SEXUAL TRANSMISSION REQUIRES that the agent be present in one partner, the other partner be susceptible to infection with that agent, and that the sex partners engage in sexual practices able to transmit the pathogen. High‐risk behavioral profiles for sexually transmitted diseases (STD) have generally included factors associated with increased risk of exposure to an infected partner, such as large numbers of sex partners, or chance of high‐risk partners, or factors that modify risk of transmission of the agent following exposure to an infected partner, such as condom use. However, there are less well characterized health behaviors and sexual practices that may influence risk of acquiring or transmitting an STD,1 such as dry sex, sex during menses, douching practices, and anal intercourse. The frequency of these practices, and how they relate to each other and known STD risk factors, is the subject of this study.
The removal of vaginal secretions before engaging in vaginal intercourse(dry sex) is apparently a common practice in many African countries.2–4 Among women using an antenatal clinic in Malawi, 13% reported using intravaginal agents for a tightening effect.4 These women had an increased risk of acquiring STD, including human immunodeficiency virus (HIV), perhaps because of the associated abrasions and ulcers that result from this practice.4 Although dry sex may be practiced in the United States among some women, this has not been confirmed in any population‐based studies.
Sex during menses has been associated with increased risk among women of acquiring gonorrhea, trichomoniasis, and HIV.5–7 In a national probability sample, sex during menses was more common among better‐educated, young white women and their sex partners, and was associated with a twofold increased risk of self‐reported STD history.5
Douching may be used to remove vaginal secretions and thus be a preliminary to engaging in sex during menses or dry sex. About half of African‐American women report douching compared with one fifth of white women.1 Douching has been associated with an increased risk of pelvic inflammatory disease (PID),8,9 tubal pregnancy,10 infertility,11 and cervicitis.12
Anal intercourse can transmit most STD, and receptive anal sex is associated with an increased risk, compared with oral or vaginal or insertive anal sex, of acquiring hepatitis B virus13 or HIV infection.14
In this study, we describe the frequency of dry sex, sex during menses, douching, and anal sex in a random‐digit dialing (RDD) sample of Seattle residents and a listed sample of African Americans (AA) from census tracks with an AA density of 40% or more. The associations between these practices and self‐reported STD history are presented.
Methods
Study Population
Two samples were selected: an RDD sample of all Seattle residents aged 18 to 39 years, and a sample of AA from people with listed telephone numbers ("listed sample"). We chose to select an additional sample of AA because they have higher rates of STD both nationally15 and in Seattle,16 and national data suggest that STD morbidity and sexual practices differ between AA and white Americans (WA).17 The RDD sample was obtained from Survey Sampling, Inc. of Westport, Connecticut and contained both phone numbers listed and not listed in current telephone directories. The listed AA oversample was selected from listings of telephone directories from census block groups in Seattle with an AA population density of 40% or more. In both the sampling frame and in the interview, racial category was defined by the respondent. Only telephone numbers were available to the interviewer for both samples. Up to six attempts were made to contact each number, including up to four attempts during evening hours, and one morning and one afternoon call. A brief message about the study and a toll‐free number were left on answering machines whenever calling resulted in three responses by answering machines. The household member aged 18 to 39 years with the most recent birthday was selected for the interview. The telephone survey was conducted by the Social and Economic Sciences Research Center, at Washington State University in Pullman, Washington.
For the RDD sample, from 3,600 telephone numbers, we were able to contact 808 potentially eligible individuals; 1,077 numbers were either disconnected, business or government numbers, or electronic devices. We reached only tape machines for 245 numbers, and were unable to contact 229. The remaining 1,241 were ineligible because the potential respondents did not meet eligibility criteria for either age or residence, did not speak English, or had a language or other disability. The cooperation rate (the ratio of the number of completed interviews to the total number of completed interviews, partially completed interviews, and refusals) was 67.5%, for a total sample size of 544. For the purposes of this analysis, the sample was limited to the 422 WA respondents and 44 AA respondents.
For the AA oversample, we had 2,800 numbers and were able to contact 510; 312 were either disconnected, business or government numbers, or electronic devices. We reached only tape machines for 160 numbers and were unable to contact 206. The remaining 1,612 were ineligible because the potential respondents did not meet eligibility criteria for either age, race, or residence, did not speak English, or had a language or another disability. The cooperation rate was 28.2% for a total sample size of 135. For the purposes of this analysis, the 44 AA respondents to the RDD survey were combined with the AA oversample, for a total of 179 respondents. The 44 AA respondents selected through RDD were similar to the AA oversample with respect to age, lifetime number of sex partners, age at coitarche, and STD history, although they were slightly younger, reported somewhat fewer sex partners, and were less likely to report history of any STD. Excluding the 44 AA respondents selected through RDD did not change the study findings, but decreased the precision of the results.
