In the United States, reproductive-aged women represent 40% of patients undergoing bariatric surgery (ie, at least 50,000 inpatient cases annually).1–3 Clinical practice guidelines cosponsored by the American Association of Clinical Endocrinologists, the Obesity Society, and the American Society for Metabolic & Bariatric Surgery in 2013 recommend that women avoid conception for 12–18 months after bariatric surgery.4 The delay is intended to optimize the likelihood of maternal weight stability during fetal growth. This recommendation precedes recent evidence suggesting conception within 2 years of bariatric surgery increases the risk for prematurity, small-for-gestational-age neonates, and neonatal intensive care unit admissions.5
We previously reported that future pregnancy was important to 30% of reproductive-aged women at the time of bariatric surgery.6 One third of these women reported an intention to conceive within 2 years of surgery.6 Despite the magnitude of this public health concern, little is known regarding the postoperative reproductive behavior of these patients.
The Longitudinal Assessment of Bariatric Surgery-2 study is a multicenter, prospective cohort study that has previously reported the preoperative reproductive health status6 as well as the safety and efficacy of bariatric surgery procedures7 in adults undergoing bariatric surgery. The aim of this study is to report postsurgical contraceptive practices and conception rates among reproductive-aged women who undergo bariatric surgery and determine factors related to 1) postsurgical unprotected intercourse while not trying to conceive, 2) early (less than 18 months) postsurgical conception, and 3) delayed (18 or greater to 42 months) postsurgical conception.
MATERIALS AND METHODS
Participants were recruited into Longitudinal Assessment of Bariatric Surgery-2 , a multicenter prospective cohort study, at 10 hospitals within six clinical centers throughout the United States between 2005 and 2009. Patients were at least 18 years old undergoing a first bariatric surgical procedure by participating surgeons as part of routine clinical care. Recruitment methodology and study design have previously been described.6–10 Before data collection, the institutional review boards at each center approved the protocol and all participants gave written informed consent to participate. Research assessments were conducted by Longitudinal Assessment of Bariatric Surgery-2-trained and certified personnel independent of surgical care within 30 days before scheduled surgery dates and annually postoperatively until January 2015 for up to 7 years. Details of the Longitudinal Assessment of Bariatric Surgery-2 cohort have previously been published.6,9
This report was restricted to women who were 18–44 years old and reported no history of surgical or natural menopause, hysterectomy, or estrogen and progestin therapy (indicated as “hormone replacement therapy” on the data collection form) before their preoperative or first follow-up reproductive health assessment. Data collected after any of these criteria were met were excluded. Of 1,931 female participants, 740 women met eligibility requirements, 710 (95.9%) of whom reported postsurgical contraceptive practices or conception (Fig. 1).
The preoperative Reproductive Health Questionnaire has been described (Appendix 1, available online at http://links.lww.com/AOG/B20).6 A postsurgical version was self-administered annually to obtain outcomes in the year before assessment (Appendix 2, available online at http://links.lww.com/AOG/B20).10 Primary outcomes were frequency of contraceptive need and use (as defined subsequently) and conception rates. Secondary outcomes included identification of risk factors for postsurgical unprotected intercourse while not trying to conceive, early postsurgical conception, and delayed postsurgical conception.
Items assessing frequency of contraceptive use during sexual intercourse with a male partner and attempts to conceive were used to categorize women as having 1) no sexual intercourse with a male partner; 2) protected intercourse only (ie, used contraception during intercourse with a male partner “all of the time”); 3) unprotected intercourse with a male partner; or 4) tried to conceive in the past year. This classification scheme reflects the fact that the proportion of women reporting contraceptive use “most of the time” or “approximately half of the time” was small and that such practices increase risk of pregnancy. Thus, these categories were grouped with “rarely” and “never” as “unprotected intercourse.”6,11 Women could only be positive for one of these four dichotomous (yes or no) variables; trying to conceive overrode the other options. Items assessing contraception for any reason, by method, were used to determine past year use of any contraception, more than one method of contraception, oral contraception alone, and specific methods of contraception alone or in combination.
In addition to the annual postsurgical Reproductive Health Questionnaire, an annual Short Form (administered starting March 2010) and an Event and Complications Form (completed at the 4- or 5-year postsurgical assessment) assessed pregnancies “since surgery” to address missing data. If a postsurgical pregnancy was reported on any form, participants were asked to complete a Pregnancy Questionnaire (Appendix 3, available online at http://links.lww.com/AOG/B20). If unavailable, conception date was estimated based on other data points (due date, length of pregnancy, pregnancy outcome, and outcome date). Inconsistent data related to conception date were resolved through independent adjudication by two authors (M.N.M., G.E.W.) with resolution by a third author (W.C.K.) (Appendix 4, available online at http://links.lww.com/AOG/B20).
