Oral contraceptive pills (OCPs) containing a combination of ethinyl estradiol (E2) and a progestin remain the most popular form of reversible contraception in the United States.1 Oral contraceptive pill failure rates, based on data from clinical trials, are less than 2% per year of use.2 In contrast, OCP failure rates based on data from the National Survey of Family Growth are as high as 8%.1 Many differences between clinical trial participants and the nationally representative National Survey of Family Growth sample may contribute to the reported differences in OCP failure rates.3 One such difference is that past OCP clinical trials have typically included only women who are close to ideal body weight, even though now more than 60% of reproductive-age women in the United States are either overweight or obese.4 If the OCP is less effective among overweight or obese women, then clinical trial data from mainly normal-weight women will overestimate OCP effectiveness for that segment of the U.S. population.
In two studies, Holt et al5,6 found that self-reported OCP failure varied by body weight or body mass index (BMI). In those studies, heavier women were more likely to report an unintended pregnancy during OCP use than women who weighed less. Most other studies, as reviewed by Trussell et al7 and Kaneshiro et al,8 reported little or no association between obesity and OCP failure. More recently, Dinger et al3 assessed OCP failure in a prospective European cohort of 59,510 OCP users and found little association between BMI and OCP failure. Westhoff et al9 found no association between BMI and OCP failure in a clinical trial of 2,185 U.S. women, one-quarter of whom were obese.
Previous studies do not allow us to decide whether reported differences in OCP effectiveness between normal-weight and obese women are physiological factors or behavioral differences in OCP use. The present study was designed to isolate differences in ovarian suppression among normal-weight compared with obese OCP users. Ovarian suppression and concomitant anovulation is the main mechanism of OCP effectiveness. If ovarian suppression is similar among normal-weight and obese OCP users, then reported differences in OCP failure may be attributed to differences in pill-taking behaviors.
PATIENTS AND METHODS
This clinical trial compared ovarian follicle development and ovulation between normal-weight and obese women randomly assigned to one of two OCP formulations. Participant-related activities were conducted between July 2006 and December 2008 in New York City. The Columbia University Institutional Review Board approved the study, and all participants gave informed consent. We recruited participants via online advertisements, newspapers, and fliers.
Eligible women were aged 18 to 35 years, with a recent history of regular, spontaneous menstrual cycles, and agreed to use an OCP for 3 to 4 months and undergo eight biweekly study visits during the third or fourth OCP cycle. We excluded women with medical contraindications to use of combined OCP based on the World Health Organization Medical Eligibility Criteria,10 enrolling only women in World Health Organization category 1. We also specifically excluded women recently pregnant or breastfeeding, those with preexisting renal disease, those with diabetes, thyroid, pituitary, adrenal, or ovarian disorders, and those smoking more than 10 cigarettes per day. Women with a history of gastric bypass or who had used depot medroxyprogesterone acetate within 12 months were ineligible. Women using medications known to affect the CYP450 system, including antiseizure medicines, rifampin, griseofulvin, or fluconazole, were ineligible. All participants had normal-appearing ovaries on baseline sonogram using a TITAN (SonoSite, Bothell, WA) with a 7.5-MHz transvaginal probe and were either normal weight (BMI 19.0–24.9) or obese (BMI 30.0–39.9) based on standardized height and body weight measurements at enrollment. We measured standing height (cm) using a stadiometer. We measured body weight (kg) and composition (total body water [kg], percent body fat, fat mass [kg], and fat-free mass [kg]) by bioelectrical impedance analysis using the BC-418 (Tanita, Tokyo, Japan), which is an eight-contact electrode, single-frequency, 50-kHz body composition analyzer. We used the same instrument for all body weight and composition measurements.
We initially excluded current OCP users, but because of slow enrollment, after May 2007, we extended recruitment to current users if they had no history of irregular menstrual cycles. In February 2008, after we had enrolled 100 normal-weight participants and 57 obese participants, we limited the enrollment of normal-weight women to Hispanic and African American women so the ethnic breakdown of the normal-weight stratum would more closely resemble the obese stratum.
