Field-practice study of sorafenib therapy for hepatocellular carcinoma: A prospective multicenter study in Italy


  • Massimo Iavarone,

    1. A.M. & A. Migliavacca Center for Liver Disease, 1st Division of Gastroenterology, Fondazione IRCCS Ca' Granda Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author
  • Giuseppe Cabibbo,

    1. Sezione di Gastroenterologia, DIBIMIS, University of Palermo, Palermo, Italy
    2. Dipartimento di Biopatologia e Metodologie Biomediche, University of Palermo, Palermo, Italy
    Search for more papers by this author
  • Fabio Piscaglia,

    1. Division of Internal Medicine, Department of Digestive Disease and Internal Medicine, General and University S. Orsola-Malpighi Hospital, Bologna, Italy
    Search for more papers by this author
  • Claudio Zavaglia,

    1. Struttura Complessa di Epatologia e Gastroenterologia, Azienda Ospedaliera Niguarda Ca' Granda, Milan, Italy
    Search for more papers by this author
  • Antonio Grieco,

    1. Institute of Internal Medicine, School of Medicine, Catholic University of the Sacred Heart, Rome, Italy
    Search for more papers by this author
  • Erica Villa,

    1. Dipartimento di Medicina Interna, UO Gastroenterologia, Università di Modena & Reggio Emilia, Modena, Italy
    Search for more papers by this author
  • Calogero Cammà,

    1. Sezione di Gastroenterologia, DIBIMIS, University of Palermo, Palermo, Italy
    Search for more papers by this author
  • Massimo Colombo,

    Corresponding author
    1. A.M. & A. Migliavacca Center for Liver Disease, 1st Division of Gastroenterology, Fondazione IRCCS Ca' Granda Maggiore Hospital, University of Milan, Milan, Italy
    • 1st Division of Gastroenterology, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Università degli Studi di Milano, Via F. Sforza 35, 20122 Milan, Italy
    Search for more papers by this author
    • fax: 39-0250320410

  • on behalf of the SOFIA (SOraFenib Italian Assessment) study group

    1. A.M. & A. Migliavacca Center for Liver Disease, 1st Division of Gastroenterology, Fondazione IRCCS Ca' Granda Maggiore Hospital, University of Milan, Milan, Italy
    Search for more papers by this author

  • Potential conflict of interest: Fabio Piscaglia receives consulting and speaking support from Bayer and Bracco. Massimo Colombo has received grant and research support from Schering-Plough, Roche, Bristol-Myers Squibb, Gilead, and Bayer and is on the Advisory Boards of Schering-Plough, Roche, Novartis, Vertex, Bristol-Myers Squibb, Gilead, Bayer, and Tibotec. He also receives speaking and teaching support from Schering-Plough, Roche, Novartis, Vertex, Bristol-Myers Squibb, Gilead, and Bayer. The other authors have nothing to report. Travel support provided by Gilead (M. I.) and Bayer (G. C.).

  • SOFIA study group: Milan1—Massimo Iavarone, Angelo Sangiovanni, Sara Vavassori, Raffaella Romeo, and Massimo Colombo; Palermo2,3—Giuseppe Cabibbo, Vito Di Marco, Calogero Cammà, and Antonio Craxì; Bologna4—Alberto Borghi, Alessandro Granito, Fabio Piscaglia, and Luigi Bolondi; Milan5—Claudio Zavaglia, Aldo Airoldi, and Giovambattista Pinzello; Rome6—Marco Biolato, Simona Racco, Maurizio Pompili, and Antonio Grieco; Modena7—Barbara Lei, Nicola De Maria, and Erica Villa.


A multicenter randomized controlled trial established sorafenib as a standard of care for patients with advanced hepatocellular carcinoma (HCC). Because the study was prematurely interrupted due to survival benefits in the sorafenib arm, we conducted an observational study to adequately assess risks and benefits of this regimen in field practice. Starting in 2008, all clinically compensated patients with advanced HCC and those with an intermediate HCC who were unfit or failed to respond to ablative therapies were consecutively evaluated in six liver centers in Italy, for tolerability as well as radiologic and survival response to 800-mg/d sorafenib therapy. Treatment was down-dosed or interrupted according to drug label. Two hundred ninety-six patients (88% Child-Pugh A, 75% Barcelona Clinic Liver Cancer [BCLC]-C, and 25% BCLC-B) received sorafenib for 3.8 months (95% CI 3.3-4.4). Two hundred sixty-nine (91%) patients experienced at least one adverse event (AE), whereas 161 (54%) had to reduce dosing. Treatment was interrupted in 103 (44%) for disease progression, in 95 (40%) for an AE, and in 38 (16%) for liver deterioration. The median survival was 10.5 months in the overall cohort, 8.4 months in BCLC-C versus 20.6 months in BCLC-B patients (P < 0.0001), and 21.6 months in the 77 patients treated for >70% of the time with a half dose versus 9.6 months in the 219 patients treated for >70% of the time with a full dose. At month 2 of treatment, the overall radiologic response was 8%. Eastern Cooperative Oncology Group performance status, macrovascular invasion, extrahepatic spread of the tumor, radiologic response at month 2, and sorafenib dosing were independent predictors of shortened survival. Conclusion: Overall, safety, effectiveness, and generalizability of sorafenib therapy in HCC was validated in field practice. The effectiveness of half-dosed sorafenib may have implications for tailored therapy. (HEPATOLOGY 2011)

