Invest New Drugs DOI 10.1007/s10637-015-0292-9
SHORT REPORT
Liver function assessment according to the Albumin–Bilirubin (ALBI) grade in sorafenib-treated patients with advanced hepatocellular carcinoma Sadahisa Ogasawara 1 & Tetsuhiro Chiba 1 & Yoshihiko Ooka 1 & Eiichiro Suzuki 1 & Naoya Kanogawa 1 & Tomoko Saito 1 & Tenyu Motoyama 1 & Akinobu Tawada 1 & Fumihiko Kanai 1 & Osamu Yokosuka 1
Received: 15 August 2015 / Accepted: 30 September 2015 # Springer Science+Business Media New York 2015
Abstract Background: The Albumin–Bilirubin (ALBI) grade has been proposed as a new, simple, and objective method of assessing liver function. However, there is lack of data in sorafenib-treated patients with advanced hepatocellular carcinoma (HCC). Methods: We evaluated the correlations between the ALBI grade and Child–Pugh score, adverse events, and survival in 89 patients with advanced HCC who were prospectively treated with sorafenib. Results: Majority of patients with ALBI grade 1 (14/15 patients, 93 %) had a Child– Pugh score of 5. Patients with ALBI grade 2 had a wide range of liver function according to the Child–Pugh scores, with scores of 5, 6, 7, and ≥8. We divided ALBI grade 2 patients into ALBI grade 2A and 2B groups according to the median ALBI score among patients with ALBI grade 2. Although no significant difference was observed, the incidence of liver dysfunction in sorafenib-treated patients with ALBI grades 1, 2A, and 2B was 7 %, 19 %, and 35 %, respectively. Overall survival in the ALBI grade 2B group was significantly shorter than that in the ALBI grade 1 and 2A groups. Thus, ALBI grade 2B was an independent predictor of poor prognosis in addition to elevated serum aspartate aminotransferase levels, increased serum alpha-fetoprotein level, and macrovascular invasion. Conclusion: Sorafenib may be indicated for all patients with advanced HCC and ALBI grade 1 Electronic supplementary material The online version of this article (doi:10.1007/s10637-015-0292-9) contains supplementary material, which is available to authorized users. * Tetsuhiro Chiba
[email protected] 1
Department of Gastroenterology and Nephrology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan
and for some with ALBI grade 2. The subdivision of patients with ALBI grade 2 increases the utility of ALBI in identifying patients indicated for sorafenib therapy. Keywords Hepatocellular carcinoma . Sorafenib . The Albumin–Bilirubin grade . Child–Pugh B . Liver function
Introduction Hepatocellular carcinoma (HCC) is the sixth commonest cancer and the third leading cause of cancer-related deaths worldwide [1]. When choosing treatment strategies for HCC, it is necessary to assess not only tumor factors [number, size, presence of macrovascular invasion (MVI), or extra-hepatic metastasis (EHM)] but also liver function [2–4]. The Child–Pugh score is widely used for liver function assessment as it is convenient and easy to understand. The Child–Pugh score is based on a score derived from five parameters, including liver function tests, ascites, and hepatic encephalopathy. Both Eastern and Western guidelines for the management of HCC use the Child–Pugh score for liver function assessment. In addition, most clinical trials for HCC have adopted the Child–Pugh score for inclusion and exclusion criteria [5–11]. Sorafenib (Nexavar; Bayer Yakuhin, Osaka, Japan) is an oral, small-molecule multi-kinase inhibitor that blocks tumor cell proliferation via targeting Raf/MEK/ERK signaling at the level of Raf kinase and exerts an anti-angiogenic effect by targeting vascular endothelial growth factor receptor-2/−3 (VEGFR-2/−3) and platelet-derived growth factor receptorbeta (PDGFR-β) tyrosine kinase [12]. The significant survival benefit of sorafenib was demonstrated in two phase III trials in patients with advanced HCC [6, 7]. These studies included patients classified as Child–Pugh A. Thus, the global guidelines
Invest New Drugs
currently recommend sorafenib as the first-line therapy for patients with advanced HCC and Child–Pugh A [2–4]. However, the Child–Pugh score was originally developed to assess the prognosis of patients with cirrhosis and portal hypertension undergoing surgery for variceal bleeding [13, 14]. Moreover, the Child–Pugh score includes two parameters (ascites and encephalopathy) that can be difficult to estimate objectively. Recently, Johnson et al. proposed a new, simple liver function assessment for HCC; the Albumin–Bilirubin (ALBI) grade [15]. The ALBI grade is an evidence-based, objective method, allowing the stratification of Child–Pugh A patients receiving sorafenib for advanced HCC. However, further data are required to validate the utility of the ALBI grade as a measure of liver function of sorafenib-treated patients with advanced HCC in clinical practice or research trials. The aim of the present study was to evaluate the correlation between the ALBI grade and Child–Pugh score, adverse events, and survival in Child–Pugh A and B patients with advanced HCC receiving sorafenib.
