Vaginal Douching and Reduced Fertility Donna Day Baird, PhD, Clanice R. Weinberg, PhD, Lynda F. Voigt, PhD, andJanet R. Daling, PhD Introduction tubal infertility found an association be- tween douching and infertility among Vaginal douching has been practiced both those with and those without prior throughout recorded history.' A wide diagnosis of pelvic inflammatory disease variety of douching preparations have but no increase in risk with increased been used: garlic and wine (noted in the frequency of douching. Fertility impair- Egyptian Papyrus of 1500 B.C. as a ment is a difficult medical problem to treatment for menstrual disorders2), di- study with standard epidemiologic meth- lute Lysol (advertised in the early 20th ods because only a self-selected minority century in the United States for regular of women with fertility problems seek feminine hygiene and for contraception'), treatment.26 To avoid the problems inher- and water and vinegar (commonly used ent in studying infertility patients, we today35). Historically, the medical commu- compared the fertility of douchers and nity recommended douching for treat- nondouchers by collecting information on ment of specific gynecologic conditions6,7 the number of months parous women but was divided about the efficacy and required to become pregnant. This method safety of regular douching as a hygiene of studying fertility in women who are practice among healthy women.819 able to conceive27 has been used previ- Vaginal douching remains a common ously to investigate effects of such factors practice among women in the United as oral contraceptive use,28 cigarette smok- States. In a national survey conducted in ing,29'30 and occupational exposures.31'32 1988, 37% of women between the ages of The method will tend to underestimate an 15 and 44 years reported douching; 18% adverse effect of douching on fertility douched at least once a week.'0 Black because sterile women are excluded by women were about twice as likely to design. douche as White women (67% vs 32%), and those who did tended to douche more frequently. Methods Douching may not be benign. It has We compared the monthly probabil- been linked to increased risk of chlamydia ity of conception for douchers and non- infection,""2 pelvic inflammatory dis- douchers in married, parous women. The ease,4'0">I7 and ectopic pregnancy5"120 (but see Phillips et al.2'). Independent Donna Day Baird and Clarice R. Weinberg are associations were demonstrated even in with the Division of Intramural Research, studies that controlled for the reason for National Institute of Environmental Health douching and for sociobehavioral corre- Sciences, Research Triangle Park, NC. Lynda lates including sexual behavior.5"5 Re- F. Voigt and Janet R. Daling are with the and Department of Epidemiology, School of Public ports linking douching to vaginal Health and Community Medicine, University cervical cancer raise further concerns of Washington, Seattle. about the safety of the practice.22-24 Requests for reprints should be sent to A relationship between douching Donna Day Baird, PhD, Epidemiology Branch, and reduced a common sequela MD A3-05, National Institute of Environmen- fertility, tal Health Sciences, PO Box 12233, Research ofboth clinical and silent pelvic inflamma- Triangle Park, NC 27709. tory disease, has been addressed in only This paper was accepted November 1, one report.25 This case-control study of 1995. June 1996, Vol. 86, No. 6 Douching and Fertliy women studied were controls from a Cumulative Percent Pregnant population-based case-control study of infertility conducted in King County, Cumulative Percent Pregnant Washington.33 Married mothers were se- by Frequency of Douching lected from birth certificate files matched to cases on race, age (within 5 years), census tract, and gravidity at the time of 100 the pregnancy attempt. Of the women selected, 1026 (74%) were interviewed. 90- Women who were under the age of 18 years (n = 9), who had diabetes (n = 9), or who were on medication for seizures 80- (n = 6) were excluded. No meaningful time to pregnancy could be derived for 133 additional women, most of whom E~ 70 reported that they conceived while using contraception. Because the detailed ques- tions from which we derived our douching 60- categories were not included in the first a) few interviews, another 29 women were 0.