Abstract
Background
Perineal talc use and douching could affect risk of uterine cancer through several possible pathways, including inflammation response, microbiota changes, or endocrine disruption. Two previous cohort studies of the association between talc use and uterine cancer have reported weak positive associations, but we know of no previous evaluations of the relationship between douching and uterine cancer.
Methods
Using a large prospective cohort, we examined the relationship between incident uterine cancer and self-reported use of talc or douche using Cox proportional hazards models.
Results
After excluding those with prior hysterectomy, 271 of 33,609 women reported incident uterine cancer (mean follow-up=8.3 years in non-cases; maximum 12.6 years). Overall, 26% of women reported ever using talc and 15% reported ever having douched. Ever talc use was associated with an increase in risk of uterine cancer (adjusted hazard ratio [HR]=1.2; 95% confidence interval [CI]: 0.94–1.6), with some evidence of a dose–response for frequency of talc use (p-for-trend=0.07). Ever douching was not associated with uterine cancer risk (HR=1.0; 95% CI: 0.72–1.5), with no evidence of a frequency dose–response (p=0.96). The estimates were similar when we restricted to invasive endometrial cancers, but not when we further restricted to endometroid adenocarcinomas.
Conclusions
The positive association we observed between talc use and uterine cancer risk is consistent with findings from previous prospective cohort studies of endometrial cancer. The relationships between uterine cancer and both douching and talc use merit further consideration, particularly as both exposures are preventable.
Keywords: uterine cancer, endometrial cancer, talc, douching
Introduction
Women may apply talc in the perineal area to help absorb moisture, manage odors, or prevent skin irritation.1 This common practice may abrade the surfaces of the vagina and reproductive tract, inciting an inflammatory response2 and potentially facilitating carcinogenesis in the uterus and other organs.3–5 Douche is another personal care product that women use to address vaginal odors, to prevent pregnancy, or to provide a feeling of cleanliness, particularly during or just after menses.6–9 Douching may also be associated with cancers of the reproductive tract via inflammatory pathways8 and could additionally affect cancer risk by altering vaginal or uterine microbiota9–11 or through exposure to hormone-disrupting chemicals such as phthalates, which are included in some products.12,13
The association between perineal talc use and ovarian cancer has been studied extensively,14 but few studies have examined the association between talc and uterine cancer, even though the uterus is more proximal to the exposure site. Due to the strong possibility of recall bias in a time of active and highly publicized lawsuits,15–18 it is particularly important to study such associations using prospective designs, with exposure to talc assessed prior to the development of the disease of interest. So far, none of the existing, prospective studies has provided clear evidence of an association between talc use and either ovarian19–21 or endometrial cancer.22,23 A retrospective case–control study of talc use and endometrial cancer also reported null results.24 In a previous analysis, we observed an inverse association between talc use in the last year and ovarian cancer, but a positive association between recent douching and ovarian cancer.21 We know of no other prior studies of the effects of douching on cancers of the reproductive tract.
The purpose of this analysis is to examine the association between perineal talc use, douching, and risk of uterine cancer in a large, prospective cohort. We hypothesized that both talc use and douching would be associated with increased risk of the disease and were particularly interested in whether the association would vary by uterine cancer subtype or be modified by factors such as race/ethnicity, menopausal status or obesity.
Methods
Study sample
We used data from the Sister Study, a prospective study (2003–2009) that enrolled 50,884 breast cancer-free women who had a sister previously diagnosed with the disease. Sister Study participants had to live in the US or Puerto Rico and be between the ages of 35 and 74 at enrollment. Baseline questionnaires consisted of a computer-assisted telephone interview and self-completed surveys retrieved during an in-home examiner visit (including one on use of personal care products). Active cohort follow-up includes basic health data collected annually and more detailed follow-ups completed every 2–3 years. Women reporting incident cancer diagnoses were asked to provide pathology reports or permission to retrieve their pathology reports. Data for this analysis is complete through September 2016 (data release 6.0), at which point more than 90% of participants had responded to their most recent follow-up request. More detailed information about the study can be found elsewhere.25 The study was reviewed and approved by the Institutional Review Boards of the National Institute of Environmental Health Sciences and the Copernicus group. All participants provided written informed consent.
For this study of uterine cancer, we excluded women who were diagnosed with uterine cancer before completion of the baseline interview (n=366), women with an uncertain uterine cancer history (n=67), women who withdrew their consent (n=2), women with pre-baseline hysterectomies (n=15,602), and women who did not fill out the personal care product questionnaire (n=989). After censoring follow-up time for women who reported a hysterectomy during follow-up but before any uterine cancer diagnosis (n=2,414), we identified 274 incident invasive uterine cancer cases among 33,858 women.
Exposure assessment
The baseline questionnaires included questions about personal care product use in the 12 months prior to enrollment. Specifically, participants were asked to report their frequency of douching (did not use, used <1 time per month, 1–3 times/month, 1–5 times/week, or >5 times/week) and about their frequency of genital talc use in the form of powder or spray applied to a sanitary napkin, underwear, diaphragm, cervical cap, or vaginal area (same categories). The baseline questionnaire also included questions about perineal talc and douche use at age 10–13 (never, sometimes, frequently). For our main analyses, we categorized women as ever users if they reported use in the last 12 months or at ages 10–13. We also categorized women based on frequency of use, with frequent users defined as those who reporting douching or applying perineal talc at least once a month in the last year or “frequently” at ages 10–13. We also considered timing of use and the combined use of both douche and talc. We used “long-term” use to describe those who reported use at both ages 10–13 and in the last year.
