Surgically confirmed severe endometriosis is associated with lower risk of steatotic liver disease: Evidence of estrogen‐linked hepatic–reproductive crosstalk

Abstract

Objective

The prevalence of steatotic liver disease (SLD) is increasing globally. Endometriosis, affecting up to 10% of women, is a major cause of infertility. While associations between endometriosis and metabolic syndrome have been reported, its relationship with SLD remains unclear. Our study aimed to evaluate the association between endometriosis severity and SLD.

Methods

We retrospectively analyzed women who underwent surgery for endometriosis with documented revised American Society for Reproductive Medicine scores at a tertiary hospital (January 1, 2009, to February 28, 2025) and had undergone abdominal computed tomography (CT) within 6 months as part of the clinical evaluation for endometriosis. Hepatic steatosis was assessed on abdominal CT using the liver‐to‐spleen Hounsfield unit (LS) ratio with predefined cutoff values, where lower values indicate greater hepatic fat accumulation, and SLD and severe SLD were defined as LS <1.0 and <0.8, respectively. The hepatic steatosis index (HSI) was used for sensitivity analysis.

Results

Among 455 patients, 232 had severe endometriosis (SE, stage 4) and 223 had non‐severe endometriosis (NSE). Baseline characteristics were similar between groups, except for higher estradiol levels in the SE group (P < 0.001). The prevalence of SLD was significantly lower in SE (48.3%) compared to NSE (62.3%) (P = 0.002), and severe SLD was less common in SE (8.2% vs 16.1%, P = 0.009). Patients with SLD exhibited lower estradiol levels than those without, suggesting a potential protective effect of estradiol on hepatic steatosis. Multivariate analysis confirmed that SE was independently associated with a reduced risk of SLD, irrespective of age, body mass index, and metabolic comorbidities.

Conclusions

SE was associated with a lower prevalence of SLD, potentially mediated by higher estradiol levels.

Severe surgically confirmed endometriosis was associated with a lower prevalence of steatotic liver disease than non‐severe disease. Higher estradiol levels may partly explain this inverse hepatic–reproductive association, supporting integrated gynecologic and hepatologic risk assessment.

Abbreviations

ALT

alanine aminotransferase

AMH

Anti‐Mullerian Hormone

BC

blood chemistry

BMI

body mass index

CA 125

Cancer Antigen 125

CBC

complete blood count

CT

Computed Tomography

DM

diabetes mellitus

E2

Estradiol

FIB‐4

Fibrosis‐4

FSH

Follicle‐stimulating hormone

GnRHa

gonadotropin‐releasing hormone agonists

HDL‐C

high‐density lipoprotein cholesterol

HSI

Hepatic steatosis index

HTN

hypertension

LDL‐C

Low‐Density Lipoprotein Cholesterol

LS ratio

Liver‐to‐spleen Hounsfield unit ratio

MASLD

metabolic dysfunction‐associated steatotic liver disease

NAFLD

non‐alcoholic fatty liver disease

NSE

non‐severe endometriosis

ORs

odds ratios

rASRM

revised American Society for Reproductive Medicine

SDs

standard deviations

SE

severe endometriosis

SLD

Steatotic liver disease

1. INTRODUCTION

Steatotic liver disease (SLD), encompassing a spectrum of conditions characterized by excessive hepatic fat accumulation, has become a major global health concern and a leading cause of chronic liver disease. ¹ When SLD, characterized by hepatic steatosis, coexists with at least one metabolic abnormality (e.g., type 2 diabetes mellitus [DM], overweight/obesity, or other demonstrable metabolic dysregulations) the condition is defined as metabolic dysfunction‐associated steatotic liver disease (MASLD), formerly known as non‐alcoholic fatty liver disease (NAFLD). SLD in the South Korean population has been the subject of numerous investigations that consistently demonstrate a progressively increasing prevalence and an association with various comorbidities. ² , ³ Lifestyle modification is critical in the management of SLD, encompassing dietary changes, increased physical activity, and reducing alcohol consumption. ⁴ , ⁵ In addition, the augmentation of appendicular skeletal muscle mass has emerged as an important therapeutic target. ⁶ However, achieving meaningful improvement can be challenging, as it requires consistent and substantial individual effort. Recently, pharmacologic agents such as resmetirom and semaglutide have shown efficacy in improving steatohepatitis and fibrosis. ⁷ Nonetheless, their limited accessibility due to high cost highlights the importance of prioritizing patients for treatment and developing precise surveillance strategies.

