TYPE Original Research PUBLISHED 02 August 2022 DOI 10.3389/fonc.2022.829483 Survival and adverse events of OPEN ACCESS elderly patients treated with EDITED BY Jiang Chen, Zhejiang University, China sorafenib for hepatocellular REVIEWED BY Hui Guo, carcinoma First Affiliated Hospital of Xi’an Jiaotong University, China Paul Ross, Anna Soria 1, Mariona Calvo 2, Meritxell Casas 1, Zara Vidales 2, Guy’s and St Thomas’NHS Foundation Trust, United Kingdom Sergio Muñoz-Martı́nez 3,4, Victor Sapena 3,5,6, Marc Puigvehi 7, *CORRESPONDENCE Lidia Canillas 7, Raquel Guardeño 8, Adolfo Gallego 9, Mercedes Vergara Beatriz Mı́nguez 4,6, Diana Horta 4,10, Ariadna Clos 11, mvergara@tauli.cat † Silvia Montoliu 12, Mercè Roget 13, Maria Reig 3,4† These authors share senior authorship and Mercedes Vergara 1,4*† SPECIALTY SECTION This article was submitted to 1 Department of Digestive Diseases, Liver Unit, Parc Taulı´ University Hospital, Investigation and Gastrointestinal Cancers: Hepato Innovation Institute Parc Taulı´ I3PT, Universitat Autònoma of Barcelona, Sabadell, Spain, 2 Medical Pancreatic Biliary Cancers, Oncology Department, Catalan Institute of Oncology, Hospitalet, Barcelona, Spain, 3 Barcelona a section of the journal Clinic Liver Cancer (BCLC) Group, Liver Unit, Clinical Insitute of Digestive and Metabolic Diseases Frontiers in Oncology (ICMDiM), Clinic Hospital, Barcelona University, Barcelona, Spain, 4 Biomedical Research Center in Digestive and Liver Diseases Network (CIBERehd), Instituto Carlos III, Madrid, Spain, 5 Medical RECEIVED 05 December 2021 Statistics Core Facility, Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital ACCEPTED 28 June 2022 Clinic Barcelona, Barcelona, Spain, 6 Liver Unit, University Hospital Vall d’Hebron, Liver Diseases PUBLISHED 02 August 2022 Research Group, Vall d’Hebron Institut of Research (VHIR), Universitat Autònoma of Barcelona, CITATION Barcelona, Spain, 7 Department of Digestive Diseases, Liver Unit, Hospital del Mar Medical Research Soria A, Calvo M, Casas M, Vidales Z, Institute, Barcelona, Spain, 8 Medical Oncology Department, Josep Trueta University Hospital, Muñoz-Martı´nez S, Sapena V, Catalan Institute of Oncology, Girona, Spain, 9 Department of Digestive Diseases, Sant Pau i Santa Puigvehi M, Canillas L, Guardeño R, Creu University Hospital, Barcelona, Spain, 10 Department of Gastroenterology, University Hospital Gallego A, Mı´nguez B, Horta D, Clos A, Mútua Terrassa, Barcelona University, Terrassa, Spain, 11 Department of Digestive Diseases, Germans Montoliu S, Roget M, Reig M and Trias i Pujol University Hospital, Badalona, Spain, 12 Department of Digestive Diseases, Joan XXIII Vergara M (2022) Survival and adverse University Hospital, Tarragona, Spain, 13 Department of Digestive Diseases, Liver Unit, Consorci events of elderly patients treated Sanitari de Terrassa, Terrassa, Spain with sorafenib for hepatocellular carcinoma. Front. Oncol. 12:829483. doi: 10.3389/fonc.2022.829483 Introduction: The first-line treatment for advanced hepatocellular carcinoma COPYRIGHT © 2022 Soria, Calvo, Casas, Vidales, (HCC) is atezolizumab plus bevacizumab, but its availability is not universal and Muñoz-Martı´nez, Sapena, Puigvehi, elderly patients are underrepresented in clinical trials. There is little evidence of Canillas, Guardeño, Gallego, Mı´nguez, efficacy and tolerability in elderly patients under systemic treatment. The aims Horta, Clos, Montoliu, Roget, Reig and Vergara. This is an open-access article of this study were to characterize the profile of elderly patients treated with distributed under the terms of the sorafenib, assess their survival and safety profile in order to extrapolate their Creative Commons Attribution License (CC BY). The use, distribution or eligibility for systemic treatment. reproduction in other forums is permitted, provided the original Methods: Retrospective multicentre study of HCC patients aged ≥75 years old author (s) and the copyright owner(s) are credited and that the original treated with sorafenib from January 2008 to December 2019. Demographic publication in this journal is cited, in data, baseline characteristics, and variables related to HCC and sorafenib were accordance with accepted academic recorded. Overall survival (OS) and safety were analyzed. practice. No use, distribution or reproduction is permitted which does not comply with these terms. Results: The study included 206 patients from 11 hospitals, median age 77.9 years; 71.4% men and 62.6% stage Barcelona Clinic Liver Cancer- C (BCLC-C). The main causes of cirrhosis were hepatitis C (60.7%) and alcohol (14.7%). Most patients (84.5%) started with sorafenib 800mg and 15.5% at lower dosage. Arterial hypertension (AHT) (74.2 vs 62.2%; standardized mean differences (STD): 26) and baseline ECOG-PS>0 Frontiers in Oncology 01 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 (45.3 vs 34.7%; STD: 38.2) differed significantly between patients receiving low and full doses. Median OS was 15.4 months (18.2 in BCLC-B vs 13.6 in BCLC-C). OS was not modified by comorbidities, age or period with more expertise. Conclusions: Sorafenib appears to be safe in elderly patients with HCC. This is the first study to characterize the profile of elderly patients to be considered for systemic treatment. These findings could be used as the reference profile for elderly candidates for atezolizumab-bevacizumab. KEYWORDS sorafenib, elderly patients, hepatocellular carcinoma, overall survival, safety, outcome Introduction similar in the elderly and the general population (6, 7, 11–13). However, others