BMI is an independent prognostic factor for late outcome in patients diagnosed with early breast cancer: A landmark survival analysis

Abstract

•BMI negatively affects breast cancer survival.•Higher BMI is associated with older age, post-menopausal status and advanced stages.•Obese patients experience a poorer iDFS and OS at 15 years.•Obesity is an independent prognostic factor for late outcome.•The distribution of late events varies significantly according to BMI class. Breast cancer (BC) is the most common malignancy in women all over the world. Early detection combined to progress in cancer treatment has considerably improved BC outcome [[1]Kohler B.A. Sherman R.L. Howlader N. et al.Annual report to the nation on the status of cancer, 1975-2011, featuring incidence of breast cancer subtypes by race/ethnicity, poverty, and state.J Natl Cancer Inst. 2015; 107https://doi.org/10.1093/jnci/djv048Crossref PubMed Scopus (553) Google Scholar]. However, still some patients experience disease recurrence. Timing of disease relapse varies according to molecular subtypes, with triple negative and HER2-positive tumors showing the highest rate within the first 3–5 years after diagnosis. In contrast, for patients with hormone-receptor (HR) positive tumors, the risk of recurrence may persist longer [[2]Cochrane Database Syst Rev Early breast cancer Triaslists’ Collaborative Group Review update in cochrane database syst.Rev. vol. 2008. 2001Google Scholar,[3]Pan H. Gray R. Braybrooke J. et al.20-year risks of breast-cancer recurrence after stopping endocrine therapy at 5 years.N Engl J Med. 2017; 377: 1836-1846https://doi.org/10.1056/NEJMoa1701830Crossref PubMed Scopus (743) Google Scholar]. The excess of weight is a major public health problem in industrialized countries, with 39% of the world's adult population being overweight and about 13% obese in 2016 (15% of adult women) [[4]Word Health Organization Obesity and Overweight. Fact sheet n.311.http://www.who.int/en/news-room/fact-sheets/detail/obesity-and-overweightDate accessed: May , 2019Google Scholar]. Obesity in the general population is related to high mortality and it represents a risk factor for many diseases, as cardiovascular illness, diabetes and cancer, including BC [5Bhaskaran K. Douglas I. Forbes H. et al.Body-mass index and risk of 22 specific cancer: a population-based cohort study of 5.24 million UK adults.Lancet. 2014; 384: 755-776https://doi.org/10.1016/S0140-6736(14)60892-8Abstract Full Text Full Text PDF PubMed Scopus (981) Google Scholar, 6Benedetto C. Salvagno F. Canuto E.M. et al.Obesity and female malignancies.Best Pract Res Clin Obstet Gynaecol. 2015; 29: 528-540https://doi.org/10.1016/j.bpobgyn.2015.01.003Crossref PubMed Scopus (35) Google Scholar, 7Xia X. Chen W. Li J. et al.Body mass index and risk of breast cancer: a non-linear dose-response meta-analysis of prospective studies.Sci Rep. 2014; 7480https://doi.org/10.1038/srep07480Google Scholar, 8Lahmann P.H. Hoffmann K. Allen N. et al.Body size and breast cancer risk: findings from the European prospective investigation into cancer and nutritio (EPIC).Int J Cancer. 2004; 111: 762-771Crossref PubMed Scopus (455) Google Scholar]. Indeed, the relative risk for BC is 1.1 (95% Confidence Interval (CI) 1.1–1.2) in postmenopausal women for each 5 body mass index (BMI) units increase [[9]Lauby-Secretan B. Scoccianti C. Loomis D. et al.Body fatness and cancer—viewpoint of the IARC working group.N Engl J Med. 2016; 375: 794-798https://doi.org/10.1056/NEJMsr1606602Crossref PubMed Scopus (1622) Google Scholar]. The biological mechanisms underlying the association between increased BMI and BC incidence are not well established but may be related to endocrine and metabolic alterations [[10]McTiernan A. Weight physical activity and breast cancer survival.Proc Nutr Soc. 2018; 77: 403-411https://doi.org/10.1017/S0029665118000010Crossref PubMed Scopus (21) Google Scholar]. Hyper-adiposity may intensify estrogen production and cause a chronic subclinical inflammation with elevated circulating levels of pro-inflammatory proteins related to cancer development [[11]Iyengar N.M. Chen I.C. Zhou X.K. et al.Adiposity, inflammation and breast cancer phatogenesis in Asian woman.Cancer Prev Res. 2018; 11: 227-236Crossref PubMed Scopus (30) Google Scholar]. Another potential mechanism may involve the prolonged hyperinsulinemia that can increase the synthesis of adipokines and cytokines involved in breast tumorigenesis [[12]Sinicrope F.A. Dannenberg A.J. Obesity and breast cancer prognosis: weight of evidence.J Clin Oncol. 