Hair relaxer use and risk of uterine cancer in the Black Women’s Health Study

Abstract

Background

Chemical hair relaxers, use of which is highly prevalent among Black women in the US, have been inconsistently linked to risk of estrogen-dependent cancers, such as breast cancer, and other reproductive health conditions. Whether hair relaxer use increases risk of uterine cancer is unknown.

Methods

In the Black Women’s Health Study, 44,798 women with an intact uterus who self-identified as Black were followed from 1997, when chemical hair relaxer use was queried, until 2019. Over follow-up, 347 incident uterine cancers were diagnosed. We used multivariable Cox proportional hazards regression models, adjusted for age and other potential confounders, to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for associations of hair relaxer use with risk of uterine cancer.

Results

Compared to women who never used hair relaxers or used them infrequently (<4 years and ≤1–2 times/year), the HR for uterine cancer associated with heavy use (≥15 years and at least 5 times/year) was 1.18 (95% CI: 0.81, 1.71). However, among postmenopausal women, compared to never/light use, the HR for moderate use was 1.60 (95% CI: 1.01, 2.53), the HR for heavy use was 1.64 (1.01, 2.64), and the HR for ≥20 years of use regardless of frequency was 1.71 (1.08, 2.72). Results among premenopausal women were null.

Conclusions

In this large cohort of Black women, long-term use of chemical hair relaxers was associated with increased risk of uterine cancer among postmenopausal women, but not among premenopausal women. These findings suggest that hair relaxer use may be a potentially modifiable risk factor for uterine cancer.

Introduction

Chemical hair relaxers are heavily marketed to and commonly used by Black women to straighten curly or tightly coiled hair.¹ Historically, up to 95% of self-identified adult Black women in the U.S. reported ever use of hair relaxers;² in a more contemporary cohort queried 2014–2022, 84% of non-Hispanic Black women reported current or past use.³ While specific formulations are proprietary, potentially harmful toxicants and carcinogens such as formaldehyde and heavy metals may be present in chemical hair relaxers.^4,5 In addition, these products are known to contain phthalates and parabens and other endocrine disrupting chemicals.^6,7

Endocrine disrupting chemicals may cause hormonal imbalance through either estrogenic or antiestrogenic activity and have been associated with a variety of adverse effects related to development and other physiological processes, especially in the reproductive system.⁸ For example, prenatal exposure to diethylstilbestrol (DES) has been associated with uterine structural abnormalities and adenocarcinoma of the vagina.⁹ In animal studies, DES exposure increased risk of uterine adenocarcinoma.¹⁰ Population-based studies have linked biomarkers of phthalates, parabens, and other endocrine disrupters to a wide range of women’s reproductive health outcomes.¹¹

High concentrations of metabolites of phthalates and parabens have been detected in urine samples from women who use chemical hair relaxers^12–14 while use of chemical hair relaxers has been associated with earlier puberty, infertility, and uterine fibroids.^3,15,16 Some studies have supported an association for breast cancer, while others have not.^2,17–19 A single study to date, in the prospective Sister Study cohort, evaluated associations of hair relaxers and uterine cancer. In that study, frequent hair relaxer use (>4 times) in the 12 months before study enrollment was associated with increased risk of uterine cancer compared to never use (HR: 2.55; 95% CI: 1.46, 4.45) in the whole cohort. Among 2,523 Black participants, the corresponding HR was 2.12 (95% CI: 0.83, 5.39); however, there were only 17 exposed cases among Black women in that study.²⁰ Thus, we sought to evaluate the possible association of hair relaxer use with uterine cancer risk in a large prospective cohort of Black women.