Outcome assessment
We were able to confirm 207 (76%) of the reported cases using data abstracted from pathology reports or medical records (n=193) or from death certificates/National Death Index (NDI) Plus (primary or underlying cause of death; n=14). For the remaining 67 cases, data were self-reported (n=65) or reported by next of kin (n=2), but we could not obtain medical confirmation. Among women who self-reported uterine cancer and provided medical records, 88% of cases were confirmed. Age at diagnosis was based on the best available information from medical records (n=141) or self-report (n=121). In rare cases, we relied on their next of kin’s report (n=9) or NDI Plus (n=3), filling in the midpoint of ages where the participant was last known to not have uterine cancer and the age at which she was first known to have had uterine cancer if exact diagnosis age was not provided (n=5). When possible, we also collected information about histologic subtype, with 188 of the 207 medically confirmed cases classified as invasive endometrial cancers (International Classification of Disease [ICD]-10 code C54.1) and 146 of those 188 classified as invasive endometroid adenocarcinomas (ICD for Oncology-3 histology code 8380, 8570, 8282, or 8382).26 Medical record confirmation status was similar within exposure categories with (76% of self-reported cases confirmed among non-users versus 73% and 74% of those exposed to douching or talc, respectively).
Statistical Analysis
We evaluated the association between uterine cancer and douching or talc use using a Cox proportional hazards model with age as the primary time scale. We estimated hazard ratios (HRs) and 95% confidence intervals (CIs), using robust sandwich estimates27 to account for within-family clustering due to the inclusion of some multiple sisters from the same family (7.8% of families had >1 sister participating). Participants contributed person–time from the age they completed enrollment until age at which they developed uterine cancer, with censoring at age of hysterectomy, last follow-up, or death. By censoring individuals with competing events at the age they occurred, we are estimating the cause-specific HR for the association between talc use or douching and uterine cancer.28 We conducted trend tests treating frequency of use as an ordinal variable and checked proportional hazards assumptions by performing likelihood ratio tests of models with and without age-by-exposure interaction terms.
The following covariates were included as confounders, in addition to the timescale adjustment for age: race/ethnicity (Non-Hispanic white, African-American, Hispanic, other), highest level of completed education (high school or less, some college, bachelor’s degree, graduate degree), body mass index (BMI; restricted cubic spline with knots at the 5th, 35th, 65th, and 95th percentile), menopause status at baseline (pre- or post-menopausal), parity (nulliparous, parous), duration of oral contraceptive use (none, <2 years, 2-<10 years, ≥10 years), ever hormone therapy use (none, estrogen alone, or estrogen plus progestin but never estrogen alone), smoking (never, former, current), alcohol use (never/former, current <1 drink/day, current ≥1 drink/day), and age at menarche (continuous). These were selected based on a priori knowledge of possible risk factors for uterine cancer that may also be related to talc use or douching. All were assessed during the baseline interview. For the analysis of use at ages 10–13, we only adjusted for race/ethnicity, age at menarche and relative weight at age 10 (lighter than, same as, or heavier than peers), as none of the other covariates could be confounders. After excluding those with missing data for one or more confounders, our analysis included 33,609 women, 271 of whom developed uterine cancer during follow-up.
To assess possible effect modification, we examined the effects of douching and talc use within strata of menopausal status (as a time-varying factor updated throughout follow-up), race/ethnicity, parity, BMI, and unopposed estrogen use. We tested for heterogeneity by conducting likelihood ratio tests of the interaction between talc or douching and the effect modifier of interest.
We performed various sensitivity analyses, including 1) adjusting for talc use in the assessment of the effect of douching (and vice versa), as these exposures often co-occur and may confound one another; 2) adjusting for waist circumference (restricted cubic spline) or 3) weight (restricted cubic spline) instead of BMI, as being overweight or obese is a strong risk factor for use of personal care products (especially talc) and uterine cancer, and it is not clear which anthropometric measure best controls this confounding; 4) excluding the first year of follow-up to assess the possibility of reverse causation (i.e., whether those with preclinical uterine cancer are more likely to douche or use talc); 5) adjusting for baseline tamoxifen use, as tamoxifen is a risk factor for uterine cancer and has the potential to act as a confounder of the measured relationships. Additionally, we conducted analyses looking at 6) only medically confirmed cases (n=207), censoring unconfirmed cases at the time of diagnosis; and 7) performing analysis limited to medically confirmed endometrial cancers (n=190) or medically confirmed endometroid adenocarcinomas (n=147). In a separate sensitivity analysis, we looked specifically at the association between regular talc use (i.e., use least once per week) and uterine cancer among postmenopausal women. This was done to facilitate a direct comparison between our results and a similar analysis done in the Nurses’ Health Study.23
Results
We followed participants for an average of 8.3 years (maximum 12.6 years). The Sister Study cohort is predominately non-Hispanic white (85%; Table 1) and well educated (26% had a graduate degree). Compared to the full cohort, incident cases were more likely to be non-Hispanic white (89%) and highly educated (30% with graduate degree). They were also older on average (58.7 versus 54.2 at enrollment). One of the major differences between cases and the full cohort was the distribution across categories of BMI (43% with BMI <25.0 kg/m2, 31% with BMI 25–29.9, and 27% with BMI ≥30 for the cohort, with corresponding proportions of 27%, 28% and 45% among cases.)
TABLE 1.