2011; 29: 4-7https://doi.org/10.1200/JCO.2010.32.1752Crossref PubMed Scopus (89) Google Scholar]. Moreover, insulin has important effect on AKT/mTOR signaling, an important pathway often involved in tumor growth [[13]Dietze E.C. Chavez T.A. Seewaldt V.L. Obesity and triple-negative breast cancer: disparities, controversies, and biology.Am J Pathol. 2018; 188: 280-290https://doi.org/10.1016/j.ajpath.2017.09.018Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar]. The same mechanisms may explain the association between increased BMI and worse prognosis after the diagnosis of BC described in literature. However, in the majority of these studies, the follow-up period was not sufficiently long to permit an accurate evaluation of events occurring later than 10–15 years from diagnosis [10McTiernan A. Weight physical activity and breast cancer survival.Proc Nutr Soc. 2018; 77: 403-411https://doi.org/10.1017/S0029665118000010Crossref PubMed Scopus (21) Google Scholar, 11Iyengar N.M. Chen I.C. Zhou X.K. et al.Adiposity, inflammation and breast cancer phatogenesis in Asian woman.Cancer Prev Res. 2018; 11: 227-236Crossref PubMed Scopus (30) Google Scholar, 12Sinicrope F.A. Dannenberg A.J. Obesity and breast cancer prognosis: weight of evidence.J Clin Oncol. 2011; 29: 4-7https://doi.org/10.1200/JCO.2010.32.1752Crossref PubMed Scopus (89) Google Scholar, 13Dietze E.C. Chavez T.A. Seewaldt V.L. Obesity and triple-negative breast cancer: disparities, controversies, and biology.Am J Pathol. 2018; 188: 280-290https://doi.org/10.1016/j.ajpath.2017.09.018Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar, 14Kaminemi A. Anderson M.L. White E. et al.Body mass index, tumor characteristics, and prognosis following diagnosis of early stage breast cancer in a mammographically screened population.Cancer Causes Control. 2013; 24: 305-312https://doi.org/10.1007/s10552-012-0115-7Crossref PubMed Scopus (71) Google Scholar, 15Azrad M. Demark-Wahnefried W. The association between adiposity and breast cancer recurrence and survival: a review of the recent literature.Curr Nutr Rep. 2014; 3: 9-15Crossref PubMed Scopus (78) Google Scholar]. Considering the rising obesity rates and its association with pathological conditions with a potential impact on life expectancy, a better understanding of the association between BMI and long-term prognosis in BC patients is needed. In this study, we aimed at evaluating the association between BMI and outcome in a population of BC patients with a long follow up allowing to focus on late events. All patients with early BC diagnosed at Istituto Oncologico Veneto (Padua, Italy) between 2000 and 2007 that underwent primary surgery and with available information on body weight and height at the time of diagnosis were retrospectively identified. Body weight and height were measured after surgery and BMI was calculated as weight in kilograms divided by height in square meters. According to WHO standard, the following BMI categories were identified [[16]World Health Organisation (WHO) Global strategy on diet, physical ativity and health.https://www.who.int/dietphysicalactivity/childhood_what/en/Google Scholar]:-Underweight: BMI<18.5 kg/m2-Normal: BMI 18.5–24.9 kg/m2-Overweight: BMI 25–29.9 kg/m2-Obese: BMI ≥ 30 kg/m2Obese class I: 30.0–34.9 kg/m2Obese class II: 35.0–39.9 kg/m2Obese class III: ≥ 40.0 kg/m2 Length of follow up was not considered as inclusion criteria. However, in the final cohort of patients included in this study, <1% had follow up data up to 1 year after diagnosis and 96% had follow up > 36 months after diagnosis. The following data were collected: age, menopausal status, largest pathological tumor diameter, nodal status, pathological stage according to AJCC 7th edition [[17]Edge S.B. Compton C.C. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the futureof TNM.Ann Surg Oncol. 2010; 17: 1471-1474https://doi.org/10.1245/s10434-010-0985-4Crossref PubMed Scopus (6211) Google Scholar], hormone receptors (HR) and HER2 status, type of surgical and systemic treatment, type and date of survival events (if any), date of last follow up or death [[18]Hammond M.E.H. Hayes D.F. Dowsett M. et al.American Society of Clinical Oncology/College of American Pathologists guideline reccomendations for immunohistochemical testing of estrogen and progesteron receptros in breast cancer.Arch Pathol Lab Med. 2010 Jul; 134: e48-72PubMed Google Scholar]. HR status was defined by immunohistochemistry. HR positivity was defined as ≥1% of tumor cells positively stained for estrogen and/or progesterone receptor. HER2 was evaluated by immunohistochemistry and in situ hybridization in case of a 2 + score by immunohistochemistry. HER2 was defined positive if immunohistochemistry score 3 + and/or ISH amplified. Immunohistochemistry scoring and ISH evaluation were performed according to the guidelines in force at the time of analysis [[19]Bast R.B. Ravdin P. Hayes D.F. et al.2000 Update of reccomandation for the use of tumor markers in breast and colorectal cancer: clinical practice guidelines of the American Society of Clinical Oncology.J Clin Oncol. 