Sexuality and obesity, a gender perspective: results from French national random probability survey of sexual behavioursBMJ 2010; 340 doi: http://dx.doi.org/10.1136/bmj.c2573 (Published 15 June 2010) Cite this as: BMJ 2010;340:c2573
- Nathalie Bajos, research director123, honorary professor4,
- Kaye Wellings, professor4,
- Caroline Laborde, research assistant123,
- Caroline Moreau, research fellow123
- for the CSF Group
- 1INSERM (Institut National de la Santé et de la Recherche Medicale), CESP Centre for Research in Epidemiology and Population Health, U1018, Gender, Sexual and Reproductive Health Team, F-94276, Kremlin Bicetre, France
- 2Université Paris Sud 11, UMRS 1018, F-94807, Villejuif, France
- 3Institut National des Etudes Demographiques, F-75020, Paris
- 4London School of Hygiene and Tropical Medicine, Department of Public Health and Policy, London WC1E 7HT
- Correspondence to: N Bajos
- Accepted 18 April 2010
Objectives To analyse the association between body mass index (BMI) and sexual activity, sexual satisfaction, unintended pregnancies, and abortions in obese people and to discuss the implications for public health practices, taking into account the respondents’ and their partners’ BMI.
Design Random probability survey of sexual behaviours.
Setting National population based survey of 12 364 men and women aged 18-69 living in France in 2006.
Participants Random selection of 5535 women and 4635 men, of whom 3651 women and 2725 men were normal weight (BMI 18.5-<25), 1010 women and 1488 men were overweight (BMI 25-<30), and 411 women and 350 men were obese (BMI >30).
Results Obese women were less likely than normal weight women to report having a sexual partner in the past 12 months (odds ratio 0.71, 95% confidence interval 0.51 to 0.97). Obese men were less likely than normal weight men to report more than one sexual partner in the same period (0.31, 0.17 to 0.57, P<0.001) and more likely to report erectile dysfunction (2.58, 1.09 to 6.11, P<0.05). Sexual dysfunction was not associated with BMI among women. Obese women aged under 30 were less likely to seek healthcare services for contraception (0.37, 0.18 to 0.76) or to use oral contraceptives (0.34, 0.15 to 0.78). They were also more likely to report an unintended pregnancy (4.26, 2.21 to 8.23).
Conclusion There is a link between BMI and sexual behaviour and adverse sexual health outcomes, with obese women less likely to access contraceptive healthcare services and having more unplanned pregnancies. Prevention of unintended pregnancies among these women is a major reproductive health challenge. Healthcare professionals need to be aware of sensitivities related to weight and gender in the provision of sexual health services.
Obesity is emerging as one of the fastest growing pandemics in modern times. The current economic downturn might lead to reduced nutritional quality and physical activity, further increasing the prevalence of obesity and related health costs in the coming years.1 2 The adverse effects of obesity on health are, for the most part, well documented.3 The consequences for sexual health, however, are less well understood.4 5 Obesity is associated with erectile dysfunction in men,6 7 8 9 but the evidence in relation to other sexual health outcomes has been equivocal. Some surveys have reported a higher prevalence of sexually transmitted infections among obese people.10 11 12
To date, there have been no comprehensive population based studies of the impact of overweight and obesity on sexual activity and sexual health outcomes such as sexual satisfaction, unintended pregnancy, and abortion. Most research has focused on sexual dysfunction and has been conducted among clinical samples of aging or morbidly obese men.2 9 13 14 There have been few studies on women.15 16 17 18
The French national survey of sexual behaviours contains questions on sexual partnerships and sexual practices, sexual satisfaction and sexual norms, unwanted pregnancy, abortion, and sexually transmitted infections and is unique in collecting data on weight and height for both respondents and their sexual partners. We investigated sexual behaviours and sexual health by body mass index (BMI) and the implications for public health practices. Our primary goal was not to provide estimates of sexual behaviour of obese people but to compare sexual health behaviours and risks between obese and normal weight individuals on the one hand, and between overweight and normal weight individuals on the other hand. Our hypothesis was that overweight individuals, especially obese people, could be at greater risk of negative sexual outcomes compared with normal weight individuals because of social stigma and lack of adequate medical follow-up. We further hypothesised that the effect of BMI on sexual health outcomes would be different between men and women because the social stigma of overweight and obesity is stronger in women.19 20 21 22
The CSF (Contexte de la Sexualité en France) survey was a national random population based survey of sexual behaviours. The sample comprised 12 364 French speaking men (n=5540) and women (n=6824) aged 18-69 living in France who were interviewed by telephone between September 2005 and March 2006. Two pilot surveys were undertaken in November 2004 and in June 2005 to test the questionnaire and the sampling design.