Baseline characteristics of the Sister Study (2003–2009)a
| Full Cohort (N=33,858) | Uterine Cancer Cases (N=274) | |
|---|---|---|
| Age; Mean (std) | 54.2 (8.9) | 58.7 (8.1) |
| Follow-up time; Mean (std) | 8.3 (2.2) | 4.9 (2.8) |
| Age at menarche; Mean (std) | 12.7 (1.5) | 12.5 (1.4) |
| Race; N (%) | ||
| Non-Hispanic White | 28,915 (85) | 243 (89) |
| Non-Hispanic Black | 2,476 (7.3) | 17 (6.2) |
| Hispanic | 1,611 (4.7) | 10 (3.7) |
| Other | 849 (2.5) | 4 (1.5) |
| Education; N (%) | ||
| High school or less | 4,535 (13) | 36 (13) |
| Some college | 10,501 (31) | 87 (32) |
| Bachelor’s degree | 9,910 (29) | 70 (26) |
| Graduate degree | 8,906 (26) | 81 (30) |
| Weight relative to peers, age 10; N (%) | ||
| Lighter | 11,731 (35) | 90 (33) |
| Same weight | 15,776 (47) | 127 (47) |
| Heavier | 6,253 (19) | 54 (20) |
| Body Mass Index (BMI); N (%) | ||
| <25.0 kg/m2 | 14,444 (43) | 75 (27) |
| 25–<30 kg/m2 | 10,351 (31) | 77 (28) |
| 30 kg/m2 | 9,052 (27) | 122 (45) |
| Postmenopausal; N (%) | 19,323 (57) | 202 (74) |
| Hormonal Contraception Use; N (%) | ||
| None | 5,148 (15) | 64 (23) |
| <2 years | 5,005 (15) | 54 (20) |
| 2–<10 years | 14,459 (43) | 113 (41) |
| ≥10 years | 9,216 (27) | 42 (15) |
| Hormone Therapy Use; N (%) | ||
| None | 23,442 (69) | 170 (62) |
| Estrogen Alone | 2,322 (6.9) | 32 (12) |
| Estrogen plus Progestin | 8,011 (24) | 71 (26) |
| ≥1 live birth; N (%) | 27,099 (80) | 212 (78) |
| Smoking Status; N (%) | ||
| Never | 19,289 (57) | 143 (52) |
| Former | 11,921 (35) | 122 (45) |
| Current | 2,639 (7.8) | 9 (3.2) |
| Alcohol Use; N (%) | ||
| Never or Former | 5,603 (17) | 49 (18) |
| Current, <1 drink/day | 23,296 (69) | 186 (68) |
| Current ≥1 drink/day | 4,903 (15) | 39 (14) |
| Ever Used Tamoxifen; N (%) | 526 (1.6) | 4 (1.5) |
Missing: Race (7 from cohort), education (6 from cohort), relative weight age 10 (98 from cohort, including 3 cases), BMI (11 from cohort), menopausal status (28 from cohort), hormone contraception (30 from cohort, including 1 case), hormone therapy (83 from cohort, including 1 case), parity (22 from cohort, including 1 case), smoking (9 from cohort), alcohol (56 from cohort), Tamoxifen (58 from cohort, including 1 case).
Excludes women who withdrew (N=2), were diagnosed with uterine cancer before baseline (N=366), had uncertain uterine cancer status (N=67), had a hysterectomy before baseline (N=15,602), or did not complete the personal care products questionnaire (n=989).
Overall, 15% of participants reported ever douching, and 26% of participants reported using talc. The practices are not independent and 5.6% reported using both. Frequent use was fairly rare for both (2.8% for douching and 9.3% for talc). Less than 6% of women reported long-term use of talc. As less than 1% of women reported long-term use of douche, we did not include this exposure in our assessment of uterine cancer risk. Of the women who reported ever using talc, 48% reported using at ages 10–13 but not in the last year, 30% reported using in the last year but not at ages 10–13, and 21% used at both times. Among women who ever used douche, 14% reported using at ages 10–13 but not in the last year, 80% reported using in the last year but not at ages 10–13, and 5.5% used at both times.
We observed a positive association between uterine cancer and ever talc use (HR=1.2; 95% CI: 0.94–1.6; Table 2). There was no association between uterine cancer and ever douching (HR=1.0; 95% CI: 0.72–1.5). The hazard ratios from models that adjusted for age alone were higher (HR=1.4, 95% CI: 1.1–1.8 for talc and HR=1.1, 95% CI: 0.76–1.5 for douching), suggesting that the unadjusted analyses were likely confounded up and away from the null. There was no evidence that the proportional hazards assumption was violated in any analyses.
TABLE 2.