We stratified participants by BMI group at enrollment and randomly assigned them to one of two monophasic OCP formulations: 30-microgram ethinyl E2 and 150-microgram levonorgestrel or 20-microgram ethinyl E2 and 100-microgram LNG, both packaged with 21 active and 7 placebo tablets (Portia and Lessina, respectively, Barr Laboratories, Montvale, MJ). The sequence for the 1:1 treatment allocation was determined using a random-number table constrained by the use of randomly permutated blocks. An investigator not involved with participant contact generated the allocation schedule. After consent and screening, each participant received a sequentially numbered opaque envelope containing four OCP packs. Product labeling was obliterated by an opaque adhesive sticker placed onto each pill pack to blind participants to OCP formulation. All participants and clinicians who performed and interpreted the sonograms were blinded to treatment assignment for the duration of the study. Laboratory personnel were blinded to treatment assignment and to whether specimens belonged to normal-weight or obese participants. Participants completed at least two OCP cycles before the study cycle.
During the study cycle, participants underwent biweekly vaginal ultrasonograms for 4 weeks to measure ovarian follicle-like structures; we measured all follicles in two perpendicular diameters and recorded the dimensions of all follicles with a mean diameter of at least 8 mm. We also noted the presence of a corpus luteum or ruptured follicle and measured anterior–posterior endometrial thickness. A single investigator (J.P.L.) trained the others and then reviewed all sonograms with any perpendicular diameters larger than 13 mm for quality control. Participants underwent venipunctures on the same days as the ultrasonograms to measure endogenous E2, progesterone, and levonorgestrel. Specimens were allowed to clot for at least 10 minutes at room temperature and then separated and stored at −80°C until analysis.
Participants recorded daily bleeding and spotting throughout the study using paper diaries. The loss of blood requiring the use of sanitary pad or tampon was noted as “bleeding,” and the loss of blood requiring a panty liner or no protection as “spotting.”
The Core Laboratory of the Irving Institute of Clinical and Translational Research at Columbia University Medical Center performed E2 and progesterone assays using the radioimmunoassay kits Double Antibody E2 and Coat- A-Count Progesterone (Siemens Medical Solution Diagnostics, Malvern, PA). All specimens from an individual participant were analyzed in the same run. To the extent possible, we created quartets such that each assay contained specimens from two normal-weight and two obese participants, with each randomly assigned to one of the OCPs. We measured levonorgestrel in specimens collected on days 2 to 21 of the study cycle in the Reproductive Endocrine Research Laboratory at the University of Southern California. Levonorgestrel levels in serum samples were quantified by specific and sensitive radioimmunoassays as previously described.11
To evaluate follicular development and ovulation during OCP use, we excluded participants who used the OCP inconsistently or not at all. We considered study participants to be consistent OCP users, inconsistent OCP users, or nonusers based on the results of 903 levonorgestrel assays from 181 participants during days 2 to 21 of the study cycle. We considered a participant an OCP nonuser during the study cycle if all levonorgestrel values were less than 0.16 ng/mL, which was the sensitivity limit of the assay. We classified a participant to be a consistent OCP user if she had no more than one specimen with an levonorgestrel value less than 1.0 ng/mL. Finally, we considered a participant to have used the OCP inconsistently if at least two of her levonorgestrel values were less than 1.0 ng/mL, but not all of her values were less than 0.16 ng/mL. We chose the 1.0-ng/mL threshold based on the results of a pharmacokinetic substudy.12 We defined compliance and assigned participants to compliance subgroups without knowledge of treatment group, participant characteristics, or study outcomes. We used χ2 tests to assess whether consistent OCP use was differential by BMI group or OCP dose.
The primary measure of ovarian suppression in this study was the proportion of participants in each group who had a maximum perpendicular diameter of 18 mm or more during OCP use. After excluding participants who used the OCP inconsistently or not at all, we calculated the proportion of participants within each BMI group and each OCP group with perpendicular diameter larger than 18 mm; we also assessed other prespecified thresholds (8, 10, and 13 mm).13–20 We used χ2 tests to assess the association between obesity or OCP dose and maximum perpendicular diameter. We also used logistic regression to investigate the combined effect of dose and obesity on having an perpendicular diameter of 18 mm or more. We considered race/ethnicity, education, the interaction between dose and BMI group, and all other possible two-way interaction terms as potential variables in the model. We were interested in percent body fat, but the distributions of percent body fat were completely separated for the normal-weight and obese groups, which would make a model with percent body fat uninterpretable.