Hepatocellular carcinoma (HCC) is a lethal condition with an inexorably severe prognosis in the majority of patients who are diagnosed at an advanced, symptomatic stage when the chances for an effective treatment are limited.1-3 Prognosis is also grim for those patients with an intermediate, less severe tumor disease who are contraindicated for or do not respond to standard-of-care locoablative treatments like transarterial chemoembolization (TACE) or radiofrequency ablation (RFA).3 In 2008, this scenario was partially subverted by the advent of sorafenib (Nexavar, Bayer Healthcare Pharmaceuticals-Onyx Pharmaceuticals), an oral multikinase inhibitor that, by blocking cell proliferation and angiogenesis of the tumor, resulted in extended survival of patients with an advanced HCC.4-6 At variance with cytotoxic chemotherapy aimed at tumor cell necrosis, the clinical benefits of sorafenib stem from the ability of the drug to attenuate tumor cell proliferation and progression, occurring at the expense of adverse events (AEs) like diarrhea, fatigue, and skin lesions, whose management, however, was only rarely a challenge for clinicians.4-7

A registration trial in the United States and Europe and a confirmatory study in Asia in patients with advanced HCC highlighted a large variability in the clinical activity and tolerability of sorafenib, which was interpreted as a result of large interindividual variability in drug kinetics, in part related to differences in drug bioavailability.4, 8 However, because the registration trial was prematurely stopped at the second interim analysis due to significant survival advantage in the active drug arm (10.7 versus 7.9 months, P < 0.001), benefits and risks of sorafenib therapy could not be adequately assessed, as they were in other registration trials evaluating therapeutic regimens.9 Additional data on the safety of sorafenib are also required because registration trials are generally underpowered to evaluate drug tolerability and safety: they rely on rigid protocols and strict exclusion criteria, whereas field practice observational studies, notwithstanding several biases inherent to the traditional cohort studies, offer a chance to properly unravel the clinical effectiveness and toxicity of a drug like sorafenib when administered in real-life patients who have comorbidities.10-12 This is not a trivial point, because sorafenib therapy is costly and still in search of optimization due to the lack of serum biomarkers of early response that are deemed necessary to generate response-guided therapeutic algorithms. To bridge the chasm between efficacy and effectiveness of sorafenib therapy, we conducted a field practice observational study in six referral centers in Italy; we recruited patients with advanced HCC as well as patients with an intermediate tumor who failed to respond or had contraindications to locoablative treatments.


AE, adverse event; AFP, α-fetoprotein; AASLD, American Association for the Study of the Liver Disease; BCLC, Barcelona Clinic for Liver Cancer; CI, confidence interval; CT, computed tomography (contrast-enhanced); ECOG, Eastern Cooperative Oncology Group; FNB, fine-needle biopsy; HBV, hepatitis B virus; HCV, hepatitis C virus; HCC, hepatocellular carcinoma; HFSR, hand-foot skin reaction; HR, hazard ratio; MRI, magnetic resonance imaging; PS, performance status; RECIST, response evaluation criteria in solid tumors; RFA, radiofrequency ablation; TACE, transarterial chemoembolization.

Patients and Methods

Study Design.

This was a multicenter, investigator driven, observational, noninterventional study to assess the safety and effectiveness of sorafenib in patients with advanced HCC and patients with an intermediate HCC who were not eligible to or failed ablative therapies. The study, which had no industry support, started in July 2008 in six referral centers in Italy with known experience in the treatment of HCC. A kick-off meeting was held to establish common criteria for compliance, type/degree of AE, treatment interruption, and dose reduction. Assignment to sorafenib was not mandated by protocol but reflected the participating physician practice, thereby avoiding additional diagnostic or follow-up interventions. The primary objective was sorafenib safety, and the secondary objectives were treatment effectiveness in terms of overall survival (OS), early radiologic response, and time to radiologic progression. Treatment duration and cumulative dose were additionally evaluated.