Associations between douching, talc use, and uterine cancer in the Sister Study
| Full cohort, N(%) | Uterine cancer cases, N(%) | Adjusted Hazard Ratiob | |
|---|---|---|---|
| N=33,609a | N=271a | ||
| Douching | |||
| Ever | |||
| No | 28,459 (85) | 231 (86) | 1.0 |
| Yes | 4,896 (15) | 37 (14) | 1.0 (0.72, 1.5) |
| Frequencyc | |||
| None | 28,459 (85) | 231 (86) | 1.0 |
| Sometimes | 3,969 (12) | 31 (12) | 1.1 (0.73, 1.5) |
| Frequently | 927 (2.8) | 6 (2.2) | 0.86 (0.37, 2.0) |
| p-trend= 0.96 | |||
| Ever in past 12 months | |||
| No | 29,139 (87) | 234 (88) | 1.0 |
| Yes | 4,201 (13) | 33 (12) | 1.1 (0.75, 1.6) |
| Ever, ages 10–13 | |||
| No | 32,183 (97) | 259 (98) | 1.0 |
| Yes | 961 (2.9) | 5 (1.9) | 0.67 (0.28, 1.6)d |
| Talc use | |||
| Ever | |||
| No | 24,250 (74) | 178 (66) | 1.0 |
| Yes | 8,686 (26) | 90 (34) | 1.2 (0.94, 1.6) |
| Frequencyc | |||
| None | 24,250 (74) | 178 (66) | 1.0 |
| Sometimes | 5,610 (17) | 51 (19) | 1.1 (0.81, 1.5) |
| Frequently | 3,076 (9.3) | 39 (15) | 1.4 (0.99, 2.0) |
| p-trend= 0.07 | |||
| Ever in past 12 months | |||
| No | 28,299 (86) | 221 (82) | 1.0 |
| Yes | 4,625 (14) | 47 (18) | 1.1 (0.82, 1.6) |
| Ever, ages 10–13 | |||
| No | 25,940 (82) | 198 (77) | 1.0 |
| Yes | 5,847 (18) | 59 (23) | 1.2 (0.90, 1.6)d |
| Long-term use (age 10–13 and last 12 months) | |||
| No | 29,989 (94) | 244 (94) | 1.0 |
| Yes | 1,786 (5.6) | 16 (6.2) | 1.1 (0.65, 1.8)d |
| Combined use | |||
| Ever | |||
| Neither | 21,344 (65) | 159 (60) | 1.0 |
| Both | 1,840 (5.6) | 18 (6.8) | 1.3 (0.79, 2.1) |
| Talc/no douching | 6,715 (21) | 20 (26) | 1.2 (0.90, 1.6) |
| Douching/no talc | 2,783 (8.5) | 18 (6.8) | 0.98 (0.60, 1.6) |
| Ever in past 12 months | |||
| Neither | 25,023 (77) | 195 (74) | 1.0 |
| Both | 968 (3.0) | 9 (3.4) | 1.2 (0.64, 2.4) |
| Talc/no douching | 3,586 (11) | 37 (14) | 1.1 (0.79, 1.6) |
| Douching/no talc | 3,082 (9.4) | 23 (8.7) | 1.1 (0.68, 1.7) |
Missing: Ever douching and douching frequency (254 from cohort, including 3 cases), douching last 12 months (269 from cohort, including 4 cases), douching age 10–13 (465 from cohort, including 4 cases), ever talc use and talc use frequency (673 from cohort, including 3 cases), talc last 12 months (685 from cohort, including 3 cases), talc use ages 10–13 (1,822 in cohort, including 11 cases), long-term talc use (1,834 in cohort, including 11 cases), both ever (927 in cohort, including 6 cases), both in last 12 months (950 in cohort, including 7 cases)
Participants with complete confounder information.
Adjusted for race/ethnicity, education, body mass index (restricted cubic spline), menopausal status, parity, duration of oral contraceptive use, hormone therapy use, smoking, alcohol use, and age at menarche.
Sometimes users used <1/month in the last year or “sometimes” ages 10–13, regular users used ≥1/month in the last year or “frequently” ages 10–13
Adjusted for race/ethnicity, age at menarche and relative weight at age 10 only, had to be non-missing for other covariates
We observed a positive trend for frequency of talc use and uterine cancer (p-for-trend 0.07, with HR=1.4, 95% CI: 0.99–2.0 for frequent users). Talc use in the last year was associated with a small increase in risk (HR=1.1, 95% CI: 0.92–1.6, as were use at ages 10–13 (HR=1.2, 95% CI: 0.90–1.6) and long-term use (HR=1.1, 95% CI: 0.65–1.8, respectively).
Self-reported douching in the last year was associated with a small increase in uterine cancer risk (HR=1.1, 95% CI: 0.75–1.6), while douching at ages 10–13 was associated with a reduction in risk (HR=0.67, 95% CI: 0.28–1.6). We could not estimate the effect of long-term douching due to sample size limitations. Reported ever use of both douche and talc was positively associated with uterine cancer, though the confidence interval was wide (HR=1.3, 95% CI: 0.79–2.1).
The stratified results were in line with the main findings (Table 3). A possible exception was the uterine cancer–talc association stratified by parity, where we observed a positive association among nulliparous women (HR=2.1, 95% CI: 1.3–3.5), and a null association among parous women (HR=1.0, 95% CI: 0.76–1.4; p-for-heterogeneity=0.02). Though confidence intervals were quite wide due to few non-white cases, we also observed possible racial difference in the douching–uterine cancer association (HR=0.86, 95% CI: 0.55–1.3 in non-Hispanic whites and HR=1.8, 95% CI: 0.88–3.7 in all other racial groups combined; p-for-heterogeneity=0.08). We did not observe differences by menopausal status.
TABLE 3.
