The Relationship Between Tamoxifen-associated Nonalcoholic Fatty Liver Disease and the Prognosis of Patients With Early-stage Breast Cancer

Accepted Manuscript The relationship between tamoxifen-associated nonalcoholic fatty liver disease and the prognosis of early stage breast cancer patients Meiying Yan, Jingxuan Wang, Qijia Xuan, Tieying Dong, Juan He, Qingyuan Zhang PII:

S1526-8209(16)30569-9

DOI:

10.1016/j.clbc.2016.12.004

Reference:

CLBC 552

To appear in:

Clinical Breast Cancer

Received Date: 26 July 2016 Revised Date:

16 December 2016

Accepted Date: 16 December 2016

Please cite this article as: Yan M, Wang J, Xuan Q, Dong T, He J, Zhang Q, The relationship between tamoxifen-associated nonalcoholic fatty liver disease and the prognosis of early stage breast cancer patients, Clinical Breast Cancer (2017), doi: 10.1016/j.clbc.2016.12.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT The relationship between tamoxifen-associated nonalcoholic fatty liver disease and the prognosis of early stage breast cancer patients Meiying Yan1, Jingxuan Wang1, Qijia Xuan1, Tieying Dong, Juan He, Qingyuan Zhang*

University, Harbin 150040, China.

Meiying Yan, Jingxuan Wang are the first authors.

SC

1

RI PT

Department of Medical Oncology, The Third Affiliated Hospital of Harbin Medicial

Corresponding author:

M AN U

Qingyuan Zhang, M.D., Ph.D., Department of Medical Oncology, The Third Affiliated Hospital of Harbin Medicial University, Harbin, 150040, China. Tel & Fax: +86-(0) 451-86298336, E-mail address:

AC C

EP

TE D

[email protected].

ACCEPTED MANUSCRIPT MicroAbstract Total 646 eligible breast cancer patients enrolled in our retrospective study in Tumor Hospital of Harbin medical university. Finally, we found that NAFLD development might have a negative effect on DFS of early-stage breast cancer patients during the treatment of tamoxifen. The risk score we created

Abstract

RI PT

might be used for decision making of treatment.

Purpose: To investigate the relationship between tamoxifen-associated non-alcoholic fatty liver disease (NAFLD) and survival outcomes in breast cancer patients.

SC

Methods: Patients with early-stage invasive breast cancer after curative resection from Jan 2009 to June 2011 were selected. 646 patients taken tamoxifen were included finally. Patients diagnosed with

M AN U

NAFLD on ultrasonography were classified into the NAFLD and non-NAFLD groups. Results: Finally, the NAFLD group included 221 patients and the non-NAFLD group included 425 patients. Patients in the NAFLD group had significantly higher BMI than those in the non-NAFLD group (P < 0.001). DFS was significantly longer in the non-NAFLD group than the NAFLD group (P = 0.006). However, there were no significant statistical differences between these two groups on OS (P =

TE D

0.387). As regards to body mass index (BMI), total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), Alanine aminotransferase (ALT) and high-density lipoprotein cholesterol (HDL-C), the optimal cutoff point was 21.06, 4.28, 1.22, 3.13, 27.50 and 1.29, respectively,

EP

which can be identified as risk factors for distinguishing NAFLD developed patients from non-NAFLD developed patients (P < 0.05). Moreover, a risk score ≥ 3 was at a high risk of development of NAFLD

AC C

(OR = 3.03, CI 1.11 - 8.28, P = 0.037). Conclusion: NAFLD development had a negative effect on survival outcomes of breast cancer patients. The risk score created ≥ 3 had a high-level risk of developing NAFLD and it might be used for physician to evaluate each patient and give instructive advice to the further treatment. Keywords: Tamoxifen, NAFLD, Breast cancer Abbreviations: BMI = body mass index, CI = confidence interval, DFS = disease-free survival, ER = estrogen receptor, HER2 = human epidermal growth factor receptor 2, HR = hazard ratio, NAFLD = nonalcoholic fatty liver disease, OR = odds ratio, OS = overall survival, PR = progesterone receptor, HR+ = hormone receptor +, NASH = non-alcoholic steatohepatitis, TC = total cholesterol, TG = triglyceride, LDL-C = low-density lipoprotein cholesterol, HDL-C = high-density lipoprotein

ACCEPTED MANUSCRIPT cholesterol, AST = Aspartate aminotransferase, ALT = Alanine aminotransferase. Introduction Breast cancer has been proved to be one of the most frequent causes of cancer death for women both in developing and developed countries.1 The majority of female patients with breast cancer

RI PT

(approximately 70% worldwide) have hormone receptor-positive (HR+) tumors.2 Tamoxifen is a selective oestrogen receptor modulator and binds to oestrogen receptor as partial agonist or antagonist in a manner depending on target tissue.3 It is one of the most widely used endocrine therapies and has been proved to be effective in both early and advanced stages of breast

SC

cancer.4 More than 67% breast cancer patients have been reported to be sensitive to tamoxifen therapy.5,6 However, adjuvant endocrine therapies for breast cancer span at least 5 years, creating

M AN U

long-lasting side-effects, which often results in non-adherence or change of medication.7 Severe side effects are unusual with tamoxifen, but it is associated with an increased risk of thromboembolism, endometrial cancer, vaginal bleeding and other adverse reactions, such as development of NAFLD.8 NAFLD covers a spectrum of chronic liver diseases, ranging from hepatic steatosis or fatty liver to non-alcoholic steatohepatitis (NASH), liver fibrosis, liver cirrhosis and eventually hepatocellular