Endometriosis is a benign gynecological disease characterized by the presence of endometrial glands and stroma in ectopic locations, primarily the pelvic peritoneum, ovaries, and rectovaginal septum. ⁸ It affects 6 to 10% of reproductive‐age women and up to 50% of symptomatic women with pelvic pain and is recognized as a leading cause of female infertility. ⁹ The prevalence of endometriosis in South Korea was reported to be up to 3.56 per 1000 persons in 2013. ¹⁰ Given South Korea's persistently low total fertility rate, which has been approximately 0.7 births per woman in 2023 and 2024, the importance of early detection, appropriate treatment, and diligent management of endometriosis is increasingly emphasized. When surgical intervention for endometriosis is undertaken for pain relief or diagnostic confirmation, disease severity can be objectively graded using the revised American Society for Reproductive Medicine (rASRM) classification system, which relies on intraoperative observations. ¹¹

Endometriosis has been increasingly recognized as a condition associated with systemic metabolic dysfunction. Epidemiological studies have demonstrated a significantly higher prevalence of metabolic syndrome—characterized by insulin resistance, dyslipidemia, central obesity, and hypertension—among women with endometriosis. ¹² , ¹³ Further, endometriosis patients often manifest dyslipidemia, with one case–control study showing significantly higher total cholesterol, low‐density lipoprotein cholesterol (LDL‐C) and triglycerides, and lower high‐density lipoprotein cholesterol (HDL‐C), compared to women without endometriosis. ¹⁴ Mechanistically, chronic inflammation and hormonal imbalance in endometriosis might contribute to metabolic disturbances by impairing insulin signaling and vascular function through elevated cytokines and oxidative stress. ¹⁵ These findings suggest a potential link between endometriosis and SLD.

Although a few studies have investigated the potential link between MASLD or SLD and endometriosis, none have undertaken a detailed examination of this association, particularly in relation to endometriosis severity. ¹⁶ , ¹⁷ Therefore, this study aims to evaluate the association between endometriosis and SLD according to the disease severity. Further, this study aims to explore the potential mediating role of estradiol levels in this relationship.

2. MATERIALS AND METHODS

2.1. Patients

This study included patients who underwent any surgeries in records of the rASRM endometriosis scoring system from January 2009 to February 2025 in a university‐affiliated tertiary hospital in Korea. Surgeries included ovarian cystectomy, myomectomy, and hysterectomy in either open, laparoscopic, or robotic laparoscopic methods. The patients meeting the following inclusion criteria were included in the study analysis: (1) age > 19 years, (2) patients who underwent surgical procedure due to endometriosis, and (3) patients who underwent an abdominal computed tomography (CT) scan within 6 months before surgery. In line with the Declaration of Helsinki 1975, the study protocol was approved by the Institutional Review Board of the Catholic University of Korea (IRB number: KC25RASI0499). The requirement for informed consent was waived by the Institutional Review Board due to the retrospective design of the study.

2.2. Assessment of hepatic steatosis

Hepatic steatosis was quantitatively assessed using the liver‐to‐spleen Hounsfield unit ratio (LS ratio) derived from preoperative abdominal CT scans. The LS ratio reflects relative hepatic attenuation, with lower values indicating greater hepatic fat accumulation. SLD was defined by an LS ratio less than 1.0, while severe SLD was indicated by an LS ratio below 0.8. ¹⁸ Preoperative CT scans were independently reviewed by two investigators, and the LS ratio was calculated accordingly. For sensitivity analysis, the hepatic steatosis index (HSI), a non‐invasive surrogate marker based on clinical and biochemical parameters, was additionally evaluated, with cutoff values of 30 and 36 applied to rule out and rule in SLD, respectively. ¹⁹

2.3. Endometriosis severity classification

Among various endometriosis surgical scoring tools, the rASRM and the endometriosis fertility index are most widely used worldwide. ²⁰ The scores are categorized into four stages based on intraoperative findings: I (1–5 points), II (6–15 points), III (16–40 points), and IV (over 40 points). In this study, stage IV was defined as severe endometriosis (SE).

2.4. Clinical variables collected for the analysis

Age, body mass index (BMI), and underlying diseases, including liver cirrhosis, DM, and hypertension (HTN), were collected at the time of surgery. Within 1 month prior to surgery, serum complete blood count (CBC), blood chemistry (BC) results, anti‐Müllerian hormone (AMH), and cancer antigen 125 (CA 125) levels were measured. Revised ASRM scores based on the operation records were analyzed.