warn that survival and safety are lower than in Since 2020, the treatment of choice in non-resectable the general population, and that dose adjustment is frequently hepatocellular carcinoma (HCC) has been the combination of required (8, 9, 14, 15). atezolizumab and bevacizumab, which has shown better overall In view of this lack of consensus surrounding this subgroup survival (OS) than sorafenib and has become the standard of patients, the aims of this multicentre study were to assess the treatment for HCC (1). Previously, since 2007, sorafenib had OS of elderly patients with HCC treated with sorafenib and to been the standard of care in advanced HCC (2, 3). In 2018 analyse the side effects and dosage adjustment needs to lenvatinib was also accepted as a first-line treatment in advanced determine different subgroups in view to selecting patients HCC owing to the results of the non-inferiority study with sorafenib (4). Age is a crucial risk factor for the development of HCC. Materials and methods However, the study population in pivotal studies of sorafenib included patients with mean ages of 65 and 55 years old Study design respectively (2, 3). Similarly, the clinical trial of atezolizumab and bevacizumab included patients with a mean age of 64 (1). In Multicentre, retrospective, observational study involving a recent years, the definition of elderly people has changed; although cohort of patients aged ≥75 years with HCC treated with the World Health Organisation defined the elderly as the sorafenib. The inclusion period was from January 2008 to population over 65 years old (5), due to the increase in life December 2019. Eleven hospitals in Catalonia (north-eastern expectancy, most studies in this field today define the elderly as Spain) participated in the study. those over 75 or even 80 years old (6–9). Elderly patients frequently have more comorbidity and altered pharmacokinetics Study population and pharmacodynamics due to polypharmacy (10). For all these reasons, physicians are more reluctant to start systemic therapy in Applying the population age thresholds used in previous studies these patients and so the clinical experience in this subgroup is (6, 8), our study population included HCC patients aged ≥75 years limited. Management of HCC with sorafenib has received robust starting sorafenib treatment. HCC was diagnosed by non-invasive support from a large number of published papers since 2007. criteria as determined by the American Association Study Liver However, data in elderly patients remain scarce. Cancer (16), or by histology. None of the patients had previously The results regarding sorafenib dosage and safety in elderly received systemic therapy. Patients started sorafenib at a dosage of patients are controversial. Some observational studies suggest 400 mg, 600mg or 400 mg twice daily (800 mg daily) depending on that progression-free survival, disease control and side effects are local clinical judgement. Abbreviations: HCC, hepatocellular carcinoma; BCLC, Barcelona Clinic Study variables Liver Cancer; ECOG, Eastern Cooperative Oncology Group; PS, performance Status; AE, Adverse Events; HFSR, hand food skin reactions; Patient data collected included demographic, medical history OS, Overall survival; STD, standardized mean differences. and baseline variables, such as features related to hepatic disease and Frontiers in Oncology 02 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 its decompensations and cardiovascular comorbidities (AHT, variables were described with median times and 95% confidence diabetes mellitus, peripheral vasculopathy or ischemic intervals (95% CI), calculated using the Kaplan-Meier method, cardiopathy). Regarding the HCC, baseline variables such as date and compared using a log-rank test. The balance between groups of diagnosis, Barcelona Clinic Liver Cancer (BCLC) stage at the (initial sorafenib dosage of 400mg vs 800mg and initiation of time of diagnosis and previous treatments (surgery, percutaneous sorafenib treatment <2013 vs ≥2013) was assessed using ablation or trans-arterial chemoembolization) were recorded. When standardized mean differences (STD). STD >10% was sorafenib was started, levels of alpha-fetoprotein and Eastern considered unbalanced. Cooperative Oncology Group (ECOG) performance status (PS) The level of significance was set at two-sided 5%. All were registered. The starting date of sorafenib treatment was calculations and analysis were performed with SAS 9.4 recorded, as were the doses administered over the course of the software (SAS Institute, Cary, NC, USA). therapy. Sorafenib-related adverse events (AE), including hand–foot skin reaction (HFSR), diarrhoea, rash and fatigue were evaluated using the National Cancer Institute Common Terminology Criteria Results for Adverse Events 4.03. Early presentation of an AE was considered if it appeared during the first month of treatment. Baseline characteristics of the Information on whether further dose reductions were required, study population the date of treatment discontinuation, the cause and the start-day of the second line treatment (if administered) was recorded. Overall A total of 206 patients aged 75 or older attended at the 11 survival (OS) was calculated from the day of the first dose of participating centres from January 2008 to December 2019 were sorafenib to the last day of follow-up or death. The time from included in the study. Baseline characteristics of the study sorafenib discontinuation to death was also assessed. In order to see population are represented in Table 1. The