Lifestyle factors for polycystic ovary syndrome: a case-control study

Abstract

Background

Lifestyle interventions have been proposed to be the first-line treatment for polycystic ovarian syndrome (PCOS), but our knowledge of PCOS-related lifestyle factors is still limited. This study aimed to explore PCOS-related lifestyle factors in reproductive-age women in China.

Methods

A case-control study was carried out in Shandong, China, from June to October 2014. 350 PCOS-related infertile women were recruited as cases, while 308 women who were seeking the treatment of fallopian tube obstruction or assisted reproduction due to male infertility were recruited as controls. Several specific lifestyle and behavioral factors associated with EDC exposure were examined. Multivariate logistic regression analysis was used to estimate the odds ratios (ORs) and 95% confidence intervals (CI).

Results

There were no differences in socio-demographic characteristics, smoking, and alcohol consumption between the cases and controls. PCOS women had significantly more obesity than the controls (P < 0.01). After controlling for potential confounders, using foam or plastic cups at home (daily vs. never: adjusted odds ratio [aOR] 2.6, 95% confidence interval [CI] 1.4–4.5), using foam or plastic tableware at home (occasionally vs. never: aOR 1.7, 95% CI 1.2–2.5), eating melons and fruits without pericarp (for occasionally vs. always/often: aOR 1.6, 95% CI 1.0–2.5), drinking tap water vs. spring water/bottled water (aOR 1.5, 95% CI 1.1–2.2), cooking (daily vs. ≤ 1 time/week: aOR 1.8, 95% CI 1.2–2.8), contacting liquid detergent (1.5 to 3 h/week vs. < 1.5 h/week, aOR 2.5, 95% CI 1.3–4.9), and contacting soap or shampoo (3–6 vs. < 3 h/week, aOR 2.1, 95% CI 1.2–3.8) were significantly associated with PCOS. The association between mild cooking oil fumes and PCOS risk was attenuated after adjustment, but it remained borderline significant (aOR 1.4, 95% CI 0.97–1.9).

Conclusions

Our study suggests that PCOS may be associated with certain lifestyle factors. Future prospective studies are needed to confirm our findings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13048-026-02034-9.

Introduction

Polycystic ovary syndrome (PCOS) is one of the most prevalent endocrine abnormalities in women of reproductive age. A recent systematic review reported that the prevalence ranges from 5% to 18% worldwide according to the 2003 Rotterdam criteria [1]. Current estimates using the Rotterdam criteria indicate a PCOS prevalence of 7.8% among Chinese women aged 20–49 in 2020, a 65% increase over the past decade. Notably, the rate exceeds 14% among women aged 20–30 [2]. This syndrome is often associated with multiple health conditions, such as metabolic disturbance, cardiovascular diseases, reproductive abnormalities, infertility, and psychological problems [3, 4]. Maternal PCOS has also been linked to an increased risk of perinatal and long-term adverse outcomes in the offspring [5, 6].

Despite its prevalence, the etiology of PCOS remains unclear. Lifestyle factors have been associated with PCOS. Recent research has reviewed the current knowledge about major modifiable risk factors in PCOS development [7]. Women with PCOS were more likely to have inappropriate dietary composition, less physical activity, inadequate sleep hygiene, elevated stress or depression, as well as higher incidences of smoking and alcohol use compared to healthy women [7, 8]. Lifestyle and behavioral interventions such as managing weight, developing healthy dietary habits (e.g., lower calorie intake), and keeping physical exercise, have been recommended to improve reproductive functions, endocrine parameters, cardiometabolic profiles, and psychological health in PCOS women [9]. It was even considered as the first-line therapy for PCOS [10]. However, these interventions may not address all underlying lifestyle factors that contribute to the development and long-term persistence of PCOS [9].

Accumulating evidence indicates a connection between endocrine-disrupting chemicals (EDCs) and the pathogenesis and/or more severe symptoms of PCOS [11]. EDCs are exogenous compounds comprising a wide range of natural and synthetic chemicals such as bisphenol A (BPA), some pesticides, triclosan (TCS), and polycyclic aromatic hydrocarbons (PAHs) [12, 13]. Humans can be widely exposed to EDCs through dietary ingestion, absorption through the skin and mucous membranes, and inhalation [14]. Mounting evidence demonstrated the adverse effects of EDCs on the endocrine system and reproductive functions [14]. These exogenous compounds may also interfere with metabolic homeostasis, impair insulin sensitivity, induce changes in epigenetic modifications, and exert a regulatory effect on genetic susceptibility [14].