All patients with advanced HCC, i.e., classified as Barcelona Clinic Liver Cancer (BCLC) stage C and those with a BCLC-B stage who were unfit for any or failed to respond to locoablative treatments were consecutively enrolled.13 Tumors were diagnosed by the criteria of the American Association for the Study of Liver Disease3 and staged by abdominal dynamic contrast-enhanced computed tomography (CT) or gadolinium-enhanced magnetic resonance imaging (MRI). Each patient underwent also chest radiographic/CT scan and bone scanning as requested by the attending physicians. Exclusion criteria were those of the Italian Drug Regulatory Agency (AIFA),14 i.e., a performance status (PS) score >2 and clinical decompensation. HCC recurrence after orthotopic liver transplant was also defined as advanced HCC, being therefore indicated for sorafenib therapy. Concomitant anti-hepatitis B therapy was allowed. Cirrhosis, which was diagnosed either by histology or clinically, was graded according to the Child-Pugh score as a measure of liver impairment.15 PS was scored according to the Eastern Cooperative Oncology Group (ECOG).16

Written informed consent was obtained from each patient according to the ethics committee and the ethical guidelines of the 1975 Declaration of Helsinki, as updated in 2004.

Routine Chemistries.

Blood cell count, serum chemistries, and serum α-fetoprotein (AFP) levels were measured by standard laboratory procedures. Serum hepatitis B surface antigen and antibodies to hepatitis surface antigen, to hepatitis delta virus, to human immunodeficiency virus, and to hepatitis C virus (HCV) were detected by standardized test systems (Ortho DS, Raritan, NJ).

Treatment Schedule, Dose Modification, and Interruption.

Sorafenib was administered at a dose of 400 mg twice daily following the indications provided by the manufacturer.7 Grade 3/4 AEs resulted in dose modification or treatment interruption whenever the AE was clinically relevant. A dose reduction (400 mg once daily) or temporary interruption was maintained until the symptoms resolved to grade 1 or 2 according to the guidelines provided by the manufacturer, followed by a re-escalation to the full dose. The dose was also modified in patients showing grade 2 toxicity on the patient's request or whenever a grade 2 AE was judged clinically relevant. Hepatic deterioration was another criterion for dose modification or interruption. Therapy was discontinued whenever the patient developed unacceptable toxicity, when radiologic or symptomatic progression of HCC occurred, or, by the investigator's judgment, when the patient was unlikely to benefit from further treatment with sorafenib.

Outcome Measures


In each patient, the medical history, physical examination, blood cell count, serum chemistries, coagulation, and AFP level were obtained at baseline and every 4 weeks thereafter. Each visit included recording of AEs, clinical laboratory tests, physical examination, and assessment of vital signs. Safety was assessed in all patients who received at least one dose of sorafenib; AEs were graded according to the National Cancer Institute's Common Terminology Criteria (version 3.0).17 Hepatic function deterioration was defined as a Child-Pugh score increase of ≥2 points; this was evaluated at each visit and at the predefined time points of weeks 12 and 24 of therapy.


OS was measured from the date of starting sorafenib therapy until the date of death from any cause or date of last visit. Tumor response was assessed peripherally in each center every 2 months during treatment and at follow-up using CT or MRI, according to modified response evaluation criteria in solid tumors (mRECIST) criteria.18, 19 Each participating center had published expertise in diagnosis and management of HCC, as well as expertise in the application of both RECIST and mRECIST criteria to assess a radiologic response to therapy.20-25 The time to radiologic progression was defined as the time that elapsed from baseline to disease progression (using mRECIST criteria) according to the radiologic films collected in each center.

Causes of Death.

When patients died with radiologic and/or clinical evidence of tumor progression at any time point, these deaths were classified as due to HCC progression. By the same token, when patients died with clinical decompensation such as jaundice, hemorrhage, encephalopathy, or sepsis lacking radiologic signs of cancer progression, these deaths were classified as due to liver failure.

Statistical Analysis.

Data were collected by experienced medical personnel involved in the study using a common electronic database. The primary outcomes were assessed by intention-to-treat, while continuous variables were expressed as mean ± standard deviation. Categorical variables were analyzed as frequency and percentages. The following baseline features were considered for univariate analysis: age, gender, ECOG PS, etiology of liver disease, platelet count, albumin level, bilirubin level, alkaline phosphatase, creatinine levels, international normalized ratio, AFP level, esophagogastric varices, macrovascolar invasion, extrahepatic spread, tumor size, number of neoplastic lesions, and treatment center; treatment dose, response, and time to radiologic progression were also considered. Data from patients receiving an anti-HCC treatment other than sorafenib were censored at the time of entry in the second-line regimen. Survival was analyzed by a Kaplan-Meier test, and differences in the survival rates were assessed by the log-rank test.