Patients and methods The present study design has been described previously [16]. Briefly, the present study was a single-arm trial evaluating the safety and efficacy of sorafenib in Child–Pugh A and B patients with advanced HCC. Inclusion criteria consisted of failure of a treatment of established efficacy, such as resection or local ablation; a Child–Pugh score of A or B or an Eastern Cooperative Oncology Group Performance Status (ECOGPS) of 0, 1, or 2; and a life expectancy of at least 12 weeks. Patients received 400 mg sorafenib twice a day and were allowed up to two dose reductions (from 400 mg once daily to 400 mg every 2 days) for drug-related adverse events. The endpoints of the present study included overall survival (OS; measured from the date of registration to the date of death) and safety. Toxicity was graded according to the Common Terminology Criteria for Adverse Events, version 3.0. The present study was conducted in accordance with the principles of the Declaration of Helsinki and its amendments were in line with Good Clinical Practice. All patients provided written informed consent prior to participation. The present study was registered in the UMIN Clinical Trial Registry under the number UMIN000004215. Of the 101 patients assessed in the present study, 89 patients were enrolled (intention-to-treat population) of whom 88 patient were receiving sorafenib (safety population). In the present analysis, ALBI was calculated according to the baseline cecum albumin and bilirubin levels [ALBI = (log10 bilirubin × 0.66) + (albumin × 0.085), where bilirubin is in mol/L and albumin in g/L] [15]. According to previous literature, ALBI grades 1, 2, and 3 were defined as less than −2.60, between −2.60 and −1.39, and more than −1.39, respectively.
Statistical analysis The populations of the survival and safety analyses were the intention-to-treat and safety populations, respectively. Timeto-event data were estimated from Kaplan–Meier plots and presented as medians [95 % confidence interval (95 % CI)]. Adverse events were compared between ALBI grades using Fisher’s exact test. ALBI scores in terms of the emergence of adverse events were analyzed using the unpaired Student’s t-test. Univariate and multivariate Cox proportional-hazard regression models were used to estimate hazard ratios for risk factors in relation to OS. We included all enrolled patients (intention-to-treat population) in the analysis. In the analysis of predictors of OS, we took into account several parameters, including gender, age, etiology, ECOG-PS, ALBI grade, serum aspartate aminotransferase (AST) level, serum alpha-fetoprotein (AFP) level, liver tumor burden, MVI, EHM, and pretreatment. P-value less than 0.05 were considered statistically significant. All statistical analyses were performed using Statistical Package for the Social Sciences statistical software (version 22; SPSS-IBM, Chicago, IL, USA).
Results Correlation between the ALBI grade and Child–Pugh score Of the 89 patients enrolled in the present study, 15 patients (17 %) were classified as ALBI grade 1 and 74 patients (83 %) were classified as ALBI grade 2 (Table 1). Majority of the patients with ALBI grade 1 (14 of 15 patients: 93 %) had a Child–Pugh score of 5. Patients with Child–Pugh scores of 5, 6, 7, and ≥8 were included in ALBI grade 2. No patients in the present study were classified as ALBI grade 3. The median scores of ALBI grades 1 and 2 in the present study population were −2.903 (range, −3.073−−2.615) and −2.118 (range, −2.596−−1.439), respectively. Although majority of the patients were classified as ALBI grade 2 in the present study, we subdivided patients with ALBI grade 2 into two groups according to the ALBI score (grade 2A, ≤ − 2.118 and grade 2B, > − 2.118). The cutoff value for ALBI 2A and 2B (ALBI score, −2.118) was based on the median ALBI grade 2 score in the present study population. Correlations between ALBI grades 2A and 2B and Child–Pugh scores are shown in Table 1. ALBI grade 2A only included patients with Child–Pugh scores of 5, 6, and 7; with no patients with Child–Pugh scores 8 included in the ALBI grade 2A group. On the other hand, ALBI grade 2B included patients with Child–Pugh scores of 6, 7, and ≥8.