The number of pregnancies and follow-up time were used to calculate conception rates. To account for variation in follow-up, woman-years of follow-up were calculated using the time from date of surgery to the last date pregnancy status was known through the 7-year assessment window (90 months postoperatively). Rates are reported per 1,000 woman-years, which is representative of the percentage of participants who would conceive over 10 years if each woman conceived no more than once. Postsurgical rates stratified by 1) the early postsurgical period during which pregnancy is not recommended (0 to less than 18 months, ie, “early”); 2) the next 2 years, during which pregnancy is no longer contraindicated (18 to less than 42 months, ie, “delayed”); and 3) the remaining follow-up period (42–90 months) were also determined.
The Pregnancy Questionnaire allowed the outcome to be reported as: live birth, stillbirth (fetus lost after 20 weeks or 5 months of gestation), ectopic or tubal pregnancy, miscarriage (fetus lost before 20 weeks or 5 months of gestation), or abortion.
At the preoperative assessment only, women rated the importance of being able to become pregnant in the future (on a scale from 0 to 10) and anticipated the timeframe in which they would first try to become pregnant after surgery. Based on their responses, postsurgical pregnancy was categorized as 1) important (rating of 8–10)6; 2) importance unclear (rating of 3–7); or 3) unimportant (rating of 0–2) or not planned (ie, “never”). History of polycystic ovary syndrome was determined by self-reported past diagnosis by a health care professional. Menstrual regularity was defined as a history of 10–12 menstrual periods lasting between 1 and 7 days on average, a usual cycle length of 21–35 days, and no spotting or bleeding at times other than the menstrual period within the last 12 months.
Age, sex, race, ethnicity, education, medical insurance, and marital status were assessed using self-administered questionnaires. Race was considered missing for participants who did not report their race as at least one of the following: white or Caucasian, black or African American, Asian, Native American or Alaska Native, or Native Hawaiian or other Pacific Islander. When more than one type of insurance was reported, insurance type was coded according to the following hierarchy: government, private, other, or unknown. Medical history was determined using a combination of laboratory values, physical examination measures, participant-reported medication use, and comorbid diagnoses from health care providers and medical records review.9,12 Medical contraceptive risk was defined as any level 3 or 4 risk for contraceptive use as reported by the U.S. Medical Eligibility Criteria for Contraceptive Use (eg, ischemic heart disease); type of bariatric surgery was not included in this variable.13
Analyses were conducted using SAS 9.4. All reported P values were two-sided; P values ≤.05 were considered statistically significant. Longitudinal analyses were performed with mixed models with a person-level random intercept with control for site, preoperative education, and preoperative body mass index (BMI, calculated as weight (kg)/[height (m)]2), which were associated with missing follow-up contraception data, as fixed effects. Sensitivity analysis, performed to examine the robustness of results with respect to the missing at random assumption, is reported in Appendix 5, available online at http://links.lww.com/AOG/B20.
Poisson mixed models with robust error variance were used to estimate and test for change in prevalence of contraception outcomes. If the P value for variables regarding contraceptive need and use during intercourse (assessed postoperatively only) demonstrated a significant overall difference between follow-up time points, a pairwise comparison between years 1 and 2 was made to assess changes after the first postsurgical year, when a delay in conception is recommended.4 A linear trend test between years 2 and 7 was also performed to examine the presence of change after the first postsurgical year. For variables regarding contraceptive use for any reason (assessed preoperatively and postoperatively), a significant P value for an overall difference between time points was followed by a comparison between preoperative and year 1 (to assess preoperative-to-postoperative change) and a linear trend test between years 1 and 7 (to assess change across follow-up). This analysis was repeated for the two most common surgical procedures in the cohort, Roux-en-Y gastric bypass and laparoscopic adjustable gastric band. Modeled frequencies, 95% CIs, and P values adjusted to control for overall type I error are reported.14
Poisson mixed models with robust error variance were used to examine associations between participant characteristics and risk of postsurgical unprotected intercourse while not trying to conceive. Postsurgical assessments were excluded from this analysis if participants reported no intercourse, trying to conceive, or pregnancy in the past 12 months. The following independent variables were included: preoperative age,15 race,16 ethnicity,16 education,15 BMI, contraceptive use, history of polycystic ovary syndrome,17 surgical procedure, postsurgical marital status,16 medical insurance,16 menstrual regularity,18 and nonsurgical contraception risk.13 An interaction with time and each covariate was considered and retained if significant. Adjusted relative risks, 95% CIs, and P values are reported.