Lifestyle modifications and behavioral interventions addressing EDCs exposure will undoubtedly play an important role in comprehensive PCOS management. The current key recommendations [15] for the general population include avoiding plastic food/beverage containers by using glass, ceramic, or stainless steel ceramic alternatives; reducing canned food consumption, as the internal coatings of metal cans may leach BPA; prioritizing organic, home-cooked meals instead of processed/packaged options that may contain bisphenols and phthalates; selecting natural cleaners like vinegar and baking soda rather than chemical ones; choosing personal care items formulated without parabens, phthalates, or triclosan; and improving indoor air quality through ventilation to reduce airborne EDCs levels.

Most previous studies primarily focus on detecting EDC biomarkers, rather than establishing a clear link between the specific lifestyle behaviors that may lead to EDC exposure and PCOS. This study aimed to explore the association between lifestyle-mediated environmental exposures (i.e., various household lifestyle factors in China that may contribute to EDCs exposure) and PCOS among reproductive-age Chinese women-specifically including the use of foamed cups and tableware that often contain BPA, contact with soap, shampoo or liquid detergent containing TCS, habits of eating melons and fruits with pericarp and/or eating vegetables without soaking which may contain pesticide residues, consumption of drinking water which may contain pesticides or other chemicals, and cooking-related factors, such as oil fumes and the combustion process of fuels used, which may release harmful compositions such as polycyclic aromatic hydrocarbons (PAHs). The findings may provide targeted lifestyle intervention recommendations for PCOS in the Chinese population.

Materials and methods

Study design and patients

From June to October 2014, a case-control study was conducted at the Centre for Reproductive Medicine, Shandong University, Shandong, China. All potential subjects were couples seeking medical care for infertility (couples had regular unprotected intercourse within the past year, but had failed to conceive). Both females and their partners had undergone a complete workup for the diagnosis of infertility.

We included women into the case group with PCOS-related infertility if they were married, aged 20–40 years, diagnosed as PCOS based on the Rotterdam criteria, namely meeting two or three of the following item: (1) oligo-ovulation or anovulation; (2) biochemical hyperandrogenism and/or clinical hyperandrogenism while other disorders related to androgen excess excluded; (3) polycystic ovaries [16]; had no tubal obstruction or other specific infertility factors; and their husbands had normal fertility. The control group included women who were married, 20–40 years old, and came to the center for the treatment of fallopian tube obstruction or male infertility. A total of 350 women who had PCOS-related infertility were recruited as cases, and 308 women were included as controls. The method for sample size calculation was described in the supplementary file. Patients with PCOS-related infertility in the case group were diagnosed when they attended infertility clinics for treatment. Infertility is defined as the inability of couples to achieve a clinical pregnancy after 12 consecutive months of regular sexual intercourse without contraception [17].

Interviews

All cases and controls were interviewed by trained staff. Information collected was the socio-demographic characteristics, menstrual and reproductive history, work and lifestyle factors, personal medical history and family history. The collection of lifestyle factors was based on lifestyle patterns associated with environmental pollutant exposure previously reported in the literature, and it also took into account the prevalence of these behaviors among the Chinese population. The participants and the interviewer were unaware of the hypothesis of the present study, which examined the association between lifestyle factors and PCOS. This study was approved by the Institutional Review Board at Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China. A written consent form was signed by all participants.