Invest New Drugs Table 1 Correlation between the Albumin–Bilirubin (ALBI) grade and Child–Pugh (C-P) score
Any
C-P score 5
C-P score 6
C-P score 7
C-P score ≥ 8
15 74
14 25
0 20
1 18
0 11
A. ALBI grade 1 and 2 ALBI grade 1 ALBI grade 2
B. ALBI grade 2A and 2B ALBI grade 2A
37
25
9
3
0
ALBI grade 2B
37
0
11
15
11
Drug-related adverse events according to ALBI grade The most frequent adverse events in the present study population were increased serum amylase levels (64 %), hand-foot skin reaction (HSFR) (61 %), increased serum AST levels (61 %), increased serum bilirubin levels (55 %), decreased platelet count (52 %), increased serum alanine aminotransferase (ALT) levels (49 %), hypertension (47 %), fatigue (44 %), and diarrhoea (40 %). Adverse events according to the ALBI grade are shown in Table 2. Although there was no significant difference in adverse events between ALBI groups, the rate of HFSR was high in the ALBI grade 1 group (grade 1, 80 % vs grade 2, 58 %) and rash or desquamation occurred less frequently in the ALBI grade 2 group (grade 1, 53 % vs grade 2, 26 %). The incidence of liver dysfunction (defined as any grade encephalopathy, ≥grade 3 ascites, or ≥grade 3 bilirubin increase) was high in the ALBI grade 2 group (grade 1, 7 % vs grade 2, 27 %). Moreover, the rate of liver dysfunction development was higher in the ALBI grade 2B group (35 %) than in the ALBI grade 2A group (19 %; Table 2). Next, we analyzed the incidence of the adverse events listed in Table 2 in patients with ALBI grade 2 (Supplementary Fig. 1). ALBI scores were significantly higher in patients with HFSR, rash or desquamation, and voice change. In contrast, ALBI scores in patients with liver dysfunction caused by sorafenib treatment were significantly worse compared with those in patients without liver dysfunction. Overall survival according to the ALBI grade The median OS of the present study population was 11.1 months (95 % CI, 8.4–13.8 months). The median OS of patients with ALBI grades 1 and 2 were 23.6 and 10.1 months, respectively (Fig. 1a), with no significant difference in OS observed between the two groups (P = 0.158). Fifty-nine of the 89 patients were classified as Child–Pugh A. Among Child–Pugh A patients, the median OS of ALBI grades 1 and 2 were 23.6 and 13.4 months, respectively (Fig. 1b), with no significant differences in OS observed between the two groups (P = 0.322). The median OS of patients with ALBI grades 2A and 2B were 14.5 and 6.6 months, respectively (Fig. 1c), with a significant difference in OS observed between
the ALBI 1 and 2B groups (P = 0.029) and between the ALBI 2A and 2B groups (P = 0.004). Next we investigated whether similar results were obtained with the use of a range of ALBI cutoff values. To determine additional cutoff values, we calculated median ALBI scores in patients with Child–Pugh scores of 5, 6, 7, and ≥8 (Supplementary Fig. 2). The median OS of ALBI grade 2 patients was analyzed according to the median score of patients with ALBI grade 2 (−2.000), a Child–Pugh score of 6 (−2.080), and a Child–Pugh score of 7 (−1.924) (Supplementary Fig. 3). These results were similar to those with the use of cutoff values determined by the median score of ALBI grade 2 (ALBI score, −2.118). Table 3 shows the univariate and multivariate survival analyses according to baseline characteristics. The risk of death was significantly high in the ALBI grade 2B group (hazard ratio, 2.265; 95 % CI, 1.021–5.022; P = 0.044). Multivariate analysis consistently identified baseline serum AST level, serum AFP level, and MVI as well as ALBI grade as independent predictors of OS.
