Hepatic steatosis and fibrosis are associated with metabolic factors rather than endocrine therapy in breast cancer patients.

Introduction
Steatotic liver disease (SLD) is a recently proposed overarching term that encompasses the entire spectrum of liver conditions characterized by hepatic steatosis. This includes metabolic dysfunction-associated steatotic liver disease (MASLD), metabolic dysfunction and alcohol-related liver disease (MetALD), alcohol-related liver disease (ALD), and secondary causes of steatosis, such as drug-induced liver injury (1). MASLD, formerly known as nonalcoholic fatty liver disease (NAFLD), is estimated to affect approximately 38% of the global adult population, with projections suggesting an increase to over 55% by 2040 (2,3). Recent studies have reported a high prevalence of NAFLD among patients with breast cancer (4,5). Postmenopausal women often experience metabolic changes such as increased visceral adiposity, dyslipidemia, and glucose intolerance, which may contribute to the pathogenesis of SLD. Endocrine therapy for breast cancer has also been proposed as an independent risk factor for SLD (6,7). It is estimated that 40–50% of patients using tamoxifen—a selective estrogen receptor modulator—develop SLD (6,8-10). This finding was recently corroborated by a meta-analysis, especially in patients with a high body mass index (BMI) and hypercholesterolemia (11), although cardiometabolic risk factors can be associated with SLD independently from selective estrogen receptor modulators (12). Furthermore, other modalities of hormone therapy, such as androgen inhibitors, have also been associated with SLD occurrence (6,13).
The prevalence of progressive forms of SLD is still unknown, as the frequency of steatohepatitis and staging of fibrosis have not been properly investigated in prior studies. Given that hepatic decompensation occurs almost exclusively in the setting of cirrhosis, identifying patients with advanced fibrosis is critical (14). Therefore, the clinical relevance of SLD in patients with breast cancer remains uncertain. This information is essential to inform recommendations regarding SLD screening and to identify high-risk SLD subgroups among patients receiving endocrine therapy. With recent advances in breast cancer diagnosis and treatment leading to improved survival, evaluating the long-term impact of liver disease has become increasingly relevant for comprehensive patient management.
This study aims to evaluate the prevalence and severity of SLD in women with breast cancer and to investigate pharmacological and cardiometabolic risk factors associated with SLD occurrence and progression in this population. We present this article in accordance with the STROBE reporting checklist (available at https://tgh.amegroups.com/article/view/10.21037/tgh-25-48/rc).

Methods

Study design and participant selection
A cross-sectional study was conducted at the oncology outpatient clinic of Hospital das Clínicas from the Federal University of Minas Gerais, a tertiary University hospital in Brazil. Female patients older than 18 years with a histopathological diagnosis of breast cancer were prospectively enrolled. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The Federal University of Minas Gerais Ethics Committee Board approved the study (No. CAAE 48790721.5.0000.5149), and individual informed consent was obtained from all patients.
Data collected included demographic information, follow-up time since cancer diagnosis, metabolic comorbidities, and laboratory exams (i.e., transaminases, serum albumin, international normalized ratio, total cholesterol and lipoproteins, triglycerides and complete blood cell count). Patients were divided into four groups: (I) patients not exposed to hormone therapy at the time of enrollment (because it was not indicated, i.e., absence of hormone receptor in cancer immunohistochemistry, or because treatment had no yet been initiated); (II) patients exposed only to tamoxifen, either currently or previously; (III) those exposed only to anastrozole, either currently or previously; and (IV) those exposed to both drugs—hormone therapy modality was changed due to adverse effects or intolerance. Duration of hormone therapy was recorded and only patients who had received at least 3 months of hormone therapy were included in groups 2, 3 and 4. Metabolic syndrome was defined based on established criteria (15). All patients underwent abdominal ultrasound performed by a single examiner to assess hepatic steatosis and liver morphology. Ultrasonographic features suggestive of steatosis included increased hepatic echogenicity relative to the renal cortex, posterior attenuation of the ultrasound beam, and focal areas of fatty sparing (16). Grading the severity of steatosis by ultrasound was not conducted, as this modality is not reliable for quantification and has limited clinical utility (16). Transient elastography (Fibroscan®) was performed by trained examiners to assess liver stiffness. The severity of SLD was defined by the presence of advanced fibrosis. Liver stiffness thresholds of ≥8 and ≥12 kPa were used to define advanced fibrosis and cirrhosis, respectively, in accordance with previously validated criteria (14). Exclusion criteria were as follows: ALD, chronic viral hepatitis B or C, current or past diagnosis of other cancer, metastatic breast cancer, right heart failure, alcohol intake >20 g/day, aminotransferases >5 times the upper limit of normal, inability to perform elastography, refusal to undergo study’s procedures and use of other drugs known to be associated with SLD, such as amiodarone, valproic acid and methotrexate. In addition to those mentioned, other causes of liver disease were not routinely investigated prior to inclusion in the study.