TE D

carcinoma (HCC).9 It has been shown that 43% breast cancer patients treated with tamoxifen develop steatosis within the first two years of treatment.10 An Italian multi-centre prospective, randomised, double-blind, placebo-controlled trial of 5408 women with breast cancer, in which half of the women

EP

were assigned to tamoxifen 20 mg and the other half to placebo for 5 years, showing that tamoxifen was associated with higher risk of developing NASH only in overweight and obese women with characteristics of metabolic syndrome and other predictors of NAFLD include severe baseline

AC C

hypercholesterolaemia.11 However, the relationship between NAFLD development and survival outcomes for breast cancer patients has not been reported. It is believed that long-term use of tamoxifen is associated with various complications, including

hypoglycemia, hypertriglyceridemia, changes in plasma cholesterol levels and liver diseases, such as NAFLD.12 Nevertheless, the pathogenesis and risk factors of tamoxifen-induced NAFLD are still not fully understood. During tamoxifen treatment, unexplained multiple elevations of alanine aminotransferase ≥1.5 times the upper normal limit are associated with steatohepatitis.11 So far, a large cohort study of 2425 female with non-metastatic invasive breast cancer has demonstrated that patients with the most unfavorable liver function tests levels had worst 5-year OS.13 Moreover, a study of 520

ACCEPTED MANUSCRIPT women with early-stage breast cancer found that potential weak associations were seen in multivariate analysis between TC and LDL-C and recurrence.14 However, beyond that, obesity has been shown in a majority of studies to have exerted adversely influence on the prognosis of breast cancer in both preand post-menopausal women.15,16 A subsequent retrospective study proved that breast cancer survivors

RI PT

with a higher BMI who also had elevated total cholesterol levels at the time of mastectomy had a shorter 5-year survival relative to those with higher BMI alone, implying that cholesterol added further prognostic information in those with an elevated BMI.9 Thus, the relationship between BMI, lipid profiles and liver function tests and NAFLD development has attracted our attention.

SC

In this retrospective study, we attempted to evaluate the relationship between NAFLD development and survival outcomes of breast cancer patients. Moreover, factors of predicting the

M AN U

outcome of NAFLD would be also identified and we investigated the risk factors within a comprehensive risk index through using independent variables to stratify patients developing NAFLD. Material and methods Patients

This was a retrospective study by searching electronic medical records of patients with surgery

TE D

from January 2009 to June 2011. Our study protocol was approved by the Ethics Committee of the Tumor Hospital of Harbin Medical University. Written consent was obtained from each patient. Their information was collected in the hospital database and only used for the purpose of research. The

EP

inclusion criteria include (1) immunohistochemically confirmed ER and/or PR-positive and HER2 negative breast cancer, (Studies have shown that amplification and/or overexpression of the HER2 is

AC C

associated with a poor DFS rate in breast cancer patients, especially axillary node-positive patients.17 Thus, we chose HER2 negative patients.) (2) normal baseline liver function test, (3) negative antihepatitis C virus (HCV) antibodies and hepatitis B surface antigen (HbsAg) tests and (4) treated with tamoxifen at least 3 months. Exclusion criteria include (1) if they had a history of liver diseases, evidence of liver disease on physical examination, (2) history of alcohol abuse or other hepatotoxic drugs, (3) had a second primary cancer, (4) bilateral primary breast cancer, (5) stage Ⅳ at diagnosis. At last, 646 HR+ and HER2- female patients with breast cancer after surgery were included. None of them had NAFLD according to baseline ultrasonography. The inclusion and exclusion criteria are shown in Fig. 1. Study design

ACCEPTED MANUSCRIPT In this study, the clinical endpoints were disease-free survival (DFS) and overall survival (OS). DFS was defined as the time from the date of surgery until the date of the earliest documentation of disease recurrence, metastasis, or death from any cause, or until the end of follow-up. OS was defined as the time from the date of surgery until the date of death from any cause, or until the end of follow-up.

RI PT

Clinicopathological characteristics were evaluated, including age at diagnosis, body mass index (BMI), tumor size, nodal status, ER, PR, P53 and Ki-67 expression and treatment regimen (neoadjuvant/adjuvant chemotherapy, radiotherapy). BMI was defined as weight/height2 (kg/m2). Diagnosis of NAFLD

SC

All patients underwent abdominal ultrasonography, laboratory tests (liver function tests and lipid profiles) at baseline and at least once at the annual follow-up visit. At the beginning of the first year of

M AN U

treatment, patients take systematic medical examination every 3 months while at the second year, patients take systematic medical examination every 6 months. Ultrasonography was performed by experienced sonographers who were blinded to the study aims and data. The diagnosis of NAFLD was based on an increase in liver echogenicity, discrepancies between hepatic and renal echoes, and echo loss from the portal vein walls.18

TE D

Statistical analysis

Statistical analysis was performed through using chi-square tests to assess the differences in the frequency of the outcomes. The Kaplan–Meier method was adopted to calculate the cumulative

EP

survival rate and to plot DFS and OS curves. The log-rank test was used to compare the differences in DFS and OS between the study groups. Univariate Cox regression analysis was used to identify