2001 March; 19: 1865-1878https://doi.org/10.1200/JCO.2001.19.6.1865Crossref PubMed Google Scholar]. Descriptive statistics for tumor clinicopathological features and treatment data in terms of absolute and relative values were provided. The chi-square tests for the association with BMI category were calculated. Survival endpoints were: invasive-Disease Free Survival (iDFS, defined as the time from diagnosis to: local/distant relapse, contralateral invasive BC, second-primary invasive BC, second-primary non-BC or death without any event) and Overall Survival (OS, defined as the time from BC diagnosis to death from any cause). Late-iDFS was calculated from the landmark of 10 years after diagnosis. All patients who were alive and free from iDFS-event at 10 years from diagnosis were included in late-iDFS analysis. Log-rank test was used to compare survival between groups. Hazard ratios (HRs) and the 95% CI were calculated by cox regression models. Threshold for statistical significance was established at p < 0.05. All tests were two-sided. A total of 992 patients diagnosed with early BC between January 2000 and December 2007 were identified. The characteristics of study population stratified by BMI classes are shown in Table 1. According to BMI, 30 patients (3%) were underweight, 482 (49%) were normal weight, 319 (32%) were overweight and 161 were obese (respectively 11% grade I, 4% grade II and 1% grade III obesity). Due to the low number of underweight patients, we considered the following categories: underweight/normal weight, overweight, obese (any grade, BMI≥30 kg/m2). Higher BMI was found to be statistically associated with older age at diagnosis with a median age of 60 years (range 48–71) for obese patients, 58 years (range 47–70) for overweight patients and 50 years (range 43–62) for underweight/normal weight patients (p < 0.001). Consistently, overweight and obese patients were more frequently postmenopausal (77.4% and 81.1% respectively) as compared to underweight/normal patients (51.5%). More advanced stages (II and III) and nodal involvement were also significantly associated with higher BMI categories (both with p < 0.001). No significant difference was observed according to histological grade (p = 0.903), HR status (p = 0.398), HER2 status (p = 0.662), type of surgical treatment (breast conservative surgery or mastectomy, p = 0.700), receipt of chemotherapy (p = 0.463) and endocrine-therapy (p = 0.064).Table 1Baseline patients’ characteristics according to BMI category.Underweight/NormalN (%)OverweightN (%)ObeseN (%)TotalN (%)PAll512 (52)319 (32)161 (16)992 (100)Age median (yrs), Q1-Q3<0.00150 (43–62)58 (47–70)60 (48–71)53 (45–67)Menopausal StatusPOST262 (51)246 (77)129 (81)637 (65)<0.001PRE247 (49)72 (23)30 (19)349 (35)StageI262 (54)125 (40)48 (30)435 (46)<0.001II182 (37)143 (47)79 (50)404 (42)III45 (9)39 (13)31 (20)115 (12)Nodal statusNegative321 (65)182 (59)75 (47)578 (60)<0.001Positive172 (35)127 (41)84 (53)383 (40)Grade1/2343 (71)208 (69)105 (71)656 (70)0.9033143 (29)93 (31)44 (29)280 (30)HRNegative50 (10)41 (13)18 (11)109 (11)0.398Positive457 (90)277 (87)139 (89)873 (89)HER2NegativePositive297 (82)65 (18)181 (85)32 (15)95 (83)19 (17)573 (83)116 (17)0.662Type of surgeryBCS352 (69)216 (68)115 (72)683 (69)0.700Mast158 (31)101 (32)45 (28)304 (31)CTNo157 (31)103 (32)43 (27)303 (31)0.463Yes350 (69)215 (68)117 (73)682 (69)ETNo93 (18)80 (25)35 (22)208 (21)0.064Yes409 (82)235 (75)124 (78)768 (79)Type of ETTamAIOther/switch151 (38)82 (20)170 (42)64 (28)62 (27)106 (45)38 (31)35 (28)50 (41)253 (33)179 (24)326 (43)0.090Duration of ETUp to 5 yrs>5 yrs199 (62)121 (38)100 (54)85 (46)61 (61)39 (39)360 (69)245 (40)0.189Abbreviations: BMI, Body Mass Index; yrs, years; Q1, first quartile; Q3, third quartile; HR, hormone receptors; HER2, human epidermal growth factor 2; BCS, Breast Conservative Surgery; Mast, mastectomy; CT, chemotherapy; ET, Endocrine Therapy. Open table in a new tab Abbreviations: BMI, Body Mass Index; yrs, years; Q1, first quartile; Q3, third quartile; HR, hormone receptors; HER2, human epidermal growth factor 2; BCS, Breast Conservative Surgery; Mast, mastectomy; CT, chemotherapy; ET, Endocrine