Respondent Recruitment
Respondents were screened for eligibility before recruiting for the study. The purpose of the study was described as "to learn how often people engage in behaviors that might put them at risk of acquiring an STD."
Questionnaire
The survey was developed using questions used in previous telephone surveys. Questions were pretested on a sample of the study population before finalization. The telephone survey contained questions regarding sexual history, partnership characteristics, partner characteristics, STD history, and demographics. Dry sex was defined to all respondents as "intercourse after removing all vaginal secretions." Interviews averaged 21 minutes in length.
Data Analysis
We described the data using contingency tables. Differences between groups were tested using the chisquare test. To estimate the association of each behavior of interest and history of each STD measured adjusting for lifetime number of vaginal sex partners, a series of separate logistic regression models were fit. Each model had the STD of interest as the dependent variable; lifetime number of vaginal sex partners and the behavior of interest were independent variables. We calculated odds ratios (OR) and 95% confidence intervals(CI) for the association between each behavior and the STD of interest. CI excluding 1.0 are statistically significant at α = 0.05. In addition, to assess whether dry sex, sex during menses, douching, and anal sex are associated with self‐reported STD history, we fit a stepwise logistic regression model including all four behaviors and lifetime number of vaginal partners, condom use, and marital status. The significance level for entry into the model was set at 0.10. All analyses are presented separately by sample and gender, to contrast the sexual behavior and STD rates in the two samples. Data management and all other analyses were performed using PC‐SAS 6.0.18
Results
Study Population
Respondents to both the WA and AA sample were evenly distributed by age and gender (Table 1). In the WA sample, men were younger than women (P = 0.06), more often single (P = 0.03), and more often employed (P = 0.01). Men in the AA sample tended to be older than women (P = 0.08). Almost all respondents had at least a high school education and most were employed. Most were single, and almost all had ever engaged in sexual activity, defined as vaginal, oral, or anal intercourse.
Sexual History
Although there was no difference by gender in lifetime number of vaginal sex partners in the WA sample, in the AA sample the reported lifetime number of vaginal sex partners was significantly higher for men than for women (P = 0.002; Table 2). Dry sex, douching, and anal sex were reported more frequently in the AA sample than the WA sample. By contrast, sex during menses was reported more frequently in the WA sample. Most STD were reported more frequently by women than by men in both samples (Table 3). Although self‐reported history of most STD were similar in the two samples by gender, women in the AA sample reported chlamydial infection (P << 0.001), gonorrhea (P << 0.001), and PID (P << 0.001) significantly more frequently than women in the WA sample.
Association With Sociodemographic and Sexual History Variables
The frequency of dry sex, sex during menses, douching, and anal sex varied by sociodemographic variables, number of vaginal sex partners, and condom use in the WA sample (Table 4), but there was much less variation in reported practices in the AA sample (Table 5). In both samples, women and men who reported a history of both same‐ and opposite‐sex partners compared with those with opposite‐sex partners only were more likely to report engaging in anal sex.
Association of Douching and Sexual Practices With Each Other
Douching and sexual practices were highly associated with each other for both men and women in the WA sample (Table 6). For example, WA women engaging in dry sex were more likely than those not engaging in dry sex to report sex during menses (91% vs. 55%), douching(50% vs. 22%), and anal sex (27% vs. 2%). However, only sex during menses and dry sex among men were interrelated in the AA sample (Tables 7).
Association With Self‐Reported Sexually Transmitted Disease History
To estimate the association of each practice with self‐reported STD history, we fit a series of logistic regression models with the STD of interest as the dependent variable and lifetime number of vaginal sex partners and the practice of interest as the independent variables for each sample (Tables 8 and 9). Of particular interest, sex during menses was associated with self‐reported history of chlamydial infection among women in both samples. In the WA sample, the practice most strongly associated with self‐reported history of an STD was anal sex among both women and men. Anal sex was associated with self‐reported history of herpes and gonorrhea among women, and genital warts, hepatitis, gonorrhea, and nongonococcal urethritis (NGU) among men. Douching was associated with a twofold increase in self‐reported PID, but the CI are wide. Because of the smaller sample size, the CI are quite wide in the AA sample, and it is harder to interpret the findings. However, anal sex was associated with chlamydial infection and gonorrhea among women, dry sex and sex during menses were both associated with gonorrhea among men, and douching was associated with PID.