AC C

prognostic factors which are associated with DFS. The multivariate Cox proportional hazards model with a forward stepwise selection of significant univariate factors or probably important confounding variables was applied to determine which factors acted as independent prognosticators. To further evaluate the prediction performance of BMI, blood lipid parameters and liver function level, receiver operating characteristic (ROC) analysis was performed. Univariate and multivariate Cox analysis were used to confirm factors related to nonalcoholic fatty liver disease development. Multivariate logistic regression models were created to analyze risk scores associated with NAFLD development. In this work, all statistical analyses were conducted using SPSS 20.0 for Windows. A probability value less than 0.05 (P < 0.05) was considered significant. Results

ACCEPTED MANUSCRIPT Patients and tumor characteristics Baseline characteristics of patients for both groups are shown in Table 1. The median time from diagnosis was 67 months (ranging from 7 to 91 months). All patients had been programmed to receive oral tamoxifen (14 patients received ovariectomy treatment and 8 patients taken medical ovarian

RI PT

ablation in addition to tamoxifen were included). Among the 646 participants who were diagnosed with NAFLD on ultrasonography, 221 of them were classified into the NAFLD group and 425 in the non-NAFLD group. The rate of developing NAFLD after tamoxifen treatment was 34.2%. The median time of developing NAFLD is 46 months in the NAFLD group. Patients in the NAFLD group had

SC

significantly high BMI compared with those in the non-NAFLD group (P < 0.001). There were no significant differences in age, tumor size, menopause status, lymphatic metastasis, hormone receptor,

M AN U

P53 and Ki-67 expression or treatment regimens between the two groups. The effect of NAFLD development on DFS and OS

Finally, 111 patients (50.2%) had a breast cancer recurrence in the NAFLD group and 163 patients (38.4%) in the non-NAFLD group. Total 62 patients died (58 from breast cancer, 4 from other causes). Kaplan–Meier survival curves analysis showed that patients who did not develop NAFLD significantly

TE D

had better DFS than those who developed (P = 0.006) (Fig. 2A). There were no significant statistical differences between the two groups on OS (P = 0.387) (Fig. 2B). According to the univariate and multivariate analysis, prognostic factors related to disease-free survival in patients with breast cancer

EP

were performed (Table 2). In univariate analysis, tumor size, lymphatic metastasis, P53 and Ki-67 expression and NAFLD development were associated with decreased DFS. In multivariate analysis,

AC C

tumor size, lymphatic metastasis, P53 and Ki-67 expression and NAFLD development were proved to be independent prognosticators for DFS. Blood lipid parameters, BMI and liver function analysis Lipid profiles and liver function tests did not demonstrate significant differences between the two

groups at baseline (Mean±SD: TC: 4.43±0.50 vs 4.11±0.57mmol/L, TG: 1.48±0.38 vs 1.35±0.40mmol/L, LDL-C: 2.89±0.49 vs 2.11±0.63mmol/L, HDL-C: 1.43±0.37 vs 1.47±0.41mmol/L, ALT: 25.47±10.55 vs 21.54±8.03 U/L and AST: 25.75±10.72 vs 24.58±6.93 U/L in the NAFLD group vs non-NAFLD group). In the NAFLD group, TC, TG, LDL-C and ALT demonstrated a statistically significant increase from baseline to 6 months (P < 0.05), whereas, HDL-C demonstrated a significant reduction from baseline to 6 months (P < 0.001). In addition, in the non-NAFLD group, TC, LDL-C,

ACCEPTED MANUSCRIPT AST and ALT demonstrated a statistically significant increase from baseline to 6 months (P < 0.05), while HDL-C demonstrated a significant reduction from baseline to 6 months (P < 0.01). Lipid profiles, BMI and liver function tests did not demonstrate statistically significant changes from 6 months to 12 months in either the NAFLD or non-NAFLD groups. The results excluded 2 patients whose OS were

RI PT

less than 12 months (Fig. 3). Optimal cutoff points of BMI, blood lipid profiles and liver function tests

We selected lipid profiles, liver function tests and BMI at 6 months of treatment. The reason why we selected BMI at 6 months is because BMI was significantly higher in the NAFLD group than those

SC

in the non-NAFLD group (23.64±2.68 kg/m2 vs 22.45±3.13 kg/m2, P < 0.001). ROC analyses were applied to determine the optimal cutoff point. Regarding BMI, TC, TG, LDL-C, ALT and HDL-C, the

M AN U

optimal cutoff point was 21.06, 4.28, 1.22, 3.13, 27.50 and 1.29, respectively which was identified in order to distinguish NAFLD developed patients from non-NAFLD developed patients (P < 0.05). Patients with higher BMI, TC, TG, LDL-C and ALT level at 6 months had a high risk of NAFLD development during the period of tamoxifen treatment. However patients with lower HDL-C level at 6 months had a high risk of NAFLD development (Fig. 4).

TE D

Risk score related to NAFLD development

We created a comprehensive index associated with NAFLD developed or not and based on the multivariate analyses (Table 3), a score was attributed to each significant variable according to its HR.

