Abstract
When powder is applied to the genital area, it has the potential to reach internal reproductive organs and promote carcinogenesis by irritating and inflaming exposed tissues. While many studies have considered the association between genital powder use and ovarian cancer risk, the relationship between genital powder use and uterine cancer is less well-studied. We pooled data from four large, prospective cohorts (the Nurses’ Health Study, the Nurses’ Health Study II, the Sister Study, and the Women’s Health Initiative – Observational Study). We used Cox proportional hazards models to estimate hazard ratios (HRs) and 95% confidence intervals (CI), adjusting for pre-specified confounders. In total, 209,185 women were included, with 37% reporting ever genital powder use. Over a mean 14.5 years of follow-up, 3,272 invasive uterine cancers were diagnosed. There was no overall association between ever genital powder use and uterine cancer (HR=1.01, 95% CI: 0.94-1.09), with little difference observed for frequent (≥1 times/week) versus never use (HR=1.05, 95% CI: 0.95-1.16; p-for-trend=0.46). Long-term use (>20 years; HR=1.12, 95% CI: 0.96-1.31; p-for-trend=0.14) was associated with a small, but not statistically significant, increase in risk, compared to never use. There were not clear differences by uterine cancer histologic subtypes or across strata of relevant covariates, including race/ethnicity, follow-up time, menopausal status and body mass index. The results of this large, pooled analysis do not support a relationship between use of genital powder and uterine cancer, though the positive associations observed for long-term use may merit further consideration.
Introduction
Some women apply powder to their genital area, including underwear, sanitary napkins, tampons, and diaphragms, to help absorb moisture, reduce odor, and prevent irritation.1 The powder can enter the vagina and travel up the reproductive tract where it reaches the cervix, uterus, fallopian tubes and ovaries.2 It has been hypothesized that the powder, which often includes some talc,3 may irritate and inflame exposed tissues, with the potential to promote a carcinogenic response.4,5 Further, some powder users may also have been exposed to asbestos,6-8 a natural mineral and established carcinogen9 sometimes present in talc.
Studies of the association between powder use and ovarian cancer risk have provided some support for the existence of a link between genital powder exposure and increased cancer risk. Several ovarian cancer case-control studies reported positive associations,10,11 and our recent pooled analysis of prospective studies showed a positive association among women with no prior hysterectomy or tubal ligation.12
Uterine cancers, most of which are endometrial cancers, are the most common gynecologic malignancies in the United States (US).13 The uterus is closer to the application site than the fallopian tubes or ovaries, and therefore could be more highly exposed. On the other hand, detrimental effects of powder in the uterus could be mitigated by menstruation, which causes the uterine lining (endometrium) to shed and regenerate. Relatively few studies have examined the association between genital powder use and uterine cancer risk. We identified four,14-17 including three prospective cohort studies (the Nurses’ Health Study [NHS],14 the Women’s Health Initiative – Observational Study [WHI-OS],15 and the Sister Study [SIS]16) and one retrospective Australian case-control study.17 Though none met standard criteria for statistical significance, all of the prospective studies reported small, positive associations (hazard ratios [HRs] ranging from 1.06-1.2).
To address this research question with increased statistical power, we combined updated data (with extended follow-up) from the three previously published cohort studies (NHS, WHI-OS and SIS) with newly released data from a fourth cohort (the Nurses’ Health Study II [NHSII]). We examined self-reported ever use as well as frequency and duration of use. We additionally considered cancer subtypes and examined how associations differed across subgroups.
Methods
Study sample
Information about the four included cohorts has been published previously.18-20 All participants provided informed consent. Briefly, NHS and NHSII recruited volunteers from among registered nurses throughout the US. NHS enrolled 121,529 women in 1976, but as genital powder use was not assessed until 1982, we only considered person-time and cases accrued 1982-2016. NHSII enrolled 116,429 women in 1989 and assessed genital powder use in 2013; our follow-up included 2013-2017. Both studies have conducted detailed follow-ups every 2 years and were approved by the institutional review boards of the Brigham and Women’s Hospital, the Harvard T.H. Chan School of Public Health and participating state cancer registries, as required.
SIS includes 50,884 women who had a sister previously diagnosed with breast cancer, but who had never been diagnosed themselves at baseline (2003-2009). Genital powder use was assessed at enrollment and follow-up continued annually through September 2018 (data release 8.1), with detailed questionnaires distributed every 2-3 years. SIS was approved by the institutional review boards of the Copernicus Group and the National Institute of Environmental Health Sciences.
WHI-OS enrolled 93,676 US women (1993-1998), with genital powder use assessed at baseline. All participants were postmenopausal and resided near one of forty clinical centers. Detailed follow-up surveys were administered every 1-2 years for the first 10 years, with our analysis also incorporating basic health information collected through February 2019. The protocol was approved by the institutional review boards of all participating institutions.
We excluded women with a history of uterine cancer or a hysterectomy, who did not answer questions about their genital powder use, or who did not have adequate follow-up data. After these exclusions, we had a pooled sample size of 209,185 (67,724 in NHS; 53,589 in NHSII; 33,837 in SIS; and 54,035 in WHI-OS; Table 1).
Table 1.