Conception rates with 95% CIs constructed using the Poisson distribution are reported. The mid-P exact test was used to determine whether conception rates differed by postoperative timeframe (early, delayed, and the remaining follow-up period).
Poisson mixed models with robust error variance were used to examine associations between participant preoperative characteristics with 1) early conception; and 2) delayed conception. The following independent variables were included in each model: age,19 race,20 ethnicity,21 education,22 marital status, BMI,23 current or recent smoker,24 menstrual regularity, contraceptive use, history of polycystic ovary syndrome,25 importance of postsurgical pregnancy, and surgical procedure.
Descriptive statistics were used to report pregnancy outcomes.
This report includes 710 (95.9%) women who provided conception data, 670 (90.5%) of whom also provided contraception data at one or more follow-up assessments (Fig. 1). Excluding women who were ineligible for analysis at the time of follow-up or were due for an assessment after study data collection ended, contraception data attainment was 70% (516/740), 60% (417/691), 60% (382/638), 57% (340/599), 59% (326/556), and 60% (197/326) in years 1, 2, 3, 4, 5, and 7, respectively, among those eligible for follow-up (n=740).
Preoperative characteristics are shown in Table 1. The median (interquartile range) age was 34.0 (30.0–39.0) years and BMI was 46.3 (42.4–51.4). The majority of participants underwent Roux-en-Y gastric bypass (72.7%) followed by laparoscopic adjustable gastric banding (23.3%). Slightly under one third of women (30.0%) indicated that postsurgical pregnancy was important; the importance was unclear to an additional 15.2%.
Modeled prevalence of contraceptive need and use during intercourse by postsurgical time point is shown in Figure 2; supporting data are reported in supplemental materials (Appendix 6, available online at http://links.lww.com/AOG/B20). In the first year, 12.7% (95% CI 9.4–16.0) of women had no intercourse, 40.5% (95% CI 35.6–45.4) protected intercourse only, and 41.5% (95% CI 36.4–46.6) unprotected intercourse while not trying to conceive. The prevalence of these three groups did not significantly differ throughout the remainder of follow-up (P for all >.05). At each annual assessment, more than 75% of women who reported unprotected intercourse used contraception never or rarely.
In the first postsurgical year, 4.3% (95% CI 2.4–6.3) of women tried to conceive. More women tried to conceive by year 2 (13.1% [95% CI 9.3–17.0], P<.001). The prevalence decreased over time thereafter (years 2–7 linear trend P=.02) although still remained higher than year 1 (10.6% [95% CI 6.0–15.3] at year 7, P<.01).
Observed and modeled prevalence and CIs by time point as well as results of linear trend analysis are reported (Appendices 7 and 8, available online at http://links.lww.com/AOG/B20). Among all women, the prevalence of using any method of contraception increased from 52.3% (95% CI 47.9–56.6) at the preoperative assessment to 60.3% (95% CI 55.5–65.1) in the first postoperative year, (P <.001). In particular, the prevalence of intrauterine device use (6.4%, [95% CI 4.5–8.4] to 8.8% [95% CI 6.2–11.4]) and sterilization (13.6% [95% CI 10.5–16.6] to 17.6% [95% CI 14.0–21.1]) significantly increased from the preoperative assessment to year 1. Both continued to increase from years 1 to 7 (P for linear trend for both <.001). No change was observed in use of nonoral hormonal contraceptives (overall P=.57). Among women who reported contraceptive use in the year before and after surgery (47% [223/470]), 93.3% (208/223) used at least one of the same methods at both time points.
During the first year after Roux-en-Y gastric bypass, 21.1% (95% CI 16.6–25.7) of women used oral contraceptives, a category 3 risk under U.S. Medical Eligibility Criteria for Contraceptive Use for women who undergo this procedure.13 For half of these women (10.6% [95% CI 6.9–14.2] of all women who underwent Roux-en-Y gastric bypass), it was their only form of contraception (Appendix 8, http://links.lww.com/AOG/B20).