Variables with a P value < 0.10 at univariate analysis were included in the final multivariate model. Cox's proportional-hazard model was used to identify prognostic factors for mortality in a multiple regression analysis. Multiple logistic regression models were used to assess the relationship of both early radiologic progression and adherence to therapy with the demographic, laboratory, clinical, and tumor staging characteristics of patients. For all analyses, P ≤ 0.05 was considered statistically significant. All analyses were carried out with SAS version 8.1 software (SAS Institute, Cary, NC).



By July 2010, 296 patients had been consecutively evaluated (Table 1). There were no relevant differences among the patients enrolled in the six centers with respect to demography, cause or severity of liver disease, previous antitumor therapies, ECOG performance status, and BCLC stage. Chronic HCV infection was the dominant etiology, followed by chronic hepatitis B virus (HBV) infection and alcohol consumption. Two hundred and sixty (88%) patients were in the Child-Pugh A class. None of the 36 Child-Pugh B patients had ascites, clinically overt jaundice, or hepatic encephalopathy. At baseline, 115 patients (39%) had macroscopic vascular invasion by the tumor, whereas 104 (35%) had extrahepatic spread of the tumor, mainly in the abdominal lymph nodes, lungs, and skeleton. Sorafenib was the first treatment modality in 112 (38%) patients, whereas 18 (6%) had HCC recurrence after liver transplantation and 166 (56%) had previously failed locoablative treatment. Two hundred twenty-two (75%) patients were in BCLC-C stage and 74 (25%) in BCLC-B stage, including 26 (35%) who were unfit for locoablative treatment. At the time of analysis (October, 2010), 125 (42%) patients were still alive (55 on treatment) and 171 (58%) had died. Seven (2%) patients who were lost during follow-up were censored at the last visit. The median duration of treatment was 3.8 months (95% CI, 3.3-4.4) including 90 (30%) patients treated for ≥6 months and 41 (14%) treated for ≥12 months. The median treatment duration was 4.2 months (95% CI, 3.4-5.0) in the 260 Child-Pugh A patients compared with 2.0 months (95% CI, 0.2-3.8) in the 36 Child-Pugh B patients.

Table 1. Baseline Demographic, Laboratory, and Clinical Characteristics of 296 HCC Patients Treated With Sorafenib, According to BCLC Tumor Stage
  • Abbreviations: AFP, α-fetoprotein; BCLC, Barcelona Clinic Liver Cancer; CLIP, Cancer of the Liver Italian Program; ECOG, Eastern Cooperative Oncology Group; HCV, hepatitis C virus; HBV, hepatitis B virus; n/a, not applicable.

  • *

    Mean ± SD.

  • Baseline data on esophageal varices available in 256 patients.

Patients, no. (%)222 (75)74 (25)296
Age, years*66 ± 1069 ± 1067 ± 10
Male, no. (%)181 (82)61 (82)242 (82)
Etiology, no. (%)   
 HCV only116 (52)36 (49)152 (51)
 HBV only43 (19)15 (20)58 (20)
 Alcohol abuse only21 (10)10 (13)31 (10)
 Multiple22 (10)6 (8)28 (10)
 Other20 (9)7 (10)27 (9)
ECOG performance status, no. (%)   
 092 (41)74 (100)166 (56)
 1120 (54)n/a120 (41)
 210 (5)n/a10 (3)
Child-Pugh class, no. (%)   
 A192 (86)67 (91)259 (88)
 B30 (14)7 (9)37 (12)
Esophageal varices, no. (%)   
 Absent76 (41)35 (49)111 (43)
 Small87 (47)30 (42)116 (46)
 Medium/large22 (12)6 (9)28 (11)
Arterial hypertension, no. (%)91 (41)29 (39)120 (41)
AFP ≥ 200 ng/dL, no. (%)88 (40)16 (22)104 (35)
Disease burden, no. (%)   
 Macroscopic vascular invasion115 (52)n/a115 (39)
 Extrahepatic spread104 (47)n/a104 (35)
 One or both170 (77)n/a170 (57)
CLIP score   
 0n/an/a7 (2)
 1n/an/a82 (28)
 2n/an/a102 (34)
 3n/an/a71 (24)
 4n/an/a29 (10)
 5n/an/a5 (2)
Previous therapy97 (44)48 (65)145 (49)


