Recent studies estimate that two thirds (66.3 %) of all Americans are overweight or obese.1 Obesity has resulted in serious health consequences for many individuals and for the health care system as a whole. Physicians of all specialties must factor body weight into the clinical decision-making process on a daily basis.
In addition to being a risk factor for multiple chronic diseases, obesity can affect emotional health as well as psychosocial functioning.2–4 A widespread perception that physical attractiveness is correlated with a slender stature has resulted in a billion dollar weight-loss industry.5 Although clinical experience would suggest that weight can affect sexual behavior, the relationship between body mass index and sexual behavior has not been well studied. In adolescent girls, lower body fat indices were associated with increased frequency of sexual intercourse.6 An improvement in sexual functioning has been noted in formerly obese patients who undergo weight loss.7,8 A few studies suggest that obese women have more impairment in sexual function and sexual quality of life than normal weight women,9,10 and this impairment seems to be greater in women compared with men.10 An analysis of the National Health and Nutrition Examination Survey demonstrated that obese and overweight women had fewer sexual partners in the last year than individuals with a normal body mass index, but did not have a significant difference in the number of lifetime partners.5
Although some studies indicate that normal weight women may be more sexually active than obese and overweight women, other studies have indirectly alluded to the fact that the contrary may be true. Indeed, other studies have suggested that obese and overweight women may have a higher rate of unintended pregnancy than normal weight women.11–14 Although multiple factors, such as contraceptive use and efficacy, could be responsible for this observation, sexual behavior and the frequency of sexual intercourse could also be a factor. As the weight demographic in this country continues to increase, understanding how weight affects sexual behavior is important, because sexual behavior affects the risk of pregnancy as well as the risk of sexually transmitted infections.
The primary objective of this study was to characterize the relationship between body mass index and sexual behavior. Specific measures included sexual orientation, ever having had sex with a male partner, age at first intercourse, number of partners, and the frequency of intercourse.
MATERIALS AND METHODS
The 2002 National Survey of Family Growth is a validated population-based database that includes information on a variety of reproductive health outcomes and behaviors.15 Data for the 2002 National Survey of Family Growth was collected between January 2002 and March 2003 using in-person interviews with 7,643 women who were between the ages of 15 years and 44 years. The 2002 National Survey of Family Growth is a weighted database. All respondents were assigned a weight based on national averages of race, ethnicity, and age provided by the U.S. Census Bureau.16,17 The 7,643 women in the 2002 National Survey of Family Growth represented the 61 million women aged 15–44 years in the U.S. household population in 2002.15,16
Although most of the questions were asked by an interviewer, some of the more sensitive questions, particularly those pertaining to sexual behavior were asked at the end of the interview using an Audio Computer Assisted Self-Interviewing system. During the Audio Computer Assisted Self-Interviewing system portion, the computer was turned over to the respondent and they were allowed to answer questions without the interviewer knowing the question or the response. The entire interview process took an average of 85 minutes.14 In general, results from the Audio Computer Assisted Self Interviewing system are considered to be more accurate than results from the interviewer-recorded portion and are presented in this article and in other published articles using the National Survey of Family Growth.14,17 Analysis of this database was approved by the Institutional Review Board of Oregon Health and Science University (Portland, OR).
For the first time in 2002, the National Survey of Family Growth collected information on sexual orientation.16 Individuals were asked whether they identified with being heterosexual, homosexual, bisexual, or “something else.” To determine whether respondents ever had sexual intercourse with a male, they were asked, “at any time in your life, have you ever had sexual intercourse with a man, that is, made love, had sex, or gone all the way.” They were instructed not to count oral sex, anal sex, sex with a female partner, “heavy petting,” or other forms of sexual activity that did not involve vaginal penetration.
In the interviewer-recorded portion of the survey, women who reported that they had not had sexual intercourse with a male, were asked to choose the most important reason why they were virginal. Respondents could select from the following options: “don’t want to get pregnant,” “don’t want to get a sexually transmitted disease,” “against religion or morals,” “haven’t found the right person,” “in a relationship but waiting for the right time,” or “other.”
Respondents were asked to report the age at which they first had intercourse with a male, the number of lifetime male partners, the number of male partners in the last 12 months, the number of current male partners, and the number of times the respondent had male intercourse in the last 4 weeks. Data for these outcomes were collected by an interviewer as well as by the Audio Computer Self-Interviewing system.