Exposure information

Women’s occupations were categorized into five groups: office and service workers/soldiers/teachers/doctors/students (reference group); farmers/herdsmen/fishermen; engineers/chemical factory workers/clothing or food processing workers; unemployed; and others/unknown. The frequency of using foam or plastic cups at home and the frequency of using plastic tableware were categorized as: never (reference), occasionally, most of the time, daily, and unknown. The categorization of other exposures were as follows: eating melons and fruits without pericarp (always/often [reference], occasionally, never, unknown), soaking vegetables before eating (yes [reference], no, unknown), source of drinking water (spring water/bottled water [reference], river water/well water, tap water, others/unknown), cooking frequency (rarely [reference], 2–3 times/week, 4–6 times/week, daily, unknown), type of cooking fuel (natural gas [reference], coal gas, others, unknown), cooking with coal gas (rarely cook [reference], 2–3 times/week, 4–6 times/week, daily), cooking with natural gas (rarely cook [reference], 2–3 times/week, 4–6 times/week, daily), cooking oil fumes (COFs: no [reference], mild, heavy), cooking with no COFs (rarely cook [reference], 2–3 times/week, 4–6 times/week, and daily), cooking with mild COFs (rarely cook [reference], 2–3 times/week, 4–6 times/week, and daily). Infrequent cooking was defined as cooking ≤ 1 time per week, regardless of whether there were COFs or the type of cooking fuel (CF) used. Contacting soap or shampoo (≤ 3 h/week [reference], 3–6 h/week, > 6 h/week, unknown), contacting liquid detergent (≤ 1.5 h/week [reference], 1.5–3 h/week, > 3 h/week, unknown), using air freshener (never [reference], occasionally, most of the time, daily, unknown), average outdoor time per week (< 10 h [reference], 10–20 h, > 20 h, unknown), average sedentary time per day (< 4 h [reference], 4–8 h, > 8 h, unknown) and residential traffic flow (small [reference], medium, large, unknown).

Other covariates

Sociodemographic characteristics included female age (≤ 25, 25–30 [reference], 30–35, > 35 years and unknown), female educational level (≤ 6, 6–9 [reference], 10–12, 13–16, ≥ 17 years and unknown), female body mass index (BMI, < 18.5, 18.5–25 [reference], and ≥ 25 kg/m², unknown), annual income (< 30,000 [reference], 30,000–50,000, 50,000–100,000, ≥ 100,000 RMB/person, unknown).

Statistical analysis

Differences in socio-demographic characteristics (all categorical) between cases and controls were compared using Chi-square tests. Multivariable logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the association between exposure variables and PCOS, controlling for potential confounders such as age, women’s educational level, household income, and BMI. In the primary analysis, univariate and multivariate logistic regression analyses were repeated after excluding the missing data for exposure variables. In the sensitivity analysis, missing data for all exposure variables and covariates were imputed using multiple imputation by fully conditional specification (MI-FCS) [18, 19]. The results were pooled by concatenating ten imputed datasets that had no missing data. All the analyses were conducted using SAS version 9.4.

Results

Table 1 shows that women with PCOS had a higher proportion of obesity than the controls. There were no significant differences in socio-demographic characteristics, smoking, and alcohol consumption between the cases and controls.

Table 1.

Demographic characteristics in PCOS cases and controls (n = 658)

Characteristics	Cases n (%)	Controls n (%)	P
Female age (years)			0.8
≤ 25	49 (14.0)	51 (16.6)
25–30	190 (54.3)	165 (53.6)
30–35	89 (25.4)	72 (23.4)
≥ 35	18 (5.1)	18 (5.8)
Unknown	4 (1.1)	2 (0.7)
Female educational level (years)			0.9
≤ 6	28 (8.0)	19 (6.2)
6–9	169 (48.3)	152 (49.4)
10–12	87 (24.9)	74 (24.0)
13–16	57 (16.3)	53 (17.2)
≥ 17	3 (0.9)	2 (0.7)
Unknown	6 (1.7)	8 (2.6)
Female occupation			0.3
Office and service workers/soldiers, teachers/doctors/students	144 (41.1)	138 (44.8)
Farmers/herdsmen/fishermen	31 (8.9)	22 (7.1)
Engineers/chemical factory workers/clothing or food processing workers	36 (10.3)	41 (13.3)
Unemployed	106 (30.3)	88 (28.6)
Unknown	33 (9.4)	19 (6.2)
Annual income (RMB/person)			0.1
< 30,000	253 (72.3)	197 (64.0)
30,000–50,000	53 (15.1)	70 (22.7)
50,000–100,000	17 (4.9)	20 (6.5)
≥ 100,000	8 (2.3)	7 (2.3)
Unknown	19 (5.4)	14 (4.5)
*BMI (kg/m2)			< 0.0001
< 18.5	13 (3.7)	27 (8.8)
18.5–25	150 (42.9)	170 (55.2)
≥ 25	163 (46.6)	67 (21.7)
Unknown	24 (6.9)	44 (14.3)
Smoking status			0.1
Never	326 (93.1)	297 (96.4)
Have quit smoking for half a year	0 (0.00)	1 (0.3)
Occasionally or current smoking	3 (0.9)	0 (0.0)
Unknown	21 (6.0)	10 (3.3)
Alcohol consumption			0.6
Never	307 (87.7)	280 (90.9)
Have quit drinking for a year	7 (2.0)	6 (2.0)
Occasionally or regular drinking	24 (6.9)	14 (4.5)
Unknown	12 (3.4)	8 (2.6)