Discussion This present study demonstrated a correlation between the ALBI grade and Child–Pugh score in patients with HCC. Majority of the patients with ALBI grade 1 had a Child– Pugh score of 5 (14/15 patients, 93 %). Namely, 36 % (14/ 39) of patients had a Child–Pugh score of 5 and 24 % (14/59) were Child–Pugh A in the present study. These results may be expected according to the previously reported calculation formula or heat map for the rapid assessment of the ALBI grade [15]. For instance, majority of the patients with a serum albumin of 3.5 g/dL or less (i.e., patients with an increased Child– Pugh score based on serum albumin) were not classified as ALBI grade 1. Importantly, patients classified as Child–Pugh score 6 because of decreased serum albumin were predominantly ALBI grade 2. Johnson et al. evaluated data from a number of different cohorts (Japanese, European, Chinese, and US) and compared patient ALBI grade and Child–Pugh class at initial diagnosis, although no detailed data regarding Child–Pugh scores were reported. The analysis performed in their study indicated that the number of patients at ALBI grade
Invest New Drugs Table 2 Drug-related adverse events according to the Albumin– Bilirubin (ALBI) grade (safety population)
A. ALBI grade 1 and 2 ALBI grade 1 (n = 15) Any (%) Grade ≥ 3 (%)
ALBI grade 2 (n = 73) Any (%) Grade ≥ 3 (%)
P Any
Grade ≥ 3
HFSR
12 (80)
1 (7)
42 (58)
16 (22)
0.147
0.285
Rash or desquamation Alopecia
8 (53) 4 (27)
1 (7) –
19 (26) 9 (12)
6 (8) –
0.062 0.224
1.000 –
Fatigue Anorexia
6 (40) 4 (27)
2 (13) 0 (0)
33 (45) 22 (30)
2 (3) 3 (4)
0.781 1.000
0.133 1.000
Diarrhoea Voice change
7 (47) 4 (27)
0 (0) 0 (0)
28 (38) 31 (42)
0 (0) 0 (0)
0.574 0.731
1.000 1.000
Hypertension
7 (47)
0 (0)
31 (42)
4 (5)
0.782
1.000
White blood cell decrease
4 (27)
0 (0)
27 (37)
6 (8)
0.560
0.584
Anaemia Platelet count decrease
5 (33) 5 (33)
0 (0) 1 (7)
26 (36) 41 (56)
3 (4) 13 (18)
1.000 0.156
1.000 0.448
AST increase
11 (73)
4 (27)
42 (58)
18 (25)
0.368
1.000
ALT increase
8 (53)
1 (7)
35 (48)
3 (4)
0.781
0.533
Bilirubin Amylase
6 (40) 11 (73)
0 (0) 4 (27)
40 (55) 45 (62)
7 (10) 17 (23)
0.397 0.558
0.598 0.749
Liver dysfunction*
1 (7)
–
20 (27)
–
0.106
–
HFSR
ALBI grade 2A (n = 36) Any Grade ≥ 3 25 (69) 10 (28)
ALBI grade 2B (n = 37) Any Grade ≥ 3 17 (46) 6 (16)
P Any 0.059
Grade ≥ 3 0.269
Rash or desquamation Alopecia
11 (31) 5 (14)
4 (11) –
8 (22) 4 (11)
2 (5) –
0.432 0.736
0.430
Fatigue Anorexia Diarrhoea
16 (44) 14 (39) 10 (28)
1 (3) 2 (6) 0 (0)
17 (46) 8 (22) 18 (49)
1 (3) 1 (3) 0 (0)
1.000 0.131 0.093
1.000 0.615 1.000
B. ALBI grade 2A and 2B
Voice change
12 (33)
0 (0)
3 (8)
0 (0)
0.010
1.000
Hypertension White blood cell decrease
18 (50) 12 (33)
3 (8) 4 (11)
13 (35) 15 (41)
1 (3) 2 (5)
0.241 0.630
0.358 0.430
Anaemia Platelet count decrease AST increase ALT increase
13 (36) 21 (58) 26 (72) 17 (47)
2 (6) 5 (14) 9 (25) 2 (6)
13 (35) 20 (54) 16 (43) 18 (49)
1 (3) 8 (22) 9 (24) 1 (3)
1.000 0.815 0.018 1.000
0.615 0.543 1.000 0.615
Bilirubin Amylase Liver dysfunction*
18 (50) 19 (53) 7 (19)
4 (11) 8 (22) –
22 (59) 26 (70) 13 (35)
3 (8) 9 (24) –
0.484 0.153 0.190
0.711 1.000 –
*Liver dysfunction is defined as any grade encephalopathy, grade 3 or more ascites, or grade 3 or more increase in serum bilirubin level Abbreviations: ALBI, Albumin–Bilirubin, HFSR, hand foot skin reaction; AST, aspartate aminotransferase; ALT, alanine aminotransferase