Description of Participating Cohorts
| Study | Sample size | Follow-up years (mean duration) |
Ages at baseline (mean) |
All uterine cancer cases |
Medically- confirmed uterine cancer cases |
% ever genital powder use |
Long-term genital powder use |
Frequent genital powder use |
|---|---|---|---|---|---|---|---|---|
| Nurses’ Health Studya | 67,724 | 1982-2016 (25.2) | 35-62 (47) | 1,675 | 1,331 | 41% | -- | 27% |
| Nurses’ Health Study IIb | 53,589 | 2013-2017 (3.6) | 48-68 (58) | 206 | 109 | 26% | 6% | 26% |
| Sister Studyc | 33,837 | 2003-2018 (10.0) | 35-76 (54) | 346 | 249 | 26% | 5% | 6% |
| Women’s Health Initiatived | 54,035 | 1993-2019 (14.7) | 49-81 (63) | 1,045 | 1,045 | 52% | 15% | -- |
| TOTAL | 209,185 | 1982-2019 (14.5) | 35-81 (55) | 3,272 | 2,734 | 37% | 9% | 22% |
Genital powder use assessed in 1982 follow-up questionnaire; participants excluded if they did not respond to genital powder questionnaire (n=28,581), had uterine cancer prior to 1982 (n=487), had a hysterectomy prior to 1982 questionnaire (n=24,736), or did not contribute any person-time after 1982 (n=1). Frequent use was defined as at least once per week. Women who had a hysterectomy or a diagnosis of in situ uterine cancer during follow-up were censored at the minimum of their age of hysterectomy or in situ diagnosis. Follow-up is complete through May 2016.
Genital powder use assessed in 2013 follow-up questionnaire; participants excluded if they did not respond to genital powder questionnaire (n=41,141), had uterine cancer prior to 2013 (n=706), had a hysterectomy prior to 2013 questionnaire (n=20,993). Long-term use was defined as use of powder in the genital area for 20 years or longer. Frequent use was defined as at least once per week. Women who had a hysterectomy during follow-up were censored at age of hysterectomy. Follow-up is complete through May 2016.
Participants excluded if they withdrew from the study (n=3), did not respond to the genital powder question (n=1,196), had uterine cancer prior to baseline or unclear uterine cancer status at baseline (n=449), had a hysterectomy before baseline (n=15,244), or did not contribute any person-time after enrollment (n=155). Ever genital powder use defined as use in 12 months prior to baseline or at ages 10-13. Participants were considered long-term users if they reported using at ages 10-13 and within last 12 months. Frequent use was defined as use at least once per week (in the last 12 months) or “frequently” at ages 10-13. Women who had a hysterectomy during follow-up were censored at age of hysterectomy. Follow-up is complete through September 2018.
Participants excluded if they did not complete the genital powder questionnaire (n=342), had uterine cancer before baseline (n=1,575), had a hysterectomy before baseline (n=37,473), or had no follow-up information (n=251). Participants were considered long-term users if they reported using genital powder for 20 or more years. Women who had a hysterectomy during follow-up were censored at age of hysterectomy and women with in situ uterine cancer were censored at their age of diagnosis. Follow-up is complete through February 2019.
Exposure assessment
The only exposure measure common to all four cohorts was ever (versus never) use of powder in the genital area. This included application to sanitary napkins, underwear, tampons, diaphragms, cervical caps, or directly to the vulvovaginal/rectal area. NHS and NHSII asked about “talcum, baby powder or deodorizing powder”, SIS asked about “talcum powder” and WHI-OS asked about “powder” generally (Supplement).
NHSII, SIS and WHI-OS collected data on duration of use. Long-term users were defined as women who reported >20 years of use (NHSII and WHI-OS) or those who reported use at ages 10-13 and in the year prior to enrollment (SIS). NHS, NHSII and SIS collected data on frequency of use. Frequent users were defined as those who used powder at least once per week (NHS, NHSII, SIS) or “frequently” when they were 10-13 (SIS).
Outcome assessment
We included all self-reported invasive uterine cancers, as medical confirmation could not be obtained for all cases (83% for NHS, 62% for NHSII, 70% for SIS, 100% for WHI-OS). As NHS, NHSII and SIS did not ask women to differentiate between endometrial and other uterine cancers, we considered them as a single outcome. Among women who self-reported a diagnosis and for whom we received medical documentation, the positive predictive value was approximately 80% (82% for NHS, 78% for NHSII, and 87% for SIS). Women who self-reported uterine cancer were not counted as cases if the diagnosis was later self-disavowed or shown to be inaccurate upon medical review. We ran sensitivity analyses restricted to medically-confirmed cases and conducted analyses of specific histologic subtypes (including epithelial endometroid, epithelial non-endometroid, and sarcoma21) only among those who were medically confirmed.
Covariates
All four studies had rich details on relevant covariates. These were harmonized as follows: self-identified race/ethnicity (non-Hispanic white, Black, other), education (≤high school, some college, ≥college graduate), body mass index (BMI in kg/m2; as a restricted cubic spline), parity (0, 1, 2, ≥3 births), smoking status (never, former, current), oral contraceptive use (never, ever), menopausal status (premenopausal, post-menopausal), menopausal hormone therapy use (none, any unopposed estrogen, other), age at menarche (continuous), bilateral oophorectomy (yes/no), and years since most recent birth (nulliparous or >20 versus ≤20 years ago).