*BMI body mass index

Table 2 shows that after adjusting for potential confounders, PCOS was significantly associated with using foam or plastic cups daily at home (adjusted odds ratio [aOR] 2.6, 95% confidence interval [CI] 1.4–4.5), using foam or plastic tableware occasionally at home (aOR 1.7, 95% CI 1.2–2.5), eating melons and fruits without pericarp (for occasionally vs. always/often: aOR 1.6, 95% CI 1.0–2.5), drinking tap water vs. spring water/bottled water (aOR 1.5, 95% CI 1.1–2.2), cooking (daily vs. ≤ 1 time/week, aOR 1.8, 95% CI 1.2–2.8). In the group of cooking with natural gas, cooking frequency was not associated with PCOS. Among women who cooked with coal gas, cooking every day increased the risk of PCOS (aOR 2.4, 95% CI 1.5–3.9) compared with those who rarely cooked. The association between mild COFs and PCOS risk was borderline significant (aOR 1.4, 95% CI 0.97–1.9). In women who reported cooking without COFs, cooking frequency was not associated with PCOS. However, among women who reported cooking with mild COFs, cooking every day was associated with an elevated risk of PCOS (aOR 2.3, 95% CI 1.4–3.7) compared with women who rarely cook. Contacting liquid detergent (1.5 to 3 h/week vs. < 1.5 h/week, aOR 2.5, 95% CI 1.3–4.9), and contacting soap or shampoo (3–6 vs. < 3 h/week, aOR 2.1, 95% CI 1.2–3.8) were associated with increased risks of PCOS. PCOS was not significantly associated with air freshener use, eating vegetables without soaking them, average daily sitting time, or occupation.

Table 2.