We defined ovulation as having at least one serum progesterone level of 3 ng/mL or greater. We used perpendicular diameter, E2, and progesterone values to create Hoogland Scores, which have been used by other investigators to assess ovarian suppression during OCP use.14,15,21 The Hoogland Score comprises six grades: 1) no activity, 2) potential activity, 3) nonactive follicle-like structure, 4) active follicle-like structure, 5) luteinized unruptured follicle, and 6) ovulation. Because of small numbers we combined grades 5 and 6.
For each participant, we calculated mean endometrial thickness, maximum E2, and number of bleeding and spotting days during the study cycle. We used t tests or Wilcoxon rank sum tests to compare these variables between each BMI and OCP dose group.
A sample size of 100 participants in each group was planned a priori to have 80% power to differentiate between 20% and 38% of participants having a maximum perpendicular diameter of 18 mm or more. To detect differences in the proportion of participants with smaller (and more prevalent) follicles, the study has greater power (for all comparisons, two-sided alpha=0.05). In the event, the analyzable dataset included 150 participants, giving 66% power to detect the same difference in proportions. Statistical analyses were performed using SAS statistical packagev.9 (SAS Institute, Cary, NC).
Figure 1 shows the flow of participants through this trial. Of the 1,028 women who underwent telephone screening, 252 were presumed eligible and invited for screening. During screening, 26 participants did not meet eligibility criteria; thus, we randomly assigned 226 participants (128 normal weight and 98 obese) and assigned 114 to 20-microgram ethinyl E2/100-microgram levonorgestrel OCP and 112 to 30-microgram ethinyl E2/150-microgram levonorgestrel OCPs. Twenty participants were lost to follow-up after the enrollment visit, and 25 withdrew before the study cycle; these 45 do not contribute any results to the analyses. Overall, 19.9% dropped out (17.2% in the normal-weight group, 23.4% in the obese group, P=.24). Women who completed the study reported more education (P=.002), were less likely to have a previous pregnancy (P=.06), and were more likely to be hormonal contraceptive users at enrollment (P=.05) than participants who dropped out. These findings persisted within each weight and dose stratum. One hundred eighty-one participants completed the study cycle.
Thirty-one of the 181 participants who participated in the study cycle were noncompliant with OCP use based on levonorgestrel assays. One hundred fifty participants were consistent OCP users during the study cycle; therefore, analyses of the primary outcomes include these 150 (66%) of the 226 randomly assigned participants: 96 (75%) of the 128 normal-weight and 54 (55%) of the 98 obese participants, 70 (61%) of the 114 participants assigned to 20-microgram ethinyl E2/100-microgram levonorgestrel OCP and 80 (71%) of the 112 participants assigned to 30-microgram ethinyl E2/150-microgram levonorgestrel OCP.
Follow-up visits occurred during the third or fourth cycle for 145 (97%) of 150 participants and during the fifth to seventh cycle for five of 150 (3%) participants. One hundred thirty-four (89%) participants completed at least seven biweekly sonograms.
Baseline characteristics of 150 participants who completed the study and who were compliant with OCP use are presented in Table 1. Marked differences between the subgroups in body composition measurements were attributable to deliberate stratification. Despite efforts to balance the two groups, normal-weight and obese participants differed with respect to race/ethnicity. Obese women were more likely to be parous. Women randomly assigned to the 30-microgram ethinyl E2/150-microgram levonorgestrel OCP were more likely to be current OCP users than women randomized to the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP. The few participants who smoked cigarettes reported smoking, on average, three or fewer cigarettes per day.
Participants who did not complete follow-up and those who were noncompliant with OCP use were less likely to be using the OCP at enrollment, more likely to be parous, and had lower educational attainment than the participants who completed follow-up (data not shown). Table 2 compares the 31 noncompliant and the 150 compliant participants by obesity status. The noncompliant participants were more likely to be obese (P<.005), but compliance did not differ by treatment assignment (P=.13).