median age was whether training in sorafenib management acquired over the years 77.9 [IQR 76.3-80] years, and 147 (71.4%) were men. The main might improve patient survival and follow up, we analysed OS baseline comorbidity was AHT, recorded in 130 (63.7%), Most dividing the patients into two six-year periods according to the start patients (191, 92.7%) had liver cirrhosis and the main aetiology of sorafenib treatment: from 2008 to 2013, and from 2014 to 2019. was chronic hepatitis C, in 116 (60.7%). A group of 15 non- cirrhotic patients (7.3%), most with a diagnosis of non-alcoholic fatty liver disease, were also diagnosed with HCC. Regarding Ethical considerations complications of cirrhosis, 46 (23%) presented detectable ascites in abdominal ultrasound (grade 1) without diuretic treatment. The study complied with the principles of the Declaration of Helsinki. The Clinical Investigation Ethics Committee of all participating centres reviewed and approved the study with HCC baseline characteristics of regard to the adequate fulfilment of Good Clinical Practice study population principles. Confidentiality was preserved in agreement with current Spanish legislation on data protection (1999). Data HCC baseline characteristics at the moment of initiating confidentiality was carefully preserved. Each hospital included sorafenib are displayed in Table 1. The PS of 127 (63.8%) in the study was assigned a code, which was used to identify each patients was 0. The BCLC stage at the moment of starting patient registered. The principal investigators at each centre had sorafenib was C in 129 patients (62.6%) and B in 75 (36.4%). a separate list that was not accessible to other investigators, in One hundred and twenty-four patients (60.7%) had a history of which these codes were related to each patient’s clinical history previous HCC treatment and none had a history of liver transplant. number. Once the data were recorded, those clinical history numbers were dissociated from the initially assigned code. The need for informed consent was waived due to the retrospective Sorafenib treatment and related adverse and observational nature of the study, and because patients had events advanced HCC whose treatment would have been the same regardless of whether the patients were included in the study. The starting dose was 800mg/day in the majority of patients (84.5%); the lower dose was administered at the discretion of the physician due to renal insufficiency, AHT, intermittent Statistical analysis coagulopathy or anticoagulant treatment. Table 2 displays the differences in baseline characteristics between the cohorts of Continuous data were presented as medians and patients regarding starting sorafenib dose. In general, patients interquartile intervals [IQR: percentile 25th – 75th], and receiving 400mg were older and had more comorbidities categorical data as frequencies and percentages. Time-to-event (diabetes mellitus and AHT). Frontiers in Oncology 03 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 TABLE 1 Patient baseline characteristics at initiation of sorafenib. The median follow-up from the initiation of sorafenib to discontinuation or death of the patient was 13.4 months [IQR Variable N= 206 8-25]. Sorafenib safety profile is displayed in Table 3. During Age (years), median [IQR]1 77.9 [76.3 - 80] sorafenib treatment, 155 (75.2%) patients developed at least Gender (male), n (%) 147 (71.4) one AE, and 121 (58.7%) within the first 30 days. The most Cirrhosis, n (%) frequent AE were fatigue in 123 (59.7%). Dose adjustment was Yes 191 (92.7) required in 103 (52.4%) patients with a median time until the No 15 (7.3) first modification of 3.9 months [IQR 1.1-5.2]. Sorafenib was Cirrhosis aetiology, n (%) definitively discontinued in 182 (88.3%) patients, almost half of HCV2 116 (60.7) these (84, 46.2%) due to symptom progression. The median Alcohol 28 (14.7) overall duration of sorafenib treatment was 5.6 months [IQR NAFLD3 18 (9.4) 1.9-12.3]: 3.4 months [IQR 1.4 - 16.8] in those with a starting Alcohol + HCV 13 (6.8) dose of 400mg, and 5.9 months [IQR 2 - 12.2] in those with a Others 8 (4.2) starting dose of 800mg (STD 1.5, p value 0.5). During follow- 4 HBV 5 (2.6) up, 175 (85%) patients died and the median time from Alcohol + HBV 3 (1.6) treatment discontinuation to death was 5.4 months [IQR 1.6- Comorbidities, n (%) 12.6]. Twenty-two patients (10.7%) were able to initiate AHT 130 (63.7) second-line treatment. Comparison of treatment duration Diabetes Mellitus 78 (38.2) from 2008 to 2013 (5.2; [IQR 1.5 - 12]) and from 2014 to Ischemic cardiopathy 15 (7.4) 2019 (5.6; [IQR2 - 12.9]) did not reveal any Peripheral vasculopathy 8 (3.9) significant differences. Prior history of cirrhosis decompensations The descriptive analysis of the development of AE according Variceal haemorrhage, n (%) 6 (3) to initial sorafenib dose and time of occurrence is shown in Ascites, n(%) Table 4. HSFR and diarrhoea were substantially more frequent Grade I 46 (23) in the 800mg group than in the 400mg group (76 [43.7%] vs 8 No 154 (77) [25.8%] and 43 [24.7%] vs 1 [3.2%] respectively). However, Prior HCC5 treatment, n (%)* 124 (60.2) fatigue was higher in the 400mg group (20 [64.5%] vs Surgical resection 20 (10.1) 103 [59.2%]). Percutaneous treatment 54 (26.6) Chemoembolization 91 (45) ECOG-PS6, n (%) 0 127 (63.8) Overall survival and associated factors 1 70 (35.2) 2 2 (1) The median OS of the whole cohort was 15.4 months [IQR BCLC7 stage, n (%) 12.9-18.4]. The univariate analysis is displayed in Table 5. The A+ 2 (1) only prognostic factor related to OS was BCLC stage (Figure 1): B 75 (36.4) OS was significantly longer in patients with BCLC-B stage (18.2 C 129 (62.6) months [95% CI; 12.9-22.5]) than in those with BCLC-C stage AFP8 (ng/mL), median [IQR] 17.9 [5.4 - 229] (13.6 months [95% CI; 11.7-18.4]) with a p=0.04. There was no Bilirubin (mg/dL), median [IQR] 1.1 [0.8 - 3.1] statistically significant relation between OS and comorbidities. INR9, median [IQR] 1.1 [1 - 1.4] When dividing the population into three groups according to Albumin (g/L), median [IQR] 39.1 [36 - 42] age (≥75-80 years, >80-85 years and ≥85 years), there were no Initial sorafenib dose (mg), n(%) significant differences in median OS (Table 5). 