Comparison of Treatment Safety and Patient Survival in Elderly versus Nonelderly Patients with Advanced Hepatocellular Carcinoma Receiving Sorafenib Combined with Transarterial Chemoembolization: A Propensity Score Matching Study

Aims This retrospective study was carried out to compare the outcomes between elderly (≥70 years of age) and nonelderly patients (<70 years of age) with advanced hepatocellular carcinoma (HCC) who received sorafenib combined with transarterial chemoembolization (TACE). Methods 88 patients with a confirmed diagnosis of advanced HCC were enrolled in this study. Of these, 24 elderly patients were matched with 48 nonelderly patients at a 1:2 ratio using propensity score matching to minimize selection bias. The related adverse events and survival benefits were compared between the two groups. Results Sorafenib combined with TACE was equally well tolerated in both age groups, and grade 3 or 4 adverse events were similarly observed in 54.2% of elderly and 50.0% of nonelderly patients (P = 0.739). There were no significant differences in survival time between the elderly and nonelderly patients (P = 0.876). Significant prognostic factors for overall survival as identified by multivariate analysis were the Child–Pugh score and portal vein invasion. Conclusions Sorafenib combined with TACE may be well tolerated and effective in elderly patients with advanced HCC. Age alone is not a parameter for the treatment of advanced HCC patients.

retrospective comparative study was approved by the ethics committee of the first affiliated hospital of Nanjing medical university. Written informed consent was obtained from each patient before treatment. All patients' medical records were accessed using their real names, which were anonymized before analysis. After the pretreatment investigations, our multidisciplinary treatment team of radiologists, hepatologists, and oncologists selected suitable patients to receive sorafenib combined with TACE. The diagnosis of HCC was reconfirmed in all patients by histological or cytological examination or on the basis of the American Association for the Study of Liver Diseases practice guidelines [25].
The inclusion criteria were advanced HCC (the presence of Barcelona Clinic Liver Cancer (BCLC) stage C disease involving vascular invasion and/or extrahepatic spread), an Eastern Cooperative Oncology Group performance status of 0 to 2, and Child-Pugh class A or B liver disease. The exclusion criteria were as follows: a history of liver transplantation; only nodal or distant metastases without viable liver lesions; any secondary malignancies; a history of concomitant use of another targeting agent, chemotherapy, or immunotherapy; and loss to followup. The TACE procedure and sorafenib treatment have previously described [26].

Follow-up
All TACE and sorafenib-related adverse events (AEs) were graded according to the National Cancer Institute Common Toxicity Criteria Adverse Events version 3.0. Laboratory tests performed during follow-up included hematologic and biochemical evaluations (e.g., complete blood cell count), prothrombin time measurement, and liver function evaluation. To assess liver function, the aspartate aminotransferase, alanine aminotransferase, alpha-fetoprotein, bilirubin, and albumin levels were measured every 4 weeks. Dynamic liver computed tomography or magnetic resonance imaging was performed every 6 to 8 weeks to evaluate the treatment response. If necessary, chest computed tomography and/or a bone scans were also performed to diagnose extrahepatic metastasis.

Statistical analysis
A 1:2 propensity score-matched analysis was performed to adjust for potential bias. This method is frequently used in observational studies because it allows for nonrandomized group assignments. We used nearest-neighbor matching with no replacement (a single participant could not be selected multiple times) to match patients in the nonelderly group with those in the elderly group (Stata command psmatch2; StataCorp LP, College Station, TX, USA) [27].
Differences between the two age groups were evaluated with Fisher's exact test or the chisquared test for categorical variables and Student's t-test or the Wilcoxon-Mann-Whitney test for continuous variables. We compared overall survival (OS) between the two propensity scorematched cohorts using the Kaplan-Meier method and log-rank test. Multivariable Cox proportional hazards models were used to determine the effect of age on OS after adjusting for the prognostic variables, and hazard ratios with 95% confidence intervals (CIs) were calculated.
All analyses were two-sided and performed at the 5% significance level. Statistical analyses were performed with SPSS software package version 13.0 (SPSS, Inc., Chicago, IL, USA) and Stata 12.0 (StataCorp LP).