Therapy. With a median follow up of 152 months (95% CI 149–155 months), a total of 358 iDFS and 212 OS events have occurred. Fig. 1A shows Kaplan-Meier iDFS curves according to BMI categories. Rates of iDFS at 15 years were 61%, 54% and 34% for underweight/normal, overweight and obese patients, respectively (log-rank p = 0.003). Univariate cox-regression model (Table 2) showed a significant worse iDFS for obese as compared to underweight/normal patients (HR 1.56, 95% CI 1.19–2.03, p = 0.001), whereas there was no significant difference between overweight and underweight/normal weight patients (HR 1.04, 95% CI 0.82–1.32, p = 0.767). With regards to OS, rates at 15 years were 78%, 70% and 61% for underweight/normal, overweight and obese patients, respectively (log-rank p = 0.075), as shown in Fig. 1B. Univariate cox-regression model (Table 2) showed a significantly worse outcome for obese patients compared to underweight/normal weight patients (HR 1.49; 95% CI 1.04–2.12, p = 0.029) and no significant difference between overweight and underweight/normal weight patients (HR 1.23, 95% CI 0.91–1.68, p = 0.182).Table 2Univariate and Multivariate Cox proportional Hazard Models for 15-years iDFS, 15-years OS and Late iDFS at 5 years after the landmark of 10 years.iDFSUnivariatepMultivariatepHazard Ratio (95%CI)Hazard Ratio (95%CI)Age (continuous)1.02 (1.01–1.03)<0.0011.01 (1.00–1.02)0.027MenopausePRERef1.89 (1.48–2.41)<0.001Ref1.47 (1.08–2.00)0.014POSTStageIRef1.38 (1.09–1.75)2.59 (1.90–3.51)0.007<0.001Ref1.35 (1.06–1.71)2.46 (1.79–3.37)0.014<0.001IIIIIGrade1/2Ref1.17 (0.93–1.47)0.188NA3BMIUnder/NormalRef1.04 (0.82–1.32)1.56 (1.19–2.03)0.7670.001Ref0.84 (0.65–1.08)1.16 (0.87–1.55)0.1770.307OverObeseOSUnivariatepMultivariatepHazard Ratio (95%CI)Hazard Ratio (95%CI)Age (continuous)1.04 (1.03–1.05)<0.0011.03 (1.02–1.05)<0.001MenopausePRERefRefPOST2.91 (2.03–4.17)<0.0011.54 (0.98–2.41)0.062StageIRefRefII1.65 (1.19–2.28)0.0031.63 (1.17-2-26)0.004III3.31 (2.32–4.90)<0.0013.25 (2.16–4.89)<0.001Grade1/2Ref1.35 (1.01–1.80)0.042Ref1.45 (1.07–1.95)0.0153BMIUnder/NormalRef1.23 (0.91–1.68)1.49 (1.04–2.12)0.1820.029RefOver0.94 (0.68–1.30)0.706Obese0.96 (0.65–1.41)0.820Late iDFSUnivariatepMultivariatepHazard Ratio (95%CI)Hazard Ratio (95%CI)Age (continuous)1.02 (1.00–1.04)0.0251.00 (0.98-1-02)0.972MenopausePRERefRefPOST2.25 (1.36–3.71)0.0021.87 (1.04–3.37)0.036StageIRefII1.18 (0.74–1.88)0.4791.07 (0.67–1.70)0.793III2.31 (1.17–4.56)0.0161.58 (0.76–3.26)0.223Grade1/2Ref0.59 (0.34–1.03)0.062NA3BMIUnder/NormalRefRefOver1.69 (1.01–2.83)0.0471.34 (0.78–2.31)0.286Obese3.73 (2.25–6.20)<0.0012.81 (1.64–4.83)<0.001Abbreviations: yrs, years; iDFS, Invasive Disease Free Survival; CI, Confidence Interval; Ref, reference; BMI, Body Mass Index; OS, Overall Survival. Open table in a new tab Abbreviations: yrs, years; iDFS, Invasive Disease Free Survival; CI, Confidence Interval; Ref, reference; BMI, Body Mass Index; OS, Overall Survival. In multivariate models for iDFS and OS including the other factors that resulted significantly associated with survival in univariate analysis, BMI was not an independent predictor of outcome, as shown in Table 2. At 10 years from diagnosis, 612 patients were alive and free from iDFS event. At a median follow up of 39 months (95% CI 37–41 months), a total of 87 late-iDFS events have occurred. At 5 years after the 10-years landmark, late-iDFS rates were 85%, 74% and 50% for underweight/normal, overweight and obese patients, respectively (log-rank p < 0.0001), as shown in Fig. 2A. Univariate cox-regression model (Table 2) evidenced a significantly worse outcome not only for obese, but also for overweight patients as compared to underweight/normal weight patients (HR 3.73; CI 2.25–6.20, p < 0.001 and HR 1.69; 95% CI 1.01–2.83, p = 0.047, respectively). In a multivariate analysis including other variables that resulted significantly associated with late-iDFS, BMI category maintained an independent prognostic value for the comparison between obese and underweight/normal patients (HR 2.81; 95% CI 1.64–4.83, p < 0.001; Table 2). Table 3 shows type of late-iDFS events distribution stratified by BMI. Event type distribution was significantly different according to BMI categories (p < 0.0001), with distant relapse, second breast cancer, second non-breast cancer and death from any cause more frequently occurring among obese patients.Table 3Late iDFS events distribution across BMI classes.Underweight/NormalN (%)OverweightN (%)ObeseN (%)pLocal relapse7 (2)4 (2)3 (3)<0.0001Distant relapse3 (1)1 (1)6 (7)Second BC6 (2)1 (1)5 (5)Second non-BC7 (2)8 (4)5 (5)Death w/o