Because dry sex, sex during menses, douching practices, and anal sex are correlated with each other and with lifetime number of vaginal sex partners, gender of sex partner, condom use, and marital status, we included all seven variables in a series of stepwise logistic regression models predicting history of each individual STD. Using this procedure, no behavioral practice of interest was able to enter the model with a 0.10 significance level for genital warts, hepatitis, syphilis, PID, or NGU. Sex during menstruation did enter for chlamydial infection among WA women but not among WA men or in the AA sample; sex during menstruation also entered the model for gonorrhea among AA men. Anal sex entered the model for herpes among women in both groups; and anal sex entered the model for gonorrhea among AA women but no other group. A self‐reported history of both same‐ and opposite‐sex partners was associated with genital warts among WA women and WA and AA men, gonorrhea among WA men, and NGU among AA men.
Discussion
We describe the frequency of douching practices and three potentially risky sexual practices‐dry sex, sex during menses, and anal sex‐among WA and AA local urban population samples. Of these practices, sex during menses was most frequently reported. Sex during menses was reported more frequently in our sample than in a national sample (57% of women in the WA sample compared with 26% in the National Survey of Women), although this may reflect differences in how the questions were asked. We asked about how often the respondent engaged in sex during menses with the most recent sex partner, whereas the National Survey reports on usual practice and vaginal intercourse during last menses. We did not find an association of sex during menses with younger age, but the practice increased with lifetime number of sex partners. Consistent with the National Survey, sex during menses was more commonly reported among WA than among AA.
By contrast, douching and dry sex were reported more frequently among AA than WA. Two thirds of AA compared with one fifth of WA women reported douching, and 16% of AA versus 6% of WA women reported engaging in dry sex. Vaginal tightening was reported by 13% of Malawian women attending a antenatal clinic,4 and dry sex is reputed to be common in Zaire2 and Zimbabwe,3 but prevalence estimates for the United States are not available. We defined dry sex as "intercourse after removing all vaginal secretions," but did not ask further about drying agents or the reason for this practice. Although the frequency of douching practices and association with racial group that we observed are consistent with previous reports,1 we are unaware of previous reports describing the prevalence of dry sex in the United States or the prevalence by racial group.
Similar to a national probability sample survey,17 anal intercourse was reported more frequently by men than women, and by AA than WA. The estimates are not strictly comparable because we asked about anal intercourse during a typical 4‐week period with the most recent sex partner, whereas Laumann et al. measured anal intercourse during last event and last year. In contrast to the national sample, we found nonmarried people to be more likely than married ones to report anal intercourse. We also found WA men with fewer sex partners to be more likely to have engaged in anal intercourse with their current partner; further, WA men who always used a condom were also more likely to report anal intercourse.
Dry sex, douching, and sex during menses were interrelated, possibly because of logical or temporal linkages. Women may remove vaginal secretions before engaging in sex during menses; douching is one way to remove secretions, "drying" the vagina. In the WA sample, people engaging in dry sex were more likely to engage in sex during menses, douching, or anal intercourse. AA men engaging in dry sex were also more likely to report sex during menses.
The prevalence of self‐reported STD history in this study is similar to that reported for adults enrolled in a multicenter study of cardiovascular disease,19 and a national multistage probability household survey of women selected as controls for a cancer study.20 The lack of proper time sequence and validation of self‐reported history of STD limits the interpretation of the associations between the practices described here and STD history. Therefore, the consistency of several of the relationships with the literature gives somewhat more weight to newly identified associations. The associations of anal sex with self‐reported history of hepatitis13 and that of douching with PID8 are consistent with previous studies. Sex during menses has been associated previously with PID,21 and chlamydial salpingitis with menses within the previous week22; our finding of an association of sex during menses with self‐reported history of chlamydial infection underscores the need for further research in this area. Although the numbers are very small, dry sex was associated with self‐reported history of a number of STD‐but only in the AA population, where the prevalence of STD and of dry sex was more common. Future studies should determine whether the associations between STD history and dry sex can be confirmed, as well as describing the drying agents used and the reasons for engaging in the practice in the United States.
The association of anal sex with gonorrhea and herpes among women suggests that women engaging in anal sex might be at higher risk of contacting infected partners rather than the practice being the mechanism of transmission: rectal gonorrhea is rare among women. Anal sex was reported most frequently by single, widowed, or divorced women, and was associated with increased numbers of lifetime vaginal sex partners, never or only sometimes using a condom, and having both same‐ and opposite‐sex partners. In an analysis of the sexual repertoires in these data (manuscript in preparation), having nonmutually monogamous relationships was a predictor of engaging in the combination of oral, vaginal, and anal sex. Although further study is needed, should women with same‐ and opposite‐sex partners tend to have opposite‐sex partners who also have same‐sex partners, these women would be, through their partners, mixing with the gay population, which has a higher prevalence of STD.