Based on self-reported height and weight, respondents were divided into three body mass index categories: normal (body mass index less than 25 m/kg2), overweight (body mass index 25–30 m/kg2), and obese (body mass index more than 30 m/kg2). Age, cohabitation status, education, type of residence (large metropolitan city, urban area, or other), race/ethnicity, total household income, gravidity, parity, and general health status were included in the analysis as potential confounding factors.
Respondents were excluded from the analysis if they did not report their height or weight. Because pregnancy can affect both weight and sexual behavior, respondents were also excluded if they reported being pregnant. Descriptive statistics, including frequency measures, were computed to elucidate differences in demographic, socioeconomic, and health-related variables between body mass index groups. The significance of association between body mass index and the sexual behavior outcomes as well as sociodemographic outcomes was determined using χ2 tests for categorical variables and analysis of variance for continuous factors. A Bonferroni adjustment was used when making pair-wise comparisons between body mass index groups. Unadjusted odds ratios (ORs) and 95% confidence intervals (CIs) were obtained when measuring the association between body mass index and sexual behavior outcomes using logistic regression. A multiple logistic regression model was created with body mass index category and any potential confounders and predictor variables that were associated with the outcome using a P<.20 cutoff for inclusion in the model. Through backward elimination, confounders or predictor variables were removed from the model at a significance level of P<.05.
To assess for confounding, each variable was removed from the model during the above process, and the percent change in the OR between body mass index and a particular sexual behavior outcome was assessed. If the percent change in OR was more than 10%, the variable was considered to be a confounder and was not removed from the model.18 Effect modification was assessed by including the appropriate interaction terms in this analysis. All analyses were performed using SPSS 16.0 for Windows (SPSS Inc., Chicago, IL) with the complex samples module to account for the complex sampling design used by the 2002 National Survey of Family Growth.
A total of 7,643 women participated in the 2002 National Survey of Family Growth. Of these respondents, 138 did not report their height or weight, and 815 identified themselves as being pregnant, leaving a total of 6,690 participants.
The mean (±standard deviation) body mass index for the study population was 25.84 m/kg2 (±0.09) with 53.6% classified as normal, 25.0% as overweight, and 21.4% as obese (Table 1). The mean age at the time of the interview for the study population was 30.12 years (±0.16). Selected demographic and socioeconomic characteristics of the study population are presented in Table 1. No statistically significant differences among weight groups were found for age, cohabitation status, race/ethnicity, education (high school and college), and total household income. There were statistically significant differences among body mass index groups for general health, gravidity, and parity (Table 2). Notably, a higher proportion of normal weight women were nulliparous and a higher proportion of overweight and obese women had three or more children. Given the effect of body mass index on general health and the effect of parity on weight, differences between body mass index groups would be expected. Because these health-related factors could play a role in sexual behavior, they were controlled for in the multiple logistic regression portion of our analysis as potential confounders.
The majority (90.5%) of all respondents reported that they were heterosexual, with 9.5% reporting that they were homosexual, bisexual or “something else.” Within the 9.5% of women who were not heterosexual, 1.3% reported being homosexual, 3.3% reported being bisexual, and 4.8% identified with “something else.” There were no statistically significant differences in sexual orientation, frequency of sexual intercourse, and number of current partners between body mass index groups (Table 3). These findings persisted when multiple logistic regression was performed and age, race/ethnicity, cohabitation status, education, total household income, place of residence, general health, gravidity, and parity were included in the analysis as potential confounders.
In addition, there were no statistically significant differences between body mass index groups for age at first intercourse, the number of lifetime male partners, and the number of male partners in the last 12 months. These findings were consistent when analyzing both self-recorded Audio Computer Self-Interviewing and interviewer-recorded data. As would be expected, women reported a slightly younger age at first intercourse as well as a higher number of male partners in the self-recorded portion of the survey compared with the interviewer-recorded data. The results for the Audio Computer Self-Interviewing portion are displayed in Table 4.