Unadjusted and adjusted relative risks for unprotected intercourse while not trying to conceive by preoperative and postoperative factors are shown in Appendix 9 (available online at http://links.lww.com/AOG/B20). Nonwhite race (adjusted relative risk [RR] 1.33 [95% CI 1.08–1.62]), no contraception in the year before surgery (adjusted RR 2.12 [95% CI 1.74–2.58]) and having a postoperative nonsurgical contraceptive risk (adjusted RR 1.20 [95% CI 1.02–1.40]) were associated with a higher risk of unprotected intercourse postoperatively. Interactions with time were not significant (data not shown) indicating associations did not differ by time point. Risk was not significantly different by surgical procedure.
Among 710 women, 237 pregnancies were reported by 154 women over a median (interquartile range) of 6.5 (5.9–7.0) years of follow-up, translating to a conception rate of 53.8 (95% CI 40.0–71.1) per 1,000 woman-years. Conception rates were not significantly different in the early postsurgical period during which pregnancy is not recommended (0 to less than 18 months), the next 2 years, during which pregnancy is no longer contraindicated (18 to less than 42 months), and the remaining follow-up period (42–90 months), 42.3 (95% CI 30.2–57.6), 60.9 (95% CI 46.9–79.2), and 46.5 (95% CI 34.5–61.4) per 1,000 woman-years, respectively (Appendix 10, available online at http://links.lww.com/AOG/B20).
Older age was independently associated with a lower risk of early conception (adjusted RR 0.41 [95% CI 0.19–0.89] per 10 years). Reported status of married or living as married was independently associated with a higher risk (adjusted RR 4.76 [95% CI 20.2–11.21]). Preoperative importance of a postoperative pregnancy was also independently associated with a higher risk of early conception (adjusted RR 8.50 [95% CI 2.92–24.75] for important and adjusted RR 5.78 [95% CI 1.74–19.21] for importance unclear compared with unimportant or not planned). Age and preoperative importance of a postoperative pregnancy were also significantly associated with risk of delayed conception. Other preoperative variables and type of surgical procedure were not independently associated with risk of conception during either timeframe. Unadjusted and adjusted associations between all participant characteristics and risk of early conception and delayed conception, respectively, are reported (Appendix 11, available online at http://links.lww.com/AOG/B20).
As an exploratory analysis, we examined whether percent weight loss was associated with postoperative conception. Percent weight loss from preoperative assessment to 6 months was not significantly associated with early conception (adjusted RR 1.40 [95% CI 0.96–2.06] for 5% weight loss; P=.08). Likewise, percent weight loss from preoperative measurement to 12 months was not significantly related to delayed conception (adjusted RR 1.05 [95% CI 0.87–1.29] for 5% weight loss; P=.57).
Outcomes for 183 of 237 pregnancies (77.2%) were reported by 154 women over a median (interquartile) of 6.5 (5.9–7.0) years. Among pregnancies with reported outcomes, 68.9% (126/183) were reported as live births and 21.9% (40/183) were reported as miscarriages. The frequency of all reported pregnancies outcomes and the range of potential values accounting for missing data (missing n=54) are provided in the supplemental material (Appendix 12, available online at http://links.lww.com/AOG/B20).
This large-scale, multicenter report provides postsurgical contraceptive practices and conception rates through a prospective cohort study of reproductive-aged female patients undergoing bariatric surgery. During the first postsurgical year, a delay in conception is recommended as a result of concerns for fetal health4; however, 4% tried to conceive and an additional 42% of women reported unprotected intercourse. Unprotected intercourse while not trying to conceive remained common throughout follow-up. Among women not trying to conceive, nonwhite race, no contraception use in the year before surgery, and medical contraceptive risk were associated with greater risk of postsurgical unprotected intercourse while not trying to conceive. Younger age, being married or living like married, and rating postsurgical pregnancy as important before surgery were associated with early (less than 18 months) postsurgical conception.
Despite concerns regarding absorption of oral contraceptives in patients undergoing malabsorptive procedures,26 risk of early or delayed conception was not associated with type of surgical procedure. However, among women who underwent Roux-en-Y gastric bypass, just more than one tenth reported using only oral contraception in the first postsurgical year despite the category 3 risk under the U.S. Medical Eligibility Criteria for Contraceptive Use for women who have undergone malabsorptive procedures.13 A study by Chor et al27 suggests that approximately two thirds of bariatric surgeons refer patients to an obstetrician–gynecologist or primary care physician to obtain contraception; only 5% were comfortable or very comfortable prescribing contraception. However, this same study noted that approximately 40% of bariatric surgeons do not require early postoperative contraception.27
A major strength of this study lies in the standardized and detailed assessment of a large, multicenter cohort of geographically diverse participants over 7 years with relatively high retention.28 Although missing data are a concern, our analyses of contraceptive use controlled for preoperative factors related to missing follow-up data. With imputed data not assumed missing at random, sensitivity analysis yielded similar estimates of unprotected intercourse over time. Although we had limited statistical power to provide precise estimates of conception rates (ie, CIs were wide) and thus detect differences in these rates by postoperative timeframe, our initial sample size and retention rate ensured sufficient statistical power to evaluate contraceptive need and use over time and to identify several factors associated with unprotected intercourse while not trying to conceive and early conception.