Participants were informed that the survey was anonymous. The interviewer did not know the telephone number he or she was calling and the telephone numbers were deleted from the interview data file immediately after the first question was answered. The research team trained interviewers for two days. Researchers were present every day to listen to the interviews.
Participants were selected with a two stage probability sampling design, post-stratified for age. An initial sample of households was randomly selected from the telephone directory. One eligible respondent per household was then selected with the next birthday method. In addition, we randomly selected a subsample of people who exclusively used mobile phones to achieve adequate representation of the population. Of those who initially selected, 74.6% agreed to complete the questionnaire. At the end of the interview, 4957 individuals aged 18-44 were asked to participate in a complementary module (response rate 76%) involving a chlamydia screening test carried out at the person’s home (a vaginal smear test in women and a urine test in men).
The mean duration of the interviews was 49 minutes. Information collected included the sociodemographic and health related characteristics of respondents and their living partners, including weight, height, living conditions, and chronic health problems. Respondents who had ever had sexual intercourse were asked about their partnership histories, including the total number of sexual partners and the number of partners in recent periods, their sexual practices, non-consensual sexual relations, health seeking behaviours, use of contraception, sexual problems, and sexually transmitted infections (table 1)⇓.
We used the WHO classification of BMI (weight (kg)/(height (m)2) and distinguished four categories: underweight (BMI <18.5), normal weight (18.5-<25), overweight (25-<30), and obese (≥30).
From the initial sample of 12 364 people, a randomly selected 10 170 completed a more detailed questionnaire that collected information on weight and height. This was our group for analyses (4635 men and 5535 women). Pregnant women were excluded from the analysis. Analyses were performed separately for men and women because of sex differences in the associations between obesity and specific sexual behaviours.
We used logistic regression models to test for the effect of obesity and overweight on sexual and contraceptive behaviours in obese and overweight men and women compared with men and women with normal BMI. In multivariate logistic regression models we included those variables significant at P<0.25 in univariate analysis (that is, education, living as a couple, health insurance, income, chronic disease, limitation of activity, number of lifetime sexual partners). We also looked at trends over BMI to test the hypothesis that there is a linear relation between BMI and sexual outcomes. In some cases, however, the relations seemed non-linear so we added a polynomial term (BMI2) to account for the non-linearity of the associations.
Our estimates were weighted to take into account unequal probabilities of selection in the survey. These probabilities depended on the number of eligible people in the household and on the age of the respondent (we oversampled those aged under 25 to study sexual risk practices at sexual debut).23 Weights were equal to the inverse of the probability of being selected in the sample, as used for Horwitz-Thompson estimators.24 Post-stratification weights were then applied to adjust the sample to marginal distributions of sex, age, level of education, and region according to the census data to limit non-participation bias.25 The composition of the final sample corresponded well to that of the French population as described in the most recent national census.23
We present weighted estimates of percentages, means, and odds ratios and unweighted numbers of individuals who answered the questions. All univariate and multivariate analyses were performed with the svy command in Stata 10SE, which takes into account the sampling design for point estimates and variance estimation, with the Taylor linearisation method.
Although we present repeated analyses, we chose not to adjust for multiple comparisons in our study. Indeed, these adjustments increase the frequency of incorrect statements asserting no relation when in fact the association in the data is not a result of chance.26 Our analysis is based on a theoretical framework on gender inequalities and builds on existing empirical studies. Therefore, we believe that the universal null hypothesis is unlikely to apply to our study. The reported P values should nevertheless be interpreted with caution because of multiple comparisons. They should be considered as merely indicative of likely relations between BMI and sexual activity and health.
A total of 9% of 5535 women and 9% of 4635 men were obese (BMI ≥30); 21% of women and 35% of men were overweight (BMI 25-29), and 63% of women and 54% of men were of normal weight (BMI 18.5-24).
Respondents who were obese differed from normal weight individuals (BMI 18.5-24) in terms of age, level of education, and medical condition. We found no differences according to cohabitation status or parity. Individuals categorised as obese were significantly older (mean age in obese v normal weight 48.5 v 41.1, P<0.001, in women, and 50.2 v 37.7, P<0.001, in men). They were also more likely to have no education beyond 16 years (odds ratio 3.31 (95% confidence interval 2.42 to 4.63) for women and 2.81 (1.89 to 4.18) for men).