Materials and Methods

Study population

The US-based Black Women’s Health Study (BWHS) includes 59,000 women who self-identified as Black and were between the ages of 21 and 69 at enrollment in 1995.²¹ Participants completed a self-administered questionnaire about their personal and family medical history as well as behavioral and other factors that might influence health, such as medication use, cigarette smoking, and diet. Since 1997, participants have completed follow-up questionnaires every two years to update information and to identify incident diagnoses of cancer and other health conditions. Deaths are ascertained from next-of-kin, the U.S. Postal Service, and annual searches of the National Death Index. Follow-up is complete for >80% of potential person-years.

For the current analyses, we excluded women who did not complete the 1997 questionnaire (n=5,923) as well as women who had had a hysterectomy (n=6,164) and those with a history of cancer (other than non-melanoma skin cancer) (n=1,342) before 1997, when we asked about hair relaxer use (see below). We also excluded those who did not answer the hair relaxer questions (n=773). In total, 44,798 BWHS participants were included in these analyses.

The Boston University Medical Campus Institutional Review Board (IRB) and the IRBs of participating cancer registries approved this study. Informed consent was implied by return of the baseline and follow-up questionnaires.

Case ascertainment

Incident cases of uterine cancer were mainly identified via self-report on follow-up questionnaires. Additional cases were identified via linkage to state cancer registries covering 95% of participants and the National Death Index. Diagnoses were confirmed by medical record and cancer registry record review. Through 2019, 347 incident uterine cancer cases were diagnosed during 862,507 person-years of follow-up. Among 285 cases with known histology, 264 (93%) were endometrial cancer and 21 (7%) were uterine sarcomas (histology codes 8800, 8890, 8891, 8900, 8901, 8930, 8931, and 8935).

Exposure and covariate assessment

Exposures of interest included ever use of chemical hair relaxers as well as frequency of use, duration of use, age at first use, type most commonly used, and number of burns. These questions were asked on the 1997 questionnaire, which is available on the study website (www.bu.edu/bwhs). Because only 5% of women reported no past use of hair relaxers,² we combined reported frequency and duration to create three categories of exposure: never/light use (reference category), moderate use, and heavy use. Specifically, women who reported never having used hair relaxers and those who reported using them for <4 years and ≤2 times per year were classified as “never/light” users (n=4,535). Women who reported ≥15 years of use and ≥5 times/year were classified as “heavy” users (n=16,535). All others were classified as “moderate” users (n=23,728).

Covariates of interest included established and suspected risk factors for uterine cancer. Current weight, parity, oral contraceptive use, menopausal status, postmenopausal hormone therapy, smoking history, alcohol consumption, physical activity, and geographic region of residence were reported on the 1997 questionnaire. Height, age at menarche, and educational attainment was ascertained on the 1995 questionnaire. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Self-reports of weight and height showed strong correlation with technician measurements in a subsample of the study population.²²

Statistical methods

Person-time was calculated from 1997 to diagnosis of any cancer, hysterectomy, death, or the end of follow-up (2019), whichever came first. Cox proportional hazards models were stratified by age and follow-up period and used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the risk of uterine cancer associated with moderate or heavy use of hair relaxers compared to the referent category of never/light use. Separate models were fit to evaluate HRs and 95% CIs for associations of duration in years and frequency of use per year (vs. the referent category of never/light use). Among moderate and heavy users, we also evaluated associations with age at first use and number of burns. We repeated these analyses for formulations reported as the most commonly used (i.e., lye or no-lye); in these analyses, some categories of exposure were collapsed due to small numbers.

Multivariable models included adjustment for potential confounders: age at menarche (≤11, 12, 13, ≥14 years), oral contraceptive use (≥5 years, <5 years, never), parity (nulliparous, 1 child, 2 children, ≥3 children), menopausal status (pre, post), postmenopausal hormone use (never, ever estrogen alone, ever estrogen plus progestin but never estrogen alone, ever progestin alone), BMI (continuous), smoking history (never, past, current), alcohol consumption (<1, 1–6, ≥7 drinks/week), physical activity (none, <5, ≥5 hours/week), residential region (Northeast, South, Midwest, West, other), educational attainment (≤12, 13–15, ≥16 years). Missing information in covariates was low (generally <2%) and included as a separate category in models.