Statistical analysis
We used study-stratified Cox proportional hazards models, with age as the time scale, to estimate HRs and 95% confidence intervals (CIs). Women were followed from age at powder assessment until age at invasive uterine cancer diagnosis, with censoring at age at death or age at end of follow-up. Those who had hysterectomies after baseline (other than for uterine cancer) were censored at age of hysterectomy, and those reporting incident in situ uterine cancer were censored at age of diagnosis.
We identified potential confounders using a directed acyclic graph approach,22 considering covariates that may be related to genital powder use and uterine cancer. We a priori hypothesized that BMI would be an especially important confounder, as it is both a strong predictor of genital powder use1,12 and a well-established risk factor for uterine cancer.23,24 To limit the possibility of residual confounding, we modeled this flexibly, using a restricted cubic spline. We assessed models adjusting for 1) age only; 2) age and BMI; and 3) age, BMI, race/ethnicity, education, parity, smoking status, oral contraceptive use, menopausal status, type of menopausal hormone therapy, age at menarche, and bilateral oophorectomy.
For all regression analyses, we excluded individuals missing one or more covariates (<5%). We tested for study heterogeneity via likelihood ratio tests comparing models with and without exposure-by-study interaction terms. We additionally calculated a heterogeneity p-value from a random effects meta-analysis.25 We tested the proportional hazards assumption with a likelihood ratio test comparing models with and without an exposure-by-time interaction term.
We estimated cumulative risk of uterine cancer by age 70 using the Breslow method to calculate the baseline hazard,26 assuming all covariates were set to their observed distributions. We additionally reported covariate-adjusted risk differences (RD) with 95% CIs.
When estimating the effects of duration or frequency of genital powder use on uterine cancer risk, we compared long-term (>20 years of use) or frequent (≥1 per week) users to never users, assigning shorter-term or infrequent users to a middle category. We tested for trend by entering the ordinal form of the variables into the regression models and reporting the p-value from a Wald test. For analyses comparing estimates across histologic subtypes we used joint Cox proportional hazards models,27 reporting likelihood ratio tests of exposure-by-subtype interaction p-values as tests of heterogeneity.28
Lastly, we conducted stratified analyses based on a priori selected covariates. These included race/ethnicity, age, menopausal status, follow-up time, BMI, parity, years since most recent birth, and menopausal hormone therapy use. For analyses of effect modification by age and menopausal status, we accounted for changes in participants’ status over time. Effect modification by time-varying BMI and menopausal hormone therapy use were also considered. To combine data on time-varying factors from studies with different follow-up intervals, we assessed each cohort as if the updates occurred every two years, carrying forward values when more timely information was unavailable.
Results
The eligible sample consisted of 209,185 women (Table 1). Of these, 3,272 were diagnosed with invasive uterine cancer over a mean of 14.5 years (median=11.0 years) of follow-up. For 2,734 cases (84%), medical documentation was received and the diagnosis confirmed. Thirty-seven percent reported ever using genital powder, with higher use in NHS (41%) or WHI-OS (52%) than NHSII (26%) or SIS (26%). In the pooled sample, 9% reported long-term use and 22% reported frequent use. The mean age at baseline was 55 years (range 35-81).
Most NHS and WHI-OS participants were born in the 1920s and 30s, while most NHSII and SIS participants were born after 1945 (Supplementary Table 1). All 4 cohorts were predominantly non-Hispanic white (93%; range 86-98%) and well-educated (93% with at least some college). At baseline, most women were parous (87%), never smokers (53%) and never-users of menopausal hormone therapy (68%). Average BMI in the pooled sample was 26.3 kg/m2, with 21% categorized as obese (BMI >30 kg/m2).
Ever and never genital powder users were approximately the same age at baseline (56.0 versus 55.5; Supplementary Table 2). Ever users had higher baseline BMI (27.1 versus 25.8 kg/m2). Compared to never users, ever users were also more likely to identify as Black (5% versus 3%), and to have less schooling (76% versus 81% completed college), and were less likely to have ever used oral contraceptives (59% versus 67%). Uterine cancer cases had slightly higher baseline BMI (mean=28.0 kg/m2) than non-cases (mean=26.3 kg/m2) and were less likely to have ever used oral contraceptives (47% versus 64%).
After excluding women missing data on one or more covariates, 199,483 participants remained. These included 3,162 incident uterine cancer cases over nearly 3 million person-years of follow-up (cumulative incidence=108/100,000 person-years). Cases were more likely to have been genital powder users (Table 2; 45%) than non-cases (37%; age-adjusted HR=1.13, 95% CI: 1.05-1.21), but no association remained after adjusting for BMI (HR=1.03, 95% CI: 0.95-1.10) or BMI and other covariates (HR=1.01, 95% CI: 0.94-1.09).
Table 2.