Transient elastography evaluation
Liver stiffness measurements (LSMs) were performed using a Fibroscan® 402 (Echosens, Paris, France). Experienced operators performed the exam in patients with at least three hours of fasting. Subjects assumed the dorsal decubitus position with the right arm in maximal abduction. Probe size M or XL was used, as appropriate, according to BMI status, and placed over the right liver lobe through intercostal spaces. The exam was considered adequate when, for each liver evaluation, 10 valid measurements were performed with a success rate ≥80% and an interquartile range (IQR) <30% of the median value. LSM was defined by the median value of the obtained measurements.

Prospective evaluation
A sub-cohort of 35 patients continuously treated with hormone therapy was prospectively followed for approximately 24 months to evaluate liver stiffness dynamics during endocrine treatment, as well as any associated anthropometric changes.

Statistical analysis
Analyses were performed using the software SPSS version 23.0, Armonk, NY. The Shapiro-Wilk test was used to assess continuous variables for normal distribution, and those with Gaussian distribution were expressed as mean ± standard deviation, whereas those with skewed distribution were described as median and IQR. Pairwise deletion was applied to missing data. Categorical variables were expressed as absolute numbers and percentages. For univariate analysis of categorical variables, we utilized the chi-square or Fisher’s exact test, as appropriate. For continuous variables, we compared the data using the analysis of variance (ANOVA) or Student t-test if the distribution was normal, and the Kruskal-Wallis or Mann-Whitney test for skewed data. For paired analysis in prospective evaluation, paired Student t-test or Wilcoxon test were utilized, as appropriate, for continuous variables, or McNemar test for categorical analysis. Variables with clinical relevance that had P value <0.20 in univariate analysis were selected for logistic regression multivariate analysis by the backwards method. Absence of multicollinearity between independent variables was a prerequisite for logistic regression. Variables included in the multivariate analysis of factors associated with fatty liver were: age, diabetes, arterial hypertension, dyslipidemia, central obesity, alanine aminotransferase, gamma-glutamyl transferase and hormonotherapy exposure time. For the multivariate analysis of factors associated with advanced fibrosis, variables included were: metabolic syndrome, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, alkaline phosphatase and hormonotherapy exposition time. Odds ratio (OR) and 95% confidence interval (CI) were reported. P values <0.05 were considered significant.

Results
The initial cohort consisted of 235 patients; however, 11 were excluded due to technical difficulties with elastography, and 53 declined to participate in imaging examinations for the assessment of steatosis. Ultimately, 171 patients, with a mean age of 58±10 years, were enrolled. The follow-up period ranged from 1 to 315 months (median 53 months; IQR, 25–102 months). Common comorbidities included diabetes mellitus (26.9%), arterial hypertension (53.2%), dyslipidemia (31.0%), and central obesity (70.2%). There was no patient in use of thiazolidinediones nor glucagon-like peptide 1 agonists. Forty-five (26.3%) patients were using statins. A detailed description of cohort characteristics is provided in Table 1.
In total, 55 (32.2%) patients did not receive hormone therapy. Of the remaining participants, 72 (42.1%) were exclusively exposed to tamoxifen for a median duration of 51 months (IQR, 23–60 months), 16 (9.4%) exclusively to anastrozole for a median duration of 25 months (IQR, 22–32 months), and 28 (16.4%) received both drugs for a median duration of 57 months (IQR, 39–65 months).
Of the 116 subjects exposed to hormone therapy, 51 (44.0%) were no longer receiving treatment at the time of evaluation. This subgroup had been treated for a median duration of 60 months (IQR, 40–68 months), with a median time since discontinuation of 25 months (IQR, 11–48 months). The proportion of participants who had discontinued hormone therapy varied by treatment group: 36.1% among those exclusively exposed to tamoxifen, 18.8% among those exclusively exposed to anastrozole, and 78.6% among those exposed to both drugs.