Association between lifestyle factors and PCOS

Exposures	Cases n (%)	Controls n (%)	OR (95% CI)	aOR# (95% CI)
Female occupation (n = 606)
Office and service workers, soldiers, teachers, doctors, students	144 (45.4)	138 (47.8)	Ref	Ref
Farmers, herdsmen and fishermen	31 (9.8)	22 (7.6)	1.4 (0.7–2.4)	1.1 (0.5–2.1)
Engineers, chemical factory workers, clothing or food processing workers	36 (11.4)	41 (14.2)	0.8 (0.5–1.4)	0.9 (0.5–1.7)
Unemployed	106 (33.4)	88 (30.4)	1.2 (0.8–1.7)	1.0 (0.7–1.6)
Using foam or plastic cups at home (n = 646)
Never	190 (55.4)	193 (63.7)	Ref	Ref
Occasionally	73 (21.3)	63 (20.8)	1.2 (0.8–1.7)	1.1 (0.7–1.7)
Most of the time	28 (8.2)	26 (8.6)	1.1 (0.6–1.9)	1.2 (0.6–2.1)
Daily	52 (15.2)	21 (6.9)	2.5 (1.5–4.3)	2.6 (1.4–4.5)
Using foam or plastic tableware at home (n = 642)
Never	218 (64.7)	228 (74.7)	Ref	Ref
Occasionally	110 (32.6)	68 (22.3)	1.7 (1.2–2.4)	1.7 (1.2–2.5)
Most of the time	7 (2.1)	6 (2.0)	1.2 (0.4–3.7)	1.2 (0.2–3.7)
Daily	2 (0.6)	3 (1.0)	0.7 (0.1–4.2)	0.4 (0.1–3.2)
Eating melons and fruits without pericarp (n = 646)
Always/often	68 (19.9)	84 (27.5)	Ref	Ref
Occasionally	136 (39.9)	108 (35.4)	1.6 (1.0–2.3)	1.6 (1.0–2.5)
Never	137 (40.2)	113 (37.1)	1.5 (1.0–2.3)	1.5 (1.0–2.4)
Soaking vegetables before eating (n = 641)
Yes	246 (72.6)	207 (68.5)	Ref	Ref
No	93 (27.4)	95 (31.5)	0.8 (0.6–1.2)	0.9 (0.6–1.2)
Drinking water (n = 644)
Spring water/barrelled water	92 (27.0)	107 (35.3)	Ref	Ref
River water/well water	40 (11.7)	25 (8.3)	1.8 (1.0–3.1)	1.6 (0.9–3.1)
Tap water	209 (61.3)	171 (56.4)	1.4 (1.0–1.9)	1.5 (1.1–2.2)
Cooking (n = 637)
≤ 1 time/week	77 (22.6)	96 (32.4)	Ref	Ref
2–3 times/week	46 (13.5)	38 (12.8)	1.5 (0.9–2.5)	1.6 (0.9–2.8)
4–6 times/week	29 (8.50)	40 (13.5)	0.9 (0.5–1.6)	0.8 (0.5–1.5)
Daily	189 (55.4)	122 (41.2)	1.9 (1.3–2.8)	1.8 (1.2–2.8)
Source of fuels for cooking or heating (n = 624)
Natural gas	94 (27.7)	94 (33.0)	Ref	Ref
Coal gas	187 (55.2)	127 (44.5)	1.5 (1.0–2.1)	1.4 (0.95–2.2)
Others*	58 (17.1)	64 (22.5)	0.9 (0.6–1.4)	0.9 (0.5–1.5)
Cooking using coal gas (n = 407)
Rarely cook†	77 (33.9)	96 (53.3)	Ref	Ref
2–3 times/week	22 (9.7)	17 (9.4)	1.6 (0.8–3.3)	1.7 (0.8–3.6)
4–6 times/week	20 (8.8)	14 (7.8)	1.8 (0.8–3.8)	1.6 (0.7–3.4)
Daily	108 (47.6)	53 (29.4)	2.5 (1.6–4.0)	2.4 (1.5–3.9)
Cooking using natural gas (n = 315)
Rarely cook†	77 (53.9)	96 (55.8)	Ref	Ref
2–3 times/week	15 (10.5)	14 (8.1)	1.3 (0.6–2.9)	1.3 (0.6–3.2)
4–6 times/week	5 (3.5)	17 (9.9)	0.4 (0.1–1.0)	0.4 (0.1–1.1)
Daily	46 (32.2)	45 (26.2)	1.3 (0.8–2.1)	1.1 (0.6–2.0)
Cooking oil fume (n = 612)
None	128 (38.6)	132 (47.1)	Ref	Ref
Mild (not smoky)	178 (53.6)	131 (46.8)	1.4 (1.0–2.0)	1.4 (0.97–1.9)
Heavy (smoky)	26 (7.8)	17 (6.1)	1.6 (0.8–3.0)	1.5 (0.8–3.1)
Cooking with no oil fume (n = 341)
Rarely cook†	77 (48.1)	96 (53.0)	Ref	Ref
2–3 times/week	14 (8.8)	18 (9.9)	1.0 (0.5–2.1)	0.9 (0.4–2.1)
4–6 times/week	8 (5.0)	19 (10.5)	0.5 (0.2–1.3)	0.5 (0.2–1.3)
Daily	61 (38.1)	48 (26.5)	1.6 (1.0–2.6)	1.5 (0.9–2.6)
Cooking with mild oil fume (n = 407)
Rarely cook†	77 (34.5)	96 (52.2)	Ref	Ref
2–3 times/week	26 (11.7)	16 (8.7)	2.0 (1.0–4.0)	2.3 (1.0–4.9)
4–6 times/week	17 (7.6)	18 (9.8)	1.2 (0.6–2.4)	1.1 (0.5–2.4)
Daily	103 (46.2)	54 (29.4)	2.4 (1.5–3.7)	2.3 (1.4–3.7)
Contacting liquid detergent (n = 640)
≤ 1.5 h/week	281 (82.7)	275 (91.7)	Ref	Ref
1.5–3 h/week	36 (10.6)	14 (4.7)	2.5 (1.3–4.8)	2.5 (1.3–4.9)
> 3 h/week	23 (6.8)	11 (3.7)	2.0 (1.0–4.3)	1.8 (0.8–3.9)
Contacting soap or shampoo (n = 641)
≤ 3 h/week	290 (85.0)	279 (93.0)	Ref	Ref
3–6 h/week	44 (12.9)	20 (6.7)	2.1 (1.2–3.7)	2.1 (1.2–3.8)
> 6 h/week	7 (2.1)	1 (0.3)	6.7 (0.8–55.0)	4.9 (0.6–42.3)
Using air freshener (n = 646)
Never	291 (85.1)	265 (87.2)	Ref	Ref
Occasionally	49 (14.3)	39 (12.8)	1.1 (0.7–1.8)	1.1 (0.7–1.8)
Most of the time	1 (0.3)	0 (0.0)	--	--
Daily	1 (0.3)	0 (0.0)	--	--
Average outdoor time per week (n = 633)
< 10 h	209 (63.0)	199 (66.1)	Ref	Ref
10–20 h	69 (20.8)	74 (24.6)	0.9 (0.6–1.3)	0.9 (0.6–1.3)
> 20 h	54 (16.3)	28 (9.3)	1.8 (1.1–3.0)	1.6 (0.9–2.7)
Average sitting time per day (n = 635)
< 4 h	120 (36.0)	118 (39.1)	Ref	Ref
4–8 h	149 (44.7)	120 (39.7)	1.2 (0.9–1.7)	1.2 (0.8–1.7)
> 8 h	64 (19.2)	64 (21.2)	1.0 (0.6–1.5)	0.9 (0.6–1.4)
Residential traffic flow (n = 646)
Small	97 (28.3)	84 (27.7)	Ref	Ref
Medium	165 (48.1)	146 (48.2)	1.0 (0.7–1.4)	1.0 (0.7–1.5)
Large	81 (23.6)	73 (24.1)	1.0 (0.6–1.5)	0.9 (0.6–1.5)