Pooled hazard ratios and 95% confidence intervals for the association between ever genital powder use and risk of uterine cancer
| Person-time at riska |
Non- casesa |
Uterine cancer casesa |
Prevalence of genital powder use Non-cases / cases |
Age-Adjusted Hazard Ratio (95% CI) |
Age- and BMI- Adjusted Hazard Ratio (95% CI) |
Adjusted Hazard Ratiob (95% CI) |
|
|---|---|---|---|---|---|---|---|
| Ever used genital powder | |||||||
| Nurses’ Health Study | 1,662,438 | 64,208 | 1,637 | 41% / 43% | 1.10 (0.99, 1.21) | 0.97 (0.88, 1.07) | 0.96 (0.87, 1.06) |
| Nurses’ Health Study II | 172,365 | 48,328 | 178 | 26% / 38% | 1.67 (1.23, 2.26) | 1.31 (0.96, 1.78) | 1.31 (0.96, 1.78) |
| Sister Study | 332,107 | 32,746 | 334 | 26% / 32% | 1.35 (1.07, 1.70) | 1.17 (0.93, 1.48) | 1.17 (0.93, 1.48) |
| Women’s Health Initiative | 770,399 | 51,039 | 1,013 | 52% / 54% | 1.06 (0.94, 1.20) | 1.01 (0.90, 1.15) | 0.99 (0.87, 1.12) |
| Pooled Estimate | 2,937,309 | 196,321 | 3,162 | 37% / 45% | 1.13 (1.05, 1.21) | 1.03 (0.95, 1.10) | 1.01 (0.94, 1.09)c |
| Long-term genital powder use | |||||||
| Nurses’ Health Study II | 172,365 | 48,328 | 178 | 6% / 10% | 1.98 (1.20, 3.25) | 1.48 (0.90, 2.46) | 1.47 (0.89, 2.44) |
| Sister Study | 332,107 | 32,746 | 334 | 5% / 7% | 1.46 (0.95, 2.24) | 1.18 (0.77, 1.82) | 1.18 (0.77, 1.82) |
| Women’s Health Initiative | 767,639 | 50,843 | 1,008 | 15% / 17% | 1.17 (0.98, 1.39) | 1.07 (0.90, 1.28) | 1.07 (0.90, 1.27) |
| Pooled Estimate | 1,272,112 | 131,917 | 1,520 | 9% / 14% | 1.28 (1.10, 1.49) | 1.13 (0.97, 1.32) | 1.12 (0.96, 1.31)c |
| p-trendd | 0.0009 | 0.10 | 0.14 | ||||
| Used genital powder at least ≥1/week | |||||||
| Nurses’ Health Study | 1,662,438 | 64,208 | 1,637 | 27% / 30% | 1.17 (1.05, 1.30) | 1.01 (0.90, 1.12) | 1.00 (0.90, 1.12) |
| Nurses’ Health Study II | 172,365 | 48,328 | 178 | 26% / 38% | 1.67 (1.23, 2.26) | 1.31 (0.96, 1.78) | 1.31 (0.96, 1.78) |
| Sister Study | 332,107 | 32,746 | 334 | 6% / 9% | 1.53 (1.05, 2.24) | 1.24 (0.84, 1.82) | 1.22 (0.83, 1.80) |
| Pooled Estimate | 2,166,911 | 145,282 | 2,149 | 22% / 28% | 1.24 (1.12, 1.37) | 1.05 (0.95, 1.16) | 1.05 (0.95, 1.16)c |
| p-trendd | <0.0001 | 0.40 | 0.46 |
Reference group is never users. All pooled models are stratified by study.
Limited to participants with complete covariate information. Includes all self-reported invasive uterine cancer cases.
Adjusted for BMI (as a restricted cubic spline), race/ethnicity (non-Hispanic white, Black, other), education (≤high school, some college, ≥college graduate), parity (0, 1, 2, 3+ births), smoking status (never, former, current), oral contraceptive use (never, ever), menopausal status (pre or post-menopausal), a menopause status by age at menopause interaction term, type of hormone therapy use (none, any unopposed estrogen, other), age at menarche (continuous), and bilateral oophorectomy (yes/no).
P-values for tests of between study heterogeneity were p=0.16 (ever use), p=0.40 (long-term use), p=0.14 (frequent use).
P-for-trend is a Wald test of the β coefficient when treating duration (never, non-long-term use, long term use) or frequency (never, non-frequent, frequent) of use as ordinal variables.
The estimated covariate-adjusted risk difference was 0.03% (95% CI: −0.14% to 0.19%), with approximately 2.3% of both exposed and unexposed women developing uterine cancer before age 70. Associations were positive, albeit imprecise, in the younger cohorts (fully adjusted NHSII HR=1.31, 95% CI: 0.96-1.78 and SIS HR=1.17, 95% CI: 0.93-1.48), but not in the older cohorts (NHS HR=0.96, 95% CI: 0.87-1.06 and WHI-OS HR=0.99, 95% CI: 0.87-1.12). Tests for between-study heterogeneity were not statistically significant (p=0.16 for the likelihood ratio test in the pooled sample and p=0.15 for the random effect meta-analysis).
In fully adjusted models, there was a small positive, but not statistically-significant association between long-term versus never genital powder use and uterine cancer (HR=1.12, 95% CI: 0.96-1.31; p-for-trend=0.14). The association between frequent versus never genital powder use and uterine cancer was weaker (HR=1.05, 95% CI: 0.95-1.16; p-for-trend=0.46). As with the analysis of ever versus never use, a comparison of the age- versus BMI-adjusted models indicated that there was likely strong confounding by BMI but not the other covariates.