2.5. Hormonal level analysis

The prescription history of hormonal medical treatments before endometriosis surgery was reviewed; it included Dienogest, gonadotropin‐releasing hormone agonists (GnRHa), and continuous‐use oral contraceptives.

Serum follicle‐stimulating hormone (FSH) and estradiol (E2) levels were measured during the early follicular phase (menstrual cycle Days 2–4), within 1 month prior to surgery. In patients who had initiated hormonal medical therapy before surgery, a washout period of 3 weeks was applied, after which serum FSH and E2 measurements were obtained within the same preoperative timeframe.

To improve the reliability of the estradiol analysis, patients with FSH levels exceeding 40 mIU/mL were excluded. Such levels are indicative of postmenopausal status, during which circulating estradiol remains consistently low regardless of endometriosis severity, thereby potentially confounding the assessment of their association. ²¹

2.6. Temporal stratification of endometriosis treatment

Treatment era was defined a priori according to the calendar year of surgery to reflect changes in clinical practice following the publication of the 2014 European Society of Human Reproduction and Embryology (ESHRE) guideline. ²² This guideline recommended hormonal therapy, including Dienogest, as a first‐line option for the management of endometriosis‐associated pain, which was accompanied by increased adoption of medical therapy in routine clinical practice. ²³ Based on this temporal framework, patients who underwent surgery between 2009 and 2015 were classified as the pre‐Dienogest era, whereas those treated from 2016 onward were categorized as the post‐Dienogest era.

2.7. Statistical analysis

All statistical analyses were performed using R software (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria; http://cran.r‐project.org; accessed July 15, 2025). Categorical variables were analyzed using either the χ ²‐test or Fisher's exact test depending on the sample size. For continuous variables, group differences were assessed using two‐sided Student's t‐tests, and results are presented as means with corresponding standard deviations. Linear regression analysis was conducted to evaluate the association among LS ratio, HSI, and endometriosis severity based on the rASRM scoring system, with correlation coefficients (r) used to quantify the strength of association. Logistic regression analysis was used to identify factors associated with SLD. A P‐value <0.05 was considered statistically significant.

3. RESULTS

3.1. Baseline characteristics

Among 2483 patients who underwent surgeries with rASRM scoring records, 455 patients were extracted with an inclusion criterion of any abdominal CT scan performed within 6 months before surgery (Figure S1). A total of 455 patients were included in the analysis and stratified into two groups: 232 patients with SE in rASRM stage IV and 223 patients with non‐severe endometriosis (NSE) in rASRM stage I to III. The association between these groups and SLD was subsequently evaluated. For the hormonal level analysis, 130 patients with serum FSH and E2 level were included.

Mean age was 36.4 ± 8.8 years old in the NSE group and 37.1 ± 8.3 in the SE group without statistical significance (P = 0.409) (Table 1). BMI difference was non‐significant between the two groups (22.1 ± 3.4 kg/m² in the NSE group and 22.0 ± 3.6 kg/m² in the SE group, P = 0.799). The proportion of patients with underlying diseases, including DM, HTN, and dyslipidemia, did not differ significantly between the two patient groups. While the two patient cohorts had comparable baseline values for most serum laboratory findings, including CBC and BC profiles, a statistically significant difference was observed in alanine aminotransferase (ALT) levels. The NSE group showed a higher baseline ALT level than SE group (18.1 ± 13.5 IU/L vs 13.7 ± 8.3 IU/L, P = 0.001). The proportion of patients with a history of hormonal medical treatment (including GnRHa, Dienogest, and oral contraceptives) was significantly higher in the SE group than in the NSE group (91.4% vs 75.3%). Baseline serum CA 125 levels were comparable between the SE and NSE groups (34.8 ± 66.2 U/mL vs 46.6 ± 73.4 U/mL, P = 0.162). Complete cul‐de‐sac obliteration observed during surgery was present in 95.7% of patients in the SE group compared with 41.7% in the NSE group, representing a statistically significant difference between the groups (P < 0.001).

TABLE 1.

Baseline characteristics.