400 31 (15) Table 6 describes the differences between the basal 600 1 (0.5) characteristics of the cohort according to the period of 800 174 (84.5) treatment (2007-2013 and 2014-2019). Significant differences 1. Interquartile range; 2. Chronic hepatitis C infection; 3.Non-alcoholic fatty liver disease; are found in almost all basal characteristics except for certain 4. Chronic hepatitis B infection. Missing values: Diabetes Mellitus (n=2), AHT (n=2), comorbidities (diabetes mellitus, AHT, and peripheral Ischemic cardiopathy (n=3), Peripheral vasculopathy (n=3), Variceal haemorrhage (n=5), Ascites (n=6). 5.Hepatocellular carcinoma; 6. Eastern Cooperative Oncology Group vasculopathy). OS in the 60 patients who started sorafenib performance status; 7. Barcelona Liver Cancer Clinic; 8. Alpha-fetoprotein; 9. between 2008 and 2013 was 13.3 months [IQR 11.5-17.6], and International Normalized Ratio. + 2 BCLC-A: 1 treatment migration, 1 not available reason, Missing values: surgical resection (n=8), percutaneous treatment (n=3), in the 115 patients who initiated this treatment after January Chemoembolization (n=4), ECOG-PS (n=7).*Regarding history prior to HCC 2013 was 17.4 months [IQR 12.9-21.2], though the differences treatment, some patients received more than one alternative during follow-up. were not significant (Figure 2). Frontiers in Oncology 04 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 TABLE 2 Population baseline characteristics according to initial sorafenib dose. Parameter All 400mg 800mg STD (%)1 Patients, n 205* 31 174 Age (years), median [IQR2] 77.9 [76.3 - 80] 79.2 [77.5 - 80.9] 77.7 [76.2 - 79.7] 30.2 Gender (male), n (%) 146 (71.2) 21 (67.7) 125 (71.8) 8.9 Cirrhosis, n (%) 190 (92.7) 27 (87.1) 163 (93.7) 22.5 Cirrhosis aetiology, n (%) 41.3 HCV3 116 (61.1) 17 (63) 99 (60.7) Alcohol 28 (14.7) 5 (18.5) 23 (14.1) NAFLD4 18 (9.5) 1 (3.7) 17 (10.4) Alcohol + HCV 13 (6.8) 2 (7.4) 11 (6.7) Others 8 (4.2) 2 (7.4) 6 (3.7) HBV 4 (2.1) 0 (0) 4 (2.5) Alcohol + HBV5 3 (1.6) 0 (0) 3 (1.8) Comorbidities, n (%) Diabetes Mellitus 78 (38.4) 16 (51.6) 62 (36) 31.8 AHT 130 (64) 23 (74.2) 107 (62.2) 26 Ischemic cardiopathy 15 (7.4) 1 (3.2) 14 (8.2) 21.5 Peripheral vasculopathy 8 (4) 2 (6.5) 6 (3.5) 13.6 Prior history of cirrhosis decompensations Variceal haemorrhage, n (%) 6 (3) 1 (3.2) 5 (3) 1.5 Ascites, n(%) Grade I 46 (23.1) 1 (3.2) 45 (26.8) 69.9 No 153 (76.9) 30 (96.8) 123 (73.2) Prior HCC6 treatment, n (%) Surgical resection 20 (10.2) 4 (12.9) 16 (9.6) 10.3 Radiofrequency ablation 54 (26.7) 3 (9.7) 51 (29.8) 52.3 Chemoembolization 91 (45.3) 14 (45.2) 77 (45.3) 0.3 ECOG-PS7, n (%) 38.2 0 126 (63.6) 17 (54.8) 109 (65.3) 1 70 (35.4) 12 (38.7) 58 (34.7) 2 2 (1) 2 (6.5) 0 (0) BCLC8 stage, n (%) 24.5 A 1 (0.5) 0 (0) 1 (0.6) B 75 (36.6) 9 (29) 66 (37.9) C 129 (62.9) 22 (71) 107 (61.5) Alphafetoprotein (ng/mL), median [IQR] 17 [5.4 - 229] 19 [4.9 - 182] 17 [5.8 - 270] 24.2 1. Standardized mean differences; 2. Interquartile range; 3. Chronic hepatitis C infection; 4. Non-alcoholic fatty liver disease; 5. Chronic hepatitis B infection; 6. Hepatocellular carcinoma; 7. Eastern Cooperative Oncology Group performance status; 8. Barcelona Liver Cancer Clinic. Missing values: Diabetes Mellitus (n=2), AHT (n=2), Ischemic cardiopathy (n=3), Peripheral vasculopathy (n=3), Variceal haemorrhage (n=5), Ascites (n=6). Surgical resection (n=8), percutaneous treatment n=3), Chemoembolization (n=4), ECOG-PS (n=7). A patient with an initial dosage of 600mg was excluded from the analysis. The values in bold were considered unbalanced (STD>10). Discussion initiated treatment with a lower dose of sorafenib (15%) presented more comorbidities (diabetes, AHT, ischemic A number of publications have explored the impact of age and cardiopathy and peripheral vasculopathy) than those who initial sorafenib dose on outcomes in HCC patients (9, 15, 17–19). started at full-dose. This information reflects the presence of However, our study is the first to use STD analysis to establish two different profiles of elderly patients who may be candidates whether there were significant baseline differences in elderly HCC for systemic treatments. Overall, 63.7% of all elderly patients had patients in terms of initial sorafenib dose and treatment period that a history of AHT but this proportion differed significantly might account for the differences in OS reported in the literature. between the groups (74.2% in low-dose and 62.2% in full-dose; Patient characterization is a key factor in the discussion of STD: 26). The same pattern was observed with the ECOG-PS>0 the treatment options for HCC. In our cohort, the patients who (low-dose: 45.2% vs full-dose: 34.7%; STD: 38.2). These data Frontiers in Oncology 05 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 TABLE 3 Sorafenib treatment characteristics and adverse events. atezolizumab-bevazizumab where the most frequent related