Twenty-one participants ovulated during the study cycle; only four of these were consistent OCP users (Table 3). Inconsistent OCP use or nonuse during the study cycle was associated with more ovulation than consistent OCP use (P<.001). Three of the consistent OCP users who ovulated were normal weight and one was obese; two were assigned to 20-microgram ethinyl E2/100-microgram levonorgestrel OCP and two were assigned to 30-microgram ethinyl E2/150-microgram levonorgestrel OCP. One normal-weight participant ovulated based on serum progesterone of 5.8 ng/mL at cycle day 25; another ovulated based on serum progesterone of 12.4 ng/mL at day 13. A third ovulated based on serum progesterone of 13.9 ng/mL at cycle day 4. One obese participant ovulated based on serum progesterone of 10.1 ng/mL at cycle day 9. Based on the same criteria, 5 of 13 inconsistent OCP users ovulated: two of four normal-weight and three of nine obese women.
During the study cycle, a substantial proportion of participants had perpendicular diameter 18 mm or larger; however, this did not differ between normal-weight and obese participants (25 of 96 [26%] compared with 8 of 54 [14.8%], P=.11). The prevalence of smaller follicles was likewise similar between these groups (data not shown). In addition, we found little difference in perpendicular diameter 18 mm or larger by OCP dose (24.3% compared with 20% for 20-microgram ethinyl E2/100-microgram levonorgestrel OCP compared with 30-microgram ethinyl E2/150-microgram levonorgestrel OCP, respectively, P=.53). Based on the likelihood criterion, our final logistic regression model included only BMI group and OCP dose, and neither obesity (odds ratio 0.5, 95% confidence interval 0.2–1.2) nor dose (odds ratio 1.2, 95% confidence interval 0.6–2.7) was a significant factor for failure of ovarian suppression. Race or ethnicity and education were not associated with failure of ovarian suppression. There was also insufficient evidence to suggest a statistically significant interaction between OCP dose and obesity (odds ratio 0.2, 95% confidence interval 0.03–1.5).
By comparing endogenous E2 results during the entire cycle, the maximum E2 levels were low and similar for normal-weight and obese participants (median 19.7 pg/mL compared with 18.1 pg/mL, respectively, P=.48), indicating substantial ovarian suppression. The maximum E2 levels during the study cycle were somewhat higher among the participants using the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP than the 30-microgram ethinyl E2/150-microgram levonorgestrel OCP (median 24.5 pg/mL compared with 14.9 pg/mL, respectively, P<.01). The sensitivity limit of this assay was 3 pg/mL. Normal-weight and obese participants had similar E2 levels at all visits (data not shown). In contrast, the higher levels E2 achieved by women randomly assigned to the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP appear to occur during the first and last week of the cycle (data not shown). Using perpendicular diameter, E2 and progesterone levels together to calculate Hoogland scores, Tables 4 and 5 show no meaningful differences between the normal-weight and obese participants, or between the participants randomly assigned to the 20-microgram ethinyl E2/100-microgram levonorgestrel or 30-microgram ethinyl E2/150-microgram levonorgestrel OCP.
During the study cycle, 133 participants completed at least 25 days in their bleeding diaries; these participants reported a mean of 3.8 bleeding days and 2.5 spotting days. The reported bleeding and spotting days did not differ between normal-weight and obese participants (P=.80); however, participants randomly assigned to the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP reported more spotting days than those randomly assigned to the 30-microgram ethinyl E2/150-microgram levonorgestrel OCP (mean 3.0 compared with 1.9, P=.02). The mean endometrial stripe thickness during the study cycle varied between 4 and 5 mm and did not differ by OCP dose or by obesity.
This study compares ovarian effects of the OCP in normal-weight compared with obese women. Perpendicular diameter itself, and even ovulation, however, are not clinically relevant outcomes; the important clinical outcome for women using the OCP is unintended pregnancy. Previous articles3,5,6,8,9,22–26 that reported on unintended pregnancy among normal-weight compared with obese OCP users cannot, however, disentangle whether these pregnancies were attributable to biologically mediated OCP failure or to behavioral differences in OCP use. Our results indicate no difference in ovarian suppression during consistent OCP use in normal-weight compared with obese participants; thus, differences in unintended pregnancy must be attributable to other factors.