event8 (3)12 (6)10 (11)No Events291 (90)171 (86)64 (69)Total322 (100)197 (100)93 (100)Abbreviations: iDFS, Invasive Disease Free Survival; BMI, Body mass Index; BC, Breast Cancer. Open table in a new tab Abbreviations: iDFS, Invasive Disease Free Survival; BMI, Body mass Index; BC, Breast Cancer. We examined the impact of BMI on outcome according to HR status by using a definition of HR-positivity of ≥1% of cells positive for estrogen and/or progesterone receptor. In the subgroup of HR positive patients, rates of iDFS at 15 years from diagnosis were: 61%, 55% and 34% for underweight/normal, overweight and obese patients, respectively (log-rank p < 0.001). Cox regression analysis showed a HR of 1.04 (95% CI 0.80–1.35, p = 0.754) for overweight vs underweight/normal weight and a HR of 1.72 (95% CI 1.30–2.28, p < 0.001) for obese vs underweight/normal weight patients. For HR positive patients, rates of late-iDFS at 5 years after the 10-years landmark were 84%, 76% and 50% for underweight/normal, overweight and obese patients, respectively (log-rank p < 0.0001, Fig. 2B). Cox regression analysis showed a HR of 1.61 (95% CI 0.92–2.81, p = 0.097) for overweight vs underweight/normal weight and a HR of 3.86 (95% CI 2.25–6.62, p < 0.001) for obese vs underweight/normal weight patients. In this subgroup of patients we also analyzed the impact of type and duration of endocrine therapy on late-iDFS, overall and according to BMI category. We observed no association between type of ET with late outcome. There was a numerical difference favoring a better outcome with longer duration of ET, however this was not statistically significant. More details can be found as Supplementary Table S1. In the subgroup of HR-negative patients, with only 107 patients included, we observed no impact of BMI on iDFS (15-year rates 63%, 54%, 43% for underweight/normal weight, overweight and obese, p = 0.662). The sample size was too limited and did not allow for late-iDFS analysis (n = 65 patients, n = 10 events). Supplementary Fig. S1 shows the impact of BMI on iDFS and late-iDFS in HER2-positive patients (n = 116 cases of 689 patients with available HER2 status). BMI was not significantly associated with iDFS (log-rank p = 0.537). In late-iDFS analysis, 67 HER2-positive patients were included. Obesity was significantly associated with worse late outcome (log-rank p = 0.006). This study examined the impact of BMI on outcome, focusing on late events, in a large cohort of early BC patients. Obesity conferred a poor prognosis in term of both iDFS and OS. The added value of this work is the long follow-up, allowing to specifically assess the impact of BMI on late outcome. We demonstrated that obesity was independently associated with worse late-iDFS. In our cohort, a higher BMI was correlated with advanced stage and nodal involvement at diagnosis. These findings are supported by literature [20Biganzoli E. Desmedt C. Fornili M. et al.Recurrence dynamics of breast cancer according to baseline body mass index.Eur J Cancer. 2017; 87: 10-20https://doi.org/10.1016/j.ejca.2017.10.007Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar, 21Deglise C. Bouchardy C. Burri M. et al.Impact of obesity on diagnosis and treatment of breast cancer.Breast Canc Res Treat. 2010; 120: 185-193https://doi.org/10.1007/s10549-009-0459-1Crossref PubMed Scopus (51) Google Scholar, 22Eichholzer M. Huang D.J. Modlasiak A. et al.Impact of body mass index on prognostically relevant breast cancer tumor characteristics.Breast Care. 2013; 8: 192-198Crossref PubMed Scopus (24) Google Scholar] and could be related to a more difficult breast examination, especially self-exam, in obese women, which may lead to a delay in breast lumps diagnosis [23Cui Y. Whiteman M.K. Flaws J.A. et al.Body mass and stage of breast cancer diagnosis.Int J Cancer. 2002; 98: 279-283Crossref PubMed Scopus (144) Google Scholar, 24Reeves M.J. Newcomb P.A. Remington P.L. et al.Body mass and breast cancer. Relationship between method of detection and stage of disease.Cancer. 1996; 77: 301-307Crossref PubMed Scopus (67) Google Scholar, 25Destounis S. Newell M. Pinsky R. Breast imaging and intervention in the overweight and obese patient.Am J Roentgenol. 