The overall incidence of AEs was 91% (Table 2). Fatigue, weight loss, diarrhea, hand-foot skin reaction (HFSR), and arterial hypertension were the most frequent AEs, whereas hypothyroidism occurred in 5, perianal fistula in 3, acute pancreatitis in 3, alopecia in 6, and hypophosphatemia in 12 patients. In 133 (45%) patients, AEs were grade 3/4, including fatigue in 74 patients, HFSR in 25, diarrhea in 18, and gastrointestinal bleeding in 16. Myocardial ischemia occurred in three patients. Hepatic function deteriorated in 25 (15%) Child-Pugh A patients and in 2 (8%) Child-Pugh B patients after 12 weeks of therapy and in 17 (16%) Child-Pugh A patients and 6 (40%) Child-Pugh B patients after 24 weeks of therapy. During therapy, 71 (24%) patients developed ascites, 7 (2%) hepatic encephalopathy, and 7 (2%) jaundice.

Table 2. Treatment-Emergent Adverse Events in the 296 HCC Patients Treated With Sorafenib
Adverse Event*Any GradeGrade 1/2Grade 3/4
  • *

    Listed are treatment-emergent adverse events, as defined by the National Cancer Institute Common Terminology Criteria (v3.0), that occurred in at least 5% of patients.

Overall, no. (%)269 (91)136 (46)133 (45)
Constitutional symptoms, no. (%)   
 Fatigue195 (66)121 (41)74 (25)
 Weight loss115 (39)97 (33)18 (6)
Dermatological events, no. (%)   
 Hand-foot skin reaction82 (28)57 (19)25 (9)
 Rash15 (5)8 (3)7 (2)
Gastrointestinal events, no. (%)   
 Diarrhea103 (35)85 (29)18 (6)
 Nausea/vomiting34 (11)25 (8)9 (3)
 Constipation18 (6)18 (6)0
 Stomatitis17 (6)17 (6)0
 Bleeding26 (9)10 (3)16 (5)
Arterial hypertension53 (18)32 (11)21 (7)
Any cardiovascular event15 (5)8 (3)7 (2)

Dose Reduction and Treatment Interruption.

Treatment was down-dosed in 161 (54%) patients because of AE in 133 (83%) and liver function worsening in 28 (17%) (Table 3). The most frequent AEs leading to dose reductions were fatigue (39%), HFSR (18%), and diarrhea (14%). Treatment was permanently discontinued in 233 (79%) patients, as follows: for AEs in 94 (40%), for severe liver function deterioration in 38 (16%), and for HCC progression in 101 (44%). The main cause of treatment interruption for AE was fatigue, which occurred in 20 (6%) patients within 15 days from baseline.

Table 3. Frequency and Causes of Dose Reduction and Permanent Interruption of Treatment
 No. (%)
  • Abbreviations: AE, adverse event; HCC, hepatocellular carcinoma; HFSR, hand-foot skin reaction.

  • *

    Fatigue (n = 63), HFSR (n = 29), diarrhea (n = 23).

  • Fatigue (n = 20).

Dose reduction161 (54)
AE*133 (45)
 Deteriorated liver function28 (9)
Permanent discontinuation233 (79)
AE94 (32)
 Deteriorated liver function38 (13)
 HCC progression101 (34)

Ninety-seven (33%) patients discontinued treatment without a previous dose reduction, mainly because of AE (75%), and less frequently (25%) because of liver function deterioration. Duration of treatment was shorter in patients who interrupted for AE than in patients who interrupted for HCC progression (1.7 versus 8.7 months). By multivariate analysis, baseline predictors of premature discontinuation due to intolerance ECOG performance status, serum albumin, bilirubin levels, and age (Table 4).

Table 4. Risk Factors Associated With Treatment Interruption due to Intolerance in 296 HCC Patients Treated With Sorafenib*
Predictor (code)Univariate AnalysisMultivariate Analysis
HR (95% CI)P-valueHR (95% CI)P-value
  • Abbreviations: ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; CI, confidence interval.

  • *

    To avoid effect of colinearity with variables, BCLC and Child-Pugh scores were not included in the model.