Results limited to medically-confirmed cases were nearly identical to the main findings (Table 3), with a fully-adjusted HR=1.00 (95% CI: 0.93-1.09) for ever versus never use. Effect estimates were similar for epithelial endometroid tumors (HR=1.00, 95% CI: 0.91-1.09), which made up the largest subgroup, and there was no evidence of heterogeneity by histologic subtype (p=0.82). There was a positive association between long-term genital powder use and epithelial, non-endometroid type tumors (HR=1.46, 95% CI: 1.00-2.11; p-for-heterogeneity=0.41), based on 217 cases.
Table 3.
Pooled hazard ratios and 95% confidence intervals for medically-confirmed cases and by histologic subtype
| Casesa | Adjusted Hazard Ratiob (95% CI) |
Test for heterogeneity by type |
|
|---|---|---|---|
| Ever used genital powder | |||
| Medically-confirmed uterine cancer | 2,646 | 1.00 (0.93, 1.09) | 0.82 |
| Epithelial, endometrioidc | 2,141 | 1.00 (0.91, 1.09) | |
| Epithelial, non-endometroidd | 292 | 0.96 (0.76, 1.22) | |
| Sarcomase | 55 | 0.97 (0.56, 1.67) | |
| Otherf | 69 | 1.25 (0.77, 2.02) | |
| Long-term genital powder use | |||
| Medically-confirmed uterine cancer | 1,341 | 1.05 (0.89, 1.24) | 0.41 |
| Epithelial, endometrioidc | 1,046 | 0.99 (0.82, 1.19) | |
| Epithelial, non-endometroidd | 217 | 1.46 (1.00, 2.11) | |
| Sarcomase | 15 | 1.18 (0.31, 4.54) | |
| Otherf | 14 | 0.49 (0.06, 4.09) | |
| Frequent genital powder use | |||
| Medically-confirmed uterine cancer | 1,633 | 1.02 (0.91, 1.15) | 0.73 |
| Epithelial, endometrioidc | 1,350 | 1.01 (0.89, 1.15) | |
| Epithelial, non-endometroidd | 121 | 0.75 (0.47, 1.19) | |
| Sarcomase | 44 | 0.90 (0.44, 1.84) | |
| Otherf | 59 | 1.36 (0.76, 2.45) |
Medically-confirmed cases missing subtype information: 89 (ever use analysis), 14 (long-term use analysis), 59 (frequent use analysis).
Referent group is never genital powder users. Stratified by study and adjusted for race/ethnicity (non-Hispanic white, Black, other), education (≤high school, some college, ≥college graduate), BMI (as a restricted cubic spline), parity (0, 1, 2, 3+ births), smoking status (never, former, current), oral contraceptive use (never, ever), menopausal status (pre or post-menopausal), a menopause status by age at menopause interaction term, type of hormone therapy use (none, any unopposed estrogen, other), age at menarche (continuous), and bilateral oophorectomy (yes/no).
Includes endometroid adenocarcinomas, other adenocarcinomas, and adenosquamous cell carcinomas (histology codes 8380, 8140, 8570, 8480, 8260, 8050, 8262, 8560, 8210, 8382, 8481)
Includes serous adenocarcinoma, papillary serous carcinoma, clear cell carcinoma, carcinosarcoma, mixed cell adenocarcinoma, mullerian mixed tumor, mesodermal mixed tumor (histology codes 8441, 8460, 8310, 8980, 8323, 8461, 8950, 8951, 8255)
Includes leiomyosarcoma, endometrial stromal sarcoma, sarcoma NOS, epithelioid leiomyosarcoma, rhabdomyosarcoma, adenosarcoma (histology codes 8890, 8930, 8800, 8891, 8900, 8933).
Includes epithelial tumor, undifferentiated carcinoma, squamous cell carcinoma, sarcomatoid carcinoma, thymoma, neoplasm NOS, mixed tumor (histology code 8010, 8020, 8070, 8033, 8580, 8000, 8940)
In stratified models (Table 4), most factors did not modify the observed associations between ever genital powder use and risk of uterine cancer. However, even though the proportional hazards assumption was not violated (p=0.25), we did see differences in the HR over time, with a positive association observed for the first 10 years of follow-up (HR=1.12, 95% CI: 1.00-1.24), but not after (HR=0.93, 95% CI: 0.85-1.03; p-for-heterogeneity=0.02). The positive trend during the first 10 years of follow-up was also somewhat apparent in the analyses of long-term (Supplementary Table 3; p-for-heterogeneity=0.10) and frequent genital powder use (p-for-heterogeneity=0.13).
Table 4.