2011; 196: 296-302https://doi.org/10.2214/AJR.10.5556Crossref PubMed Scopus (25) Google Scholar]. In this population, the delay in medical consultation could also depend from psychological perception of self and low socioeconomic status, as previously reported [[26]Ferrante J.M. Chen P.H. Crabtree B.F. et al.Cancer screening in women: body mass index and adherence to physician recommendations.Am J Prev Med. 2007; 32: 525-531Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar]. Our results are consistent with main literature with regards to the association between BMI and BC patients' prognosis. Two large meta-analyses, respectively of 43 and 82 international studies, showed a poorer outcome for obese BC patients vs non-obese, independently from menopausal and HR status [[27]Protani M. Coory M. Martin J.K. Effect of obesity on survival of woman with breast cancer: systemic review and meta-analysis.Breast Canc Res Treat. 2010; 123: 627-635https://doi.org/10.1007/s10549-010-0990-0Crossref PubMed Scopus (685) Google Scholar,[28]Chan D.S. Vieira A.R. Aune D. et al.Body mass index and survival in women with breast cancer-systematic literature review and meta-analysis of 82 follow-up studies.Ann Oncol. 2014; 25: 1901-1914https://doi.org/10.1093/annonc/mdu042Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar]. Our cohort was composed for the vast majority of HR-positive patients. Most of the existing data show the strongest association of BMI and outcome in patients with HR-positive tumors. Data from a National Surgical Adjuvant Breast and Bowel Project (NSABP) trial of among 4077 HR positive, node-negative patients showed an increased risk of overall mortality, second tumors and contralateral breast cancer, but not of recurrence, among obese patients; no difference in breast cancer-specific survival was found [[29]Dignam J.J. Wieand K. Johnson K.A. et al.Obesity, tamoxifen use and outcomes in women with estrogen receptor-positive early-stage breast cancer.J Natl Cancer Inst. 2003; 95: 1467-1476Crossref PubMed Scopus (200) Google Scholar]. Similarly, Sparano et al. showed an association between high BMI at diagnosis and higher risk of recurrence and death in HR-positive and HER2-negative patients included in 3 adjuvant trials [[30]Sparano J.A. Wang M. Zhao F. et al.Obesity at diagnosis is associated with inferior outcomes in hormone receptor-positive operable breast cancer.Cancer. 2012; 118: 5937-5946https://doi.org/10.1002/cncr.27527Crossref PubMed Scopus (151) Google Scholar]. Several mechanisms have been proposed to explain the correlation between HR-positive BC and adipose tissue: accumulation of adipose tissue induces high circulating levels of sex hormones, as estrone and free estradiol, that may stimulate residual neoplastic cell to grow [[31]McTiernan A. Rajan K.B. Tworoger S.S. et al.Adiposity and sex hormones in postmenopausal breast cancer survivors.J Clin Oncol. 2003; 21: 1961-1966Crossref PubMed Scopus (222) Google Scholar]. Moreover, hyper-adiposity is related to a chronic hyperinsulinemia which reduces production of insulin-like growth factors (IGF) binding proteins, causing elevated levels of free IGF which can stimulate the synthesis of sex steroid hormones [[12]Sinicrope F.A. Dannenberg A.J. Obesity and breast cancer prognosis: weight of evidence.J Clin Oncol. 2011; 29: 4-7https://doi.org/10.1200/JCO.2010.32.1752Crossref PubMed Scopus (89) Google Scholar,[32]Yu H. Rohan T. Role of insulin-like growth factor family in cancer development and progression.J Natl Cancer Inst. 2000; 92: 1472-1489Crossref PubMed Scopus (1235) Google Scholar]. The worst prognosis among obese patients could also be influenced by treatment-related factors. In fact, the use of ideal body surface area (or a maximum of 2 kg/mq) to calculate chemotherapy dose could lead to an under-estimation of drugs. Furthermore, it has been observed that many physicians tend to reduce drugs’ doses in heavy patients concerning of adverse effects [[33]Griggs J.J. Sorbero M.E. Lyman G.H. Undertreatment of obese women receiving breast cancer chemotherapy.Arch Intern Med. 2005; 165: 1267-1273Crossref PubMed Scopus (229) Google Scholar]. With regards to endocrine therapy, some evidences suggested a lower benefit in obese patients treated with aromatase inhibitors (letrozole and anastrozole) in both advanced and adjuvant setting. The lack of complete suppression of the aromatization process of androgens to estrogens in adipose tissue in obese women with standard doses of aromatase inhibitors could be a possible explanation of this phenomenon [[34]Schmidt S. Monk J.M. Robinson L.E. et al.The integrative role of leptin, oestrogen and the insulin family in obesity-associated breast cancer: potential effects of exercise.Obes Rev. 2015; 16: 473-487https://doi.org/10.1111/obr.12281Crossref PubMed Scopus (62) Google Scholar,[35]Sestak I. Distler W. Forbes J.F. et al.Effect of body mass index on recurrences in tamoxifen and anastrozole treated women: an exploratory analysis from the ATAC trial.J Clin Oncol. 