Liver steatosis was prevalent and associated with metabolic risk factors
Liver steatosis was identified in 57.9% of the patients, with no statistically significant differences across the four groups (P=0.09), as shown in Figure 1. In univariate analysis comparing women with and without steatosis (Table 2), several factors were significantly associated with the presence of SLD: diabetes (37.4% vs. 12.5%, P<0.001), hypertension (63.6% vs. 38.9%, P=0.001), dyslipidemia (38.4% vs. 20.8%, P=0.01), metabolic syndrome (53.5% vs. 36.1%, P=0.02), BMI (31 vs. 28 kg/m2, P=0.001), waist circumference (99 vs. 90 cm, P<0.001), central obesity (80% vs. 40%, P<0.001), liver stiffness (5.9 vs. 4.6 kPa, P<0.001), alanine aminotransferase (23 vs. 19 U/L, P=0.03), gamma-glutamyl transferase (36 vs. 28 U/L, P=0.002) and triglycerides (146 vs. 123 mg/dL, P=0.009). Among patients who had discontinued hormone therapy at the time of evaluation, a shorter time since treatment suspension was significantly associated with the presence of steatosis [14 (IQR, 9–34) vs. 31 (IQR, 20–56) months, P=0.047]. In a multivariate analysis, diabetes remained independently associated with steatotic liver, even after adjusting for the duration of exposure to hormone therapy (OR =5.89, 95% CI: 1.88–18.42, P=0.002). Figure 2 highlights the relevance of diabetes association with steatotic liver. Among the 17 patients without cardiometabolic risk factors (1), seven had not been exposed to hormone therapy, seven had received tamoxifen only, one had received anastrozole only, and two had been exposed to both tamoxifen and anastrozole. The median duration of hormone therapy exposure in this subgroup was 39 months (IQR, 6–60 months). Liver steatosis was detected in only one patient, who had not received hormone therapy.

Advanced liver fibrosis was related to metabolic risk factors rather than hormonotherapy
Liver stiffness did not significantly differ across groups, with a median value of 5.4 kPa (P=0.20). Overall, 12.3% of patients exhibited liver stiffness ≥8 kPa (P=0.57), and 5.8% of patients had liver stiffness ≥12 kPa (P=0.18), as shown in Figure 1. In a univariate analysis comparing women with liver stiffness ≥8 kPa to those with <8 kPa (Table 3), factors associated with advanced fibrosis were: diabetes (47.6% vs. 24%, P=0.02), metabolic syndrome (71.4% vs. 42.7%, P=0.02), BMI (33 vs. 29 kg/m2, P=0.005), waist circumference (105 vs. 95 cm, P=0.001), central obesity (90.5% vs. 67.3%, P=0.03), aspartate aminotransferase (32 vs. 26 U/L, P=0.02), alanine aminotransferase (26 vs. 20 U/L, P=0.01) and gamma-glutamyl transferase (47 vs. 33 U/L, P=0.02). In a multivariate analysis, metabolic syndrome remained independently associated with advanced fibrosis, even after adjusting for hormone therapy duration (OR =4.63, 95% CI: 1.39–15.48, P=0.01). Figure 2 highlights the association between metabolic syndrome and liver fibrosis. As a sensitivity analysis, patients were also grouped by hormone therapy exposure (Table S1). No significant differences were observed between exposed and non-exposed groups, apart from a longer follow-up duration among those who received hormone therapy.

No significant dynamic changes in liver stiffness were observed after two years of endocrine therapy
A prospective evaluation was conducted in 35 patients undergoing hormone therapy. Treatment history included tamoxifen (n=20), anastrozole (n=1), and both agents sequentially (n=14). At baseline, comorbidities included diabetes (14.3%), hypertension (45.7%), dyslipidemia (28.6%), hypertriglyceridemia (40.0%), low HDL-cholesterol (22.9%), and central obesity (65.7%). Liver steatosis was present in 19 patients (54.3%) and advanced fibrosis in 2 patients (5.7%).
After a median follow-up of 27 months (IQR, 25–28 months), paired analysis showed a modest but statistically significant reduction in BMI (28.9 vs. 28.0 kg/m2, P=0.003) and waist circumference (92.5 vs. 91.3 cm, P=0.04). No significant change was observed in liver stiffness (5.6 vs. 5.3 kPa, P=0.47). A ≥10% increase in liver stiffness was observed in 12 patients (34.3%), and a ≥20% increase in 10 patients (28.6%). Conversely, 12 patients (34.3%) had a ≥10% decrease, and 6 patients (17.1%) had a ≥20% decrease. No clinical, anthropometric, or laboratory parameter was associated with liver stiffness variation.
During follow-up, medication adjustments included tamoxifen discontinuation (n=3; at 5, 6, and 8 months) and switching from tamoxifen to anastrozole (n=3; at 4, 6, and 12 months). Additional details on the baseline and follow-up characteristics of this subgroup are provided in Tables S2,S3.