AE are AHT and fatigue. Variable N= 206 In our study HSFR (43.7 and 25.8%) and diarrhoea (24.7 and Development of AE1, n(%) 155 (75.2) 3.2%) were more frequently reported with full-dose than with HFSR2 84 (40.8) low-dose sorafenib, but fatigue was more frequent in the low- Diarrhoea 44 (21.4) dose group (64.5% vs 59.2%). However, these results should be Fatigue 123 (59.7) considered as descriptive because the cohorts differed Development of early AE, n(%) 121 (58.7) significantly in most baseline characteristics related to prognosis. HFSR2 64 (31.1) Our study did not have a control cohort of younger patients, Diarrhoea 19 (9.2) this point being a limitation of our study. Some studies consider 70 Fatigue 80 (38.8) years to consider patients as elderly. One limitation of our study is Sorafenib treatment duration (months), median [IQR3] 5.6 (1.9-12.3) not having included patients between the ages of 70-74. Several Need for sorafenib dose adjustment, n(%) 108 (52.4) retrospective studies (albeit with small samples) have previously Definitive sorafenib discontinuation, n(%) 182 (88.3) ruled out differences in tolerability between groups with HCC of Reason for sorafenib discontinuation, n(%) different ages (6, 7, 12, 13, 19, 20). Only one small retrospective Symptomatic progression 84 (46.2) study reported more AE (bleeding) in patients over 75 years old due Adverse events 65 (35.7) to a greater exposure to antithrombotic drugs, and also a higher rate Radiological progression 23 (12.6) of treatment interruption in this subgroup due to AE (15). During Others 10 (5.5) the sorafenib era, there was a tendency to administer a lower Second-line treatment, n(%) 22 (10.7) starting dose in the elderly in order to avoid severe AE and Regorafenib 4 (18.2) complications. However, in the setting of atezo-beva dose Cabozantinib 2 (9.1) adjustment is not possible and therefore a delay strategy has to be Nivolumab 5 (22.7) considered. This issue is an unmet need and should be addressed Clinical trial 10 (45.5) separately. Morimoto et al. (17) compared patients who initiated Others 1 (0.004) sorafenib treatment with half doses or standard doses. The logistic Death, n(%) 175 (85) regression showed that older patients were frequently selected to Time from discontinuation to death (months), median [IQR] 5.4 (1.6-12.6) receive half doses, associated with less severe AE but similar OS. In 1. Adverse event; 2. Hand-foot skin reaction; 3. Interquartile range. our study only 15% of patients started at half dosage so it was not possible to establish differences in OS. Dose adjustment during treatment was needed in half of the patients, mainly in the first four show that elderly patients with a lower dose from the beginning months. Edeline and Williet et al. support tapering off sorafenib present an unfavourable profile that may be associated with a dosage in the elderly from the beginning, especially in those >80 higher incidence of AE and cirrhosis complications. An years, since two thirds of their patients of this age experienced grade important information for the analysis would have been to IV AE, which in turn increased the discontinuation rate (9, 15). have the number of medications taken by each patient. This Tovoli et al. also found that tailoring dosage on an individualized data was taken into account when making decisions by the basis in response to AE, lengthened treatment duration, increased clinician individually in each patient. Normally the number of the cumulative dose and improved OS (18). In contrast, in a cohort comorbidities is associated with the number of medications. of 792 patients Hajiev et al. reported that a starting dose of 800mg vs Unfortunately, these data were not included in the database that it was created, so it cannot be analyzed individually. In the TABLE 4 Adverse events according to initial dosage and time of IMBRAVE trial, the rate of patients who developed AHT under occurrence. sorafenib and atezolizumab-bevazizumab was 24.4 and 29.8% Initial sorafenib dosage (mg). N (%) (for any grade) and 12.2 and 15.2% (for grade 3-4), respectively (1). Regarding the rate of fatigue in patients treated with Development of AE1, n (%) 400 800 sorafenib and atezolizumab-bevazizumab, it was 18.6 and 2 HFSR No 23 (74.2) 98 (56.3) 20.4% for any grade, and 3.2 and 2.4% for grade 3-4. Yes 8 (25.8) 76 (43.7) Nevertheless, this trial did not consider safety as a primary or Diarrhoea No 30 (96.8) 131 (75.3) secondary end-point and elderly patients were not adequately Yes 1 (3.2) 43 (24.7) represented, since the median age was 66 (IQR: 59–71) in the Fatigue No 11 (35.5) 71 (40.8) sorafenib cohort and 64 (IQR: 56–71) in the atezolizumab- Yes 20 (64.5) 103 (59.2) bevazizumab cohort. Our study did not find any differences in terms of AE profile under sorafenib in this elderly population, 1. Adverse events; 2. Hand-foot skin reaction. One patient started with 600mg daily and but this issue must be analyzed in elderly patients treated with was excluded from the table. Frontiers in Oncology 06 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 TABLE 5 Prognostic factors related to overall survival. Group Categories Events Patients at risk OS1, months (95%CI) p-value All 175 206 15.4 (12.9 - 18.4) BCLC2 B 63 75 18.2 (12.9 – 22.5) 0.04 C 110 129 13.6 (11.7 – 18.4) Year of treatment initiation <2013 60 60 13.3 (11.5 - 17.6) 0.4 ≥2013 115 146 17.4 (12.9 - 21.2) AHT3 No 65 74 19.6 (12.6 - 26.6) 0.06 Yes 108 130 13.9 (11.3 - 17.4) DM4 No 110 126 14.8 (11.3 - 21) 0.8 Yes 63 78 15.4 (12.7 - 19) Age ≥75 - <80 137 154 14.7 (12.6 - 19.1) 0.6 (Years) ≥80 - <85 35 48 14.5 (10.2 - 18.5) ≥85 3 4 23.8 (21.2 - NE) 1. Overall survival; 2. Barcelona Liver Cancer Clinic; 3.AHT; 4. Diabetes mellitus. BCLC A patients were excluded from the analysis. 