After we controlled for age, obese respondents were more likely to have chronic disease (such as diabetes, cardiovascular disease, depression, or back pain) (odds ratios 1.85 (1.45 to 2.37) for women and 1.85 (1.40 to 2.44) for men). After adjustment for age and chronic disease, obese women were more likely to report being physically limited in their daily activities (1.97, 1.46 to 2.66). This was not so in the case of obese men.
Half of 411 obese women (51%) and a quarter of 350 obese men considered themselves to be “very overweight” (P<0.001). In addition, a third of 3651 women (34%) with normal BMI thought they were somewhat overweight compared with 11% of 2725 men (P<0.001).
Among those with a regular heterosexual partner, the BMI of the two partners was strongly correlated (table 2).⇓ The association was stronger for obese women than for obese men; 67% of obese women reported a weight and height for their partner indicating that he was overweight or obese compared with 39% of obese men (P<0.001).
BMI and sexual activity
After adjustment for potential confounding factors (age, education level, chronic disease, and limitation of daily activity), obese women were 30% less likely than normal weight women to report a sexual partner in the past 12 months (0.71, 0.51 to 0.97). Obese women were as likely as normal weight women to be living with a sexual partner in the past 12 months but were significantly less likely to have an occasional sexual partner (data not shown). The same was not true for women who were overweight. Results showed no association between obesity or being overweight and having a sexual partner in the past 12 months among men (table 3)⇓. Nonetheless, obese and overweight men were about 70% and 30%, respectively, less likely to report more than one sexual partner in the past 12 months (table 3).⇓ In fact, the likelihood of reporting more than one sexual partner in the past 12 months decreased with increasing BMI (P=0.003 for women and P=0.001 for men).
Obese women aged 18-29 were three times as likely to report having met a sexual partner through the internet than those with a normal BMI. The same was not true for overweight women. The likelihood of having met a partner through the internet, however, significantly increased with BMI (P=0.001). Obese women were also twice as likely to report having seen a pornographic film in the past 12 months. These associations were not observed among overweight women and among men.
Among those with a sexual partner in the past year, 14% of women (4871) and 15% of men (4265) reported no sexual intercourse in the past month, with no difference according to BMI status. Likewise, among individuals who had a partner at the time of the survey, results showed no variation in the frequency of sexual intercourse (table 3), in the length of last sexual intercourse, or in their sexual practices in the past 12 months (data not shown) according to the respondent’s BMI.
Obese women were only half as likely to consider that sexuality was important for their “personal life balance” than normal weight women (P<0.001). The difference was not significant for overweight women, but the results show a significant trend towards lesser importance of sex for personal life balance with increasing BMI (P=0.005). The difference was less marked and not significant (P=0.07) for obese men, who were 60% less likely to consider sexuality an important component of their “personal life balance.” Results also indicate that obese women were less interested in sex than obese men (2.57, 1.59 to 4.21).
Forty four percent of women and 35% of men reported they were very satisfied with their sexual life, with no difference across BMI status.
BMI and sexual health
There was no difference in sexual dysfunction (lack of sexual desire, arousal, painful intercourse) between obese or overweight women compared with women with a normal BMI (table 4)⇓. Nevertheless, results show a significant trend towards decreasing sexual desire with increasing BMI (P=0.01). Data also show a significant increase in erectile dysfunction with increasing BMI for men (P=0.002) (table 5)⇓. Compared with normal weight men, the odds of often experiencing erectile dysfunction in the past 12 months were 2.58 among obese men (P=0.03) and 2.69 among overweight men (P=0.02).
The odds of reporting an unintended pregnancy or an abortion were four times higher among obese women aged under 30 than in normal weight women in the same age group (table 6)⇓. We found no significant association for overweight women, but the likelihood of reporting an unintended pregnancy (P=0.001) or an abortion (P=0.035) significantly increased with BMI among women aged under 30. Obese women were also less likely to have seen a doctor for contraception in the past 12 months (0.37, 0.18 to 0.76, for those aged 18-29 and 0.37, 0.24 to 0.57, for those aged 30-49). There was no significant difference in the use of any method of contraception by BMI status, but obese women were 70% less likely to report using the pill and eight times more likely to use less effective methods, such as withdrawal, compared with women with normal BMI (table 6).⇓ Overweight women were less likely to have used condoms in the past 12 months than normal weight women (P=0.04). Results show a significant decrease in the use of oral contraception (P=0.03) and the use of condoms (P=0.002) and, conversely, an increase in the use of withdrawal (P=0.017) with increasing BMI among women aged under 30.