	Total (n = 455)	NSE (n = 223)	SE (n = 232)	P‐value
Age	36.7 ± 8.6	36.4 ± 8.8	37.1 ± 8.3	0.409
BMI (kg/m2)	22.0 ± 3.4	22.1 ± 3.4	22.0 ± 3.6	0.799
DM	9 (2.0%)	4 (1.8%)	5 (2.2%)	1.000
HTN	7 (1.5%)	5 (2.2%)	2 (0.9%)	0.415
Dyslipidemia	65 (14.3%)	31 (13.9%)	34 (14.7%)	0.924
WBC (1000/μL)	6.7 ± 2.1	6.7 ± 2.1	6.7 ± 2.1	0.950
Hb (g/dL)	12.2 ± 1.3	12.3 ± 1.2	12.1 ± 1.3	0.059
PLT (1000/μL)	264.0 ± 65.2	261.7 ± 63.1	266.2 ± 67.1	0.472
Creatinine (mg/dL)	0.7 ± 0.2	0.7 ± 0.2	0.7 ± 0.2	0.365
TB (mg/dL)	0.6 ± 0.3	0.6 ± 0.3	0.6 ± 0.3	0.715
AST (IU/L)	18.1 ± 6.2	18.5 ± 6.8	17.7 ± 5.6	0.240
ALT (IU/L)	15.8 ± 12.1	18.1 ± 13.5	13.7 ± 8.3	0.001
Albumin (g/dL)	4.4 ± 0.3	4.4 ± 0.3	4.4 ± 0.3	0.299
PT (INR)	1.03 ± 0.06	1.02 ± 0.06	1.03 ± 0.06	0.523
Prior hormonal tx.	380 (83.5%)	168 (75.3%)	212 (91.4%)	<0.001
E2 (pg/mL) a	141.68 ± 123.48	95.82 ± 120.84	183.5 ± 123.91	<0.001
FSH (mIU/mL) a	8.53 ± 9.26	8.00 ± 9.37	9.11 ± 9.29	0.251
CA125 (U/mL)	40.8 ± 70.1	34.8 ± 66.2	46.6 ± 73.4	0.162
AMH (ng/mL)	3.5 ± 3.2	3.4 ± 3.0	3.5 ± 3.4	0.947
Complete cul‐de‐sac obliteration	315 (69.2%)	93 (41.7%)	222 (95.7%)	<0.001

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; DM, diabetes mellitus; HTN, hypertension; PLT, platelet count; PT, prothrombin time; TB, total bilirubin; WBC, white blood cell count.

^a Hormonal data on E2 and FSH levels were available for 130 patients. Data are presented as n (%) and mean ± standard deviation.

3.2. Comparison of steatotic liver disease prevalence between severe and non‐severe endometriosis

Based on CT scan‐derived LS ratios, the prevalence of SLD was compared between SE and NSE groups (Figure 1). The prevalence of SLD differed significantly between the two groups, with 48.3% in the SE group and 62.3% in the NSE group (P = 0.002). Similarly, the prevalence of severe SLD‐defined as an LS ratio below 0.8 was significantly higher in the NSE group (16.1%) compared to the SE group (8.2%) (P = 0.009).

FIGURE 1.

Comparison of SLD prevalence between severe and non‐severe endometriosis evaluated by LS ratios. SLD defined by LS ratio less than 1.0, and severe SLD less than 0.8. (a) Prevalence of SLD between SE and NSE. (b) Prevalence of severe SLD between SE and NSE. P‐value is indicated by asterisks: **P < 0.01. LS, liver to spleen; NSE, non‐severe endometriosis; SE, severe endometriosis; SLD, steatotic liver disease.

Additional analyses using the HSI, applying two different cutoff values (HSI >30 and HSI >36) to define SLD, yielded consistent and statistically significant differences in prevalence between the SE and NSE groups for both thresholds (Figure 2). Using the HSI >30 threshold (Figure 2a), SLD was observed in 38.8% of the SE group and 52.5% of the NSE group (P = 0.003). Further, when applying the HSI >36 threshold (Figure 2b), prevalence remained significantly higher in the NSE group (18.8%) compared to the SE group (9.9%) (P = 0.006).

FIGURE 2.

Comparison of SLD prevalence between severe and non‐severe endometriosis evaluated by HSI. (a) Prevalence of SLD using the HSI >30 threshold between SE and NSE. (b) Prevalence of SLD using the HSI >36 threshold between SE and NSE. P‐value is indicated by asterisks: **P < 0.01. HSI, hepatic steatosis index; NSE, non‐severe endometriosis; SE, severe endometriosis; SLD, steatotic liver disease.