Discussion
In this study, over 50% of patients with breast cancer had SLD, and around 10% had advanced fibrosis. Diabetes and metabolic syndrome were independently associated with the diagnosis of steatotic liver and advanced fibrosis, respectively, even after adjustment for hormone therapy exposure. These findings suggest that metabolic risk factors may play a more significant role in SLD progression than hormone therapy itself.
Breast cancer is the most frequent non-skin tumor in Brazilian women, and an important cause of death (17). The expression of estrogen receptor, observed in approximately 75% of breast cancers, serves as a valuable prognostic indicator, guiding targeted hormone therapy decisions (18,19). Endocrine therapy is tailored differently for pre- and post-menopausal women, with aromatase inhibitors being the preferred choice post-menopause, unless contraindicated or poorly tolerated (20). For premenopausal women with a low recurrence risk, a 5-year therapy involving tamoxifen is recommended, while those at a high risk may be prescribed an extended 10-year therapy (21). Given the high prevalence of breast cancer and the potential for long follow-up of survivors, investigating the hepatic effects of endocrine therapy assumes critical importance.
Previous studies have demonstrated that up to 50% of patients using tamoxifen also exhibit steatotic liver (6,8-10), a finding consistent with the observations in this study. Risk factors previously described align closely with our findings: dyslipidemia, altered fasting glucose (10,22), high BMI and visceral fat (11,22,23). While the literature remains uncertain, it is hypothesized that tamoxifen may play a multifactorial role in the development of drug-induced SLD, involving increased synthesis of fatty acids and triglycerides, inhibition of mitochondrial fatty acid beta-oxidation, and suppression of estrogen synthesis. Additionally, enzymatic and protein polymorphisms associated with steatotic liver in tamoxifen users are described in genes related to its pharmacogenetics (CYP2D6, CYP1, OATP1B1, SREBP-1, and UGT) (24-28). However, it is still undefined whether tamoxifen has a relevant impact on SLD development. Even though tamoxifen is the most investigated drug, aromatase inhibitors have also been assessed for their liver-related side-effects, with a lower incidence of steatotic liver compared to tamoxifen (6). Previous studies have suggested that hormone therapy may increase the risk of developing cardiometabolic comorbidities in women with breast cancer, as estrogen is believed to exert a protective effect against metabolic disorders (29,30). Aromatase inhibitors have been associated with an increased risk of diabetes, dyslipidemia, and hypertension, while tamoxifen may also contribute to dyslipidemia and elevated blood pressure. These metabolic adverse effects appear to be more common among postmenopausal women. Previous prospective studies have shown that most cases of fatty liver associated with endocrine therapy develop within the first 2 years of treatment, and that hepatic steatosis often resolves within 1 to 2 years after discontinuation of therapy (8,9,31). Consistently, our study found that a shorter time since hormone therapy suspension was associated with the presence of steatosis. However, a recent prospective study using controlled attenuation parameter to measure liver steatosis at 3 months and 2 years after initiating tamoxifen therapy demonstrated a stable prevalence of steatosis at 48–51%, with no significant increase during follow-up (32). These findings strongly align with our results and reinforce the notion that patients’ underlying metabolic status may play a more pivotal role in the development and persistence of steatotic liver than hormone therapy exposure alone.
The impact of cardiometabolic risk factors on the development of liver steatosis, independently of hormonotherapy, has been poorly addressed in the literature, largely due to methodological variability across studies. Most prior investigations have assessed the association between SLD and cardiometabolic risk factors exclusively in patients receiving hormone therapy, without including unexposed control groups (10,11,22,31,33,34). A randomized clinical trial has compared patients exposed to tamoxifen versus placebo; however, only evaluated subgroups with altered aminotransferases (7). In a multivariate analysis, Yang et al. (12) found that baseline BMI, total cholesterol, and treatment with selective estrogen receptor modulators were independently associated with the development of SLD. In our study, among participants without any cardiometabolic risk factors, only one patient had liver steatosis, and, notably, she had not been exposed to hormone therapy. These findings reinforce the prominent role of metabolic risk in SLD pathogenesis and highlight the importance of including appropriate control groups in future studies.