In stratified analyses, we evaluated premenopausal and postmenopausal uterine cancer separately. We also repeated the overall analyses with endometrial cancer as the outcome (i.e., excluding known uterine sarcomas).

All analyses were performed using SAS 9.4 (Cary, North Carolina).

Results

Among 44,798 BWHS participants eligible for these analyses, 10% were classified as never/light users of chemical hair relaxers, 53% were moderate users, and 37% were heavy users. Moderate and heavy users were slightly younger (mean age, 40 and 39 years, respectively) than never/light users (mean age, 43 years); they were also more likely to have BMI >30 kg/m² (32%) than never/light users (26%). Moderate and heavy users were more likely to have used oral contraceptives (85% and 88%, respectively) than never/light users (81%); they were also more likely to have used oral contraceptives for at least five years. Heavy users had somewhat higher levels of educational attainment than never/light users. Moderate and heavy users were more likely to live in the South and less likely to live in the West compared to never/light users. Other characteristics such as age at menarche, parity, menopausal status, ever use of postmenopausal hormones, smoking history, alcohol consumption, and physical activity were similarly distributed across categories of hair relaxer use (Table 1).

Table 1.

Characteristics of the study population by hair relaxer use at baseline (1997)

Characteristic	Never/light (n=4,535)	Hair relaxer use Moderate (n=23,728)	Heavy (n=16,535)
	(n=4,535)	(n=23,728)	(n=16,535)
Age in years, mean (SD)*	43.4 (11.9)	39.8 (10.8)	39.1 (9.2)
Body mass index, kg/m2, %
<25	42	35	35
25–29.9	30	31	32
30–34.9	14	17	17
≥35	12	15	15
Age at menarche in years, %
<11	26	28	29
12	28	28	29
13	24	24	24
≥14	21	19	18
Parity, %
Nulliparous	34	34	35
1 child	22	23	23
2 children	22	23	23
≥3 children	21	19	18
Menopausal status, %
Premenopausal	81	81	81
Postmenopausal	18	19	19
Ever use of postmenopausal hormones, %
Never used	83	82	81
Estrogen alone	8	9	10
Estrogen + progestin	4	5	5
Progestin or unknown type	4	4	4
Duration of oral contraceptive use, %
Never	20	15	12
<5 years	54	53	49
≥5 years	27	32	39
Smoking history, %
Current	16	16	15
Past	18	19	19
Never	66	65	66
Alcohol consumption in drinks/week, %
<1	75	73	73
1–6	21	23	23
≥7	3	4	4
Educational attainment in years, %
≤12	21	18	16
13–15	34	37	35
≥16	44	45	49
Physical activity in hours/week, %
None	43	43	45
<5	46	47	46
≥5 hours per week	10	10	9
Residential region
Northeast	28	30	24
South	24	29	36
Midwest	19	22	24
West	28	18	15

Values are means (SD) or percentages and are standardized to the age distribution of the study population. Percentages may not sum to total due to missing values or rounding.

^*Value is not age adjusted

As shown in Table 2, results from age-adjusted and multivariable-adjusted models were similar. Compared to never/light use of hair relaxers, the multivariable-adjusted HRs for uterine cancer were 1.23 (95% CI: 0.86, 1.76) for moderate use and 1.18 (95% CI: 0.81, 1.71) for heavy use. Women who reported using hair relaxers for 20 or more years had a slightly increased risk of uterine cancer compared to never/light use (HR: 1.28; 95% CI: 0.89, 1.84), but this result was not statistically significant and there was no clear linear trend in associations for duration of use. Likewise, there were no apparent trends for frequency of use, age at first use, or number of burns. In analyses that excluded known uterine sarcomas, results were similar to results for all uterine cancer (Supplementary Table 1).

Table 2.