EP

Thus, a score value of 1 was given to patients whose BMI ≥21.06 kg/m2, TG ≥1.22 mmol/L, LDL-C ≥3.13 mmol/L and ALT ≥27.50 U/L at 6 months. However, a score 0 was given to patients whose BMI

AC C

<21.06 kg/m2, TG <1.22 mmol/L, LDL-C <3.13 mmol/L and ALT <27.50 U/L at 6 months. The comprehensive risk index was created through using the following formula: (BMI score value) + (lipid profile of TG and LDL-C score value) + (ALT score value), determining a range of 0–4. Multivariate logistic regression analysis of risk scores associated with NAFLD development is shown in Table 4. It was found that in the NAFLD group, 5 had a final index of 0; 19 of 1; 51 of 2, 77 of 3 and 69 of 4, while, in the non-NAFLD group, 24 had a final index of 0; 59 of 1; 129 of 2, 122 of 3 and 91 of 4. What’s more, we found the index cut-off of NAFLD development at 3, indicating < 3 a low risk of NAFLD development and ≥ 3 a high risk of NAFLD development (OR = 3.03, CI 1.11 - 8.28, P = 0.037) (Table 4). Discussion

ACCEPTED MANUSCRIPT Tamoxifen is an effective drug which has been widely used for the treatment of hormone receptor-positive breast cancer.19 Women taking tamoxifen from 5 to 10 years exhibit the reduction of the risks of breast cancer recurrence and mortality.20,21 While generally well-tolerated, tamoxifen is known to induce fatty liver in 43% of women within the first 2 years of treatment.22,23 Moreover, the

RI PT

incidence of tamoxifen induced NAFLD was estimated to be around 40% at 1 year.11 Currently, there is little information concerning the relationship between NAFLD development and survival of breast cancer patients in the literature. However, in the present study, we observed that DFS in the non-NAFLD group was significantly better than those in the NAFLD group during follow-ups,

SC

however, there were no significant statistical differences between the two groups on OS. To our knowledge, this is the first study to report that adjuvant tamoxifen associated with NAFLD had a poor

M AN U

prognosis of patients with breast cancer. Interestingly, in a retrospective study, they claimed that patients taken adjuvant tamoxifen or toremifene treatment and developed NAFLD had better DFS and OS compared with those who did not.24 The results could be explained by the differences of medication, sample sizes and HER2 expression. They used tamoxifen or toremifene or both as a postoperative endocrine therapy. In our study, we used tamoxifen only. Compared with patients who developed

TE D

NAFLD after tamoxifen use, their sample sizes were smaller (221 vs 158). Their HER2+ patients were consisted of 10.8% in the NAFLD group; in our study they were all HER2- patients. In addition to NAFLD development, we also found that tumor size, lymphatic metastasis, P53 and Ki-67 expression

EP

were independent prognosticators for DFS.

Although tamoxifen is effective in patients with HR+ breast cancer, it is associated with two fold

AC C

increased risk of developing NAFLD and NASH.11 The pathogenesis of NASH is not fully understood, most investigators agree that a baseline of steatosis (first hit) requires a second hit capable of inducing inflammation, fibrosis or necrosis for NASH to develop. Obesity, insulin resistance, increased cytokine activity, oxidative stress and mitochondrial dysfunction also play a role in the pathogenesis of NASH. Tamoxifen was shown to increase hepatic fat content, through blocking the role of estrogen in maintaining hepatic lipid homeostasis by supporting the expression of genes involved in lipid β-oxidation, albeit the exact mechanism is not known. In addition, tamoxifen has been shown to increase serum triglyceride and lower low-density lipoprotein and cholesterol levels. It is possible that tamoxifen increases serum triglyceride levels and reduces hepatic lipid β-oxidation, and ultimately this may enhance hepatic fat content.25,26 Moreover, the etiology of NASH is unclear, it is often associated

ACCEPTED MANUSCRIPT with obesity, type 2 diabetes, hyperlipidemia and insulin resistance.27 Other predictors of NAFLD include severe baseline hypercholesterolaemia and hypertension.11 A study of 70 patients of risk factors of NAFLD development during tamoxifen use reported that tamoxifen was associated with a high risk of development of non-alcoholic steatohepatitis in patients with higher triglycerides and fasting blood

RI PT

sugar (FBS) and lower HDL. However, it found no relationship with the level of BMI, LDL, hypertension, overweight and obesity.28 Predictors of developing NAFLD are still under investigation. Therefore, we examine whether lipid profiles, liver function tests and BMI influence development of tamoxifen associated NAFLD. Lipid profiles, liver function tests and BMI were detected at the baseline,

SC

6 months and 12 months of treatment. Finally, we found those evaluation above at 6 months had predictive value of NAFLD development. In agreement with the study by Qiufan Zheng et al, we found

M AN U

high BMI was positively associated with NAFLD development.24 Moreover, high TC, TG, LDL-C, ALT level and high BMI and low HDL at 6 months were predictors of NAFLD development during tamoxifen treatment.

In our study, ROC analyses were applied to determine the optimal cutoff points. In order to illustrate these factors on predicting NAFLD development, we calculated area under the

TE D

time-dependent receiver operating characteristic curves (AUC). Then, we created a comprehensive index associated to NAFLD development. A score was attributed to each variable according to the multivariate analyses. Patients with a risk score ≥ 3 had a high risk of NAFLD development during

EP

tamoxifen treatment and they may also need lifestyle and pharmacological interventions. Therefore, improvement of those predictors above may reduce the incidence of NAFLD development and may

AC C

improve the survival outcomes of breast cancer patients. In fact, intervention targets of NAFLD generally include those risk factors above. However, to date, there are no established treatment guidelines and no single approved therapy for NAFLD treatment. Historically, the principal treatment for NAFLD consisted of removal of offending drugs and toxins, weight loss, and control of associated metabolic disorders as hyperlipidemia and diabetes.29 For subjects with a BMI higher than 30 kg/m2 or with a BMI higher than 27 along with other comorbid conditions (such as sleep apnea), pharmacologic weight management with orlistat or sibutramine may be considered as these agents which could exert a significant influence on NAFLD.30 Hannah WN et al reported that weight loss of 3% to 5% is associated with decreased steatosis; however, a 7% to 10% decrease is necessary to achieve NAFLD/NASH remission and fibrosis regression.31 Vitamin E, an inexpensive potent antioxidant, has