400mg a day did not modify OS in either elderly or non-elderly significantly across the treatment centres in different countries, patients (19). So the question of whether the starting dose should be ranging from a low of 5.8 months (95% CI: 4.4-7.2) in South Korea adjusted in the elderly with comorbidities from the beginning to a high of 14.3 months in Japan (95% CI: 9.2-19.4). None of these remains unanswered. studies found differences in OS in the elderly in comparison with The OS in this study (15.4 months) is higher than that recorded the youngest, even if a higher threshold of age (80 and 85 years old) in the largest multicentre international cohort study of patients aged was considered (6, 7, 20). Similarly, the median OS in the present 75 years or more (19), which reported an OS of 7.3 months without study is still higher than in the populations in those studies any differences with respect to patients below this age. The shorter regardless of age. This difference is at least partially explained by survival in this international study could be due to the inclusion of the patient selection. The policy in Catalonia is more restrictive in an unspecified proportion of BCLC-D patients, who in our area terms of the patient’s liver function, and also because the decision to were not administered sorafenib. In the observational studies start treatment in older people may have been influenced by the published so far (6, 7, 12, 13, 15, 20–24) median OS differs general condition and frailty of the patient. FIGURE 1 Survival plot by BCLC stage. Frontiers in Oncology 07 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 TABLE 6 Population baseline characteristics at sorafenib initiation according to period of treatment. Parameter All 2008-2013 ≥2014 STD1 (%) Patients, n 206 60 146 Age (years), median [IQR2] 77.9 [76.3 - 80] 77.1 [76 - 78.8] 78.2 [76.3 - 80.2] 16.0 Gender (male), n (%) 147 (71.4) 36 (60) 111 (76) 34.9 Cirrhosis, n (%) 191 (92.7) 58 (96.7) 133 (91.1) 23.4 Cirrhosis aetiology, n (%) 70.1 HCV3 116 (60.7) 46 (79.3) 70 (52.6) Alcohol 28 (14.7) 3 (5.2) 25 (18.8) NAFLD4 18 (9.4) 2 (3.4) 16 (12) Alcohol + HCV 13 (6.8) 3 (5.2) 10 (7.5) Others 8 (4.2) 0 (0) 8 (6) HBV5 5 (2.6) 2 (3.4) 3 (2.3) Alcohol + HBV 3 (1.6) 2 (3.4) 1 (0.8) Comorbidities, n (%) AHT 130 (63.7) 37 (63.8) 93 (63.7) 0.2 Diabetes mellitus 78 (38.2) 21 (36.2) 57 (39) 5.9 Ischemic cardiopathy 15 (7.4) 2 (3.5) 13 (8.9) 22.5 Peripheral vasculopathy 8 (3.9) 3 (5.3) 5 (3.4) 9.0 Prior history of cirrhosis decompensations Variceal haemorrhage, n (%) 6 (3) 3 (5.4) 3 (2.1) 17.5 Ascites, n(%) Grade I 46 (23) 17 (31.5) 29 (19.9) 26.8 No 154 (77) 37 (68.5) 117 (80.1) Prior HCC6 treatment, n (%) Surgical resection 20 (10.1) 4 (7.1) 16 (11.3) 14.3 Radiofrequency ablation 54 (26.6) 12 (21.1) 42 (28.8) 17.9 Chemoembolization 91 (45) 31 (54.4) 60 (41.4) 26.3 ECOG-PS7, n (%) 22.6 0 127 (63.8) 36 (67.9) 91 (62.3) 1 70 (35.2) 17 (32.1) 53 (36.3) 2 2 (1) 0 (0) 2 (1.4) BCLC8 stage, n (%) 18.2 A 2 (1) 0 (0) 2 (1.4) B 75 (36.4) 20 (33.3) 55 (37.7) C 129 (62.6) 40 (66.7) 89 (61) Alpha-fetoprotein (ng/mL), median [IQR] 17.9 [5.4 - 229] 44.9 [12.9 - 673] 13.9 [4.3 - 203.9] 9.3 1. Standardized mean differences; 2. Interquartile range; 3. Chronic hepatitis C infection; 4. Non-alcoholic fatty liver disease; 5. Chronic hepatitis B infection; 6. Hepatocellular carcinoma; 7. Eastern Cooperative Oncology Group performance status; 8. Barcelona Liver Cancer Clinic. Missing values: Diabetes mellitus (n=2), AHT (n=2), Ischemic cardiopathy (n=3), Peripheral vasculopathy (n=3), Variceal haemorrhage (n=5), Ascites (n=6). Surgical resection (n=8), percutaneous treatment n=3), Chemoembolization (n=4), ECOG-PS (n=7). The values in bold were considered unbalanced (STD>10). Raoul et al. (25) analysed outcomes in two different treatment with BCLC-C. Although the longer OS in both studies may reflect periods (2007-2012 vs 2013-2017) to evaluate the impact of the the incorporation of second- or third-line treatment into the HCC physicians’ learning curve on patients’ outcomes, and found OS to landscape, only 22 patients started second-line treatment in our be significantly longer in the second period. In the present study OS study. Second- and third- line options were introduced in Spain in was similar in the two populations, although it was higher in the 2017, but they are still not widely available. second period. Raoul et al. estimated the differences between The main limitation of our study is its retrospective design. The baseline characteristics using p values rather than STD, as we did physician’s subjective interpretation of the patient’s frailty is in our study; there were also significant differences between the essential for deciding the dose to be administered in advanced populations in terms of BCLC stage, as they included 173 patients HCC in the elderly. Using frailty scales from the beginning would (92%) with BCLC-C stage while we included 129 patients (63%) help to assess all patients with HCC who are candidates for systemic Frontiers in Oncology 08 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 FIGURE 2 Survival plot by year of sorafenib initiation. treatment. In our study, the fact that the acceptance or rejection of Data availability statement frail patients for systemic treatment depended on the physician’s clinical judgement may have introduced a selection