To evaluate whether the observed association was consistent across treatment eras, endometriosis treatment was further stratified into pre‐ and post‐Dienogest eras (Figure S2). In the pre‐Dienogest era, the prevalence of both SLD and severe SLD was higher in the NSE group than in the SE group (SLD: 52.2% vs 42.0%; severe SLD: 16.4% vs 7.2%), although these differences did not reach statistical significance. In the post‐Dienogest era, the NSE group continued to show a significantly higher prevalence of SLD and severe SLD compared with the SE group (SLD: 66.7% vs 51.5%, P = 0.006; severe SLD: 16.0% vs 8.6%, P = 0.043).

3.3. Multivariate analysis of the association between severe endometriosis and steatotic liver disease

The odds ratios (ORs) for SLD according to the presence of SE were as evaluated (Table 2). In patients with SE, the OR for presence of SLD, defined by an LS ratio less than 1.0, was 0.574 (P = 0.004). In age‐adjusted model 2, the OR was 0.551 (P = 0.002). Consistent with these results, models 3 (adjusted for age, HTN, DM, and dyslipidemia) and 4 (adjusted for model 3 and BMI) also demonstrated an inverse correlation between SE and SLD, exhibiting the ORs of 0.552 and 0.556, respectively (both P = 0.003). SE showed a consistent negative association with SLD when defined by HSI at both cutoff values of 36 and 30. In patients with SE, the OR for presence of SLD with HSI cutoff over 36 was 0.474 (P = 0.007). In models 2 (age‐adjusted) and 3 (adjusted for age, HTN, DM, and dyslipidemia), the ORs were 0.468 (P = 0.007) and 0.470 (P = 0.009), respectively. Along the same line, SE showed a negative association with SLD, with an HSI cutoff over 30, presenting ORs of 0.574 (P = 0.003) in an unadjusted model, 0.557 in an age‐adjusted model (P = 0.002), and 0.561 in model 3 (adjusted for age, HTN, DM, and dyslipidemia) (P = 0.003). After adjustment for the use of hormonal therapy in model 5, the inverse association between SE and SLD remained statistically significant. Specifically, the OR for the presence of SLD defined by an LS ratio <1.0 was 0.587 (P = 0.006). Consistently, when SLD was defined using an HSI cutoff of >36; the OR was 0.471 (P = 0.008). In Model 6, which additionally adjusted for the year of surgery to account for potential temporal confounding, SE continued to show a robust negative association with SLD. The OR for the presence of SLD defined by an LS ratio <1.0 was 0.559 (P = 0.003). Similarly, SE was inversely associated with SLD defined by HSI, with ORs of 0.469 (P = 0.007) using a cutoff of >36 and 0.565 (P = 0.003) using a cutoff of >30.

TABLE 2.

Odds ratios for steatotic liver disease according to the presence of severe endometriosis.

	SLD by LS ratio of 1.0		SLD by HSI (36 as cutoff value)		SLD by HSI (30 as cutoff value)
	Odds ratio (95% CI)	P‐value	Odds ratio (95% CI)	P‐value	Odds ratio (95% CI)	P‐value
Model 1	0.574 (0.395–0.834)	0.004	0.474 (0.275–0.819)	0.007	0.574 (0.396–0.833)	0.003
Model 2	0.551 (0.377–0.805)	0.002	0.468 (0.271–0.810)	0.007	0.557 (0.382–0.812)	0.002
Model 3	0.552 (0.378–0.815)	0.003	0.470 (0.267–0.826)	0.009	0.561 (0.383–0.820)	0.003
Model 4	0.556 (0.378–0.818)	0.003	‐	‐	‐	‐
Model 5	0.587 (0.400–0.860)	0.006	0.471 (0.270–0.822)	0.008	0.620 (0.424–0.908)	0.014
Model 6	0.559 (0.383–0.816)	0.003	0.469 (0.271–0.811)	0.007	0.565 (0.389–0.822)	0.003

Note: Model 1: Unadjusted. Model 2: Age‐adjusted. Model 3: Model 2 + HTN + DM + dyslipidemia adjusted. Model 4: Model 3 + BMI adjusted. Model 5: Hormonal therapy adjusted. Model 6: Year of surgery adjusted.

Abbreviations: CI, confidence interval; HSI, Hepatic steatosis index; LS ratio, liver‐to‐spleen Hounsfield unit ratio; SLD, steatotic liver disease.

3.4. Correlation of revised American Society for Reproductive Medicine score with hepatic steatosis markers

Additionally, linear regression analyses were performed to examine the associations between rASRM score based on the operative records and the LS ratio and HSI, respectively (Figure S3). While the associations did not reach statistical significance, an increased LS ratio showed a positive relationship with rASRM scores (r = +0.007, P = 0.072), implying a potential link between endometriosis severity and SLD (Figure S3a). Conversely, higher HSI values were significantly and inversely associated with lower rASRM scores (r = −0.095, P = 0.042) (Figure S3b).