ECOG performance status (1-2 vs. 0)1.9 (1.2-2.8)0.0021.6 (1.1-2.6)0.02
Albumin, g/dL (continuous)0.4 (0.3-0.7)0.00040.6 (0.3-0.9)0.03
Total bilirubin, mg/dL (continuous)1.7 (1.2-2.6)0.0031.5 (1.01-2.3)0.03
Age, years (continuous)1.03 (1.01-1.05)0.011.02 (1.01-1.05)0.04
Extrahepatic spread (yes vs. no)0.5 (0.3-0.9)0.02

A total of 77 (26%) patients received a half-dose of sorafenib for ≥70% of the treatment period, which lasted a median of 6.8 months (95% CI 4.2-9.4). Treatment dose was reduced because of AEs in 71 and because of liver function deterioration in 6. Among the remaining patients, 136 maintained a full dose of sorafenib for a median of 3 months (95% CI 2.2-3.8), whereas 83 received a half-dose for <70% of the whole treatment period of 3 months. In the latter patients, the cause of dose reduction was any AE in 73 and liver deterioration in 10.

Radiologic Response.

At month 2 of therapy, 2 (1%) patients had a complete response, 22 (7%) had a partial response, 217 (73%) had stable disease, and 55 (18%) had progressive disease. By multivariate analysis, younger age, extrahepatic spread of tumor, and ECOG performance status were associated with an increased likehood of nonresponse (Table 5).

Table 5. Risk Factors for Early (Month 2) Radiological Progression during Sorafenib Treatment in 296 HCC Patients*
Predictor (code)Univariate AnalysisMultivariate Analysis
HR (95% CI)P-valueHR (95% CI)P-value
  • Abbreviations: ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; CI, confidence interval.

  • *

    To avoid effect of colinearity with variables, BCLC and Child-Pugh scores were not included in the model

Age, years (continuous)0.95 (0.9-0.97)0.070.95 (0.9-0.98)0.003
Extrahepatic spread (yes vs. no)2.7 (1.5-4.9)0.0012.2 (1.1-4.2)0.010
ECOG performance status (1-2 vs. 0)1.9 (1.1-3.1)0.011.8 (1.0-3.1)0.034
Previous therapy (yes vs. no)0.6 (0.3-1.1)0.003--
Macroscopic vascular invasion (yes vs. no)2.4 (1.3-4.4)0.004--

Time to Radiologic Progression.

By October 31, 2010, a radiologic progression of the tumor was demonstrated in 101 (34%) patients, with a overall median time to progression of 9.2 (range, 5.5-12.9) months (Fig. 1A). The time to radiologic progression did not differ significantly between BCLC-B and BCLC-C patients (5.5 [range, 2.8-8.5] versus 12 [8-15] months) and between Child-Pugh A and Child-Pugh B patients (10 (range, 5.5-13.5] versus 6.9 [3.5-10] months)

Figure 1.

Kaplan–Meier analysis of 296 patients treated with sorafenib. (A) Time to radiologic progression. (B) Overall survival. (C) Survival according to BCLC.


The median OS was 10.5 months (Fig. 1B), corresponding to 49% of the patients being alive at 1 year. The survival rate of the 74 BCLC-B patients was longer than that of the 222 BCLC-C patients (20.6 versus 8.4 months; P < 0.0001; Fig. 1C), as was the median survival of Child-Pugh A compared with Child-Pugh B patients (12.7, 95% CI, 10.6-14.8 versus 7.7, 95% CI, 6.2-9.3). By October 31, 2010, 171 (58%) had patients died of HCC progression (80%), end-stage liver disease (18%), massive gastroesophageal hemorrhage, (1%) or liver-unrelated causes (1%). The distribution of causes of death was the same according to BCLC and Child-Pugh status. Death from decompensation in the first 6 months of therapy occurred more frequently in patients who received a full dose of sorafenib than in half-dosed patients (29 versus 3). By univariate analysis, the variables significantly associated with an increased likehood of mortality were ECOG performance status, macroscopic vascular invasion by the tumor, extrahepatic spread of tumor, age, albumin, total bilirubin, platelet count, and early radiologic progression at month 2. By multivariate analysis, impaired ECOG performance status (hazard ratio [HR] = 1.8, 95% CI 1.4-2.4), macroscopic vascular invasion (HR = 1.7, 95% CI 1.4-2.4), extrahepatic spread (HR = 1.5, 95% CI 1.1-2.4), and early radiologic progression at month 2 (HR = 1.6, 95% CI 1.1-2.4) were the only independent predictors of mortality (Table 6).

Table 6. Risk Factors for Overall Mortality in 296 HCC Patients Treated With Sorafenib*
Predictor (code)Univariate AnalysisMultivariate Analysis
HR (95% CI)P-valueHR (95% CI)P-value
  • Abbreviations: ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; CI, confidence interval.

  • *

    To avoid effect of colinearity with variables, BCLC and Child-Pugh scores were not included in the model.