A major limitation of this study is incomplete reporting regarding pregnancy outcomes among women who conceived. Other notable limitations include lack of assessment of preoperative contraceptive counseling practices as well as postoperative contraceptive failure among those who conceived. Although a large cohort, given the number of variables under investigation, this analysis had limited power to ensure that statistical adjustment adequately corrected for baseline differences among participants, which might have influenced associations. Finally, because the study did not have a parallel control group, findings cannot be attributed to the surgery itself.
In conclusion, data from reproductive-aged women in the Longitudinal Assessment of Bariatric Surgery-2 cohort allow us to quantify contraceptive utilization during the at-risk postsurgical interval (currently defined as up to 18 months)4 as well as choice of contraceptive method by surgical procedure, for example, combined oral contraceptive use in patients undergoing Roux-en-Y gastric bypass. The early conception rate reported in our study of 42.3 per 1,000 woman-years in the 18-month postsurgical window is especially concerning given recent findings that bariatric surgery increases the risk of small-for-gestational-age neonates, preterm deliveries, and neonatal intensive care unit admissions in the first 2 years after surgery.5,29 Our findings highlight a public health concern that merits additional scrutiny regarding contraceptive counseling and provision of services for all reproductive-aged women undergoing bariatric surgery. Guidelines recommending inclusion of referral for counseling for postoperative contraceptive use as part of the preoperative evaluation would be prudent. Lack of patient awareness of current guidelines and potential postoperative reversal of preoperative conditions that affect fertility may also affect patient adherence to recommendations for contraception. Our findings also suggest that many of the risk factors for unprotected intercourse while not trying to conceive and early postsurgical conception can be identified before surgery and represent opportunities for targeted preoperative counseling.
1. Edison E, Whyte M, van Vlymen J, Jones S, Gatenby P, de Lusignan S, et al. Bariatric surgery in obese women of reproductive age improves conditions that underlie fertility and pregnancy outcomes: retrospective cohort study of UK National Bariatric Surgery Registry (NBSR). Obes Surg 2016;26:2837–42.
2. Young MT, Phelan MJ, Nguyen NT. A decade analysis of trends and outcomes of male vs female patients who underwent bariatric surgery. J Am Coll Surg 2016;222:226–31.
3. Maggard MA, Yermilov I, Li Z, Maglione M, Newberry S, Suttorp M, et al. Pregnancy and fertility following bariatric surgery: a systematic review. JAMA 2008;300:2286–96.
4. Mechanick JI, Youdim A, Jones DB, Garvey WT, Hurley DL, McMahon MM, et al. Clinical practice guidelines for the perioperative nutritional, metabolic, and nonsurgical support of the bariatric surgery patient—2013 update: cosponsored by American Association of Clinical Endocrinologists, the Obesity Society, and American Society for Metabolic & Bariatric Surgery. Obesity (Silver Spring) 2013;21(suppl 1):S1–27.
5. Parent B, Martopullo I, Weiss NS, Khandelwal S, Fay EE, Rowhani-Rahbar A. Bariatric surgery in women of childbearing age, timing between an operation and birth, and associated perinatal complications. JAMA Surg 2017;152:1–8.
6. Gosman GG, King WC, Schrope B, Steffen KJ, Strain GW, Courcoulas AP, et al. Reproductive health of women electing bariatric surgery. Fertil Steril 2010;94:1426–31.
7. Courcoulas AP, Christian NJ, Belle SH, Berk PD, Flum DR, Garcia L, et al. Weight change and health outcomes at 3 years after bariatric surgery among individuals with severe obesity. JAMA 2013;310:2416–25.
8. Belle SH, Berk PD, Courcoulas AP, Flum DR, Miles CW, Mitchell JE, et al. Safety and efficacy of bariatric surgery: longitudinal assessment of bariatric surgery. Surg Obes Relat Dis 2007;3:116–26.