Hazard ratios (95% confidence intervals) for uterine cancer in relation to hair relaxer use, 1997–2019

	Person-years	N cases	Age-adjusted HR (95% CI)	Multivariable HR* (95% CI)
Hair relaxer use
Never/light	86,446	38	1.00 (ref)	1.00 (ref)
Moderate	458,424	187	1.18 (0.83, 1.68)	1.23 (0.86, 1.76)
Heavy	317,637	122	1.02 (0.71, 1.48)	1.18 (0.81, 1.71)
Duration of use
Never/light	86,446	38	1.00 (ref)	1.00 (ref)
1–4 years	34,687	15	1.03 (0.56, 1.87)	1.00 (0.55, 1.83)
5–9 years	90,926	29	1.01 (0.62, 1.64)	1.00 (0.61, 1.64)
10–14 years	186,663	58	1.24 (0.82, 1.88)	1.35 (0.88, 2.06)
15–19 years	198,391	54	0.98 (0.64, 1.49)	1.06 (0.69, 1.63)
≥20 years	265,394	153	1.16 (0.81, 1.65)	1.28 (0.89, 1.84)
Frequency of use^
Never/light	86,446	38	1.00 (ref)	1.00 (ref)
1–2 times/year	57,533	29	1.19 (0.74, 1.94)	1.22 (0.75, 2.00)
3–4 times/year	221,857	110	1.23 (0.85, 1.78)	1.25 (0.86, 1.83)
5–6 times/year	232,257	89	1.09 (0.74, 1.59)	1.23 (0.83, 1.81)
≥7 times/year	258,866	77	0.96 (0.64, 1.42)	1.08 (0.72, 1.61)
Age at first uset†
<10 years	47,997	10	1.18 (0.55, 2.50)	1.19 (0.56, 2.54)
10–19 years	475,060	134	1.11 (0.71, 1.73)	1.13 (0.72, 1.78)
20–29 years	204,976	135	1.35 (0.89, 2.04)	1.42 (0.93, 2.15)
≥30 years	45,512	30	1.00 (ref)	1.00 (ref)
Number of burns^†
Never burned	119,074	51	1.00 (ref)	1.00 (ref)
1–2 times	211,841	101	1.33 (0.95, 1.86)	1.32 (0.94, 1.86)
3–4 times	169,247	56	0.98 (0.67, 1.44)	0.97 (0.66, 1.43)
5–9 times	110,674	42	1.25 (0.83, 1.89)	1.29 (0.85, 1.96)
>10 times	162,054	54	1.10 (0.74, 1.62)	1.08 (0.73, 1.61)

Never or light use is defined as 1–4 years, ≤2 times/year; heavy use is defined as ≥15 years, ≥5 times/year; remainder are moderate users (1–4 years, >3 times/year; 5–14 years, any frequency; or ≥15 years, <5 times/year).

HR, hazard ratio; CI, confidence interval

^*HRs adjusted for age, period, menopause status (pre, post), education (≥16, 13–15, ≤12 years), physical activity (none, <5, ≥5 hours per week), parity (no children, 1 child, 2 children, ≥3 children), smoking status (never, past, current), body mass index (continuous), alcohol consumption (<1, 1–6, or ≥7 drinks per week), oral contraceptive use (≥5 years, <5 years, never), age of menarche ( ≥ 14, 13, 12, ≤11 years), residential region (Northeast, South, Midwest, West), female hormone use (progestin alone or unknown, estrogen plus progestin but never estrogen alone, estrogen alone, never hormone use)

^{^}Frequency of use was missing for 4 cases; number of burns was missing for 5 cases

^†Analyses exclude never/light users

Results were also null in separate analyses of lye-based relaxers and no-lye relaxers. Compared to a common referent group of never/light use of any type of chemical hair relaxer, HRs were 1.04 (95% CI: 0.58, 1.87) for heavy use of lye-based relaxers and 1.24 (95% CI: 0.83, 1.84) for heavy use of no-lye relaxers (Table 3).