3.5. Estradiol level depending on the endometriosis severity

The difference in serum E2 levels between the SE and NSE groups was assessed (Figure S4). The mean serum E2 level was significantly higher in the SE group (183.5 pg/mL) than in the NSE group (95.82 pg/mL), indicating a potential association between higher E2 levels and the presence of SE (P < 0.001) (Figure S4a). This trend was further supported by linear regression analysis, which demonstrated a significant positive association between serum E2 levels and rASRM scores (r = +0.271, P = 0.002) (Figure S4b).

3.6. Steatotic liver disease and estradiol level

To further explore the potential mechanistic link between SLD and endometriosis severity, the association between E2 levels and the presence of SLD, classified based on a LS ratio threshold of 1, was investigated (Figure 3). Mean serum E2 levels were 122.0 pg/mL in the SLD group and 165.4 in the no SLD group showing statistically significant difference (P = 0.046) (Figure 3a). In addition, the linear regression analysis showed a consistent positive relationship between LS ratio and E2 level (r = +0.182, P = 0.039) (Figure 3b). Further adjustment for age using logistic regression showed that the odds ratio for SLD per 10‐unit increase in estradiol was 0.993 (95% CI, 0.987–1.000; P = 0.058), suggesting a trend toward a protective association, although the association did not reach statistical significance.

FIGURE 3.

Association between E2 level and SLD by LS ratio. (a) Serum E2 level difference between SLD and no SLD group (122.0 pg/mL vs 165.4 pg/mL, P = 0.046). (b) Linear regression analysis between serum E2 level and LS ratio (r = +0.182, P = 0.039). P‐value is indicated by asterisks: *P < 0.05. E2, estradiol; LS ratio, liver‐to‐spleen Hounsfield unit ratio; SLD, steatotic liver disease.

When defining SLD by HSI threshold over 30, serum E2 level was also significantly higher in the no SLD group than the SLD group (158.8 pg/mL vs 117.5 pg/mL, P = 0.047) (Figure S5a). The linear regression analysis showed a negative relationship between HSI and E2 level (r = −0.157, P = 0.074) (Figure S5b).

3.7. Correlation between endometriosis‐ and steatotic liver disease‐related variables

In this cross‐sectional cohort study, associations were evaluated between SLD‐related parameters—including LS ratio (LSR), HSI, and Fibrosis‐4 (FIB‐4) index—and endometriosis severity indicators, such as rASRM score, serum E2 level, CA‐125 level, and BMI (Figure 4). A negative correlation was observed between LSR and HSI with an r‐value of −0.25 (P < 0.001), which was expected given their opposing correlations with SLD. The rASRM score showed a significant negative correlation with HSI (r = −0.10) (P < 0.05), while demonstrating a weak positive correlation with LSR (r = +0.08). Further, preoperative serum E2 level was positively correlated to LSR (r = +0.18) with statistical significance (P < 0.05).

FIGURE 4.

Correlation matrix between endometriosis and SLD related variables. P‐value is indicated by asterisks: *P < 0.05, **P < 0.01, ***P < 0.001. BMI, body mass index; CA 125, cancer antigen 125; E2, estradiol; FIB‐4, Fibrosis‐4; HSI, hepatic steatosis index; LSR, liver‐to‐spleen Hounsfield unit ratio; rASRM, revised American Society for Reproductive Medicine.

4. DISCUSSION

In this study, we analyzed the association between endometriosis and SLD in 455 surgically confirmed patients over a 16‐year period in a university‐affiliated tertiary hospital in Korea. We performed an integrated assessment of the severity of both endometriosis and SLD, which were based on the rASRM score recorded in operative records and the abdominal CT‐based LS ratio, respectively. Our analysis revealed that endometriosis severity is inversely related to SLD. The prevalence of SLD, defined by an LS ratio less than 1.0, was lower in the SE group than the NSE group (48.3% vs 62.3%, P = 0.002). A multivariable analysis revealed a consistent inverse correlation between SE and SLD, even after adjusting for age, underlying metabolic diseases, BMI, use of hormonal therapy, and year of surgery. To our knowledge, this is the first report to suggest a potential inverse relationship between SLD and endometriosis severity.