ECOG Performance status (1-2 vs. 0)1.8 (1.4-2.3)<.00011.8 (1.4-2.4)<0001
Macroscopic vascular invasion (yes vs. no)2.1 (1.5-2.8)<.00011.7 (1.4-2.4)0.0009
Extrahepatic spread (yes vs. no)1.7 (1.2-2.3)0.0011.5 (1.1-2.2)0.01
Lack of radiological response (yes vs. no)2.6 (1.6-3.6)<.00011.6 (1.1-2.4)0.02
Age, years (continuous)0.98 (0.9 - 1)0.01
Total bilirubin, mg/dL (continuous)1.2 (1.1-1.3)0.03
Albumin, g/dL (continuous)0.7 (0.5-0.9)0.03
Platelet × 103/mmc (continuous)1.01 (1-1.02)0.03

The median survival of the 77 patients receiving a half-dose of sorafenib for ≥70% of the treatment period was 21.6 months (95% CI 13.6-29.6) compared with 9.6 months (95% CI 6.9-12.3) for the remaining 219 patients, who maintained full dosing or had a dose reduced for <70% of the treatment period (Fig. 2).

Figure 2.

Kaplan-Meier analysis of overall survival according to the prevalent dose of sorafenib (219 full-dosed or half-dosed <70% treatment period and 77 half-dosed ≥70% treatment period).

When sorafenib dosing was included in the multivariate analysis of the entire cohort of 296 patients, impaired ECOG performance status (HR = 1.9, 95% CI 1.5-2.5), macroscopic vascular invasion (HR = 1.9, 95% CI 1.4-2.6), extrahepatic spread (HR = 1.4, 95% CI 1.1-1.9), early radiologic progression at month 2 (HR = 1.4, 95% CI 1.1-2.1), and full dose (HR = 1.8, 95% CI 1.4-2.4) were the only independent predictors of mortality. Moreover, multivariate analyses performed excluding patients with early radiologic progression (55 patients) confirmed that full dose was independently correlated with mortality (HR = 1.6, 95% CI 1.04-2.4, P = 0.031).


This field practice study of 296 consecutively recruited patients in six Italian centers clearly validated sorafenib as a safe and effective treatment modality for advanced HCC. Strengths of this study were the large sample size, which was equal to the treated arm of the registration trial, the consecutive enrollment of real-world patients with broad eligibility criteria reflecting the complexity and diversity of our clinical practice in HCC, and the involvement of centers with well-referenced expertise in diagnosis and treatment of HCC patients, which guaranteed standardized radiologic assessment of the response to sorafenib.20-25 Another strength of our pragmatic study was the outstanding level of patient retention (98% of the enrolled patients had a final assessment) and the use of mRECIST criteria to assess the response to sorafenib; unlike classical RECIST, mRECIST captured changes in tumor vascularization caused by sorafenib in the absence of significant shrinkage of the tumor.18, 19

In our pragmatic study, sorafenib appeared to be a safe treatment modality, as in the registration and confirmatory trials and a few other uncontrolled studies in Asia and Europe.5, 6, 26-29 Fatigue, diarrhea, and HSFR were the dominant AEs in our patients; in the other studies these factors were distributed differently, fatigue being more frequently reported in our patients than in the registration and confirmatory trials (66% versus 20%-20%). In addition, we noticed a fair rate (18%) of arterial hypertension (occurring either de novo or as a worsening preexisting event) that occasionally led to dose reduction or interruption of sorafenib therapy. Although the lack of a control group prevented us from accurately weighing the role of arterial hypertension, the significant difference in the rates of arterial hypertension between sorafenib-treated and placebo-treated patients in both registration and confirmatory trials clearly argues in favor of a causal relationship between the vasoactive drug sorafenib and onset or deterioration of arterial hypertension in treated patients.30 This was even more so in a Phase III study in Japan and Korea in which patients with advanced HCC treated with sorafenib after TACE had much higher rates of arterial hypertension (33%) than placebo-treated controls (9%).31

In contrast to previous reports, in our study high rates (45%) of grade 3/4 AEs accounted for the many instances of reduction or interruption of sorafenib. Not unexpectedly, among the 133 patients who had to reduce sorafenib therapy for AEs, the most frequent causes were fatigue (39%), HSFR (18%), and diarrhea (14%); similar findings were reported in the confirmatory trial, in which 31% of sorafenib-treated patients had to down-dose sorafenib because of AEs.6 From a clinical point of view, more relevant was the finding of dose reduction and interruption of sorafenib therapy occurring as a consequence of liver function deterioration in 17% and 16% of our patients, respectively, i.e., more frequently than in sorafenib-treated patients in the registration trial (1%) and in the confirmatory trial, in which only 3.5% of the patients had to interrupt treatment due to liver dysfunction and ascites development.5, 6 We think that the higher prevalence of Child-Pugh B patients in our study compared with the registration and confirmatory trials (12% versus 3%-5%) could partially account for the high rates of discontinuation for liver impairment seen in our practice.