ACCEPTED MANUSCRIPT been examined as a treatment agent for NAFLD in many adult and pediatric studies. It was well tolerated, and most studies showed that modest improvements in ultrasonographic appearance of the liver, serum aminotransferase levels and histologic findings.32,33,34 In one published series of 11 pediatric patients with NASH who have received vitamin E (d-α-tocopherol), 400 to 1200 IU, ALT

RI PT

improved.34,35,36 The effect of lipid lowering agents is under active investigation. A novel and selective PPARα agonist (K-877) that appears to have greater triglyceride lowering activity is currently in phase 2 clinical trials. In a study of 423 patients with hypertriglyceridemia, K-877 significantly reduced liver enzymes when it was added to a statin,37,38 a promising finding that deserves further exploration as a

SC

treatment for NAFLD/NASH. At present, several anti-fibrotic, anti-apoptotic and immune therapies are in pre-clinical and phase II trials, and hopefully in the next decade and thus we will achieve a more

M AN U

efficient liver-specific treatment.39 Conclusions

To conclude, patients with early-stage breast cancer, tamoxifen associated NAFLD was an independent prognosticator with poor DFS. Risks of NAFLD development were high BMI, hyperlipemia and abnormal liver enzyme levels. The comprehensive index we created associated with

TE D

NAFLD development might be used for clinicians to persuade patients who have experienced side effects to take non-pharmacological or pharmacological interventions or even change the treatment regimen. However, the present study still had limitations because we could not differentiate

EP

steatohepatitis from steatosis, because it is not possible to use ultrasonography. Therefore, well-designed, prospective cohort studies are required to validate our findings and intervention study of

AC C

NAFLD development is needed to be conducted in the future. Clinical Practice Points


Tamoxifen was widely used as the first choice for the hormone positive breast cancer patients. Many researches have reported that patients with early-stage breast caner using long-term adjuvant tamoxifen have a high risk of developing nonalcoholic fatty liver disease (NAFLD).




NAFLD can be diagnosed conveniently by ultrasonography. In our study, we found NAFLD development during tamoxifen treatment might be related to poor prognosis in patients with early-stage breast caner. Hence, when and how to evaluate each patient drove our attention.




We found that routine examination of liver function, lipid profile and BMI for every patient at the 6 months of tamoxifen treatment had remarkable prognostic value. Then, we created a risk score

ACCEPTED MANUSCRIPT by which physician can timely inform each patient with breast cancer to have life style or pharmacological interventions. Acknowledgments I confirm that the paper is my original work which has not been published elsewhere and that it has not

RI PT

been submitted simultaneously for publication elsewhere. Besides, I also confirm that the achievement is free of plagiarism, is accurate, and does not infringe on anyone’s copyright or other rights. Disclosure The authors declare that they have no conflicts of interest.

SC

Informed consent

All procedures performed in studies including human participants were in accordance with the ethical

M AN U

standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. In addition, this paper does not contain any studies with animals performed by any of the authors. References

1. Yoshimoto FK, Guengerich FP. Mechanism of the third oxidative step in the conversion of

136:15016-15025.

TE D

androgens to estrogens by cytochrome P450 19A1 steroid aromatase. J Am Chem Soc 2014;

2. Lim E, Metzger-Filho O, Winer EP. The natural history of hormone receptor-positive breast cancer.

EP

Oncology 2012; 26:688-694, 696.

3. Yeh WL, Lin HY, Wu HM, et al. Combination treatment of tamoxifen with risperidone in breast

AC C

cancer. PLoS One 2014; 9:e98805.

4. MacGregor JI, Jordan VC. Basic guide to the mechanisms of antiestrogen action. Pharmacological reviews 1998; 50:151-196. 5. Rakha EA, El-Sayed ME, Green AR, et al. Biologic and clinical characteristics of breast cancer with single hormone receptor positive phenotype. J Clin Oncol 2007; 25:4772-4778. 6. Yang LH, Tseng HS, Lin C, et al. Survival benefit of tamoxifen in estrogenreceptor-negative and progesterone receptor-positive low grade breast cancer patients. J Breast Cancer 2012; 15:288-295. 7. Amir E, Seruga B, Niraula S, et al. Toxicity of adjuvant endocrine therapy in postmenopausal breast cancer patients: a systematic review and meta-analysis. J Natl Cancer Inst 2011; 103:1299-1309. 8. Pratt DS, Knox TA, Erban J. Tamoxifen-induced steatohepatitis. Ann Intern Med 1995; 123:236.

ACCEPTED MANUSCRIPT 9. Loomba R, Sanyal AJ. The global NAFLD epidemic. Nat Rev Gastroenterol Hepatol 2013; 10:686-690. 10. Cole LK, Jacobs RL, Vance DE. Tamoxifen induces triacylglycerol accumulation in the mouse liver by activation of fatty acid synthesis. Hepatology 2010; 52:1258-1265.

RI PT

11. Bruno S, Maisonneuve P, Castellana P, et al. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. BMJ 2005; 330:932.

12. Lee MH, Kim JW, Kim JH, Kang KS, Kong G, Lee MO. Gene expression profiling of murine

SC

hepatic steatosis induced by tamoxifen. Toxicology letters 2010; 199:416-424.