Patient characteristics
88 patients with advanced HCC were admitted to our department from March 2009 to November 2013 and received sorafenib combined with TACE. Of these patients, 64 (72.7%) belonged to the nonelderly group (<70 years old) and 24 (27.3%) belonged to the elderly group (70 years old). An Eastern Cooperative Oncology Group (ECOG) performance status of 0 (P = 0.021), prior procedures (P = 0.043), and nodular HCC (P = 0.042) were more frequent in the nonelderly group. There were no between-group differences in liver cirrhosis, etiology, Child-Pugh score, ascites, extrahepatic tumor metastasis or portal vein invasion, or serum alpha-fetoprotein levels.
After performing nearest-neighbor matching (1:2) and based on the number of elderly patients, 48 nonelderly patients were matched for the analyses. The characteristics of the patients are listed in Table 1. No significant differences in the baseline characteristics of the patients were observed between the two groups after propensity score matching. In total, 24 elderly patients (33.3%; mean age, 76 years; range, 70-83 years) and 48 nonelderly patients (66.7%; mean age, 57 years; range, 31-69 years) were included in this analysis. Most patients were male (83.8% and 79.2% in the nonelderly and elderly groups, respectively), consistent with the local epidemiology. The majority of patients in both groups had reasonable liver function (Child-Pugh class A liver disease) prior to treatment; 68.8% of patients in the nonelderly group and 70.8% of patients in the elderly group had Child-Pugh class A liver disease (P = 0.856). Most patients in both age groups had an ECOG performance status score of 0 to 1 (68.7% and 58.3% in the nonelderly and elderly groups, respectively), although more elderly patients than nonelderly patients had a performance status score of 2 (31.3% and 41.7% in the nonelderly and elderly groups, respectively). The median follow-up period for all patients was 9.6 months (range, 1.8-39.7 weeks).

Safety and tolerability
Sorafenib combined with TACE was well tolerated in all patients in both age groups ( Table 2). TACE-related grade 3 or 4 abdominal pain was more common in elderly (4.2%) than in nonelderly patients (2.1%), and the incidence of grade 3 or 4 nausea and/or vomiting was lower in elderly (0.0%) than in nonelderly patients (4.2%). However, neither difference was statistically significant (P = 0.612 and 0.310, respectively). The most common sorafenib-induced AEs of all grades were hand-foot skin reaction, diarrhea, fatigue, alopecia, hypertension, and thrombocytopenia. The observed proportions of patients with sorafenib-related grade 3 or 4 hypertension and thrombocytopenia were larger in elderly patients (12.5% and 4.2%, respectively) than in nonelderly patients (4.2% and 0.0%, respectively), and the incidence of grade 3 or 4 hand-foot skin reaction, diarrhea, and alopecia were lower in elderly patients (8.3%, 8.3%, and 0.0%, respectively) than in nonelderly patients (10.4%, 10.4%, and 4.2%, respectively). Again, these differences were not statistically significant (P = 0.778, 0.778, and 0.310, respectively). Additionally, the proportions of other AEs did not differ significantly with age. Severe AEs, including gastrointestinal bleeding, hepatic encephalopathy, and hyperbilirubinemia, were observed in 4.2%, 8.3%, and 0.0% of the elderly patients, respectively. Dosage interruption due to AEs was observed in 45.8% and 33.3% of elderly and nonelderly patients, respectively, with no significant difference (P = 0.302).

Overall survival
Kaplan-Meier survival analysis indicated no significant difference in the median OS following combination therapy between elderly and nonelderly patients (6.5 [95% CI, 5.5-7.6] vs. 8.4 [95% CI, 5.6-8.8] months, respectively; P = 0.876) ( Table 3 and Fig. 1). We performed subgroup survival analysis according to independent prognostic factors for OS identified in the multivariate analysis (Table 4). Age did not have a significant influence on OS after adjusting for other related risk factors (P = 0.507 for OS).