Pooled hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between ever genital powder use and uterine cancer within covariate strata
| Cases; n | HR (95% CI) | Heterogeneity p- valueb |
|
|---|---|---|---|
| Race/ethnicity | |||
| Non-Hispanic White | 3,005 | 1.01 (0.93, 1.08) | 0.81 |
| Black | 76 | 1.12 (0.70, 1.80) | |
| Other | 81 | 1.22 (0.76, 1.95) | |
| Agea | |||
| <60 years | 603 | 0.95 (0.80, 1.12) | 0.43 |
| ≥60 years | 2,559 | 1.03 (0.95, 1.11) | |
| Menopausal statusa | |||
| Premenopausal | 168 | 0.99 (0.72, 1.37) | 0.85 |
| Postmenopausal | 2,994 | 1.01 (0.94, 1.09) | |
| Follow-up time | |||
| 0-10 years | 1,481 | 1.12 (1.00, 1.24) | 0.02 |
| >10 years | 1,681 | 0.93 (0.85, 1.03) | |
| BMI at baseline | |||
| <25 kg/m2 | 1,300 | 1.01 (0.90, 1.13) | 0.75 |
| 25- <30 kg/m2 | 891 | 1.00 (0.87, 1.14) | |
| ≥30 kg/m2 | 971 | 1.07 (0.94, 1.21) | |
| Parity | |||
| Nulliparous | 413 | 1.16 (0.95, 1.42) | 0.16 |
| Parous | 2,749 | 1.00 (0.92, 1.08) | |
| Recency of birth | |||
| Nulliparous or >20 years | 2,086 | 1.05 (0.96, 1.15) | 0.10 |
| Gave birth ≤20 years ago | 967 | 0.92 (0.81, 1.04) | |
| Hormone therapy use | |||
| Never user | 2,102 | 0.97 (0.89, 1.06) | 0.03 |
| Any unopposed estrogen | 389 | 1.22 (0.99, 1.50) | |
| Other hormone therapy | 671 | 1.04 (0.89, 1.21) |
Hazard ratios are stratified by study and adjusted for race/ethnicity (non-Hispanic white, Black, other), education (≤high school, some college, ≥college graduate), BMI (as a restricted cubic spline), parity (0, 1, 2, 3+ births), smoking status (never, former, current), oral contraceptive use (never, ever), menopausal status (pre or post-menopausal), a menopause status by age at menopause interaction term, type of hormone therapy use (none, any unopposed estrogen, only estrogen plus progestin, age at menarche (continuous), and bilateral oophorectomy (yes/no). Stratification covariate not included in adjustment set.
Age/Menopause status allowed to change over time between baseline and end of follow-up. Number of cases is based on status at time of diagnosis.
Likelihood ratio test for interaction term of powder use and the stratified covariate.
Additionally, there was some evidence of interaction between ever powder use and menopausal hormone therapy, with an elevated HR among users of unopposed estrogen (Table 4; HR=1.22, 95% CI: 0.99-1.50; p-for-heterogeneity=0.03). This pattern was no longer present in an analysis that considered changes in hormone therapy use after baseline (Supplementary Table 4). As unopposed estrogen use is an established risk factor for uterine cancer,29 it is contraindicated in women with a uterus. In our sample, unopposed estrogen users tended to be older at enrollment, from earlier birth cohorts, and to have had earlier menopause than non-users (Supplementary Table 5).
Discussion
In this pooled analysis of four large, prospective cohorts, we did not observe an association between use of powder in the genital area and uterine cancer risk. The association did not increase among frequent users, though we did observe a slightly elevated association between long-term genital powder use and uterine cancer risk. There was no heterogeneity by tumor subtype or by most of the potential effect modifiers we considered.
The three previous prospective studies of this topic all reported positive, but not statistically significant associations (HR range = 1.06-1.2).14-16 Rather than meta-analyzing these results, we amplified our statistical power by pooling data from the three previously-published studies plus NHSII, adding follow-up time and uterine cancer events.
The additional follow-up data brought higher precision, though changes in the HR over time make the interpretation complicated. One possible explanation for the positive association seen between genital powder use and uterine cancer during the first 10 years is that the lag time between exposure and uterine cancer development is short, and thus recent exposure is more relevant than exposures that occurred long ago. Because none of the included studies asked about powder use after baseline, we could not assess changes over time. Presumably, exposure misclassification increased with longer follow-up. This is common in any study with a one-time measure of exposure, but may be more extreme here to due to the link between powder use and menstruation, which becomes less common as the cohorts age.
Consistent with the observation that the HR changed over time, we observed some between-study heterogeneity, with positive associations observed in the younger (SIS and NHSII), but not in the older (NHS and WHI-OS) cohorts. However, the relative directions of these study-specific effect estimates are not consistent with the previously stated hypothesis that asbestos contamination is the underlying link between talc use and cancer risk.30-32 More specifically, while there is no official regulation of talc-based products and some asbestos contamination may still be occurring,7 cosmetic talc manufacturers agreed to eliminate asbestos from their products beginning in 1976.8 Therefore, it seems more likely that older participants were exposed to asbestos-contaminated talc than younger participants. As such, if asbestos caused uterine cancer we would have expected to see increased risk among NHS and WHI-OS participants (rather than NHSII and SIS).
As highlighted above, a major limitation of this analysis was that the studies assessed powder use at a single time point, and some lacked data on duration or frequency of use. Additionally, we could not differentiate between types of powder used, or identify what chemicals they contained. Because medical confirmation was inconsistent, we relied on self-reported uterine cancer diagnosis. While we think accuracy is high, some outcome misclassification was certainly present, particularly as women may have difficulty distinguishing between uterine cancers and other gynecological cancers or non-invasive conditions.