Our results show similar follicular development in participants using the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP compared with the 30-microgram ethinyl E2/150-microgram levonorgestrel OCP, but participants consistently using the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP had higher E2 levels during the first week of the cycle and the placebo week, indicating somewhat less ovarian suppression than in participants using the 30-microgram ethinyl E2/150-microgram levonorgestrel OCP. Finally, those using the 20-microgram ethinyl E2/100-microgram levonorgestrel OCP also reported greater spotting. Nonetheless, participants using the lower-dose OCP had substantial ovarian suppression, consistent with its known contraceptive effectiveness. These are expected differences between the doses and thus support the validity of this study.27
We identified very few ovulations among the consistent OCP users. A strength of this study is that we defined consistent OCP use based on serum levonorgestrel levels, rather than based on self-report. The consistent users may have had some irregularity of their OCP intake—we cannot define them as perfect or correct OCP users. The participants who were nonusers or inconsistent OCP users based on levonorgestrel levels (not based on self-report) experienced substantial rates of ovulation whether they were normal weight or obese, confirming that consistent OCP use is the key to avoiding an unintended pregnancy.
In the general population, obese women may be anovulatory more often than normal-weight women and thus at lower risk of OCP failure. In this study, we tried to enroll only participants with a normal menstrual history and normal ovaries at the baseline ultrasonogram, because these are the women who would be most at risk of an OCP failure. We did not confirm ovulation before enrollment and did not evaluate more subtle difference in hypothalamic-pituitary-ovarian axis functioning.28 Nonetheless, excess OCP failures among obese women, if any, would be most likely to occur in the subset of obese women from whom we enrolled into this study.
An unexpected finding in this study was the differential rate of OCP noncompliance; our obese participants were substantially less likely than the normal-weight participants to use the OCP consistently during the study cycle (despite reporting correct use to the investigators). If we did not have direct biological markers of OCP compliance, then we would have incorrectly concluded that the obese women had more ovulation during OCP use when, in fact, the extra ovulations occurred during OCP nonuse. In the general population, obesity is associated with poverty, less education, and other characteristics of social deprivation.29,30 These factors have been associated with higher unintended pregnancy rates.31 These factors may also influence who volunteers to participate in a clinical trial, such as this one, that offers monetary compensation for time and effort across many study visits.
In conclusion, this trial shows that consistent OCP users, whether normal weight or obese, have substantial ovarian suppression. Obese women are not at a higher risk of pregnancy during consistent OCP use.
1. Chandra A, Martinez GM, Mosher WD, Abma JC, Jones J. Fertility, family planning, and reproductive health of U.S. women: data from the 2002 National Survey of Family Growth. Vital Health Stat 23 2005:1–160.
2. Trussell J. Contraceptive failure in the United States. Contraception 2004;70:89–96.
3. Dinger JC, Cronin M, Mohner S, Schellschmidt I, Minh TD, Westhoff C. Oral contraceptive effectiveness according to body mass index, weight, age, and other factors. Am J Obstet Gynecol 2009;201:263.e1–9.
4. McDowell MA, Fryar CD, Ogden CL, Flegal KM. Anthropometric reference data for children and adults: United States, 2003–2006. National health statistics reports; no 10. Hyattsville (MD): National Center for Health Statistics; 2008.
5. Holt VL, Cushing-Haugen KL, Daling JR. Body weight and risk of oral contraceptive failure. Obstet Gynecol 2002;99:820–7.
6. Holt VL, Scholes D, Wicklund KG, Cushing-Haugen KL, Daling JR. Body mass index, weight, and oral contraceptive failure risk. Obstet Gynecol 2005;105:46–52.
7. Trussell J, Schwarz EB, Guthrie K. Obesity and oral contraceptive pill failure. Contraception 2009;79:334–8.
8. Kaneshiro B, Edelman A, Carlson N, Nichols M, Jensen J. The relationship between body mass index and unintended pregnancy: results from the 2002 National Survey of Family Growth. Contraception 2008;77:234–8.
9. Westhoff C, Reape K, Shu H. Impact of body weight on observed pregnancy rates with a low-dose estrogen, 91-day extended regimen oral contraceptive. Contraception 2009;80:196–7.
10. World Health Organization Department of Reproductive Health and Research. Medical eligibility criteria for contraceptive use. 3rd ed. Geneva (Switzerland): World Health Organization; 2004.
11. Stanczyk FZ, Hiroi M, Goebelsmann U, Brenner PF, Lumkin ME, Mishell DR Jr. Radioimmunoassay of serum d-Norgestrel in women following oral and intravaginal administration. Contraception 1975;12:279–98.