2010; 28: 3411-3415https://doi.org/10.1200/jco.2009.27.2021Crossref PubMed Scopus (241) Google Scholar]. A number of studies [[13]Dietze E.C. Chavez T.A. Seewaldt V.L. Obesity and triple-negative breast cancer: disparities, controversies, and biology.Am J Pathol. 2018; 188: 280-290https://doi.org/10.1016/j.ajpath.2017.09.018Abstract Full Text Full Text PDF PubMed Scopus (63) Google Scholar,[36]Sun H. Zou J. Chen L. et al.Triple-negative breast cancer and its association with obesity.Mol Clin Oncol. 2017; 7: 935-942https://doi.org/10.3892/mco.2017.1429PubMed Google Scholar,[37]Al Jarroudi O. Abda N. Seddik et al.Overweight: is it a prognostic factor in xomen with triple-negative breast cancer?.Asian Pac J Cancer Prev APJCP. 2017; 18: 1519-1523PubMed Google Scholar] have also tested the role of obesity on outcome in TNBC patients. Even if a recent genomic analysis revealed molecular networks and biological pathways associating obesity with TNBC [[38]Mamidi T.K.K. Wu J. Tchounwou P.B. et al.Whole genome transcriptome analysis of the association between obesity and triple-negative breast cancer in caucasian women.Int J Environ Res Public Health. 2018; 15https://doi.org/10.3390/ijerph15112338Crossref PubMed Scopus (11) Google Scholar], a recent meta-analysis with 4412 TNBC patients showed no significant association between obesity and DFS (p = 0.60) or OS (p = 0.71) [[39]Mei L. He L. Song Y. et al.Association between obesity with disease-free survival and overall survival in triple-negative breast cancer: a meta-analysis.Medicine(Baltimore). 2018; 97e0719https://doi.org/10.1097/MD.0000000000010719PubMed Google Scholar]. In our analysis, the sample size of hormone receptor-negative subgroup was underpowered to allow any subgroup analysis difference (only 11% of total population). Our paper focused specifically on late outcome, with a landmark analysis after 10 years from diagnosis. In the last decades survival rates have constantly increased in women with BC [[40]Noone A, 1975-2015. National Cancer Institute. Bethesda. Available at: https://seer.cancer.gov/csr/1975_2015/. Accessed 21 Feb 2019.Google Scholar,[41]Engholm G. Ferlay J. Christensen N. et al.NORDCAN--a Nordic tool for cancer information, planning, quality control and research.Acta Oncol. 2010; 49: 725-736https://doi.org/10.3109/02841861003782017Crossref PubMed Scopus (374) Google Scholar]. However, the risk of late relapse still represents a challenge. Indeed, it has been recently demonstrated that the risk of BC recurrence continues steadily after 5–20 years from diagnosis especially for HR-positive BC patients [[3]Pan H. Gray R. Braybrooke J. et al.20-year risks of breast-cancer recurrence after stopping endocrine therapy at 5 years.N Engl J Med. 2017; 377: 1836-1846https://doi.org/10.1056/NEJMoa1701830Crossref PubMed Scopus (743) Google Scholar]. Therefore, the identification of factors associated with late relapse in BC survivors is essential in order to offer individualized treatment strategies (i.e. extended adjuvant endocrine therapy) and plan life style-related preventive measures. Moreover, the pool of long-term BC survivors is expected to further increase, thus calling for a particular attention on competing causes of mortality/morbidity (as cardiovascular disease, chronic illness and second cancers). In this perspective, the control of modifiable risk factors such as obesity should be encouraged. In an exploratory analysis we also showed the association between obesity and increased risk of late-iDFS event in HER2-positive patients. Most of the HER2-positive patients in our cohort were also HR-positive (89%). Our results are somehow in contrast with previous data in HER2-positive patients, showing a prognostic impact of BMI and in particular of obesity for HER2-positve/HR-negative patients and not for HER2-positve/HR-positive patients [[42]Kim J.Y. Lee D.W. Lee K.H. et al.Prognostic role of body mass index is different according to menopausal status and tumor subtype in breast cancer patients. vol. 176. Apr 2019: 453-460https://doi.org/10.1007/s10549-019-05249-1Google Scholar,[43]Mazzarella L. Disalvatore D. Bagnardi V. et al.Obesity increases the incidence of distant metastases in oestrogen receptor-negative human epidermal growth factor receptor 2-positive breast cancer patients.Eur J Cancer. 