Stratified hazard ratios for associations between talc use, douching, and uterine cancer in the Sister Study
| Number of cases | Douching | Talc | |||
|---|---|---|---|---|---|
| Ever HR (95% CI) | Last 12 months HR (95% CI) | Ever HR (95% CI) | Last 12 months HR (95% CI) | ||
| Time-varying menopausal status | |||||
| Premenopausal | 35 | 0.67 (0.24, 1.9) | 0.39 (0.10, 1.6) | 1.3 (0.64, 2.7) | 0.96 (0.38, 2.5) |
| Postmenopausal | 236 | 1.1 (0.75, 1.6) | 1.2 (0.83, 1.8) | 1.2 (0.92, 1.6) | 1.2 (0.82, 1.6) |
| Heterogeneity p-value | 0.39 | 0.13 | 0.85 | 0.72 | |
| Race | |||||
| Non-Hispanic White | 240 | 0.86 (0.55, 1.3) | 0.92 (0.58, 1.5) | 1.2 (0.90, 1.6) | 1.0 (0.72, 1.5) |
| Other | 31 | 1.8 (0.88, 3.7) | 1.8 (0.89, 3.8) | 1.5 (0.74, 3.1) | 2.0 (0.93, 4.4) |
| Heterogeneity p-value | 0.08 | 0.11 | 0.53 | 0.12 | |
| Parity | |||||
| Nulliparous | 61 | 0.57 (0.23, 1.4) | 0.54 (0.20, 1.5) | 2.1 (1.3, 3.5) | 1.8 (1.0, 3.3) |
| Parous | 210 | 1.2 (0.79, 1.7) | 1.3 (0.85, 1.9) | 1.0 (0.76, 1.4) | 0.96 (0.66, 1.4) |
| Heterogeneity p-value | 0.16 | 0.12 | 0.02 | 0.08 | |
| Body mass index (BMI) | |||||
| Non-obese (<30 kg/m2) | 149 | 1.0 (0.63, 1.7) | 1.2 (0.71, 2.0) | 1.4 (0.95, 1.9) | 1.1 (0.71, 1.9) |
| Obese (≥30 kg/m2) | 122 | 1.0 (0.63, 1.7) | 1.0 (0.62, 1.8) | 1.1 (0.79, 1.7) | 1.2 (0.78, 1.8) |
| Heterogeneity p-value | 1.00 | 0.70 | 0.52 | 0.90 | |
| Ever unopposed estrogen hormone therapy | |||||
| Yes | 32 | 0.82 (0.29, 2.3) | 0.99 (0.35, 2.8) | 1.1 (0.51, 2.3) | 0.77 (0.27, 2.2) |
| No | 239 | 1.1 (0.72, 1.5) | 1.1 (0.74, 1.7) | 1.2 (0.94, 1.6) | 1.2 (0.84, 1.6) |
| Heterogeneity p-value | 0.67 | 0.84 | 0.75 | 0.45 | |
Adjusted for race/ethnicity, education, body mass index (restricted cubic spline), menopausal status at baseline, parity, duration of oral contraceptive use, hormone therapy use, smoking status, alcohol use, and age at menarche. The factor being stratified on was omitted from the adjustment set for that analysis.
Sensitivity analyses did not reveal any major departures from our main results (Table 4). Importantly, when we only included cases that had been medically confirmed, the effect estimates were similar to the original estimates that included all self-reported cases (HR=1.0, 95% CI: 0.67–1.5 for douching, and HR=1.2, 95% CI: 0.88–1.6 for talc use). Effect estimates were again similar for medically confirmed invasive endometrial cancers (HR=0.97, 95% CI: 0.62–1.5 for douching, HR=1.2, 95% CI: 0.84–1.6 for talc), but the effect estimate for talc was null when we further restricted to invasive endometroid type tumors (HR=1.0, 95% CI: 0.69–1.4). The hazard ratio for the association between regular talc use and uterine cancer in postmenopausal women was 1.2 (95% CI: 0.75–1.9).
Table 4:
Sensitivity Analyses: Talc, douching, and risk of uterine cancer
| Number of cases | Ever Use HR (95% CI) | Use in Last Year HR (95% CI) | |
|---|---|---|---|
| Douching | |||
| Original estimate | 267 | 1.0 (0.72, 1.5) | 1.1 (0.75, 1.6) |
| Adjusting for talc use | 264 | 1.0 (0.71, 1.5) | 1.1 (0.73, 1.6) |
| Substitute waist circumferencea for BMI | 267 | 1.0 (0.72, 1.5) | 1.1 (0.76, 1.6) |
| Substitute weighta for of BMI | 267 | 1.0 (0.72, 1.5) | 1.1 (0.76, 1.6) |
| Excluding first year of follow-up | 240 | 1.1 (0.73, 1.5) | 1.1 (0.75, 1.7) |
| Adjusting for tamoxifen use | 267 | 1.0 (0.72, 1.5) | 1.1 (0.75, 1.6) |
| Medical record-confirmed, all | 202 | 1.0 (0.67, 1.5) | 1.1 (0.73, 1.7) |
| Invasive endometrialb | 185 | 0.97 (0.62, 1.5) | 1.1 (0.67, 1.7) |
| Endometroid adenocarcinomac | 144 | 0.89 (0.54, 1.5) | 1.0 (0.55, 1.6) |
| Talc use | |||
| Original estimate | 268 | 1.2 (0.94, 1.6) | 1.1 (0.82, 1.6) |
| Adjusting for douching | 264 | 1.2 (0.93, 1.6) | 1.1 (0.82, 1.6) |
| Substitute waist circumferencea for BMI | 268 | 1.2 (0.95, 1.6) | 1.2 (0.84, 1.6) |
| Substitute weighta for BMI | 268 | 1.2 (0.95, 1.6) | 1.2 (0.84, 1.6) |
| Excluding first year of follow-up | 243 | 1.3 (0.98, 1.7) | 1.1 (0.81, 1.6) |
| Adjusting for tamoxifen use | 267 | 1.2 (0.93, 1.6) | 1.1 (0.83, 1.6) |
| Medical record-confirmed, all | 204 | 1.2 (0.88, 1.6) | 1.1 (0.76, 1.6) |
| Invasive endometrialb | 187 | 1.2 (0.84, 1.6) | 1.1 (0.74, 1.6) |
| Endometroid adenocarcinomac | 145 | 1.0 (0.69, 1.4) | 1.0 (0.65, 1.6) |
BMI indicates body mass index. Adjusted for race/ethnicity, education, body mass index (restricted cubic spline), menopausal status at baseline, parity, duration of oral contraceptive use, hormone therapy use, self-reported history of fibroids, smoking status, alcohol use, and age at menarche.