Another potential limitation was lack of generalizability. The included women were predominately non-Hispanic white and well-educated, and thus the effects estimated here may not be representative of the general US population. Though we are not aware of strong confounders that were omitted, there may be residual confounding. Confounding by indication could have occurred if women who had conditions related to uterine cancer differentially used genital powder. Lastly, estimates could also be biased if women who used genital powder were more likely to have other conditions (e.g. uterine fibroids) that lead them to get hysterectomies.
Strengths of this analysis include its prospective design, detailed adjustment for potential confounders, and large sample size. Although the putative association between genital powder use and ovarian cancer is more widely publicized,33 recall bias34,35 may also exist among uterine cancer survivors. A prospective design is crucial for ruling out that possibility. The four included studies collected detailed information on important covariates, which we harmonized. The large sample size allowed us to obtain precise effect estimates and evaluate associations between genital powder use and histologic subtypes of uterine cancers.
This large, pooled analysis brings together data from four ongoing US cohorts. To the best of our knowledge, these are the only four prospective studies to collect data on genital powder use. We did not observe an association between uterine cancer and ever use of powder in the genital area.
Supplementary Material
What’s new?
When powder is applied to the genital area, it may promote carcinogenesis by irritating and inflaming exposed tissues. We pooled data from four large prospective studies, including 209,185 women, to examine whether genital powder use was associated with uterine cancer (3,272 cases). Thirty-seven percent of women reported ever using genital powder. Genital powder use was not associated with the risk of developing uterine cancer, though we observed a possible increase in risk among long-term users.
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); The Intramural Research Program of the National Cancer Institute; Department of Defense Ovarian Cancer Research Program Grant No. W81XWH-12-1-0561; U01 CA186107, P01 CA87969, UM1 CA176726, and R01 CA67262 (Nurses’ Health Study, Nurses’ Health Study II); and National Heart, Lung, and Blood Institute, NIH/DHHS through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C (Women’s Health Initiative). H.R.H. is supported by NIH K22 CA193860. The authors would like to thank the participants and staff of the NHS and NHSII for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full responsibility for analyses and interpretation of these data.
In all cases, the funders had no role in the design or conduct of the study; the collection, management, or analysis of data; the interpretation of the data; or the preparation, review, or approval of the manuscript. They were not included in the decision to submit the manuscript for publication.
Abbreviation:
- BMI
body mass index
- CI
confidence interval
- HR
hazard ratio
- NHS
Nurses’ Health Study
- NHSII
Nurses’ Health Study II
- SIS
Sister Study
- US
United States
- WHI-OS
Women’s Health Initiative Observational Study
Footnotes
Ethics Statement
Nurses’ Health Study and Nurses’ Health Study II were approved by the institutional review boards of the Brigham and Women’s Hospital, the Harvard T.H. Chan School of Public Health and participating state cancer registries, as required. The Sister Study was approved by the institutional review boards of the Copernicus Group and the National Institute of Environmental Health Sciences. The Women’s Health Initiative Observational Study protocol was approved by the institutional review boards of all participating institutions. All participants provided informed consent.
Conflict of Interest
No conflicts related to genital powder use or uterine cancer. AMK receives payment as a consultant for Pfizer (injectable contraception), AMAG (treatment of genital atrophy), Mithra (contraceptive and menopausal hormone products), and Merck (implantable and vaginal ring contraception). AMK also receives royalties from UpToDate and funding for clinical trials through the University of Florida from Medicines 360 (intrauterine devices), Allergan (treatment of uterine fibroids), Myovant (treatment of uterine fibroids), and Endoceutics (treatment of genital atrophy).
Data availability statement
The data that support the findings of this study are available for request from the individual cohort studies. Please refer to the following links for more information: Nurses’ Health Study and Nurses’ Health Study II (https://www.nurseshealthstudy.org/researchers), Sister Study (https://sisterstudy.niehs.nih.gov/English/data-requests.htm), and Women’s Health Initiative – Observational Study (https://www.whi.org/page/working-with-whi-data). Further details available on request to the corresponding author.