9. Belle SH, Berk PD, Chapman WH, Christian NJ, Courcoulas AP, Dakin GF, et al. Baseline characteristics of participants in the Longitudinal Assessment of Bariatric Surgery-2 (LABS-2) study. Surg Obes Relat Dis 2013;9:926–35.
10. Gourash WF, Ebel F, Lancaster K, Adeniji A, Koozer Iacono L, Eagleton JK, et al. Longitudinal Assessment of Bariatric Surgery (LABS): retention strategy and results at 24 months. Surg Obes Relat Dis 2013;9:514–9.
11. Kim C, Gebremariam A, Iwashyna TJ, Dalton VK, Lee JM. Longitudinal influences of friends and parents upon unprotected vaginal intercourse in adolescents. Contraception 2011;83:138–44.
12. Courcoulas AP, Christian NJ, O'Rourke RW, Dakin G, Patchen Dellinger E, Flum DR, et al. Preoperative factors and 3-year weight change in the Longitudinal Assessment of Bariatric Surgery (LABS) consortium. Surg Obes Relat Dis 2015;11:1109–18.
13. Curtis KM, Tepper NK, Jatlaoui TC, Berry-Bibee E, Horton LG, Zapata LB, et al. U.S. Medical Eligibility Criteria for Contraceptive Use, 2016. MMWR Recomm Rep 2016;65:1–103.
14. Edwards D, Berry JJ. The efficiency of simulation-based multiple comparisons. Biometrics 1987;43:913–28.
15. Mosher W, Jones J, Abma J. Nonuse of contraception among women at risk of unintended pregnancy in the United States. Contraception 2015;92:170–6.
16. Jackson AV, Karasek D, Dehlendorf C, Foster DG. Racial and ethnic differences in women's preferences for features of contraceptive methods. Contraception 2016;93:406–11.
17. Joham AE, Boyle JA, Ranasinha S, Zoungas S, Teede HJ. Contraception use and pregnancy outcomes in women with polycystic ovary syndrome: data from the Australian Longitudinal Study on Women's Health. Hum Reprod 2014;29:802–8.
18. Yao X, Stewart EA, Laughlin-Tommaso SK, Heien HC, Borah BJ. Medical therapies for heavy menstrual bleeding in women with uterine fibroids: a retrospective analysis of a large commercially insured population in the USA. BJOG 2016;124:322–330.
19. Female age-related fertility decline. Committee Opinion No. 589. American College of Obstetricians and Gynecologists. Obstet Gynecol 2014;123:719–21.
20. Humphries LA, Chang O, Humm K, Sakkas D, Hacker MR. Influence of race and ethnicity on in vitro fertilization outcomes: systematic review. Am J Obstet Gynecol 2016;214:212.e211–17.
21. Maalouf W, Maalouf W, Campbell B, Jayaprakasan K. Effect of ethnicity on live birth rates after in vitro fertilisation/intracytoplasmic sperm injection treatment: analysis of UK national database. BJOG 2017;124:904–10.
22. Swift BE, Liu KE. The effect of age, ethnicity, and level of education on fertility awareness and duration of infertility. J Obstet Gynaecol Can 2014;36:990–6.
23. Practice Committee of the American Society for Reproductive Medicine. Obesity and reproduction: a committee opinion. Fertil Steril 2015;104:1116–26.
24. Practice Committee of the American Society for Reproductive Medicine. Smoking and infertility: a committee opinion. Fertil Steril 2012;98:1400–6.
25. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19–25.
26. Horton LG, Folger SG, Berry-Bibee E, Jatlaoui TC, Tepper NK, Curtis KM. Research gaps from evidence-based contraception guidance: the US Medical Eligibility Criteria for Contraceptive Use, 2016, and the US Selected Practice Recommendations for Contraceptive Use, 2016. Contraception 2016;94:582–9.
27. Chor J, Chico P, Ayloo S, Roston A, Kominiarek MA. Reproductive health counseling and practices: a cross-sectional survey of bariatric surgeons. Surg Obes Relat Dis 2015;11:187–92.
28. Puzziferri N, Roshek TB III, Mayo HG, Gallagher R, Belle SH, Livingston EH. Long-term follow-up after bariatric surgery: a systematic review. JAMA 2014;312:934–42.
29. Johansson K, Cnattingius S, Naslund I, Roos N, Trolle Lagerros Y, Granath F, et al. Outcomes of pregnancy after bariatric surgery. N Engl J Med 2015;372:814–24.