Discussion
The optimal management of elderly patients with advanced HCC is an important issue because both the average age of patients and the incidence of HCC are increasing [28,29]. Limited information on the efficacy and safety of combination treatment in elderly patients with advanced HCC is currently available. The present study has addressed this issue. The OS was similar in the elderly and nonelderly patient groups. Furthermore, there was no significant evidence that age affected survival after adjustment for other prognostic factors. Although more AEs were observed in elderly patients, these events were largely manageable, and the incidence of AEs was similar in both age groups. We performed a case-control study of TACE combined with sorafenib versus TACE monotherapy to evaluate the treatment efficacy and safety in 246 patients with HCC [26]. Both univariate and multivariate analysis showed that age was not independently associated with OS. However, we did not compare safety and tolerance between the elderly and nonelderly patients. Therefore, in the present study, we comprehensively and thoroughly evaluated OS and AEs between elderly and nonelderly patients. Both the safety and tolerability in the present study population were similar to our previous experience with sorafenib and TACE.
Recent data were published concerning the safety and efficacy of sorafenib in a cohort of elderly patients from Hong Kong. In this population of >70-year-old patients with advanced HCC, the safety of sorafenib was shown to be equal to that in younger patients (grade 3 or 4 AEs, 68.6% and 62.7%, respectively; P = 0.56) [14]. More recently, the first non-Asian study [30] to investigate the use of sorafenib in a large cohort of elderly patients demonstrated that nonelderly patients who received sorafenib treatment were more susceptible to developing severe AEs than were older patients despite a lower rate of comorbidities (grade 3 or 4 AEs, 15.7% and 9.2%, respectively; P = 0.146). These results are similar to those obtained in another single-center study in Italy [31]. On the other hand, despite the poorer outcomes in elderly patients in a previous retrospective cohort study [32], a large case-control study recently performed by Yau et al. [33] and a 20-year multicenter experience from Italy indicated comparable efficacy of and tolerance to TACE for the treatment of advanced HCC in both young and elderly patients. More recently, TACE was reported to be safe and effective in selected very elderly patients with HCC [34]. However, more data are needed to confirm the safety of TACE in this population. With respect to efficacy, recent studies have demonstrated similar median survival, regardless of age, following monotherapy in elderly (70 years old) and nonelderly patients (<70 years old). Wong et al. [14] compared the efficacy of sorafenib in elderly (n = 37) and nonelderly patients (n = 135) with advanced HCC. The OS was similar in the elderly and nonelderly groups (5.32 vs. 5.16 months, respectively; P = 0.310). They concluded that the survival benefits of sorafenib are comparable in elderly and nonelderly patients with advanced HCC. Likewise, these results are similar to those obtained in another single-center study in Italy (16 and 12 months in elderly and nonelderly group, respectively) [31]. Yau et al. [33] conducted a large comparative study of 1040 patients with HCC treated with TACE (197 elderly and 843 nonelderly). They concluded that TACE had comparable efficacy in elderly and nonelderly patients with HCC. Finally, a prospective cohort study conducted by Cohen et al. [34] also confirmed that advanced age was not associated with a decreased survival rate.
In an attempt to compensate for the low cure rate obtained with TACE and the resulting hypoxia-induced angiogenic activity, combination treatment with sorafenib is currently being evaluated for unresectable HCC in two phase-II trials. Both of these studies have shown promising results with tolerable toxicity profiles [17,18]. Likewise, several continuous retrospective studies have further confirmed that combination therapy provides significantly better outcomes than does monotherapy for patients with advanced HCC [35,36]. To the best of our knowledge, however, the safety and efficacy of sorafenib combined with TACE have not been previously compared. Based on the present results, combination therapy appears to be equally effective for elderly and nonelderly patients with advanced HCC. Recent studies have identified tumor stage, tumor markers, and hepatic functional reserve as prognostic factors that affect the survival of patients with HCC [33,34,37]. Similar to the reported results, the prognosis after combination therapy largely depends on the pretreatment liver function and tumor burden, but patient age is not a significant factor. An important management strategy for combination  therapy is to ensure that no overlap occurs between the two treatments. To protect liver function in patients with HCC, oral sorafenib should not be administered within a 4-to 7-day window before or after the performance of TACE. There were several important findings in this study. First, sorafenib combined with TACE was shown to be safe in elderly patients with advanced HCC. Second, the survival benefits of combination therapy were comparable in elderly and nonelderly patients with advanced HCC. Finally, the outcomes of this retrospective cohort study provide new insights and guidance for treating elderly patients with advanced HCC.
There are limitations to the present study. First, this was a single-center experience, and the results may not be generalizable to patients with HCC in other countries. Second, it was a retrospective study involving a relatively limited number of elderly patients, potentially leading to patient selection bias. Propensity score matching was used to mitigate the potential confounding selection bias of this nonrandomized trial. Third, with respect to the timing of sorafenib administration, we applied an interrupted approach that involved placing patients on sorafenib between TACE procedures and temporarily stopping treatment around the TACE procedure. This method may be insufficient to inhibit the surge of angiogenic factors seen shortly after TACE [38]. However, it most likely provides the best balance between safety and efficacy. Finally, despite the results of the current study demonstrate the efficacy and safety of combined therapy in elderly patients with a high functioning status and satisfactory medical condition; thus, our results cannot be extrapolated to the general elderly population with advanced HCC.
In conclusion, for patient with advanced HCC who conform to the inclusion criteria and do not have concomitant disease that would hinder therapy, sorafenib combined with TACE may be well tolerated and effective in elderly patients with advanced HCC. Further prospective randomized trials are needed to confirm the potential safety and benefit of sorafenib combined with TACE in elderly patients.

Author Contributions
Conceived and designed the experiments: HH XL ZY. Performed the experiments: HH XL. Analyzed the data: HH ZD XL. Contributed reagents/materials/analysis tools: HH HS SL ZY. Wrote the paper: HH YH XT XX ZY.