Table 3.

Hazard ratios (95% confidence intervals) for uterine cancer in relation to lye vs. no-lye hair relaxer use,^{^} 1997–2019

	Most commonly used lye relaxers			Most commonly used no-lye relaxers
	Person-years	N cases	Multivariable HR* (95% CI)	Person-years	N cases	Multivariable HR* (95% CI)
Hair relaxer use
Never/light	86,446	38	1.00 (ref)	86,446	38	1.00 (ref)
Moderate	63,230	26	1.49 (0.87, 2.54)	334,210	131	1.22 (0.84, 1.77)
Heavy	64,744	19	1.04 (0.58, 1.87)	217,833	86	1.24 (0.83, 1.84)
Duration of use
Never/light	86,446	38	1.00 (ref)	86,446	38	1.00 (ref)
1–9 years	15,237	3	0.55 (0.17, 1.80)	88,397	32	1.12 (0.69, 1.82)
10–19 years	66,301	20	1.59 (0.88, 2.88)	278,601	77	1.15 (0.76, 1.73)
≥20 years	46,436	22	1.28 (0.73, 2.25)	185,045	108	1.31 (0.90, 1.93)
Frequency of use ^
Never/light	86,446	38	1.00 (ref)	86,446	38	1.00 (ref)
1–4 times/year	33,165	16	1.40 (0.75, 2.58)	207,112	101	1.26 (0.86, 1.86)
5–6 times/year	41,433	13	1.21 (0.62, 2.35)	164,504	61	1.24 (0.82, 1.89)
≥7 times/year	52,543	15	1.15 (0.61, 2.18)	176,802	53	1.13 (0.73, 1.74)
Number of burns ^ †
Never burned	13,020	6	1.00 (ref)	88478	40	1.00 (ref)
1–9 times	77,477	28	0.78 (0.31, 1.97)	355871	138	1.16 (0.8, 1.67)
≥10 times	37,032	9	0.50 (0.17, 1.48)	105914	36	1.12 (0.7, 1.79)

Never or light use is defined as 1–4 years, ≤2 times/year; heavy use is defined as ≥15 years, ≥5 times/year; remainder are moderate users (1–4 years, >3 times/year; 5–14 years, any frequency; or ≥15 years, <5 times/year).

HR, hazard ratio; CI, confidence interval

^*HRs adjusted for age, period, menopause status (pre, post), education (≥16, 13–15, ≤12 years), physical activity (none, <5, ≥5 hours per week), parity (no children, 1 child, 2 children, ≥3 children), smoking status (never, past, current), body mass index (continuous), alcohol consumption (<1, 1–6, or ≥7 drinks per week), oral contraceptive use (≥5 years, <5 years, never), age of menarche ( ≥ 14, 13, 12, ≤11 years), residential region (Northeast, South, Midwest, West), female hormone use (progestin alone or unknown, estrogen plus progestin but never estrogen alone, estrogen alone, never hormone use)

^{^}Type of hair relaxer most commonly used was unknown for 47 cases; frequency of use was missing for 3 cases; number of burns was missing for 5 cases

^†Analyses exclude never/light users

Among postmenopausal women, the HR associated with moderate use of hair relaxers was 1.60 (95% CI: 1.01, 2.53) and the HR associated with heavy use was 1.64 (95% CI: 1.01, 2.64). In particular, the HR for ≥20 years of use relative to never/light use was 1.71 (1.08, 2.72), and similar associations were observed for 10–14 and 15–19 years of use (Table 4). Among postmenopausal women, results were similarly positive for moderate or heavy use of both lye and no-lye products (Supplementary Table 2).

Table 4.