Also, breast cancer patients are commonly older women who present diseases closely related to MASLD (i.e., diabetes, hypertension and obesity) (4,5), since the perimenopausal period is associated with metabolic changes that favor fat storage, dyslipidemia and glucose intolerance. These changes are possibly associated with the reduction of estrogen protective effects (35) and are the reason why in this population, SLD is so prevalent. Moreover, common habits such as alcohol consumption can contribute to hepatic fat accumulation, inflammation, and fibrosis, or potentially exacerbate pre-existing conditions, as ethanol metabolism shares several pathogenic pathways with both alcoholic and metabolic liver disease—whether increasing the risk of liver damage with moderate to excessive consumption or with potential protective effects with mild consumption (i.e., <20 g/day) (14,36).
Steatohepatitis represents the form of SLD with a higher potential for progression to chronic liver disease. However, there is a notable absence of data on the prevalence of steatohepatitis among breast cancer patients undergoing hormone therapy. Attempts have been made to estimate this prevalence, but they often rely on the elevation of aminotransferases to indicate steatohepatitis (7,37). It is now recognized that such an approach is inadequate, as a substantial number of patients with steatohepatitis and fibrosis may present with normal liver enzyme levels (38).
The prediction of liver fibrosis is arguably the most critical component in the assessment of SLD, as the fibrosis stage is the primary determinant of long-term outcomes (38,39). Hepatic decompensation, portal hypertension, and hepatocellular carcinoma occur almost exclusively in patients with cirrhosis, making the identification of significant fibrosis a key clinical objective (14). Among noninvasive diagnostic tools, hepatic elastography stands out as the most utilized and reliable method for predicting liver fibrosis. In our study, LSMs were similar between patients exposed and unexposed to hormone therapy (40). This finding contributes to the emerging understanding that hormone therapy exposure may not be the primary factor in SLD development and progression, as was also demonstrated by Braal et al. (32) Conversely, our results underscore the central role of metabolic syndrome as a potential contributor to liver fibrosis. Furthermore, in a prospective subgroup of 35 patients followed for 24 months, no significant dynamic changes in liver stiffness were observed over time, further suggesting that endocrine therapy may not impact the progression of fibrosis or hepatic inflammation.
A subgroup of four patients without detectable SLD on imaging presented with liver stiffness values >8 kPa. Their baseline characteristics were unremarkable for underlying liver disease, with ages ranging from 44 to 68 years. One patient had not been exposed to hormone therapy, while the others had exposure durations ranging from 59 to 60 months. Two patients had no identifiable cardiometabolic risk factors, although two were overweight. No cholestatic laboratory abnormalities were observed. Potential explanations for elevated liver stiffness in the absence of steatosis include undetected fatty liver not captured by ultrasound, hepatic inflammation due to recent transient insult or subclinical systemic disease, non-compliance with fasting prior to elastography, unknown cardiac disease, or even the presence of undiagnosed or emerging liver metastases.
This study has some inherent limitations. Firstly, there was heterogeneity in the distribution of patients among various endocrine therapy modalities, largely influenced by treatment indications. As a result, women treated with anastrozole were generally older. Therefore, future studies with larger and more homogeneous samples are warranted to validate our findings. Second, although abdominal ultrasound is recommended by international guidelines as the initial imaging modality for detecting hepatic steatosis—with reported sensitivity of 85–94% for moderate to severe steatosis—its sensitivity decreases to approximately 70% for mild cases, potentially leading to underestimation of the true prevalence (16). Additionally, ultrasound is subject to inter-examiner variability, which may further affect diagnostic accuracy. There was also a discrepancy in follow-up duration between patients receiving hormone therapy and those in the non-exposed group. This difference arises from the inclusion of patients with a recent cancer diagnosis who had not initiated hormone therapy in the non-exposed group. Moreover, the smaller number of patients treated exclusively with anastrozole may reflect national public healthcare protocols, which permit the use of tamoxifen for both pre- and postmenopausal women, thereby limiting anastrozole prescriptions. Finally, the limited sample size and relatively short follow-up period in the prospective cohort restrict our ability to draw definitive conclusions regarding dynamic changes in liver stiffness over time.

Conclusions
In conclusion, among breast cancer women, more than half had SLD, and approximately 10% had advanced fibrosis by elastography. Common metabolic risk factors were independently associated with the diagnosis of liver steatosis and advanced fibrosis, even after hormone therapy exposure adjustment. These findings underscore the predominant role of metabolic dysfunction in SLD development and progression within this population.

Supplementary
The article’s supplementary files as