Our study differed from Phase III trials in terms of rates of poor tolerability, leading to dose reduction or discontinuation. In fact, only 44% of our patients were able to continue treatment with sorafenib in the absence of tumor progression, whereas a majority of the patients had to interrupt treatment because of the onset of AEs or liver dysfunction, and 26% had to reduce their dosage to a half daily dose (400 mg) of sorafenib for more than 70% of the treatment period, as a consequence of AEs, mainly fatigue. These figures were better in the registration trial, in which 76% of the patients received more than 80% of the planned daily dose of sorafenib.5

Our pragmatic study also provided indirect evidence of sorafenib effectiveness, as indicated by the median survival of 10.5 months overlapping with the 10.7 months of survival in treated patients in the registration trial, who had similar ages, rates of ECOG performance status, BCLC stage, and Child-Pugh scores as our patients.5 The differences in the prevalence of HCV-related tumors between the registration trial and the present study (29% versus 51%) have no clinical consequences, as suggested by the subgroup analysis of the registration trial, which negated any influence of etiology on the response rates to sorafenib in patients with advanced HCC.32

As in the registration trial, in our study the survival benefits in sorafenib-treated patients clearly reflected a treatment-related increase of the time to HCC progression. The fact that in our study the time to radiologic progression was significantly longer than in the registration trial (9.2 versus 5.5 months) despite similar duration of treatment (3.8 versus 5.3 months) relates to our choice to apply the more sensitive mRECIST criteria of response assessment, which evaluate changes in arterial vascularity of the tumor occurring in the absence of tumor shrinkage, as previously shown in patients undergoing RFA and TACE.33

As in previous studies,32 sorafenib therapy clearly also benefitted BCLC-B patients who failed to respond or had contraindications to the standard-of-care TACE or other local ablative therapies. A subanalysis of our cohort, in fact, highlighted a clear-cut improvement in survival of BCLC-B patients compared with BCLC-C patients (20.7 versus 8.5 months, P = 0.0004), which paralleled the background differences in the spontaneous survival rates between untreated groups of patients reported by Llovet and coworkers (16 versus 6 months).13

With all the caveats due to the lack of pretreatment patient stratification for survival predictors, post hoc analysis in our study indicated an increase in survival rates for the 77 patients who received a half-dose of sorafenib for more than 70% of the treatment period compared with the 219 patients who were either full-dosed or received a half-dose of sorafenib for less than 70% of the treatment period (21.6 versus 9.6 months). With all the caveats of a post hoc analysis, differences in treatment duration between down-dosed patients and full-dosed patients (6.8 versus 3 months) possibly reflect differences in tolerability between regimens, resulting in prolonged exposure to down-dosed sorafenib of less tolerant patients. Increased survival of the latter patients is rather surprising, but it could reflect the ability of a targeted cytostatic molecule like sorafenib to retain anticancer activity at suboptimal dosing, because targeted drugs do not follow the therapeutic paradigm of cytotoxic anticancer drugs, whose clinical activity depends strictly on optimal dosing and achievement of clinically effective levels in patient blood.34

It is noteworthy that the rates of liver dysfunction in the registration and confirmatory trials (1% and 3.5%, respectively) were definitely lower than in our pragmatic study, as was the prevalence of Child-Pugh B patients (5% and 3% versus 12%).5, 6 A recent interim report of a large multinational prospective study in patients with advanced HCC further highlights liver impairment as a cause of sorafenib discontinuation, because it demonstrated a direct relationship between the prevalence of AE-related discontinuation and severity of liver impairment assessed by Child-Pugh score.35

We also found a correlation of ECOG performance status, liver function, and patient age with an increased likelihood of stopping sorafenib therapy due to poor tolerability; this information might help with clinical counseling and pretreatment patient stratification in the design of therapeutic trials. The original contribution of our study was the identification of independent predictors of radiologic progression of HCC during the first 2 months of therapy, a finding that might help to optimize this costly therapeutic regimen through pretreatment patient stratification. Another original finding of our study was the identification of on-treatment predictors of mortality (such as early radiologic progression and dose regimen) that, if validated by independent studies, could help in identifying a stopping rule and dose tailoring to assist the clinician in cost-effective management of patients with HCC. Finally, a noteworthy byproduct of the present multicenter pragmatic study was the encouraging evidence of generalizability of sorafenib therapy, as we found no association between treatment center and outcome of therapy.