As a restricted cubic spline
ICD-10 code of C54.1 = malignant neoplasm of the endometrium
A subset of the invasive endometrial tumors (ICD-0–3 histology codes: 8380, 8570, 8282, 8382)
Discussion
We observed no association between uterine cancer and ever use of douche, but saw a positive association between ever talc use and risk of uterine cancer. For the talc-uterine cancer association, there was some evidence of a dose–response for frequency of use and of effect measure modification by parity. Our findings were robust to the inclusion of unvalidated cases or non-endometrial cases, though there was no association between talc use and medically confirmed endometroid adenocarcinomas.
To our knowledge, only two previous prospective studies assessed the relationship between perineal talc use and uterine cancer.22,23 Although both were large, well-conducted cohort studies, uterine cancer is difficult to study due to its rarity, and we hoped that a third study could provide additional insight. In the Nurses’ Health Study, Karageorgi et al. 23 reported a similar elevated association between ever talc use and type I invasive endometroid adenocarcinomas (defined as ICD-Oncology-3 histology code 8380–83; HR=1.13, 95% CI: 0.96–1.33). They observed a stronger association among postmenopausal women (HR=1.21, 95% CI: 1.02–1.44), particularly among regular, postmenopausal talc users (HR=1.24, 95% CI: 1.03–1.48 for regular versus non-regular use). This was similar to our estimated HRs. In a Women’s Health Initiative study, the association between endometrial cancer (as adjudicated by a physician panel, without reference to specific ICD codes) and ever use of perineal talc was also weakly positive (HR=1.06, 95% CI: 0.87–1.28),22 with no difference seen for those reporting >20 years of any talc use, but an elevated risk of endometrial cancer among those who used powder on a diaphragm for >20 years (HR=3.06, 95% CI: 2.00–4.70). We did not separately ask about use of powder on a diaphragm.
As with the proposed link to ovarian cancer, possible mechanistic links between talc use and uterine cancer include asbestos contamination and inflammatory response.2 Early talc products may have contained asbestos, an established carcinogen29 that is sometimes mined in the same locations as talc. Although cosmetic talc producers agreed to an asbestos ban in 1976,30 recent evidence indicates that some contamination may still be present.31 Talc on its own could also potentially facilitate carcinogenesis by irritating and inflaming the uterus and other organs in its path up the reproductive tract.3–5
We previously reported a positive association between douching in the year prior to enrollment and ovarian cancer risk,21 but know of no studies that have examined the relationship between douching and uterine cancer. There is some evidence supporting a link between douching and pelvic inflammatory disease,8,32 a possible risk factor for uterine cancer.33 We previously suggested21 that the phthalates found in douching products12 could contribute to risk of hormonally driven cancers, including ovarian and uterine, via endocrine disruption. We consider this to be biologically plausible and note that the plastic tubing and bladder of the douching apparatus could still expose women to phthalates even if they douche using water, vinegar, or other phthalate-free products.6,8,9 Another possible carcinogenic mechanism is through changes in the microbiome.34 Previous reports have demonstrated associations between douching and vaginal microbiota,9,11,35 which has been shown to be highly correlated with the uterine microbiota.34
The most relevant exposure window for uterine cancer development is unclear, but the latency period could be quite long. While we asked about use at ages 10–13 and in the last year, we are missing information on use in the intermediate period and likely do a particularly poor job of capturing long-term use. This misclassification may be worse for women who were post-menopausal at baseline, especially if they also experienced menarche after age 13, as use of talc and douching products can be connected to menstruation. Further, because early menarche and late menopause are both risk factors for uterine cancer, it may be that cases who used talc or douche are more likely to have their exposure captured than non-cases. However, without detailed validation data, the degree and direction of the bias that would result from this misclassification is unclear. Another major limitation is that we did not ask about specific talc or douche products.
Our results may have limited generalizability due to the design of the Sister Study. Though there is no established link between breast cancer family history and uterine cancer risk,36 the women in our study may be systematically different from the general population in terms of genetic or behavioral risk factors common to both diseases. Further, the Sister Study sample is predominantly non-Hispanic white and well-educated, and thus may be less likely to douche or use talc than the general population (see supplement, Gonzalez et al.21). This disparity leaves open the possibility of residual confounding due to as-yet-unknown shared causes of uterine cancer and of talc use or douching among minority populations.
Because we did not have medical records with which to validate all reported cases, we have included all self-reported cases. Although the positive predictive value was high (88%), this may have led to some misclassification in women who confused uterine cancers with reproductive system cancers (e.g., ovaries or cervix), metastases from other cancers, or benign uterine conditions (e.g. fibroids). However, sensitivity analyses restricted to medically confirmed cases showed no evidence of bias.
Strengths of this analysis include its prospective design, large sample size, and the inclusion of data on both talc and douching, as well as a number of important confounders. Only 1.5% of individuals were excluded due to incomplete data on covariates and a multiple imputation analysis of the main findings produced results nearly identical to the complete case analysis (HR=1.0, 95% CI: 0.71–1.4 for ever douching; HR= 1.2, 95% CI: 0.93–1.6 for ever talc use). As mentioned previously, case–control studies may be prone to recall bias, or a tendency to over-report exposure to perceived risk factors among cases after they are diagnosed. This type of bias is particularly problematic for an exposure like talc, which has recently been the subject of several large and well-publicized lawsuits.15–17 A prospective study design such as ours circumvents this limitation by obtaining exposure histories before cases have been diagnosed.