- excluded. The final sample for analysis consisted of 840 women. Data on time to pregnancy, douch- o 40- ing, and other factors potentially related a) to reduced fertility were obtained retro- 0. spectively in a structured interview con- 30 -0-never douche (517) ducted in each respondent's home. Time to pregnancy was ascertained by collect- 20-&- .2/year(145) ing data on the number of months of 20- unprotected intercourse required to con- > 2/year, < 1/week (155) ceive. Women who reported having planned their indexpregnancywere asked, 10 - - > 1 /week (23) "How long (in months) were you trying to become pregnant?" Women with "un- planned" pregnancies were asked when 1 01 1112 they stopped using birth control before 1 2 3 4 5 67 8 9 the index pregnancy. Douching data were derived from questions about douching Number of Months without frequency at 1, 2, 3, 4, and 5 years prior to Contraception conception of the index pregnancy. At interview, respondents reported one of four douching frequencies for each of the FIGURE 1-Unadjusted cumulative percentage pregnant each month for women five time periods: (1) never douched, (2) In four groups determined by their frequency of douching. douched no more than twice a year, (3) douched more than twice a year but no more than once a week, or (4) douched more than once a week. The five numbers sure to diethylstilbestrol; history of pelvic soda, cocoa, and alcohol intake; and prior representing douching frequencies for the inflammatory disease, including postdeliv- recreational drug use, including mari- 5 years prior to conception were averaged ery infection; endometriosis; ovarian cysts; juana, cocaine, LSD, speed, and "other"). and rounded to the nearest whole num- sexually transmitted diseases; and body These life-style variables also were col- ber, and the corresponding category was mass index), reproductive factors (any lected from each woman for her spouse. assigned as the douching variable used in prior use of oral contraceptives, use of Most variables were measured as of the the analyses presented here. A second oral contraceptives as last method of beginning of eachwoman's noncontracept- variable based on the four-level douching contraception, any prior use of an intra- ing time period, but the following were frequency for the year closest to the time uterine device, pregnancy history, age at measured as of the time just before when each woman began attempting first sexual intercourse, number of sexual conception: family income, education, pregnancy showed similar associations partners, frequency of sexual intercourse, body mass, usual menstrual cycle length, (data not shown). and breast-feeding a child born before the regularity of menstrual cycle length, fre- Other factors potentially related to index pregnancy and continuing to do so quency of sexual intercourse, number of fertility included the following: demo- during at least part of the time to prior sexual partners, number of ciga- graphic factors (age, race, marital status, pregnancy), and life-style factors (smok- rettes smoked per day, and monthly education, and family income), self- ing status and number of cigarettes coffee, tea, soda, cocoa, and alcohol reported medical factors (prenatal expo- smoked per day; monthly coffee, tea, intake. June 1996, Vol. 86, No. 6 American Journal of Public Health 845 Baird et al. ratios, calculated from the multivariate TABLE 1-Prepregnancy Characteristics of 840 Women, by Their Practice of model, provide a measure ofthe degree to Vaginal Douching during the 5 Years before the Index Pregnancy which a factor is independently associated with fecundability (analogous to risk ra- Douchers Nondouchers tios in studies of chronic diseases). For (n = 178), No. (%) (n = 662), No. (%) P example, the fecundability ratio associ- ated with douching would be 0.5 if Demographic factors douchers were half as likely as nondouch- Age,y NS ers to become pregnant in each noncontra- <20 5 (3) 15 (2) cepting month. Female factors were mod- 20-24 50 (28) 169 (26) eled first. None of the several male factors 25-29 93 (52) 328 (50) 30-34 24 (13) 137 (21) that were subsequently modeled changed 35+ 6 (3) 13 (2) the relationship between douching and Race (non-White) 16 (9) 39 (6) NS fecundability, so they were dropped in Family income, $ NS further analyses. Heterogeneity of effect < 15 000 21 (12) 116 (18) 15 000-30 000 97 (55) 325 (49) was tested by including cross-product > 30 000 59 (33) 217 (33) terms in the multivariate model. Education, y <.01 <12 6 (3) 10 (2) 12 48 (27) 135 (20) Resuls 13-15 70 (39) 226 (34) .16 54 (30) 291 (44) Times to pregnancy ranged from 1 Marital status (unmarried) 6 (3) 25 (4) NS month to more than 10 years; 13% of the women required more than a year to Medical factors conceive. Douching was reported by 38% Prenatal diethylstilbestrol exposure 2 (1) 15 (2) NS ofthe women. Seventeen percent douched Prescribed medications 5 (3) 21 (3) NS very rarely (no more than twice a year). History of pelvic inflammatory disease 8 (4) 28 (4) NS History of endometriosis 3 (2) 20 (3) NS The cumulative pregnancy rate varied by History of ovarian cyst 9 (5) 52 (8) NS frequency ofdouching (Figure 1). Women History of sexually transmitted diseases 36 (20) 111 (17) NS who douched most frequently (more than Body mass index, kg/M2 NS once a week) had the lowest cumulative < 20.0 (thin) 69 (39) 224 (34) were still not 20.0-26.9 (average) 97 (54) 395 (60) pregnancy rate: 27% preg- .