13. Liu X, Meng QH, Ye Y, et al. Prognostic significance of pretreatment serum levels of albumin,

M AN U

LDH and total bilirubin in patients with non-metastatic breast cancer. Carcinogenesis 2015; 36:243-248.

14. Bahl M, Ennis M, Tannock IF, et al. Serum lipids and outcome of early-stage breast cancer: results of a prospective cohort study. Breast Cancer Res Treat 2005; 94:135-144.

15. Goodwin PJ, Boyd NF. Body size and breast cancer prognosis: a critical review of the evidence.

TE D

Breast Cancer Res Treat 1990; 16:205-214.

16. Chlebowski RT, Aiello E, McTiernan A. Weight loss in breast cancer patient management. J Clin Oncol 2002; 20:1128-1143.

EP

17. Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 1987; 235:177-182.

AC C

18. Sanyal AJ, American Gastroenterological A. AGA technical review on nonalcoholic fatty liver disease. Gastroenterology 2002; 123:1705-1725. 19. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study. J Natl Cancer Inst 1998; 90:1371-1388.

20. Hackshaw A, Roughton M, Forsyth S, et al. Long-term benefits of 5 years of tamoxifen: 10-year follow-up of a large randomized trial in women at least 50 years of age with early breast cancer. J Clin Oncol 2011; 29:1657-1663. 21. Davies C, Pan H, Godwin J, et al. Long-term effects of continuing adjuvant tamoxifen to 10 years versus stopping at 5 years after diagnosis of oestrogen receptor-positive breast cancer: ATLAS, a

ACCEPTED MANUSCRIPT randomised trial. Lancet 2013; 381:805-816. 22. Nishino M, Hayakawa K, Nakamura Y, et al. Effects of tamoxifen on hepatic fat content and the development of hepatic steatosis in patients with breast cancer: high frequency of involvement and rapid reversal after completion of tamoxifen therapy. AJR Am J Roentgenol 2003; 180:129-134.

RI PT

23. Nguyen MC, Stewart RB, Banerji MA et al. Relationships between tamoxifen use, liver fat and body fat distribution in women with breast cancer. Int J Obes Relat Metab Disord 2001; 25:296-298.

24. Zheng Q, Xu F, Nie M, et al. Selective Estrogen Receptor Modulator-Associated Nonalcoholic Fatty Liver Disease Improved Survival in Patients With Breast Cancer: A Retrospective Cohort

SC

Analysis. Medicine (Baltimore) 2015; 94:e1718.

25. Osman KA, Osman MM, Ahmed MH. Tamoxifen-induced non-alcoholic steatohepatitis: where are

M AN U

we now and where are we going? Expert opinion on drug safety 2007; 6:1-4.

26. Ahmed MH, Osman KA. Tamoxifen induced-non-alcoholic steatohepatitis (NASH): has the time come for the oncologist to be diabetologist. Breast Cancer Res Treat 2006; 97:223-224. 27. Book Ahmed MH, Byrne CD. Non-alcoholic steatohepatitis. In: Byrne CD, Wild S (eds) Metabolic Syndrome. Wiley & Sons Ltd, England 2005; pp 279–303.

TE D

28. M Akhondi-Meybodi, MR Mortazavy-Zadah, Z Hashemian et al. Incidence and risk factors for non-alcoholic steatohepatitis in females treated with tamoxifen for breast cancer. Arab Journal of Gastroenterology 2011; 12:34–36.

EP

29. Abd El-Kader SM, El-Den Ashmawy EM. Non-alcoholic fatty liver disease: The diagnosis and management. World J Hepatol 2015; 7:846-858.

AC C

30. Brunt EM, Janney CG, Di Bisceglie AM et al. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999; 94:2467-2474. 31. Hannah WN, Jr., Harrison SA. Effect of Weight Loss, Diet, Exercise, and Bariatric Surgery on Nonalcoholic Fatty Liver Disease. Clinics in liver disease 2016; 20:339-350. 32. Hasegawa T, Yoneda M, Nakamura K et al. Plasma transforming growth factor-beta1 level and efficacy of alpha-tocopherol in patients with non-alcoholic steatohepatitis: a pilot study. Aliment Pharmacol Ther 2001; 15:1667-1672. 33. Kugelmas M, Hill DB, Vivian B, et al. Cytokines and NASH: a pilot study of the effects of lifestyle modification and vitamin E. Hepatology 2003; 38:413-419. 34. Harrison SA, Torgerson S, Hayashi P et al. Vitamin E and vitamin C treatment improves fibrosis in

ACCEPTED MANUSCRIPT patients with nonalcoholic steatohepatitis. Am J Gastroenterol 2003; 98:2485-2490. 35. Sanyal AJ, Mofrad PS, Contos MJ, S et al. A pilot study of vitamin E versus vitamin E and pioglitazone for the treatment of nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol 2004; 2:1107-1115.

2000; 136:734-738. 37.

Fruchart

JC.

Selective

peroxisome

proliferator-activated

RI PT

36. Lavine JE. Vitamin E treatment of nonalcoholic steatohepatitis in children: a pilot study. J Pediatr

receptor

alpha

modulators

(SPPARMalpha): the next generation of peroxisome proliferator-activated receptor alphaagonists.

SC

Cardiovasc Diabetol 2013; 12:82.