A statistically significant difference (P<.001) between body mass index groups was found, however, in the number of individuals who reported ever having sexual intercourse with a male in the Audio Computer Self-Interviewing portion of the survey. Women in the obese and overweight groups were more likely than women in the normal weight group to report any history of sexual intercourse with a male. The unadjusted odds ratios for obese compared with normal weight women was 1.77 (95% CI 1.56–2.31), and for overweight compared with normal weight women, it was 1.56 (95% CI 1.16–2.09). To control for potential confounders, model building was performed using backward elimination. The difference continued to be significant when age and place of residence were controlled for in our final model using multiple logistic regression. The adjusted odds ratios were essentially unchanged (obese compared with normal weight women OR 1.77, 95% CI 1.35–2.30; overweight compared with normal weight women OR 1.56, 95% CI 1.16–2.095).
There were no significant differences between body mass index groups in the reasons women who had not had intercourse with a male gave for remaining virginal (P=.86). These results are presented in Table 5. It is notable that this question was asked in the interviewer-recorded portion of the survey rather than the Audio Computer Self-Interviewing portion of the survey, where a higher proportion of women reported a history of intercourse with a male. Thirteen percent (13.4%) of women who reported that they had never had sexual intercourse with a male identified as being something other than heterosexual. This is higher than the proportion of women who reported being something other than heterosexual in the study population as a whole (9.5%). However, there were no statistically significant differences in sexual orientation between body mass index groups in the proportion of women who had never had intercourse with a male (normal 14.4%, overweight 12.7%, obese 14.4%, P=.86).
There were no statistically significant interactions. Various confirmatory analyses, including dividing the normal body mass index group into normal weight (body mass index 18–25 kg/m2) and underweight (body mass index less than 18 kg/m2) were performed, and our results did not change.
Previous studies investigating the effect of weight on the quality of sexual life indicate that obese individuals report a high frequency of sexual difficulties, including lack of sexual desire and difficulty with sexual performance.10 Findings from this study, based on a large, representative database, indicate that obese and overweight women differ little from normal weight women in some of the objective measures of sexual behavior, such as the number of partners and the frequency of intercourse.
The finding that obese and overweight women were more likely than normal weight women to report a history of ever having had male sexual intercourse with a male is somewhat unexpected. Our results indicate that obese and overweight women do not initiate intercourse earlier, and it does not seem that there are significant differences in sexual orientation between weight groups. There were also no differences between weight groups in the reasons women gave for remaining virginal, with similar proportions of women in the different weight groups reporting that they “hadn’t found the right person yet,” suggesting that the primary reason for the observed difference is not finding a partner. Although differences could be due to reporting biases, we would expect this bias to manifest in the other sexual behavior outcomes.
It is important to investigate the role of body mass index on sexual behavior, because clinicians who care for women may have preexisting perceptions about an individual’s risk of unintended pregnancy or sexually transmitted infection. Contraceptive counseling, for example, may be approached differently depending on the clinician’s preconceived notion of sexual behavior. This study indicates that women of different body mass indices should be approached and counseled similarly.
These results can be contrasted with other population-based studies that have demonstrated that normal weight men report 10 more lifetime partners than obese men.5 A study similar to ours using the National Health and Nutrition Examination Survey also did not show an association between body mass index and the number of reported lifetime male partners in women.5 It may be that body mass index has a differential effect on sexual behavior in males and females. For example, obese men may have difficulty attracting female partners, whereas obese women may not have this problem. Alternatively, men who want to be highly sexually active and have more partners may be motivated to maintain a lean physique. Women, on the other hand, who desire the same thing may not need to do so.
Underreporting or overreporting in certain groups could also be responsible for this difference. Normal weight men may feel social pressure to overreport the number of sexual partners they have had. This propensity to overreport may not affect women in the normal weight category the same way. On the other hand, normal weight women may really have more sexual partners than obese and overweight women. However, in women, the bias may be to underreport the number of sexual partners.
Potential limitations of the study should be noted. Because of the cross-sectional design of this study, it cannot be concluded that associations are due to causal relationships. Recall bias would lead to an underestimation of the exposure–outcome association. Exposure misclassification is also possible because both weight and height were self-reported rather than objectively measured. Although studies have documented that self-reported weight and height provide a reasonable representation of a woman’s body mass index,19–21 the validity of this information is a limitation of the study. Because misclassification of body mass index due to self-reporting seems to affect women of all weights equally,21,22 this would also underestimate the true exposure–outcome association if one existed. Also, in this study, current body mass index was a proxy for body mass index over time. Thus, it is likely that outcomes such as age at first intercourse are subject to some degree of exposure misclassification.