#Adjusted for female age (≤ 25, 25–30, 30–35, > 35, unknown), educational levels (≤ 6, 6–9, 10–12, 13–16, ≥ 17, unknown), annual income (< 30,000, 30,000–50,000, 50,000–100,000, ≥ 100,000 RMB, unknown), BMI (< 18.5, 18.5–25, ≥ 25 kg/m², unknown)

†Rarely cook: cooking ≤ 1 time/week and regardless of whether there is cooking oil fumes and using what kind of cooking fuels

* Others: coal, wood, others

Supplementary Table 1 shows the results of the univariate and multivariate logistic regression analyses using complete case analysis. These results were consistent with those of the previous analyses. Supplementary Table 2 shows the results of associations between lifestyle factors and PCOS stratified by BMI. No distinct differences were observed across the groups, which may be attributable to the limited sample size.

Discussion

Our case-control study showed that PCOS was significantly associated with using foam or plastic cups or tableware, contacting soap, shampoo, or liquid detergent, eating melons and fruits with pericarp, and drinking tap water. We also found that COFs and daily cooking with coal gas were associated with increased odds of PCOS even after controlling for potential confounders. These factors have not been systematically studied in PCOS. Previous studies have suggested that these lifestyles are potential sources of exposure to endocrine-disrupting chemicals (EDCs), which have been linked to the development of PCOS [20, 21]. Therefore, our findings may be partly explained by exposure to EDCs.

In the present study, women with PCOS were more likely to use foam or plastic cups and tableware at home. This finding is consistent with a previous study conducted in Shandong, China, which reported that the frequency of plastic tableware usage was significantly higher in the PCOS oligo-anovulation (PCOS-OA) group than in the PCOS non-ovulatory dysfunction (PCOS-non-OD) group and the healthy control group (38.30% vs. 28.10% vs. 25.40%, P < 0.01) [22]. BPA has been extensively used in consumer products, such as food plastic packages, bottles and microwave containers [23]. BPA release from these products can be accelerated at high temperatures or in an acidic environment [24]. The average BPA concentrations in food in containers covered with plastic film far exceed the maximum acceptable daily intake (50 ug/kg body weight/day) [25]. Epidemiological evidence suggested that women with PCOS have higher serum, urinary, or follicular fluid concentrations of BPA than healthy women [20]. The animal experiment reported that neonatal BPA exposure leads to adulthood PCOS-like symptoms [21]. BPA plays a substantial role in disrupting the estrogen-androgen balance. It interacts with classical nuclear estrogen receptors and nonclassical membrane estrogen receptors [21] and stimulates theca cells in the ovary to produce androgen [26]. BPA can also increase androgen levels by displacing testosterone from sex hormone-binding globulin and inhibiting its catabolism [27, 28]. BPA has the ability to activate the pulse generator of GnRH, leading to an increase in LH secretion and a decrease in FSH secretion [29]. It was suggested that BPA might stimulate insulin production in pancreatic beta cells, thereby promoting lipid accumulation and aggravating ovarian dysfunction [21, 26]. Furthermore, long-term exposure to BPA has been linked to epigenetic modification [21]. All these possible pathways might have contributed to the increased risk of PCOS.