Sexual and other behaviors reportedly vary by geographic area.23 The prevalences of douching in our WA and AA samples are very similar to those reported by a national probability sample,1 but, as noted earlier, there were some differences from other studies in frequency of anal sex17 and sex during menses5 by sociodemographic characteristics. A preliminary analysis comparing mixing patterns in this population to national data suggests that sexual mixing patterns in Seattle may differ from mixing patterns in the nation as a whole.24 Because STD risk is a function of mixing patterns as well as of individual behavioral practices, the nature and strength of the associations observed between self‐reported STD history and individual risk factors reported here may differ from those observed at the national level or in other local populations.
The relative impact of an individual risk factor on acquiring disease in a particular population varies with the relative prevalence of both the risk factor and the disease of interest as well as where on the disease pathway the factors acts.25 Thus, the higher risk of self‐reported chlamydial infection with sex during menses among WA women (OR = 3.9; CI: 1.1, 14.0) may reflect the fact that single WA women with more lifetime sex partners are more likely to engage in sex during menses, as well as any biologic effect of sex during menses on chlamydial transmission. By contrast, the OR for the association of sex during menses and chlamydial infection among AA women was 1.6 (CI: 0.5, 5.4). This may indicate that AA women engaging in sex during menses have a somewhat greater chance of encountering an infected partner compared with AA women not engaging in sex during menses, even after adjusting for number of partners, or that engaging in sex during menses increases susceptibility to acquiring chlamydial infection. Alternatively, some other associated factor that occurs more frequently among AA women‐such as douching‐may somehow modify the association. Although the small number of chlamydial cases precludes definitive analysis, we did find that women engaging in both sex during menses and douching were more likely to report chlamydial infection than those engaging in one or other or neither practice; the effects of the two practices together were slightly greater than additive.
As in any survey with less than 100% response rate, the respondents may have chosen to participate based on some factor under study. Thus, it is reassuring that the distribution of demographic variables is similar to that reported for Seattle as a whole.26 In surveys of sexual behavior, it is difficult to validate the accuracy of reporting‐although responses can be checked for internal consistency, repeatability, and correlations with current sex partners. Sexual reports have been shown to be repeatable,27 and reports of sexual behavior by sex partners are highly correlated for short time periods.28,29
This study identified some important methodologic issues. Most important, the refusal rate among potential respondents in the AA listed sample was twice that of the RDD sample (36% vs. 18%). Other factors, such as exclusions (except race), inability to contact, and presence of phone answering machines were similar across the two samples. One sixth of potential respondents in both samples could not be contacted because the phone was answered by a machine at each attempt to call, and the call was not returned (using a toll‐free number) after a message was left on the phone machine. In both samples, almost all respondents agreeing to be interviewed completed the interview (RDD: 95%; AA: 94%).
In conclusion, these data confirm that douching is associated with risk of PID and suggest that sex during menses may increase a woman's risk of acquiring chlamydial infection. These behaviors occur with sufficient frequency in the populations studied to warrant targeting for intervention. Anal sex was associated with risk of several STD, and with a history of same‐ and opposite‐sex partners, suggesting that anal sex not only is successful at transmitting several agents but may be a marker for mixing with higher‐risk populations. Future studies, using study designs that allow assessment of the temporal relationship of exposure to STD outcome, should address the associations of dry sex and sex during menses with STD, taking care to account for the correlations between these sexual practices and other potentially high‐risk sexual behaviors.
References
1. Aral SO, Mosher WD, Cates W. Vaginal douching among women of reproductive age in the United States: 1988. Am J Public Health 1992; 82:210-214.
2. Brown JE, Ayowa OB, Brown RC. Dry and tight: Sexual practices and potential AIDS risk in Zaire. Soc Sci Med 1993; 37:989-994.
3. Civic D, Wilson D. Dry sex in Zimbabwe and implications for condom use. Soc Sci Med 1996; 42:91-98.
4. Dallabetta GA, Miotti PG, Chiphangwi JD, Liomba G, Canner JK, Saah AJ. Traditional vaginal agents: Use and association with HIV infection in Malawian women. AIDS 1995; 9:293-297.
5. Tanfer K, Aral SO. Sexual intercourse during menstruation and self-reported sexually transmitted disease history among women. Sex Transm Dis 1996; 23:395-401.
6. Kassler WJ, Aral SO. Beyond risk groups: Behavioral correlates of HIV seroconversion in sexually transmitted disease clinic patients [Abstract]. In: Eleventh Meeting of the International Society for STD Research, New Orleans, Louisiana, August 27-30, 1995.