Neither of the previous prospective studies of the association between talc and uterine cancer asked participants about douching. Although we did not observe any evidence of synergy or confounding between talc use and douching, the two behaviors are correlated and we believe that it was important to evaluate their potential relationship. We were also able to examine the importance of many other potential confounders or effect modifiers. This included BMI, which was based on the measurements that an examiner took during an in-home visit and modeled flexibly using a spline to limit the potential for residual confounding. Sensitivity analyses that adjusted for waist circumference or weight instead of BMI produced similar results.
The association we observed between self-reported talc use and uterine cancer incidence is consistent with previous prospective cohort studies of endometrial cancer and the combined data could indicate a small but true biologic association. Though we saw no association between douching and uterine cancer, we believe this relationship may also merit further consideration given recent findings on the potentially harmful effects of douching on the female reproductive system and the fact that, to our knowledge, we were the first to investigate this association in a human population.
Acknowledgments
The authors would like to thank Drs. Kristen Upson and Mary Diaz Santana for their helpful comments on an earlier draft of this manuscript.
Funding: This work was supported by the Intramural Research Program of the National Institutes of Health, National Institute of Environmental Health Sciences (Z01-ES044005 to DPS and Z01-ES102245 to CRW).
Footnotes
Conflict of Interest: The authors declare they have no conflicts of interests.
Data requests: Requests for de-identified Sister Study data, including the data used in this manuscript, can be requested through the study website (https://sisterstudy.niehs.nih.gov/English/data-requests.htm). The Sister Study is an ongoing prospective study, and the data sharing policy was developed to protect the privacy of study participants. It is consistent with study informed consent documents as approved by the NIEHS Institutional Review Board.
References
- 1.Rosenblatt KA, Mathews WA, Daling JR, Voigt LF, Malone K. Characteristics of women who use perineal powders. Obs Gynecol. 1998;92(5):753–756. [DOI] [PubMed] [Google Scholar]
- 2.Ness RB, Grisso JA, Cottreau C, et al. Factors Related to Inflammation of the Ovarian Epithelium and Risk of Ovarian Cancer. Epidemiology. 2016;11(2):111–117. [DOI] [PubMed] [Google Scholar]
- 3.Ness RB, Cottreau C. Possible Role of Ovarian Epithelial Inflammation in Ovarian Cancer. 1999;91(17):1459–1467. [DOI] [PubMed] [Google Scholar]
- 4.Dossus L, Lukanova A, Rinaldi S, et al. Hormonal, metabolic, and inflammatory profiles and endometrial cancer risk within the EPIC cohort - A factor analysis. Am J Epidemiol. 2013;177(8):787–799. doi: 10.1093/aje/kws309 [DOI] [PubMed] [Google Scholar]
- 5.Modugno F, Ness RB, Chen C, Weiss NS. Inflammation and endometrial cancer: A hypothesis. Cancer Epidemiol Biomarkers Prev. 2005;14(12):2840–2847. doi: 10.1158/1055-9965.EPI-05-0493 [DOI] [PubMed] [Google Scholar]
- 6.Upson K, Chin HB, Baird DD. Vaginal Douching in a Cohort of Young, African American Women; 2018. [Google Scholar]
- 7.Brown JM, Poirot E, Hess KL, Brown S, Vertucci M, Hezareh M. Motivations for Intravaginal Product Use among a Cohort of Women in Los Angeles. PLoS One. 2016:1–12. doi: 10.1371/journal.pone.0151378 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Forrest KA, Washington AE, Daling JR, Sweet RL. Vaginal Douching As a Possible Risk Factor for Pelvic Inflammatory Disease. Natl Med Assoc. 81(2):159–165. [PMC free article] [PubMed] [Google Scholar]
- 9.Brotman RM, Ghanem KG, Klebanoff MA, Taha TE, Scharfstein DO, Zenilman JM. The effect of vaginal douching cessation on bacterial vaginosis: a pilot study. Am J Obstet Gynecol. 2008;198(6). doi: 10.1016/j.ajog.2007.11.043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Aslan E, Bechelaghem N. To ‘douche’ or not to ‘douche’: hygiene habits may have detrimental effects on vaginal microbiota. J Obstet Gynaecol (Lahore). 2018;0(0):1–4. doi: 10.1080/01443615.2017.1395398 [DOI] [PubMed] [Google Scholar]
- 11.Brotman RM, Klebanoff MA, Nansel TR, et al. A longitudinal study of vaginal douching and bacterial vaginosis - A marginal structural modeling analysis. Am J Epidemiol. 2008;168(2):188–196. doi: 10.1093/aje/kwn103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Branch F, Woodruff TJ, Mitro SD, Zota AR. Vaginal douching and racial/ethnic disparities in phthalates exposures among reproductive-aged women: National Health and Nutrition Examination Survey 2001–2004. Env Heal. 2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gibson DA, Saunders PT. Endocrine disruption of oestrogen action and female reproductive tract cancers. Endocr Relat Cancer. 2014;21(2):T13–31. doi: 10.1530/erc-13-0342 [DOI] [PubMed] [Google Scholar]