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.Henderson WJ, Joslin CAF, Turnbull AC, Griffiths K, Joslin AF, Griffiths K. Talc and carcinoma of the ovary and cervix. J Obs Gynaecol Br Commonw. 1971;78(3):266–272. [DOI] [PubMed] [Google Scholar]
- 3.World Health Organization International Agency for Research on Cancer. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Carbon Black, Titanium Dioxide, and Talc. Vol 93.; 2010. [PMC free article] [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.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]
- 7.Tran T, Steffen JE, Clancy KM, Bird T, Egilman DS. Talc, asbestos, and epidemiology: corporate influence and scientific incognizance. Epidemiology. 2019. doi: 10.1097/EDE.0000000000000684 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.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]
- 9.International Agency for Research on Cancer. IARC Monographs: Arsenic, Metals, Fibres, and Dusts. Vol 100C.; 2012. [Google Scholar]
- 10.Terry KL, Karageorgi S, Shvetsov YB, et al. Genital powder use and risk of ovarian cancer: A pooled analysis of 8,525 cases and 9,859 controls. Cancer Prev Res. 2013;6(8):811–821. doi: 10.1158/1940-6207.CAPR-13-0037 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.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]
- 12.O’Brien KM, Tworoger SS, Harris HR, et al. Association of Powder Use in the Genital Area With Risk of Ovarian Cancer. J Am Med Assoc. 2020;323(1):49–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Society AC. Leading Sites of New Cancer Cases and Deaths - 2019 Estimates. 2019;2019(9/6/2019):2019. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-statistics/annual-cancer-facts-and-figures/2019/leading-sites-of-new-cancer-cases-and-deaths-2019-estimates.pdf. [Google Scholar]
- 14.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]
- 15.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] [PubMed] [Google Scholar]
- 16.O’Brien KM, D’Aloisio AA, Shi M, Murphy JD, Sandler DP, Weinberg CR. Perineal Talc Use, Douching, and the Risk of Uterine Cancer. Epidemiology. 2019;30(6):845–852. doi: 10.1097/ede.0000000000001078 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.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]
- 18.Bao Y, Bertoia ML, Lenart EB, et al. Origin, methods, and evolution of the three nurses’ health studies. Am J Public Health. 2016;106(9):1573–1581. doi: 10.2105/AJPH.2016.303338 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.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]
- 20.Anderson GL, Cummings SR, Freedman LS, et al. Design of the Women’s Health Initiative Clinical Trial and Observational Study. Control Clin Trials. 1998;19:61–109. doi: 10.1016/S0197-2456(97)00078-0 [DOI] [PubMed] [Google Scholar]
- 21.Clarke MA, Devesa SS, Harvey SV, Wentzensen N. Hysterectomy-corrected uterine corpus cancer incidence trends and differences in relative survival reveal racial disparities and rising rates of nonendometroid cancers. J Clin Oncol. 2019;37(22):1895–1908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10(1):37–48. [PubMed] [Google Scholar]
- 23.Aune D, Navarro Rosenblatt DA, Chan DSM, et al. Anthropometric factors and endometrial cancer risk: A systematic review and dose-response meta-analysis of prospective studies. Ann Oncol. 2015;26(8):1635–1648. doi: 10.1093/annonc/mdv142 [DOI] [PubMed] [Google Scholar]
- 24.Cote ML, Alhajj T, Ruterbusch JJ, et al. Risk factors for endometrial cancer in black and white women: a pooled analysis from the epidemiology of endometrial cancer consortium (E2C2). Cancer Causes Control. 2015;26(2):287–296. doi: 10.1007/s10552-014-0510-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177–188. [DOI] [PubMed] [Google Scholar]
- 26.Breslow NE. Discussion of Professor Cox’s paper. J R Stat Soc Ser A (Statistics Soc. 1972;34:216. [Google Scholar]
- 27.Lunn M, Mcneil D. Applying Cox Regression to Competing Risks. 2016;51(2):524–532. [PubMed] [Google Scholar]
- 28.Xue X, Kim MY, Gaudet MM, et al. A comparison of the polytomous logistic regression and joint cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies. Cancer Epidemiol Biomarkers Prev. 2013;22(2):275–285. doi: 10.1158/1055-9965.EPI-12-1050 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Grady D, Gebretsadik T, Kerlikowske K, Ernster V, Petitti D. Hormone replacement therapy and endometrial cancer risk: A meta-analysis. Obstet Gynecol. 1995;85(2):304–313. doi: 10.1016/0029-7844(94)00383-O [DOI] [PubMed] [Google Scholar]
- 30.Cook LS, Kamb ML, Weiss NS. Perineal Powder Exposure and the Risk of Ovarian Cancer. Am J Epidemiol. 1997;145(5):459–465. doi: 10.1093/oxfordjournals.aje.a009128 [DOI] [PubMed] [Google Scholar]
- 31.Harlow BL, Hartge PA. A review of perineal talc exposure and risk of ovarian cancer. Regul Toxicol Pharmacol. 1995;21:254–260. [DOI] [PubMed] [Google Scholar]
- 32.Cramer DW, Vitonis AF, Terry KL, Welch WR, Titus LJ. The association between talc use and ovarian cancer a retrospective case-control study in two us states. Epidemiology. 2016;27(3):334–346. doi: 10.1097/eDe.0000000000000434 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Society AC. Talcum Powder and Cancer. American Cancer Society. https://www.cancer.org/cancer/cancer-causes/talcum-powder-and-cancer.html. Published 2020. Accessed May 22, 2020. [Google Scholar]
- 34.Schildkraut JM, Abbott SE, Alberg AJ, et al. Association between Body Powder Use and Ovarian Cancer: The African American Cancer Epidemiology Study (AACES). Cancer Epidemiol Biomarkers Prev. 2016;25(10):1411–1417. doi: 10.1158/1055-9965.epi-15-1281 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.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]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available for request from the individual cohort studies. Please refer to the following links for more information: Nurses’ Health Study and Nurses’ Health Study II (https://www.nurseshealthstudy.org/researchers), Sister Study (https://sisterstudy.niehs.nih.gov/English/data-requests.htm), and Women’s Health Initiative – Observational Study (https://www.whi.org/page/working-with-whi-data). Further details available on request to the corresponding author.