We also found that contacting soap, shampoo or liquid detergent more frequently was associated with a higher risk of PCOS. The cleaning household products were reported to have TCS [30], and TCS can be absorbed through skin contact [31]. There is evidence that TCS exerts significant estrogenic effects and weak androgenic effects, and might affect hormone homeostasis [20]. TCS may also affect the hypothalamus-pituitary-gonad axis, which not only increases GnRH secretion by enhancing GnRH expression but also disturbs the balance of estrogen and androgen by increasing aromatase gene expression [32]. Alternation in LH, FSH, and LH/FSH index, features of PCOS, were also documented in laboratory studies [32, 33].

In our study, women who ate melons and fruits with pericarp and drank tap water had higher odds of a PCOS diagnosis. Pesticides are widely used in agriculture at various stages to protect food crops from damage by pests or weeds [34]. A large number of pesticides have been detected in the peels of fruit and vegetables, as well as in surface water and groundwater [35, 36]; however, they showed much lower concentrations and detection rates in spring water or natural mineral water [37, 38]. Epidemiological studies have shown that serum concentrations of organochlorine pesticides (OCPs) in women with PCOS are significantly higher than those in the control group [39]. The underlying biological mechanism is not clear. It has been well documented that pesticides can interfere with estrogen and androgen pathways, acting as estrogen receptor agonists or androgen receptor antagonists [40]. A laboratory study also suggested that OCPs may disrupt the morphometry of GnRH and FSH cells and, consequently, alter circulating hormone levels [40, 41].

Our study suggests that during household cooking, oil fumes and emissions from fuel combustion may be risk factors for PCOS. Chinese-cooking styles are characterized by stir-frying, which can be a main source to increase the burden of household air pollution (HAP), which is a leading cause of disability-adjusted life years and death [42]. HAP from cooking sources was associated with adverse health conditions, such as cardiovascular diseases, diabetes, and pulmonary disease [43]. In addition, pollutants from fuel combustion during cooking prolonged the time to pregnancy and increased the risk of adverse pregnancy outcomes [44]. The aforementioned study conducted in Shandong specifically investigated the association between exposure to cooking oil fumes and PCOS ovulation status. However, its findings are inconsistent with ours, as it failed to identify significant differences in exposure frequency among the PCOS-OA, PCOS-non-OD, and healthy control groups [22]. A knowledge gap remains regarding cooking and the development of PCOS.

COFs and emissions from fuel combustion contain a diverse range of harmful compositions. Polycyclic aromatic hydrocarbons (PAHs) are a classic example [45]. The total annual emissions of PAHs for a Chinese restaurant are estimated at 2,038 kg, which is much higher than those for a Western restaurant (258 kg) and a Japanese restaurant (31.4 kg) [46]. Human data showed that women with PCOS had significantly higher serum levels of PAHs than healthy women [47]. In vivo animal model showed that PAHs exposure significantly reduced the number of fetal ovarian germ cells [48]. The aromatic hydrocarbon receptor pathway is a well-established target through which PAHs stimulate follicle apoptosis, and oocyte loss may result in premature ovarian failure, potentially leading to infertility [48, 49]. Besides endocrine disrupters, other pollutants released during cooking, such as particulate matter (PM), may also be associated with the increased incidence of PCOS. Previous studies showed that the average concentration of inhalable PM (PM_2.5, 312.4 ug/m³) during cooking time was approximately 12-fold of that during non-cooking time (26.7 ug/m³) in a Chinese food stall [50]. Thus, the role of PAH and other endocrine disruptors deserves further investigation.

