to the contrary. Contracep-tive counseling, particularly initial discussions,should stress this fact. If weight change does occur,whether gain or loss, factors such as exercise, eatingpatterns, lifestyle changes, and others potentiallyinfluencing the balance of caloric intake and expen-diture should be examined.

References1. Pratt WF, Bachrach CA. What do women use when

they stop using the pill? Fam Plann Perspect 1987;19(suppl 6):257–66.

2. Rosenberg MJ, Waugh MS, Meehan TE. Use and misuse oforal contraceptives: risk indicators for poor pill taking anddiscontinuation. Contraception 1995;51:283–8.

348 Rosenberg Contraception1998;58:345–349

3. Serfaty D. Oral contraceptive compliance during ado-lescence. Ann NY Acad Sci 1997;816:422–431.

4. Reubinoff BE, Grubstein A, Meirow D, et al. Effects oflow-dose estrogen oral contraceptives on weight, bodycomposition, and fat distribution in young women.Fertil Steril 1995;63:516–21.

5. Bannemerschult R, Hanker JP, Wunsch C, et al. Amulticenter, uncontrolled clinical investigation of thecontraceptive efficacy, cycle control, and safety of anew low dose oral contraceptive containing 20 mgethinyl estradiol and 100 mg levonorgestrel over sixtreatment cycles. Contraception 1997;56:285–90.

6. Carpenter S, Neinstein LS. Weight gain in adolescentand young oral contraceptive users. J Adolesc Hlth Care1986;7:342–4.

7. Woutersz TB, Korba VD. Five-year, multicenter studyof a triphasic, low-dose, combination oral contracep-tive. Int J Fertil 1988;33:406–10.

8. Moore L, Valuck R, McDougall C, Fink W. A com-parative study of one-year weight gain among users ofmedroxyprogesterone acetate, levonorgestrel im-plants, and oral contraceptives. Contraception 1995;52:215–20.

9. Goldzieher JW, Moses LE, Averkin E, et al. A placebo-controlled double-blind crossover investigation of theside effects attributed to oral contraceptives. FertilSteril 1971;22:609–23.

10. Diffey B, Liers LS, Soares MJ, O’Dea K. The effect oforal contraceptive agents on the basal metabolic rate ofyoung women. Br J Nutr 1997;77:853–62.

349Contraception Oral Contraceptives and Weight Gain1998;58:345–349