7. Malamba SS, Wagner HU, Maude G, et al. Risk factors for HIV-1 infection in adults in a rural Ugandan community: A case-control study. AIDS 1994: 8:253-257.
8. Wolner-Hanssen P, Eschenbach DA, Paavonen J, et al. Association between vaginal douching and acute pelvic inflammatory disease. JAMA 1990; 263:1936-1941.
9. Scholes D, Daling JR, Stergachis A, Weiss NS, Wang SP, Grayston JT. Vaginal douching as a risk factor for acute pelvic inflammatory disease. Obstet Gynecol 1993; 81:601-606.
10. Daling JR, Weiss NS, Schwartz SM, et al. Vaginal douching and the risk of tubal pregnancy. Epidemiology 1991; 2:40-48.
11. Baird DD, Weinberg CR, Voigt LF, Daling JR. Vaginal douching and reduced fertility. Am J Public Health 1996; 86:844-850.
12. Levine WC, Pope V, Farshy C, et al. Association of vaginal douching with increase in cervicitis and endocervical CD4 lymphocytes among women attending an STD clinic[Abstract]. Eleventh Meeting of the International Society for STD Research, New Orleans, Louisiana, August 27-30, 1995.
13. Osmond DH, Charlebois E, Sheppard HW, et al. Comparison of risk factors for hepatitis C and hepatitis B virus infection in homosexual men. J Infect Dis 1993; 167:66-71.
14. Samuel MC, Hessol N, Shiboski S, Engel RR, Speed TP, Winkelstein W. Factors associated with human immunodeficiency virus seroconversion in homosexual men in three San Francisco cohort studies, 1984-1989. Journal of Acquired Immune Deficiency Syndromes 1993; 6:303-312.
15. U.S. Department of Health and Human Services, Public Health Service. Sexually Transmitted Disease Surveillance 1992. Atlanta, GA: Centers for Disease Control and Prevention, July, 1993.
16. Rice RJ, Roberts PL, Handsfield HH, Holmes KK. Sociodemographic distribution of gonorrhea incidence: Implications for prevention and behavioral research. Am J Public Health 1991; 81:1252-1258.
17. Laumann EO, Gagnon JH, Michael RT, Michaels S. The Social Organization of Sexuality: Sexual Practices in the United States. Chicago: The University of Chicago Press, 1994.
18. SAS Institute, Inc. SAS for Personal Computers, Version 6 Edition. Cary, NC: SAS Institute, Inc., 1994.
19. Melnick SL, Burke GL, Perkins LL, et al. Sexually transmitted diseases among young heterosexual urban adults. Public Health Rep 1993; 108:673-679.
20. Stone KM, Zaidi A, Rosero-Bixby L, et al. Sexual behavior, sexually transmitted diseases, and risk of cervical cancer. Epidemiology 1995; 6:409-414.
21. Jossens MOR, Eskenazi B, Schacter J, Sweet RL. Risk factors for pelvic inflammatory diseases: A case control study. Sex Transm Dis 1996; 23:239-247.
22. Sweet RL, Blankfort-Doyle M, Robbie MO, Schacter J. The occurrence of chlamydial and gonococcal salpingitis during the menstrual cycle. JAMA 1986; 255:2062-2064.
23. Wadsworth J, Hickman M, Johnson AM, Wellings K, Field J. Geographic variation in sexual behavior in Britain: Implications for sexually transmitted disease epidemiology and sexual health promotion. AIDS 1996; 10:193-199.
24. Aral SO. Patterns of sex partner recruitment and types of mixing as determinants of STD transmission: Limits to the spread of sexually transmitted infections. Venereology 1995; 8:240-242.
25. Koopman JS, Weed DL. Epigenesis theory: A mathematical model relating causal concepts of pathogenesis in individuals to disease patterns in populations. Am J Epidemiol 1990; 132:366-390.
26. U.S. Department of Commerce, Bureau of the Census. 1990 Census of Population and Housing. Washington, DC: U.S. Government Printing Office, 1994.
27. Rohan TE, McLaughlin JR, Harnish DG. Repeatability of interview-derived information on sexual history: A study in women. Epidemiology 1994; 5:360-363.
28. Seage GR, Mayer KH, Horsburge CR, Cai B, Lamb GA. Corroboration of sexual histories among male homosexual couples. Am J Epidemiol 1992; 135:79-84.
29. Lagarde E, Enel C, Pison G. Reliability of reports of sexual behavior: A study of married couples in rural West Africa. Am J Epidemiol 1995; 141:1194-1200.