This study provides a systematic analysis of the association between endometriosis severity and SLD, offering new insight into their relationship. Notably, our findings contrast with those of a previous study that reported a positive association between endometriosis and either SLD or metabolic syndrome. Huang et al. observed an increased risk of NAFLD in patients with laparoscopically confirmed endometriosis. ¹⁷ However, their reliance on a nurse cohort with biannual questionnaire‐based data collection raises concerns about the reliability of the NAFLD diagnostic process. In preceding studies, the association between endometriosis and metabolic syndrome has been investigated. ¹² , ¹³ Li et al. ¹³ found a positive association before adjusting for hysterectomy or oophorectomy, whereas Naz et al. ¹² reported higher prevalence of metabolic syndrome and low HDL levels in women with endometriosis, without increased odds of DM, hypertension, or dyslipidemia. In both studies, the observed associations were attributed to heightened systemic inflammation. However, both shared the limitation of defining endometriosis based solely on self‐reported prior diagnoses.

The inverse association between SLD and endometriosis severity observed in our study might be partly explained by differences in circulating estradiol (E2) levels. In our hormonal analysis, we observed that serum E2 levels were significantly lower in patients with SLD than in those without SLD. This trend was consistent whether SLD was defined by the LS ratio or the HSI. Endometriosis, defined by the presence of ectopic endometrial tissue, is associated with a dysregulation of progesterone and estrogen signaling, frequently resulting in progesterone resistance and estrogen dominance. ²⁴ , ²⁵ Based on the established protective effect of estrogen against liver fibrogenesis and a number of clinical observations, ²⁶ , ²⁷ numerous prior studies have consistently reported an increased risk of SLD in postmenopausal women. ²⁸ , ²⁹ , ³⁰ , ³¹ The study by Wang et al. provided strong support for the protective effect of endogenous estradiol by demonstrating a higher risk of SLD in women who had undergone oophorectomy compared to those who experienced natural menopause. ³¹ In our study, the relationship between endometriosis severity and serum E2 levels was sequentially examined, revealing a positive association between E2 levels and the rASRM score. We further showed a consistent inverse association between serum E2 levels and SLD in two separate analyses using the LS ratio and the HSI.

Prior experimental and clinical studies have consistently demonstrated that estrogen exerts protective effects against the development and progression of SLD by modulating hepatic lipid metabolism, suppressing inflammatory signaling, and attenuating fibrogenesis. ³² Estrogen has been shown to influence liver–adipose tissue crosstalk, enhance insulin sensitivity, and regulate mitochondrial function and oxidative stress, thereby contributing to the well‐established sex differences in SLD prevalence and severity. ³⁰ , ³³ Although estrogen dominance is not uniformly observed across all phenotypes of endometriosis, it is widely regarded as a predominant hormonal characteristic in most affected women. ³⁴ Against this biological background, our finding of an inverse association between SLD and serum estradiol levels provides supportive evidence for the role of hormonal milieu in the observed relationship between endometriosis severity and SLD. Importantly, while prior studies have largely focused on sex‐based differences in SLD outcomes, our study extends this paradigm by demonstrating that within a population of women with endometriosis, higher endogenous estradiol levels are associated with a lower burden of SLD. While findings from our study should be interpreted with caution, given the limited availability of hormonal data, potential selection bias, and the observational nature of the study, they nonetheless underscore the importance of considering metabolic risk in patients with endometriosis, particularly in those with less severe disease.

In addition to systemic hormonal effects, emerging evidence suggests that the estrobolome—the collection of gut microbiotas capable of metabolizing estrogens—might influence systemic estrogen levels and thereby contribute to the pathophysiology of endometriosis. Alterations in the estrobolome have been proposed as a potential mechanism affecting estrogen homeostasis and disease progression. However, as microbiome‐related factors were not assessed in this study, the potential contribution of the estrobolome to the observed association remains speculative and warrants further investigation. ³⁵

To assess whether temporal changes in clinical practice influenced our findings, we also performed era‐specific analysis. The observed differences between the pre‐ and post‐Dienogest eras might partly reflect temporal changes in treatment patterns. Progestins such as Dienogest have been increasingly adopted as first‐line therapy and are known to exert heterogeneous effects on lipid metabolism and insulin sensitivity, although their direct impact on hepatic steatosis remains unclear. ³⁶ , ³⁷ In our study, the inverse association between endometriosis severity and SLD was less pronounced in the pre‐Dienogest period. However, a similar inverse trend was observed across periods. Notably, this association persisted after adjustment for hormonal factors, suggesting that it is unlikely to be solely explained by progestin use. The weaker statistical significance in the pre‐Dienogest period might be related to limited sample size and reduced statistical power. Accordingly, these findings should be interpreted with caution, as they might reflect temporal changes in clinical practice and patient characteristics rather than a direct pharmacological effect of progestin alone.