38. Arai H, Ishibashi S, Yamashita S et al. Abstract 13374: efficacy and safety of K-877, a potent and

M AN U

selective PPAR-α modulator (SPPARMα), in combination with statins in Japanese patients with dyslipidemia. European Geriatric Medicine 2013; 4:365-371.

39. Machado MV, Cortez-Pinto H. Non-alcoholic fatty liver disease: what the clinician needs to know. World J Gastroenterol 2014; 20:12956-12980. Figure legends

TE D

Fig.1 Flow diagram showing the inclusion and exclusion criteria of the study. ALT = alanine aminotransferase, AST = asparticaminotransferase, NAFLD=nonalcoholic fatty liver disease,HBsAg = surface antigen of HBV, anti-HCV-Ab =anti-HCVantibody

EP

Fig.2 Kaplan–Meier curves for disease-free survival (A) and overall survival (B) according to the nonalcoholic fatty liverdisease developed or not.

AC C

Median follow-up was 67 months.CI =confidence interval, HR =hazard ratio, NAFLD =nonalcoholicfatty liver disease. Fig.3 Changes of blood lipid profiles, BMI and liver function tests at baseline, 6 months and 12 months of tamoxifen treatment. Data are shown as the mean ± S.D. of results from the medical records. Differences were considered significant at *p< 0.05, **p< 0.01, ***p< 0.001. Fig.4 Optimal cutoff points decided by ROC curves. ROC curves analysis identified BMI at 21.06 kg/m2, TC at 4.28 mmol/L, TG at 1.22 mmol/L, LDL-C at 3.13 mmol/L, HDL-C at 1.29 mmol/L and ALT at 27.50 U/L, as the optimal cutoff points distinguishing NAFLD developed patients from no NAFLD developed patients.

ACCEPTED MANUSCRIPT Table legends Table 1 Patients baseline characteristics of NAFLD and non-NAFLD groups (n=646). Table 2 Univariate and multivariate analysis of prognostic factors related to disease-free survival in patients with breast cancer.

RI PT

Table 3 Univariate and multivariate analysis of factors related to nonalcoholic fatty liver disease developed.

AC C

EP

TE D

M AN U

SC

Table 4 Risk score related to non-alcoholic fatty liver disease developed.

ACCEPTED MANUSCRIPT Table 1. Patients baseline characteristics of NAFLD and non-NAFLD groups (n=646) NAFLD group (n=221) n %

non-NAFLD group (n=425) n %

AC C

EP

TE D

M AN U

SC

RI PT

Characteristics P value 0.386 Age,y 21 - 49 402 62.2 418 64.7 50 - 83 244 37.8 228 35.3 BMI <0.001 <18.5 2 0.9 30 7.0 18.5 - 23.9 131 59.3 282 66.4 ≥24 88 39.8 113 26.6 0.363 Tumor size ≤20 mm 90 40.7 193 45.4 2050 mm 15 6.8 34 8.0 0.932 Menopause status Pre139 62.9 265 62.4 Post82 37.1 160 37.6 0.177 Hormone receptor ER (+) and PgR (+) 162 73.3 332 78.1 ER (+) / PgR (-) 39 17.6 70 16.5 ER (-) / PgR (+) 20 9.1 23 5.4 0.061 P53 Positive 29 13.1 82 19.3 Negative 192 86.9 343 80.7 0.408 Ki-67 ≤14% 118 53.4 212 49.9 >14% 103 46.6 213 50.1 0.143 Adjuvant chemotherapy Yes 207 93.7 382 89.9 No 14 6.3 43 10.1 0.197 Neoadjuvant chemotherapy Yes 20 9.0 26 6.1 No 201 91.0 399 93.9 0.642 Adjuvant radiotherapy Yes 63 28.5 113 26.6 No 158 71.5 312 73.4 BMI = body mass index, ER = estrogen receptor, PgR = progesterone receptor, NAFLD = nonalcoholic fatty liver disease, P < 0.05 was considered statistically significant.

ACCEPTED MANUSCRIPT Table 2. Univariate and multivariate analysis of prognostic factors related to disease-free survival in patients with breast cancer Univariate analysis HR (95% CI) P value