Obese women were as likely to report a sexually transmitted infection in the past five years as women with a normal BMI (table 6), but obese men aged under 30 had a higher prevalence of self reported sexually transmitted infections in the past five years (P=0.005) (table 7)⇓. Obese men aged 30-49 with more than one sexual partner were less likely to have used condoms in the past 12 months than normal weight men (P<0.05). These associations were not true for overweight men.
The data from this French national survey of sexual behaviours show interesting links between BMI and both sexual behaviour and adverse sexual health outcomes and the importance of some of these issues for obese people, especially women. For men and women alike, obesity seems to affect the number of sexual partners in the past 12 months and, for women, on whether they have a sexual partner at all. Once with a sexual partner, obese and overweight women and men seem no different from others in terms of frequency of sexual intercourse and sexual practices, a finding consistent with other research.27 In men, higher BMI was associated with an increased likelihood of sexual dysfunction and obesity was associated with unsafe sexual behaviours and sexually transmitted infections. Obese women were less likely to use effective methods of contraception, especially the pill, and experienced more unintended pregnancies and abortions. They were also less likely to attend contraceptive healthcare services.
The data show a marked gender effect. The partners of obese men and women were more likely themselves to be obese, but the association was stronger for women than for men. There was also evidence that self image varies with gender; obese women were twice as likely to see themselves as such compared with obese men. They also downgraded the importance of sexuality to their wellbeing, which might reflect a rationalising adjustment to the lack of an available sexual partner.
Strengths and limitations
We had to rely on self reported, rather than objectively measured, weight. Compared with results from a national nutrition survey in France, where weight and height were objectively measured,28 we found a lower prevalence of obesity for both men and women. Other studies have noted underestimation of weight in self reported data.29 30 Misclassification of BMI caused by self reporting nevertheless seems to affect all women equally, regardless of BMI,31 which would lead to underestimates of the association between exposure and outcome. Misclassification, however, is unlikely to be associated with the reporting of sexual behaviours as height and weight were collected at the beginning of the questionnaire before the topics of sexual behaviours and sexual health were introduced.
Data on health issues were limited in the questionnaire because of time constraints. Further research should include information on comorbidities that could act as mediators of the relation between obesity and sexual activity (such as urinary incontinence, metabolic syndrome, musculoskeletal problems, polycystic ovary syndrome).
We focused on heterosexual relationships to study the impact of high BMI on sexual health outcomes. We hypothesised that social stigma about weight would be higher for women. Furthermore, the number of individuals who reported exclusively same sex partners was too small for us to carry out a specific analysis. Nevertheless, an in depth analysis needs to be done as the impact of obesity on sexually transmitted infections has been documented among gay men,12 as well as the impact of lipodystrophy on sexual risk taking among respondents with HIV.32 Social stigma about weight, which seems to be related to gender, might be compounded by other specific issues for people whose sexuality could also be socially stigmatised.
Comparison with other studies
The gender differences we found in terms of sexual partnerships have not been shown in analyses of data from other population based surveys. In the US survey on health and nutrition, for example, obese men and women alike reported fewer sexual partners than non-obese individuals.10 Our finding of an increased risk of erectile dysfunction among obese men has been noted in other studies,7 8 9 unlike our finding of no increase in sexual dysfunction among obese women. A study among women participating in a programme to lose weight showed a lack of sexual desire and difficulties with sexual performances among obese women.18 In a case-control study, Esposito et al also found that obese women were more likely to report arousal dysfunction and being less sexually satisfied.33 Obese women in our study, however, seemed to minimise the importance of sexuality to their wellbeing. The increased risk of unplanned pregnancy among obese women shown in our study was not found in a US study examining this association, though its focus was not specifically on sexual health.34
The relation between obesity and sexual ill health might be forged via physiological, social, and psychological mechanisms.35 It is in the social factors influencing sexual behaviour, however, that we are most likely to find insights into these findings, particularly for women.36 In industrialised countries, there is considerable evidence that women are under stronger social pressure than men regarding weight.19 20 21 22 37 There is also evidence that men are more likely than women to select their partners according to weight.19 20 21 22 38 Several of our findings—that women in our study were more likely to think that they were too fat, that fewer obese men than women reported having an overweight or obese sexual partner, and that obese women were less likely than obese men to have had a sexual partner in the past 12 months—need to be seen in this context. Psychological factors (such as poor sexual self esteem leading to difficulty in allowing or initiating sex) or biological factors (such as musculoskeletal problems) might also be involved. Nevertheless, low self esteem, although a psychological attribute, derives from social stereotypes around obesity and so is essentially social in origin, though socially different to not being able to find a partner. The greater likelihood that obese women seek partners on the internet, which might increase the chances of finding a partner without surveillance as well as affording possibilities to conceal weight, can also reflect a stronger social pressure regarding weight for women. The lower prevalence of medical consultation for contraception among obese women could also be attributed to concerns over body image and weight.