- 14.Penninkilampi R, Eslick GD. Perineal Talc Use and Ovarian Cancer. Epidemiology. 2018;29(1):41–49. doi: 10.1097/EDE.0000000000000745 [DOI] [PubMed] [Google Scholar]
- 15.Trabert B Body power and ovarian cancer risk - what is the role of recall bias? Cancer Epidemiol Biomarkers Prev. 2016;25(10):1369–1370. doi: 10.1158/1055-9965.EPI-16-0476 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.McGinley L Does talcum powder cause ovarian cancer? Washington Post; https://www.washingtonpost.com/news/to-your-health/wp/2017/08/23/does-talcum-powder-cause-ovarian-cancer-experts-are-divided/?utm_term=.9342d83c278f.Published 2017. [Google Scholar]
- 17.Hsu T Johnson & Johnson Told to Pay $4.7 Billion in Baby Powder Lawsuit. New York Times; https://www.nytimes.com/2018/07/12/business/johnson-johnson-talcum-powder.html.Published 2018. [Google Scholar]
- 18.Berfield S, Feeley J, Fisk MC. Johnson & Johnson has a baby powder problem. Bloomberg Businessweek; https://www.bloomberg.com/features/2016-baby-powder-cancer-lawsuits/. Published March 31, 2016. [Google Scholar]
- 19.Houghton SC, Reeves KW, Hankinson SE, et al. Perineal powder use and risk of ovarian cancer. J Natl Cancer Inst. 2014;106(9). doi: 10.1093/jnci/dju208 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gertig DM, Hunter DJ, Cramer DW, et al. Prospective Study of Talc Use and Ovarian Cancer. J Natl Cancer Inst. 2000;92(3):249–252. [DOI] [PubMed] [Google Scholar]
- 21.Gonzalez N, O’Brien KM, D’Aloisio AA, Sandler DP, Weinberg CR. Douching, talc use, and risk of ovarian cancer. Epidemiology. 2017;28(2):797–802. doi: 10.1097/EDE.0000000000000528 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Crawford L, Reeves KW, Luisi N, Balasubramanian R, Sturgeon SR. Perineal powder use and risk of endometrial cancer in postmenopausal women. Cancer Causes Control. 2012;23(10):1673–1680. doi: 10.1007/s10552-012-0046-3 [DOI] [PubMed] [Google Scholar]
- 23.Karageorgi S, Gates MA, Hankinson SE, De Vivo I. Perineal Use of Talcum Powder and Endometrial Cancer Risk. Cancer Epidemiol Biomarkers Prev. 2010;19(5):1269–1275. doi: 10.1158/1055-9965.EPI-09-1221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Neill AS, Nagle CM, Spurdle AB, Webb PM. Use of talcum powder and endometrial cancer risk. Cancer Causes Control. 2012;23(3):513–519. doi: 10.1007/s10552-011-9894-5 [DOI] [PubMed] [Google Scholar]
- 25.Sandler DP, Hodgson ME, Deming-Halverson SL, et al. The Sister Study: Baseline methods and participant characteristics. Env Heal Perspect. 2017;125(12):127003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.International Agency for Research on Cancer and World Health Organization. International Classification of Diseases for Oncology (ICD-0–3 online). http://codes.iarc.fr/topography/C54. Published 2018. Accessed November 1, 2018. [Google Scholar]
- 27.Lin DY, Wei LJ. The Robust Inference for the Cox Proportonal Hazards Model. J Am Stat Assoc. 1989;84(408):1074–1078. [Google Scholar]
- 28.Lau B, Cole SR, Gange SJ. Competing risk regression models for epidemiologic data. Am J Epidemiol. 2009;170(2):244–256. doi: 10.1093/aje/kwp107 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.International Agency for Research on Cancer. IARC Monographs: Arsenic, Metals, Fibres, and Dusts. Vol 100C; 2012. [Google Scholar]
- 30.Fiume MM, Boyer I, Bergfeld WF, et al. Safety Assessment of Talc as Used in Cosmetics. Int J Toxicol. 2015;34:66S–129S. doi: 10.1177/1091581815586797 [DOI] [PubMed] [Google Scholar]
- 31.Rabin RC, Hsu T. Johnson & Johnson Feared Baby Powder’s Possible Asbestos Link for Years. New York Times; https://www.nytimes.com/2018/12/14/business/baby-powder-asbestos-johnson-johnson.html. Published December 14, 2018. [Google Scholar]
- 32.Zhang J, Thomas AG, Leybovich E. Vaginal Douching and Adverse Health Effects : A Meta-Analysis. Am J Public Health. 1997;87(7):1207–1211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Yang T-K, Chung C-J, Chung S-D, Muo C-H, Chang C-H, Huang C-Y. Risk of Endometrial Cancer in Women With Pelvic Inflammatory Disease. Medicine (Baltimore). 2015;94(34):e1278. doi: 10.1097/MD.0000000000001278 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Walther-Antonio MRS, Chen J, Multinu F, et al. Potential contribution of the uterine microbiome in the development of endometrial cancer. Genome Med. 2016;8(1):1–15. doi: 10.1186/s13073-016-0368-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Yeoman CJ, Thomas SM, Miller MEB, et al. A Multi-Omic Systems-Based Approach Reveals Metabolic Markers of Bacterial Vaginosis and Insight into the Disease. PLoS One. 2013;8(2). doi: 10.1371/journal.pone.0056111 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Kazerouni N, Schairer C, Friedman H, et al. Family history of breast cancer as a determinant of the risk of developing endometrial cancer: a nationwide cohort study. J Med Genet. 2002;39:826–832. doi: 10.1136/jmg.2005.034710 [DOI] [PMC free article] [PubMed] [Google Scholar]