The main limitation of the current study is that the exposure information was self-reported by the women. It is possible that women who were diagnosed with PCOS might report exposures differently from the controls if they suspected that the exposure might affect their disease. But this may not be true in all instances [51–53]. Moreover, it was suggested that the observed association could not be due to recall bias unless participants’ beliefs changed after diagnosis [54]. For instance, as the harm of sedentary habits has been reported and physical exercise is advocated for PCOS management after diagnosis, the PCOS women may have had certain knowledge of the effects of sedentariness, which is likely to influence their answers to the question of “average sitting time per day. However, in our study, we observed comparable percentages of the average sitting time per day categorized as < 4 hours (36.0% vs. 39.1%), 4–8 hours (44.7% vs.39.7%), and > 8 hours (19.2% vs. 21.2%).

Second, the cases may tend to report more exposure if they know the study hypothesis [42]. But both the participants and the interviewer were unaware of the study’s hypothesis regarding the association between lifestyle factors and PCOS. Furthermore, our participants answered numerous lifestyle questions related to work, smoking and drinking status of individuals and their families, residential environment, and other factors in addition to household lifestyles. The questions on household and lifestyle factors were embedded among them. The associations between lifestyle factors and PCOS were unlikely to be due to recall or reporting bias.

A related, more specific methodological consideration was that our methodology did not include assessing major traditional lifestyle confounders, such as dietary intake and physical activity. In our initial case-control design, we chose not to collect these specific variables due to key methodological concerns. Unlike the household behaviors related to environmental exposures that are the focus of this study, behaviors like diet and exercise are well-established, modifiable risk factors for PCOS. Since our data were collected after diagnosis, reported behaviors could have been altered by the disease status itself (e.g., patients may change habits post-diagnosis). Measuring these factors retrospectively could thus introduce significant recall bias or reverse causality, where the disease influences the reporting of past behaviors rather than the behaviors causing the disease. Consequently, we prioritized assessing exposure-related household behaviors to test our primary hypothesis. Nevertheless, we acknowledge that these unmeasured traditional lifestyle factors remain potential sources of residual confounding.

It is also possible that after the diagnosis of PCOS, the cases had changed their behavior to avoid any suspected exposure. If this were true, our findings might have been underestimated. All these potential deficiencies call for future studies with more accurate assessments of exposures to corroborate our findings and uncover potential mechanisms.

Third, this study recruited all participants exclusively from infertility clinics, which introduces inherent selection bias. This sample cannot represent all women with PCOS; it solely represents women with PCOS-related infertility who are married and have actively sought clinical assistance. This study also did not include the following groups of women with PCOS: those with no fertility needs who are not in the preconception stage, unmarried women or those in non-heterosexual relationships, and those with occult symptoms or mild PCOS. Such limitations may limit the generalizability of the study’s findings.

Finally, the inability to clearly establish temporal associations is an inherent methodological limitation of the case-control design. Future studies should adopt multicenter, prospective designs with large sample sizes that encompass participants from diverse regions, ages, disease severities, and other relevant characteristics. This will enhance the external validity and generalizability of the findings.

Conclusions

Our study found that using foam or plastic cups or tableware, contact with soap, shampoo, or liquid detergent, eating melons and other fruits with pericarp, drinking tap water, daily cooking with coal gas, and COFs were associated with increased risks of PCOS. More research is warranted on this issue, as such findings may have important public health implications.

Abbreviations

aOR

adjusted odds ratio

BMI

body mass index

BPA

bisphenol A

CIs

confidence intervals

COFs

cooking oil fumes

CF

cooking fuel

EDCs

endocrine-disrupting chemicals

HAP

household air pollution

ORs

Odds ratios

OCPs

organochlorine pesticides

PCOS

polycystic ovarian syndrome

PAHs

polycyclic aromatic hydrocarbons

PM

particulate matter

TCS

triclosan

Authors’ contributions

Xiaona Huo: Investigation, Conceptualization, Methodology, Software, Data curation, Writing - original draft. Han Liu: Investigation, Writing - Review & Editing. Min Nan: Data curation, Writing - Review & Editing. Jun Zhang: Project administration, Funding acquisition, Supervision, Visualization, Writing - Review & Editing.

Funding

This study was supported in part by the National Natural Science Foundation of China (No. 82574101) and the National Key Research and Development Program of China (2023YFC3905203).

Data availability

Data available on request from the corresponding author.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board at Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China (approval no. XHEC-C-2015-046). A written consent form was signed by all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.