Hazard ratios (95% confidence intervals) for uterine cancer in relation to hair relaxer use by menopausal status, 1997–2019

	Premenopausal			Postmenopausal
	Person-years	N cases	Multivariable HR* (95% CI)	Person-years	N cases	Multivariable HR* (95% CI)
Hair relaxer use
Never/light	36,285	12	1.00 (ref)	42,725	23	1.00 (ref)
Moderate	248,479	58	0.84 (0.45, 1.58)	165,713	123	1.60 (1.01, 2.53)
Heavy	159,662	31	0.63 (0.32, 1.24)	127,702	85	1.64 (1.01, 2.64)
Duration of use
Never/light	36,285	12	1.00 (ref)	42,725	23	1.00 (ref)
1–4 years	15,904	8	1.36 (0.55, 3.38)	16,058	7	0.87 (0.37, 2.05)
5–9 years	52,625	12	0.90 (0.40, 2.02)	29,127	15	1.09 (0.57, 2.12)
10–14 years	122,517	18	0.66 (0.31, 1.41)	43,011	39	2.21 (1.30, 3.75)
15–19 years	122,108	20	0.65 (0.31, 1.34)	54,251	31	1.49 (0.85, 2.59)
≥20 years	94,988	31	0.74 (0.38, 1.47)	150,970	116	1.71 (1.08, 2.72)
Frequency of use ^
Never/light	36,285	12	1.00 (ref)	42,725	23	1.00 (ref)
1–2 times/year	25,326	7	0.83 (0.32, 2.13)	26,981	21	1.59 (0.87, 2.91)
3–4 times/year	104,638	35	1.00 (0.51, 1.94)	97,619	72	1.56 (0.96, 2.53)
5–6 times/year	124,694	20	0.55 (0.27, 1.14)	84,959	65	1.85 (1.13, 3.03)
≥7 times/year	150,760	25	0.65 (0.32, 1.31)	81,498	48	1.47 (0.88, 2.46)

Never or light use is defined as 1–4 years, ≤2 times/year; heavy use is defined as ≥15 years, ≥5 times/year; remainder are moderate users (1–4 years, >3 times/year; 5–14 years, any frequency; or ≥15 years, <5 times/year).

HR, hazard ratio; CI, confidence interval

^*HRs adjusted for age, period, education (≥16, 13–15, ≤12 years), physical activity (none, <5, ≥5 hours per week), parity (no children, 1 child, 2 children, ≥3 children), smoking status (never, past, current), body mass index (continuous), alcohol consumption (<1, 1–6, or ≥7 drinks per week), oral contraceptive use (≥5 years, <5 years, never), age of menarche ( ≥ 14, 13, 12, ≤11 years), residential region (Northeast, South, Midwest, West), female hormone use (progestin alone or unknown, estrogen plus progestin but never estrogen alone, estrogen alone, never hormone use).

^{^}Frequency of use was missing for 2 premenopausal cases and 2 postmenopausal cases

Discussion

Within a large cohort of Black women followed for up to 22 years, hair relaxer use was associated with an increased risk of uterine cancer among postmenopausal women, but not premenopausal women. Compared to women who never or rarely used hair relaxers, postmenopausal women who reported using chemical hair relaxers more than twice a year or for more than five years had a greater than 50% increased risk of uterine cancer. Elevated risks were most apparent for postmenopausal women who reported at least 10 years of use, regardless of frequency.

Chemical hair relaxers are commonly used by Black women.^23,24 These products, which contain hormonally active ingredients,⁷ have been inconsistently linked to reproductive health conditions as well as to hormonally-related cancers, such as breast cancer. Several studies have reported use of chemical hair relaxers to be associated with markers of early puberty.^15,25,26 In a previous analysis within the BWHS, hair relaxer use was associated with increased risk of uterine fibroids.¹⁶ More recently, in a preconception cohort, hair relaxer use was associated with lower fecundability.³ Some but not all prior studies suggest modest associations between hair relaxer use and breast cancer.^2,17–19 In addition to a positive finding for uterine cancer, the NIEHS Sister Study also reported that frequent use of hair relaxers increased risk of ovarian cancer.²⁷