12. Westhoff CL, Torgal AH, Mayeda ER, Pike MC, Stanczyk FZ. Pharmacokinetics of a combined oral contraceptive in obese and normal-weight women. Contraception 2010;81:474–80.
13. Broome M, Clayton J, Fotherby K. Enlarged follicles in women using oral contraceptives. Contraception 1995;52:13–6.
14. Schlaff WD, Lynch AM, Hughes HD, Cedars MI, Smith DL. Manipulation of the pill-free interval in oral contraceptive pill users: the effect on follicular suppression. Am J Obstet Gynecol 2004;190:943–51.
15. Hoogland HJ, Skouby SO. Ultrasound evaluation of ovarian activity under oral contraceptives. Contraception 1993;47:583–90.
16. Grimes DA, Godwin AJ, Rubin A, Smith JA, Lacarra M. Ovulation and follicular development associated with three low-dose oral contraceptives: a randomized controlled trial. Obstet Gynecol 1994;83:29–34.
17. Klipping C, Duijkers I, Trummer D, Marr J. Suppression of ovarian activity with a drospirenone-containing oral contraceptive in a 24/4 regimen. Contraception 2008;78:16–25.
18. Endrikat J, Parke S, Trummer D, Schmidt W, Duijkers I, Klipping C. Ovulation inhibition with four variations of a four-phasic estradiol valerate/dienogest combined oral contraceptive: results of two prospective, randomized, open-label studies. Contraception 2008;78:218–25.
19. Rible RD, Taylor D, Wilson ML, Stanczyk FZ, Mishell DR Jr. Follicular development in a 7-day versus 4-day hormone-free interval with an oral contraceptive containing 20 mcg ethinyl estradiol and 1 mg norethindrone acetate. Contraception 2009;79:182–8.
20. Spona J, Binder N, Höschen K, Feichtinger W. Suppression of ovarian function by a combined oral contraceptive containing 0.02 mg ethinyl estradiol and 2 mg chlormadinone acetate given in a 24/4-day intake regimen over three cycles. Fertil Steril 2009 Aug 24. [Epub ahead of print.]
21. Endrikat J, Gerlinger C, Plettig K, Wessel J, Schmidt W, Grubb G, et al. A meta-analysis on the correlation between ovarian activity and the incidence of intermenstrual bleeding during low-dose oral contraceptive use. Gynecol Endocrinol 2003;17:107–14.
22. Brunner LR, Hogue CJ. The role of body weight in oral contraceptive failure: results from the 1995 national survey of family growth. Ann Epidemiol 2005;15:492–9.
23. Brunner Huber LR, Hogue CJ. The association between body weight, unintended pregnancy resulting in a live birth, and contraception at the time of conception. Matern Child Health J 2005;9:413–20.
24. Brunner Huber LR, Hogue CJ, Stein AD, Drews C, Zieman M. Body mass index and risk for oral contraceptive failure: a case-cohort study in South Carolina. Ann Epidemiol 2006;16:637–43.
25. Brunner Huber LR, Toth JL. Obesity and oral contraceptive failure: findings from the 2002 National Survey of Family Growth. Am J Epidemiol 2007;166:1306–11.
26. Vessey M. Oral contraceptive failures and body weight: findings in a large cohort study. J Fam Plann Reprod Health Care 2001;27:90–1.
27. Gallo MF, Nanda K, Grimes DA, Lopez LM, Schulz KF. 20 microg versus >20 microg estrogen combined oral contraceptives for contraception. Cochrane Database of Systematic Reviews, 2008. Issue 4. Art. No.: CD003989.
28. Edelman AB, Carlson NE, Cherala G, Munar MY, Stouffer RL, Cameron JL, et al. Impact of obesity on oral contraceptive pharmacokinetics and hypothalamic-pituitary-ovarian activity. Contraception 2009;80:119–27.
29. Wang Y, Beydoun MA. The obesity epidemic in the United States: gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007;29:6–28.
30. Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 2010;303:235–41.
31. Finer LB, Henshaw SK. Disparities in rates of unintended pregnancy in the United States, 1994 and 2001. Perspect Sex Reprod Health 2006;38:90–6.