2013 Nov; 49: 3588-3597Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar]. However, only 21% of HER2-positive patients in our study received trastuzumab and the sample size is very limited. Therefore these data should be interpreted with caution. We observed an association between obesity and risk of developing a second primary cancer. The long exposure to a status of chronic subclinical inflammation and the increased levels of free IGF-1 connected with excess weight could explain the excess of second primary tumors in patients with high BMI. Our finding is consistent with previously reported data [[44]Druesne-Pecollo N. Touvier M. Barrandon E. et al.Excess body weight and second primary cancer risk after breast cancer: a systemic review and meta-analysis of prospective studies.Breast Canc Res Treat. 2012; 135: 647-654https://doi.org/10.1007/s10549-012-2187-1Crossref PubMed Scopus (85) Google Scholar] and also with the special report of the International Agency for Research on Cancer (IARC) which concluded that there is sufficient evidence for a cancer-preventive effect of avoidance of weight gain for cancers of the colon, esophagus, renal-cell, postmenopausal BC and corpus uteri [[9]Lauby-Secretan B. Scoccianti C. Loomis D. et al.Body fatness and cancer—viewpoint of the IARC working group.N Engl J Med. 2016; 375: 794-798https://doi.org/10.1056/NEJMsr1606602Crossref PubMed Scopus (1622) Google Scholar]. Because of its retrospective nature, our database lacked information on change in body weight after diagnosis, and this could be a limitation. In fact, during and after adjuvant therapy for BC, gaining weight represents a frequent condition, especially due to alteration in metabolism and changing in dietary habits and lifestyle [[45]Trestini I. Carbognin L. Monteversi S. et al.Clinical implication of changes in body composition and weight in patients with early-stage and metastatic breast cancer.Crit Rev Oncol Hematol. 2018; 129: 54-66https://doi.org/10.1016/j.critrevonc.2018.06.011Crossref PubMed Scopus (27) Google Scholar]. Chan et al. described the impact of weight gain in a cohort of more than 200.000 patients. In this meta-analysis, for each 5 kg/m2 increase of BMI before, <12 months after, and >12 months after diagnosis, an increased risk of total mortality and BC specific mortality was demonstrated [[28]Chan D.S. Vieira A.R. Aune D. et al.Body mass index and survival in women with breast cancer-systematic literature review and meta-analysis of 82 follow-up studies.Ann Oncol. 2014; 25: 1901-1914https://doi.org/10.1093/annonc/mdu042Abstract Full Text Full Text PDF PubMed Scopus (700) Google Scholar]. Similarly, Bradshaw et al. showed a poor prognosis among women who gained more than 10% weight after diagnosis of BC compared with those who maintained their weight [[46]Bradshaw P.T. Ibrahim J.G. Stevens J. et al.Postdiagnosis change in bodyweight and survival after breast cancer diagnosis.Epidemiology. 2012; 23: 320-327https://doi.org/10.1097/EDE.0b013e31824596a1Crossref PubMed Scopus (77) Google Scholar]. Another limitation of our study is the lack of body composition measure. Indeed, BMI definition is only related to weight and height, not necessary to fatness, and definition of obesity or leanness may not be accurate by using this index. New aspects as sarcopenia and waist-to-hip ratio could be more precise in describing body composition. In particular, sarcopenia has been shown to be associated with an increased risk of recurrence and death in early and metastatic setting [[45]Trestini I. Carbognin L. Monteversi S. et al.Clinical implication of changes in body composition and weight in patients with early-stage and metastatic breast cancer.Crit Rev Oncol Hematol. 2018; 129: 54-66https://doi.org/10.1016/j.critrevonc.2018.06.011Crossref PubMed Scopus (27) Google Scholar] and a high waist-to-hip ratio seems to be correlated with poor prognosis in post-menopausal women [[47]Borugian M.J. Sheps S.B. Kim-Sing C. et al.Waist-to-hip ratio and breast cancer mortality.Am J Epidemiol. 2003; 158: 963-968Crossref PubMed Scopus (117) Google Scholar]. However, BMI remains a simple and totally reproducible index. Further studies are needed to understand how to integrate these indexes. Our cohort represents a large mono-institutional database about impact of BMI in early BC patients. We confirmed the negative impact of being obese on BC prognosis. With a long follow-up, a high BMI was associated with increased rates of relapse, second primary tumors and death occurring in the period that started from 10 years after BC diagnosis. Considering the increasing number of women living after a BC diagnosis and the progressively growing prevalence of overweight and obesity among adults, understanding the link between BC and body size is crucial. As BMI is a modifiable risk factor, interventions to control body weight should be pursued.