Information bias is also a potential problem. Interviewers or respondents may misunderstand the questions, and there is always a bias due to a respondent’s desire to give a socially acceptable answer. This is a particular concern for information on sexual behavior.23 It is notable that the 2002 National Survey of Family Growth did employ an Audio Computer Self-Interviewing system to minimize this source of error. However, the desire to give a socially acceptable answer could have affected women of different body mass indices differently depending on their self-confidence and comfort with sexuality. There are also limitations related to generalizability when using this database, because the study only included women aged 15 years to 44 years who are part of the household population.
This is one of the few population-representative studies to investigate the association between body mass index and sexual behavior in women. Our findings are substantial and robust because of the large sample size and high quality of the National Survey of Family Growth, which also allowed the incorporation of several potential confounders into our analysis.
Because sexual behavior can affect the risk of unintended pregnancy and sexually transmitted infection, understanding the relationship between body mass index and sexual behavior is important, especially given the increasing weight demographic in this country. Although obese women may experience stigmatization and prejudice because of their weight, it is important to remember that women of all statures can have difficulties with body image, and it is an individual’s self perception and confidence that may play a larger role in sexual behavior and social functioning.
1. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006;295:1549–55.
2. Fontaine KR, Barofsky I. Obesity and health-related quality of life. Obes Rev 2001;2:173–82.
3. Kolotkin RL, Meter K, Williams GR. Quality of life and obesity. Obes Rev 2001;2:219–29.
4. Kushner RF, Foster GD. Obesity and quality of life. Nutrition 2000;16:947–52.
5. Nagelkerke NJ, Bernsen RM, Sgaier SK, Jha P. Body mass index, sexual behaviour, and sexually transmitted infections: an analysis using the NHANES 1999–2000 data. BMC Public Health 2006;6:199.
6. Halpern CT, King RB, Oslak SG, Udry JR. Body mass index, dieting, romance, and sexual activity in adolescent girls: relationships over time. J Res Adolesc 2005;15:535–59.
7. Werlinger K, King TK, Clark MM, Pera V, Wincze JP. Perceived changes in sexual functioning and body image following weight loss in an obese female population: a pilot study. J Sex Marital Ther 1997;23:74–8.
8. Hawke A, O’Brien P, Watts JM, Hall J, Dunstan RE, Walsh JF, et al. Psychosocial and physical activity changes after gastric restrictive procedures for morbid obesity. Aust N Z J Surg 1990;60:755–8.
9. Esposito K, Ciotola M, Giugliano F, Bisogni C, Schisano B, Autorino R, et al. Association of body weight with sexual function in women. Int J Impot Res 2007;19:353–7.
10. Kolotkin RL, Binks M, Crosby RD, Ostbye T, Gress RE, Adams TD. Obesity and sexual quality of life. Obesity (Silver Spring) 2006;14:472–9.
11. Holt VL, Cushing-Haugen KL, Daling JR. Body weight and risk of oral contraceptive failure. Obstet Gynecol 2002;99:820–7.
12. Brunner Huber LR, Hogue C. The association between body weight, unintended pregnancy resulting in a livebirth, and contraception at the time of conception. Matern Child Health J 2005;9:413–20.
13. 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.
14. 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.
15. 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.
16. Public use data file documentation national survey of family growth cycle 6: 2002, 2007. Hyattsville (MD): U.S. Department of Health and Human Services; 2004.
17. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults—the evidence report. National Institutes of Health [published erratum appears in Obes Res 1998;6:464]. Obes Res 1998;6:51S–209S.
18. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York (NY): John Wiley and Sons, Inc.; 2000.
19. McAdams MA, Van Dam RM, Hu FB. Comparison of self-reported and measured BMI as correlates of disease markers in US adults. Obesity (Silver Spring) 2007;15:188–96.
20. Le Marchand L, Yoshizawa CN, Nomura AM. Validation of body size information on driver’s licenses. Am J Epidemiol 1988;128:874–7.
21. Brunner Huber LR. Validity of self-reported height and weight in women of reproductive age. Matern Child Health J 2007;11:137–44.
22. Goodman E, Hinden BR, Khandelwal S. Accuracy of teen and parental reports of obesity and body mass index. Pediatrics 2000;106:52–8.
23. Morris M. Telling tails explain the discrepancy in sexual partner reports. Nature 1993;365:437–40.