1 was smaller than that of patients classified as Child–Pugh A in all cohorts [15]. Taken together, these data demonstrated that patients with ALBI grade 1 maintained excellent liver function. On the other hand, majority of the patients with Child– Pugh scores of 6, 7, and ≥8 in the present study population were categorized as ALBI grade 2. In a previous report, majority of such patients were classified as ALBI grade 2 in all
cohorts [15]. As ALBI grade 2 appears to encompass a wide range of liver function and include a heterogeneous population, we then attempted to subdivide our study patients with ALBI grade 2 into two groups (ALBI 2A and 2B) according to the ALBI score. Johnson et al. reported a clear and non-overlapping survival curve for ALBI grades 1 and 2 in a large number of patients receiving sorafenib for advanced HCC. However, our results
Invest New Drugs
Fig. 1 Kaplan–Meier estimates of OS with regard to the ALBI grade. Prognostic analyses in all patients with ALBI grades 1 and 2 (a, n = 89), Child–Pugh A patients with ALBI grade 1 and 2 (b, n = 59), and all
patients with ALBI grades 1, 2A, and 2B (c, n = 89). The OS of patients with ALBI grade 2B was significantly shorter than that in patients with ALBI grades 1 and 2A (P = 0.029 and P = 0.004, respectively)
demonstrated no significant difference between ABLI grades 1 and 2, although the OS of ALBI grade 1 was greater than that of ALBI grade 2 overall and of the Child–Pugh A dataset (Fig. 1a, b). These results may be attributable to the number of patients included in the present study population (n = 89), particularly as only 15 patients were classified as ALBI grade 1. Although a relatively small number of patients were included in the present study population, no significant difference in survival curves was observed between patients with ALBI grades 1 and 2. However, a significant difference in survival curves was observed between patients with ALBI grades 1 and 2B and between patients with ALBI grades 2A and 2B. Our multivariate survival analysis demonstrated that ALBI grade 2B was a significant indicator of poor prognosis. On the other hand, patients with ALBI grade 2A had relatively better liver function than patients with ALBI grade 2B,
indicating that this ALBI cut-off value was not a significant poor prognostic factor. Tumor-associated factors, such as serum AST level, serum AFP level, or MVI, may influence patient prognosis more than factors associated with liver function in the present study population. Most of the patient adverse events were similar between ALBI grades, although the frequency of voice change and increased serum AST levels significantly differed between the ALBI grade 2A and 2B groups. No significant differences in any severe adverse events (grade 3 or more) were observed between ALBI grades. However, the frequency of liver dysfunction was higher in the ALBI grade 2B group than in the grade 2A and grade 1 groups. Majority of the patients with ALBI grade 2B were Child–Pugh B, and all of patients with Child–Pugh ≥8 were included in the ALBI grade 2B group. The frequency of liver dysfunction in the Child–Pugh score ≥ 8 group was significantly higher than that in previous reports [16].
Table 3 Univariate and multivariate survival analyses of baseline characteristics
Variables
Univariate analysis n
ALBI grade 1 15 2A 37 2B 37 AST, > 3 × UNL Absent 69 Present 20 AFP, > 400 ng/ml Absent 52 Present MVI Absent Present
P
Hazard ratio
95 % CI
Reference 1.230 2.597
0.558–2.711 1.183–5.703
Reference 2.207
Multivariate analysis
P
Hazard ratio
95 % CI
0.608 0.017
Reference 0.982 2.265
0.438–2.199 1.021–5.022
0.964 0.044
1.241–3.924
0.007
Reference 1.985
1.078–3.656
0.028
Reference
Reference
37
2.184
1.310–3.641
0.003
2.248
1.338–3.776
0.002
66 23
Reference 3.335
1.933–5.754