This study offers several notable strengths in addition to its novelty. First, the study cohort was restricted to patients with surgically confirmed endometriosis, ensuring a higher level of diagnostic accuracy compared with studies that rely on imaging‐based diagnoses. This methodological choice also enabled classification of disease severity using the rASRM score, derived directly from intraoperative findings. Second, to our knowledge, this is the first investigation to examine the relationship between SLD and the severity of endometriosis. Rather than limiting the analysis to a simple comparison of prevalence, we adopted a more comprehensive approach by incorporating an evaluation of circulating sex hormone concentrations. To improve the precision of hormonal analysis, women with serum FSH levels exceeding 40 IU/L—consistent with postmenopausal status—were excluded from E2 comparison. Finally, the evaluation of SLD was strengthened by the combined use of both the LS ratio and the HSI, thereby improving the robustness of the assessment and reducing the risk of misclassification. Collectively, these methodological considerations enhance the validity and reliability of our findings, providing a solid foundation for further research in this area.

This study has several limitations. The retrospective cross‐sectional design precludes the establishment of causal relationships between endometriosis severity and SLD. As a single‐center study conducted in South Korea, the findings might be influenced by limited ethnic diversity, which could affect their generalizability. Further, the absence of a non‐endometriosis control group limits causal inference regarding the overall impact of endometriosis on SLD risk, as the study focused on intragroup comparisons among women with surgically confirmed endometriosis. In addition, although our cohort included some incidentally detected, low‐stage endometriosis identified intraoperatively, the study was restricted to surgically confirmed cases. Accordingly, women managed with medical therapy alone were not represented, which might introduce selection bias and limit the generalizability of our findings to the broader endometriosis population. Finally, the diagnostic process for SLD did not include hepatic ultrasound or vibration‐controlled transient elastography assessment, which could have increased diagnostic accuracy.

Despite its limitations, this study reports a previously unrecognized inverse association between endometriosis severity and SLD. The lower prevalence of SLD observed in patients with SE might be associated with higher serum E2 levels, which have been suggested to exert protective effects on the liver. The present study, by incorporating estrogen as a potential mediating factor for the association between surgically confirmed severe endometriosis and steatotic liver disease, extends the concept of a hepato–reproductive hormonal axis beyond PCOS to endometriosis. ³⁸ Our results highlight the need for integrated gynecologic and hepatologic care in the management of women with endometriosis, ensuring appropriate metabolic risk assessment and long‐term health strategies. Future prospective, multicenter studies with ethnically diverse populations and standardized hepatic assessments are warranted to confirm and expand these findings.

AUTHOR CONTRIBUTIONS

Jaejun Lee and Jung Hyun Park contributed to the study conception and design. Project administration was performed by Jaejun Lee and Mee‐Ran Kim. Jung Hyun Park collected data. Jung Hyun Park and Jaejun Lee analyzed and interpreted the collected data and wrote the manuscript. Software and visualization were done by Jaejun Lee. All authors contributed to the conceptualization, methodology, and supervision of the paper. All authors read and approved the final manuscript.

FUNDING INFORMATION

This study was supported by The Research Supporting Program of The Korean Association for the Study of the Liver and The Korean Liver Foundation (KASLKLF2025‐05) (J.L.).

CONFLICT OF INTEREST STATEMENT

All the authors declare no conflicts of interest.

ETHICS STATEMENT

Study approval statement: This study was approved by the Institutional Review Board of the Catholic University of Korea (IRB number: KC25RASI0499). The requirement for informed consent was waived by the Institutional Review Board due to the retrospective design of the study.

Supporting information

FIGURE S1. Flow chart of patient selection.

FIGURE S2. Comparison of SLD and severe SLD prevalence between severe endometriosis (SE) and non‐severe endometriosis (NSE) across treatment eras.

FIGURE S3. Association between endometriosis severity and SLD.

FIGURE S4. Association between E2 level and endometriosis severity.

FIGURE S5. Association between E2 level and SLD by HSI.

DATA AVAILABILITY STATEMENT

The original contributions presented in the study are included in the article and supplementary materials. Further inquiries can be directed at the corresponding authors.