Exposure to phthalates, parabens, and other endocrine disrupting chemicals in hair relaxers may occur via inhalation or dermal absorption. Experimental and animal studies have shown that these chemicals have estrogenic and anti-androgenic properties^28–30 and thus may influence uterine cancer development indirectly through hormonal pathways.³¹ The carcinogenic potential of endocrine disrupting chemicals is likely multifactorial, as these chemicals have also been implicated in epigenetic reprogramming, overexpression of cyclooxygenase-2, regulation of the epithelial-mesenchymal transition, and immune dysfunction.^32,33

To our knowledge, the NIEHS Sister Study is the only study to have published on hair relaxer use and uterine cancer risk. Consistent with our findings, Chang et al. reported positive associations of frequent use of hair relaxers in relation to uterine cancer among postmenopausal women (HR: 2.52; 95% CI: 1.39, 4.55), but not among premenopausal women.²⁰ There are, however, some important differences between the two studies. First, all participants in the Sister Study had a family history of breast cancer; thus, they represent a high-risk population for breast cancer, and possibly other hormonally-related cancers as well. This is not the case in the BWHS. Second, the Sister Study queried use of hair products in the year prior to study enrollment (2003–2009), whereas the BWHS queried lifetime use of hair relaxers as of 1997. With a 6- to 12-year difference in timing of exposure assessment between the studies, it is possible that the specific products on the market differed during the relevant years of use. Neither study had detailed information regarding brands or formulations of hair products used. In addition, the Sister Study assessed frequent use as >4 times in the 12 months prior to baseline whereas we assessed “heavy use” as ≥15 years at least 5 times per year as our primary exposure. Finally, while the Sister Study reported a significant positive association for frequent use of hair relaxers and uterine cancer overall, the association was most apparent for non-Hispanic white women; the estimate for the small number of Black women (n=17 exposed cases, pre- and postmenopausal women combined) was elevated but not statistically significantly (HR: 2.12; 95% CI: 0.83, 5.39).

An important limitation of the present study is that we did not update exposure information over time. Therefore, it is not known whether these results are generalizable to chemical hair relaxers on the market today or to contemporary patterns of use. That said, contemporary studies have documented the presence of parabens and phthalates, as well as formaldehyde, a known carcinogen, in modern day formulations of many hair products, including chemical relaxers.^5,7,34 We lacked information regarding brand or product ingredients and were therefore unable to identify whether observed associations may have been driven by a specific product or chemical class. Another important limitation is that because the majority of BWHS participants reported using chemical hair relaxers, we included never users and infrequent users in the same exposure referent category to achieve stable effect estimates, which may have biased results toward the null.

Major strengths of this analyses are the prospective study design and assessment of lifetime use of chemical hair relaxers. Importantly, this is the largest study of hair relaxers and uterine cancer in Black women to date, which afforded excellent statistical power for most analyses.

Results of these analyses suggest that moderate and heavy use of chemical hair relaxers may increase risk of uterine cancer among postmenopausal Black women. These products are widely marketed to and commonly used by Black women. Our findings highlight the importance of continued research regarding potential adverse health effects of exposure to chemical hair relaxers and their constituents. Uterine cancer mortality rates are substantially higher for Black women compared to non-Hispanic white women³⁵ and there are few established modifiable risk factors for uterine cancer. If the present results are confirmed, efforts to reduce exposure to these potentially harmful products – such as by identifying safer alternatives to straightening hair, stricter regulation of cosmetic products, and policies prohibiting discrimination against natural hair (e.g., the CROWN Act³⁶) – could represent important steps toward reducing racial disparities in this cancer type.

Highlights.

• Black women have higher uterine cancer mortality than non-Hispanic white women

• Use of chemical hair relaxers is common among Black women

• Long-term use of hair relaxers was associated with increased risk of uterine cancer among postmenopausal Black women

Data availability

The data underlying this article cannot be shared publicly to protect the privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.