Multivariate analysis HR (95% CI) P value

AC C

EP

TE D

M AN U

SC

RI PT

Variable Age (<50 vs ≥50 years) 1.05 (0.77 - 1.43) 0.742 0.91 (0.58 - 1.43) 0.688 BMI 18.5 - 23.9 vs <18.5 1.25 (0.61 - 2.58) 0.537 1.20 (0.58 - 2.52) 0.622 18.5 - 23.9 vs ≥24 1.27 (0.93 - 1.74) 0.134 1.02 (0.73 - 1.42) 0.923 Tumor size ≤20 mm vs 20 50 mm 2.76 (1.66 - 4.60) 2.43 (1.43 - 4.13) <0.001 0.001 Menopause status pre- vs post1.07 (0.79 - 1.45) 0.665 1.10 (0.70 - 1.72) 0.692 Lymphatic metastasis positive vs negative 2.03 (1.49 - 2.78) 1.84 (1.32 - 2.58) <0.001 <0.001 Family tumor history yes vs no 0.99 (0.68 - 1.43) 0.946 0.99 (0.68 - 1.45) 0.975 Hormone receptor PR+ vs PR1.29 (0.89 - 1.89) 0.183 1.25 (0.83 - 1.87) 0.279 ER+ vs ER1.25 (0.66 - 2.38) 0.494 1.08 (0.55 - 2.11) 0.818 P53 positive vs negative 1.60 (1.11 - 2.31) 1.52 (1.03 - 2.23) 0.011 0.034 Ki-67 ≤14% vs >14% 1.53 (1.14 - 2.07) 1.38 (1.01 - 1.89) 0.005 0.044 NAFLD developed vs no developed 1.36 (1.01 - 1.84) 1.50 (1.09 - 2.06) 0.043 0.013 BMI = body mass index, HR = hazard ratio, CI = confident interval, NAFLD = nonalcoholic fatty liver disease, P < 0.05 was considered statistically significant.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Table 3 Univariate and multivariate analysis of factors related to nonalcoholic fatty liver disease developed Univariate analysis Multivariate analysis Variable HR (95% CI) P value HR (95% CI) P value Age (years) <50 vs ≥50 1.13 (0.82 - 1.55) 0.470 0.98 (0.65 - 1.47) 0.916 Menopause status pre- vs post1.03 (0.74 - 1.43) 0.855 0.92 (0.60 - 1.40) 0.680 Tumor size (mm) ≤20 vs 20 50 0.88 (0.45 - 1.71) 0.701 0.92 (0.46 - 1.82) 0.808 Lymphatic metastasis positive vs negative 0.94 (0.68 - 1.28) 0.676 0.68 (0.48 - 0.97) 0.032 Family tumor history yes vs no 0.76 (0.50 - 1.56) 0.199 0.80 (0.52 - 1.24) 0.323 Hormone receptor PR+ vs PR1.15 (0.75 - 1.75) 0.520 1.10 (0.71 - 1.71) 0.668 ER+ vs ER2.43 (1.39 - 4.24) 1.91 (1.06 - 3.44) 0.002 0.031 P53 positive vs negative 0.73 (0.45 - 1.16) 0.180 0.79 (0.48 - 1.29) 0.347 Ki-67 ≤14% vs >14% 0.90 (0.66 - 1.24) 0.527 0.97 (0.69 - 1.35) 0.841 2) BMI cut-off (kg/m <21.60 vs ≥21.60 2.38 (1.61 - 3.52) 2.12 (1.41 - 3.19) <0.001 <0.001 TC cut-off (mmol/L) <4.28 vs ≥4.28 1.81 (1.23 - 2.66) 1.14 (0.73 - 1.77) 0.573 0.003 TG cut-off (mmol/L) <1.22 vs ≥1.22 2.93 (2.00 - 4.29) 2.21 (1.45 - 3.36) <0.001 <0.001 LDL-C cut-off (mmol/L) <3.13 vs ≥3.13 1.77 (1.28 - 2.43) 1.50 (1.04 - 2.16) <0.001 0.031 HDL-C cut-off (mmol/L) <1.29 vs ≥1.29 0.61 (0.44 - 0.85) 0.87 (0.61 - 1.25) 0.445 0.004 ALT cut-off (U/L) <27.50 vs ≥27.50 1.69 (1.23 - 2.33) 1.48 (1.07 - 2.06) 0.001 0.018 BMI = body mass index, HR = hazard ratio, CI = confident interval, TC = total cholesterol, TG = triglyceride, LDL-C = low-density lipoprotein cholesterol, HDL-C = high-density lipoprotein cholesterol, ALT = Alanine aminotransferase, P < 0.05 was considered statistically significant.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Table 4. Risk score related to non-alcoholic fatty liver disease developed Risk score NAFLD Non-NAFLD (n=646) (n=221) (n=425) OR (95% CI) P value 0 5 24 1 19 59 1.55 (0.52 - 4.61) 0.456 2 51 129 1.90 (0.69 - 5.25) 0.263 3 77 122 3.03 (1.11 - 8.28) 0.037 4 69 91 3.64 (1.32 - 10.02) 0.012 OR = odds ratio, CI = confident interval, NAFLD = nonalcoholic fatty liver disease, P < 0.05 was considered statistically significant.

ACCEPTED MANUSCRIPT

SC

RI PT

Figure 1. Flow diagram showing the inclusion and exclusion criteria of the study.

AC C

EP

TE D

M AN U

ALT = alanine aminotransferase, AST = asparticaminotransferase, NAFLD=nonalcoholic fatty liver disease,HBsAg = surface antigen of HBV, anti-HCV-Ab =anti-HCVantibody

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Figure 2. Kaplan–Meier curves for disease-freesurvival (A) and overall survival (B) according to the nonalcoholic fatty liverdisease developed or not. Median follow-up was 67 months.CI =confidence interval, HR =hazard ratio, NAFLD =nonalcoholicfatty liver disease.

ACCEPTED MANUSCRIPT

RI PT

Figure 3.Changes of blood lipid profiles, BMI and liver function tests at baseline, 6 months and 12 months of tamoxifen treatment.

AC C

EP

TE D

M AN U

SC

Data are shown as the mean ± S.D. of results from the medical records. Differences were considered significant at *p< 0.05, **p< 0.01, ***p< 0.001.

ACCEPTED MANUSCRIPT Figure

4.

Optimal

cutoff

pointsdecided

by

ROC

curves.

ROCcurves

analysisidentifiedBMI at 21.06 kg/m2, TC at 4.28 mmol/L, TG at 1.22 mmol/L, LDL-C at 3.13 mmol/L, HDL-C at 1.29 mmol/L and ALT at 27.50 U/L, as theoptimal cutoff pointsdistinguishing NAFLD developed patients from no NAFLD developed

AC C

EP

TE D

M AN U

SC

RI PT

patients.