bias and may The original contributions presented in the study are have been one of the reasons why the OS of our cohort was longer. included in the article/supplementary material. Further These results may prompt the thought that perhaps other patients inquiries can be directed to the corresponding author. who were left untreated due to frailty might have benefited from therapy. Other limitation of the study is the lack of younger control group to analyze the results. It could be interesting to apply the Ethics statement same analysis of STD in this population to compare survival and adverse events. As mentioned before, we defined two groups in This study was reviewed and approved by Ethics comitee of elderly patients depending on initial dose related to comorbidities Parc Tauli Sabadell. Written informed consent for participation and ECOG. Future prospective researches could be in this way in all was not required for this study in accordance with the national population treated with systemic therapy. legislation and the institutional requirements. Our study reinforces the idea that even though advanced age and comorbidity are intrinsic factors in elderly HCC, these factors should not bar elderly patients from receiving systemic Author contributions treatment. Rather, elderly patients should be considered as a special population. This study has identified certain confounding Garantor of the article: MV. Specific author contributions: AS factors related to the patient profile that may have a strong Data collection, analysis and interpretation of data, draft of the bearing on physicians’ decisions regarding the initiation of manuscript and critical revision of the manuscript. MaC, MeC, ZV, systemic treatment. Prospective studies in clinical practice are SM-M, MP, LC, RG, AG, BM, DH, AC, SM: data collection and needed to identify predictors of tolerability, and also to address critical review of the manuscript. VS: statistical analysis, and critical the issue of dose adjustment in this subgroup of patients. revision of the manuscript. MRo: study concept and design, data In conclusion, this is the first study to show significant collection, and critical revision of the manuscript. MRe: analysis and baseline differences in elderly HCC patients according to initial interpretation of data; drafting and critical revision of the sorafenib dose and treatment period. It has also characterized manuscript, and study supervision. MV: study concept and two groups of patients on the basis of their comorbidity profile. design, data collection, analysis and interpretation of data; Therefore, our data can be considered as a reference for drafting and critical revision of the manuscript and study defining the profile of elderly HCC patients who are supervision. All authors approved the final version of the candidates for atezolizu mab-bevacizumab as fir st - manuscript including the authorship list. All authors contributed line treatment. to the article and approved the submitted version. Frontiers in Oncology 09 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 Funding Laquente, Maica Galá n and Eduard Fort from the Institut Català d’Oncologia-L’Hospitalet; Margarita Sala from Josep AS: Travel grants from Tillots, Ferring, Norgine, Alfasigma, Trueta Hospital; Susana Coll from the Hospital del Mar Jansen, Abbvie. MC: None. ZV: None. SM-M: None. VS: Travel Medical Research Institute; Agnes Raga from Mutua Terrassa grants from Bayer. Consultancy LEO Pharma. MP: None. LC: Hospital and finally Jordi Ortiz from Terrassa Consorci Hospital None. RG: None. AG: None. BM: Consultancy: Bayer-Shering for their contributions to the study. We sincerely thank Michael Pharma, Eisai-Merck. Conferences/lectures: Eisai, MSD, Roche. Maudsley for their English correction. Research grant: Lab Viñas. Funding: BM is funded by grants PI18/00961 and PI21/00714 from Instituto de Salud Carlos III. DH. None. AC. None. SM. Conferences/lectures: Bayer. Travel Conflict of interest grants: Bayer, and Eisai. MRo. None. MRe. Consultancy: Bayer- Shering Pharma, BMS, Roche, Ipsen, AstraZeneca, Lilly. BTG/ The authors declare that the research was conducted in the Paid conferences: Bayer-Shering Pharma, BMS, Gilead, Lilly. absence of any commercial or financial relationships that could Research Grants: Bayer-Shering Pharma, Ipsen. MV. Travel be construed as a potential conflict of interest. grants: Gilead, MSD, Bayer, Abvie. Conferences/lectures: MSD, Gilead, Abvie, Eisai. Publisher’s note Acknowledgments All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated We sincerely thank Mireia Miquel, Jorge Sá nchez-Delgado, organizations, or those of the publisher, the editors and the Cristina Solé and Montserrat Gil from Parc Taulı́ Hospital; reviewers. Any product that may be evaluated in this article, or Á lvaro Dı́az, Marco Sanduzzi-Samparelli, Neus Llarch and claim that may be made by its manufacturer, is not guaranteed Gemma Iserte from Hospital Clı́n ic, Barcelona; Berta or endorsed by the publisher. References 1. Finn RS, Qin S, Ikeda M, Galle PR, Ducreux M, Kim T-Y, et al. Atezolizumab 11. Wong H, Tang YF, Yao T, Chiu J, Leung R, Chan P, et al. The outcomes and safety plus bevacizumab in unresectable hepatocellular carcinoma. N Engl J Med (2020) of single-agent sorafenib in the treatment of elderly patients with advanced hepatocellular 382(20):1894–905. doi: 10.1056/NEJMoa1915745 carcinoma (HCC). Oncologist (2011) 16(12):1721–8. doi: 10.1634/theoncologist.2011-0192 2. Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF, et al. Sorafenib 12. Marta GN, da Fonseca LG, Braghiroli MI, Moura F, Hoff PM, Sabbaga J. in advanced hepatocellular carcinoma. N Engl J Med (2008) 359:378–90. doi: Efficacy and safety of sorafenib in elderly patients with advanced hepatocellular 10.1056/NEJMoa0708857 carcinoma. Clinics (2021) 76:1–6. doi: 10.6061/clinics/2021/e2498 3. Cheng AL, Kang YK, Chen Z, Tsao CJ, Qin S, Kim JS, et al. Efficacy and safety 13. Arora SP, Ananth S, Ketchum N, Gelfond J, Michalek J, Mahalingam D. The of sorafenib in patients in the Asia-pacific region with advanced hepatocellular efficacy and safety of sorafenib in older adults with advanced hepatocellular carcinoma: a phase III randomised, double-blind, placebo-controlled trial. Lancet carcinoma: An analysis of a majority Hispanic cohort. J Geriatr Oncol (2020) 11 Oncol (2009) 10(1):25–34. doi: 10.1016/S1470-2045(08)70285-7 (7):1157–60. doi: 10.1016/j.jgo.2020.03.018 4. Kudo M, Finn RS, Qin S, Han KH, Ikeda K, Piscaglia F, et al. Lenvatinib 14. Montella L, Addeo R, Cennamo G, Vincenzi B, Palmieri R, Sperlongano P, versus sorafenib in first-line treatment of patients with unresectable hepatocellular et al. Sorafenib in elderly patients with advanced hepatocellular carcinoma: A case carcinoma: a randomised phase 3 non-inferiority trial. Lancet (2018) 391:1163–73. series. Oncology (2013) 84(5):265–72. doi: 10.1159/000345558 doi: 10.1016/S0140-6736(18)30207-1 15. Edeline J, Crouzet L, Le Sourd S, Larible C, Brunot A, Le Roy F, et al. 5. Brubaker TH, Powers EA. The stereotype of “Old” a review and alternative Sorafenib use in elderly patients with hepatocellular carcinoma: Caution about use approach. J Gerontol (1976) 31:441–7. doi: 10.1093/geronj/31.4.441 of platelet aggregation inhibitors. Cancer Chemother Pharmacol (2015) 75(1):215– 6. Ziogas DC, Papadatos-Pastos D, Thillai K, Korantzis I, Chowdhury R, Suddle 9. doi: 10.1007/s00280-014-2645-z A, et al. Efficacy and safety of sorafenib in patients with advanced hepatocellular 16. Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology carcinoma: Age is not a problem. Eur J Gastroenterol Hepatol (2017) 29(1):48–55. (2005) 42(5):1208–36. doi: 10.1002/hep.20933 doi: 10.1097/MEG.0000000000000739 17. Morimoto M, Numata K, Kondo M, Kobayashi S, Ohkawa S, Hidaka H, 7. Jo M, Yasui K, Kirishima T, Shima T, Niimi T, Katayama T, et al. Efficacy and et al. Field practice study of half-dose sorafenib treatment on safety and efficacy for safety of sorafenib in very elderly patients aged 80 years and older with advanced hepatocellular carcinoma: A propensity score analysis. Hepatol Res (2015) 45 hepatocellular carcinoma. Hepatol Res (2014) 44(13):1329–38. doi: 10.1111/hepr.12308 (3):279–87. doi: 10.1111/hepr.12354 8. Morimoto M, Numata K, Kondo M, Hidaka H, Takada J, Shibuya A, et al. 18. Tovoli F, Ielasi L, Casadei-Gardini A, Granito A, Foschi FG, Rovesti G, et al. Higher discontinuation and lower survival rates are likely in elderly Japanese Management of adverse events with tailored sorafenib dosing prolongs survival of patients with advanced hepatocellular carcinoma receiving sorafenib. Hepatol Res hepatocellular carcinoma patients. J Hepatol (2019) 71(6):1175–83. doi: 10.1016/ (2011) 41(4):296–302. doi: 10.1111/j.1872-034X.2011.00778.x j.jhep.2019.08.015 9. Williet N, Clavel L, Bourmaud A, Verot C, Bouarioua N, Roblin X, et al. Tolerance 19. Hajiev S, Allara E, Motedayеn Aval L, Arizumi T, Bettinger D, Pirisi M, et al. and outcomes of sorafenib in elderly patients treated for advanced hepatocellular Impact of age on sorafenib outcomes in hepatocellular carcinoma: an international carcinoma. Dig Liver Dis (2017) 49(9):1043–9. doi: 10.1016/j.dld.2017.06.008 cohort study. Br J Cancer (2021) 124(2):407–13. doi: 10.1038/s41416-020-01116-9 10. Eldesoky ES. Pharmacokinetic-pharmacodynamic crisis in the elderly. Am J 20. Nishikawa H, Kimura T, Kita R, Osaki Y. Treatment for hepatocellular carcinoma Ther (2007) 14:488–98 doi: 10.1097/01.mjt.0000183719.84390.4d in elderly patients: A literature review. J Cancer (2013) 4:635–43. doi: 10.7150/jca.7279 Frontiers in Oncology 10 frontiersin.org Soria et al. 10.3389/fonc.2022.829483 21. Kozyreva ON, Chi D, Clark JW, Wang H, Theall KP, Ryan DP, et al. A in the elderly. a study of the cancer of the liver Italian program (CLIP). Crit Rev multicenter retrospective study on clinical characteristics, treatment patterns, and Oncol Hematol (2006) 59(3):243–9. doi: 10.1016/j.critrevonc.2006.01.002 outcome in elderly patients with hepatocellular carcinoma. Oncologist (2011) 16 24. Hoshida Y, Ikeda K, Kobayashi M, Suzuki Y, Tsubota A, Saitoh S, et al. (3):310–8. doi: 10.1634/theoncologist.2010-0223 Chronic liver disease in the extremely elderly of 80 years or more: Clinical 22. Tsukioka G, Kakizaki S, Sohara N, Sato K, Takagi H, Arai H, et al. characteristics, prognosis and patient survival analysis. J Hepatol (1999) 31 Hepatocellular carcinoma in extremely elderly patients: An analysis of clinical (5):860–6. doi: 10.1016/S0168-8278(99)80287-6 characteristics, prognosis and patient survival. World J Gastroenterol (2006) 12 25. Raoul JL, Adhoute X, Penaranda G, Perrier H, Castellani P, Oules V, et al. (1):48–53. doi: 10.3748/wjg.v12.i1.48 Sorafenib: Experience and better manage-ment of side effects improve overall 23. Pignata S, Gallo C, Daniele B, Elba S, Giorgio A, Capuano G, et al. survival in hepatocellular carcinoma patients: A real-life retrospective analysis. Characteristics at presentation and outcome of hepatocellular carcinoma (HCC) Liver Cancer (2019) 8(6):457–67. doi: 10.1159/000497161 Frontiers in Oncology 11 frontiersin.org