Preventing unintended pregnancies among obese women is a major reproductive health challenge. Obesity is known to increase the risk of adverse outcomes of pregnancy.39 The marked increase in risk of unintended pregnancy, despite lower fecundity in obese women,40 reflects not patterns of sexual behaviour but reliance on less effective methods of contraception. Such methods do not generally require medical intervention and so this finding must be seen in tandem with obese women’s infrequent attendance at healthcare services for contraception and might, in part, reflect negative feelings towards their fat body. It might also reflect reluctance on the part of healthcare professionals to prescribe oestrogen-progestogen contraceptives for obese women because of concerns over the increased risk of cardiovascular events41 (though being overweight remains only category 2 (caution) in the WHO medical eligibility criteria).42 43 Nevertheless, the lower prescription of such contraceptives for obese women does not seem to be compensated for either by greater use of other effective methods such as an intrauterine device or progestogen only methods or by other long acting reversible contraception (which are rarely used in France). Yet these methods would be particularly suitable for obese women given their high efficacy rates and the low risk of associated venous thromboembolism. The lower prevalence of use of hormonal contraception, which persisted after adjustment for chronic disease including cardiovascular risk factors, could also result from women’s fear of weight gain, an important reason for women stopping the pill.44 45
Conclusions and policy implications
The public health impact of these findings is important. The scale of the problem and the magnitude of the effects (particularly the fourfold increase in risk of unintended pregnancy among obese women) warrants focused attention. In terms of targeting advice and care, a considerable proportion of the population is obese, is easily identified as such, and is at increased risk in terms of poorer sexual health status.46 Any temptation to think that obese women are less sexually active and therefore less in need of effective methods of contraceptive is clearly without foundation.
These data should encourage enhanced awareness of weight on the part of healthcare professionals in the provision of sexual health advice and the careful targeted and tailored advice relating to fertility control, prevention of infection, and psychosexual counselling. Given the social and gender related sensitivities, obese women and men will certainly benefit from tactful guidance and advice on preventive sexual health matters.
What is already known on this topic
Studies have shown a relation between obesity and erectile dysfunction in men, but the evidence in relation to other sexual health outcomes such as sexual satisfaction, unintended pregnancy, and abortion has been equivocal
Most research has been conducted among clinical samples of ageing or morbidly obese men; few studies have been conducted among women
What this study adds
There is a link between BMI and sexual behaviour and adverse sexual health outcomes, in particular an increased likelihood of failing to access contraceptive healthcare services and of unplanned pregnancy in obese women
There is a clear gender effect of BMI on sexual behaviour, and health professionals need to be aware that this can have an impact on sex and relationships and sexual health
Cite this as: BMJ 2010;340:c2573
We thank the participants in the study. The CSF group includes N Bajos, M Bozon, N Beltzer, A Andro, M Ferrand, V Goulet, A Laporte, C Le Van, H Leridon, S Levinson, N Razafindratsima, L Toulemon, and J Warszawski.
Contributors: NB, CM, and KW conceived and wrote the paper. CL carried out the statistical analysis and commented on the first draft of the paper and approved the final version, as did all the other members of the research team. NB is guarantor.
Funding: The survey was funded by the French National Agency of Aids Research (ANRS). The Fondation de France, the Institut National de la Prevention et d’Education pour la Sante, and the Direction de la Recherche des Etudes Economiques et Sociale (DREES) also contributed to funding. The researchers operated independently of the funders.
Competing interests: All authors have completed the unified competing interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare (1) no financial support for the submitted work from anyone other than their employer; (2) no financial relationships with commercial entities that might have an interest in the submitted work; (3) no spouses, partners, or children with relationships with commercial entities that might have an interest in the submitted work; and (4) no non-financial interests that may be relevant to the submitted work.
Ethical approval: The survey was approved by the French Commission Nationale Informatiques et Libertes (www.cnil.fr/english/). This commission is in charge of examining both ethical and anonymity issues of research protocols. There was no possible linkage between the identity of the respondents, their answers to the questionnaire, and the results of the chlamydia test. Participants who underwent chlamydia testing signed an informed consent agreeing on the anonymous computerised treatment of the data.
Data sharing: No additional data available.
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode.