Abstract
Background/Aim: Hepatic encephalopathy is an adverse event resulting from lenvatinib use in patients with hepatocellular carcinoma (HCC). We analyzed the influence of lenvatinib on portal venous flow velocity (PVV) and serum ammonia concentration. Patients and Methods: Eleven patients with unresectable HCC were enrolled, including three with modified albumin-bilirubin (mALBI) grade 1, three with grade 2a, and five with grade 2b. PVV was measured by Doppler ultrasound sonography before and on day 2 of administration. Results: Out of 11 patients, one developed hepatic encephalopathy. PVV was reduced in 10 patients, and the change from baseline was significantly correlated with lenvatinib dosage. The increase in serum ammonia concentration was affected by lenvatinib dose and baseline hepatic function as a threshold between mALBI grade 2a and 2b statistically. There was no correlation between changes in PVV and serum ammonia concentration. Conclusion: Lenvatinib might directly disturb hepatocyte metabolism to result in increased serum ammonia concentration.
Lenvatinib, a novel oral angiogenesis inhibitor that targets vascular endothelial growth factor receptors (VEGFR) 1 to 3, fibroblast growth factor receptors (FGFR) 1 to 4, platelet-derived growth factor receptor-α, and RET and KIT oncogenes, showed preclinical and clinical activity in the treatment of solid tumors, including hepatocellular carcinoma (HCC) (1-3). Based on the results of the REFLECT phase III trial, lenvatinib was approved as a first-line single agent for the treatment of unresectable HCC in Japan, showing clinical benefit and a toxicity profile similar to other tyrosine kinase inhibitors (TKIs) such as sorafenib and regorafenib (4).
Both the phase II and III studies of lenvatinib for HCC showed higher incidences of hypertension than the preceding sorafenib or regorafenib trials (4, 5). Hypertension during the administration of TKIs is thought to be caused by reduced nitric oxide and increased endothelin production, which promotes vasoconstriction and capillary rarefaction (6). These phenomena mainly depend on the inhibition of VEGFRs by sorafenib and of both VEGFRs and FGFRs by regorafenib and lenvatinib. Because the cases of most patients with HCC are complicated by liver fibrosis, the capillary changes causing hypertension can also reduce hepatic blood flow to result in liver failure. Indeed, although small in number, several patients with HCC who experienced hepatic encephalopathy (HE) during administration of sorafenib or regorafenib were reported (7-10). For lenvatinib, however, evaluation of the incidence of liver failure including reversible HE was controversial. A phase II study of patients with unresectable HCC showed 5/46 (11%) of HE whereas there was no description of HE in a phase III study (4, 5).
To date, it is unclear how HE occurs in patients with HCC during TKI treatment, and whether the mechanism triggering HE is identical for all TKIs. We supposed that if TKIs have a shared process causing HE in patients with HCC, it might be reflected in a reduce in hepatic blood flow secondary to capillary constriction. In this study, we evaluated the correlation between changes in portal flow and serum ammonia concentration in patients with HCC treated with lenvatinib to test our hypothesis.
Patients and Methods
Patients. Eleven consecutive patients with advanced unresectable HCC who agreed to lenvatinib treatment from November 2018 to December 2019 were enrolled in the study. All patients were male and aged from 54 to 84 years, with a mean age of 71.1 years. Hepatic functional reserve was evaluated by Child–Pugh classification and modified albumin-bilirubin grade (mALBI grade) prior to the administration of lenvatinib. mALBI grading was undertaken using the score calculated with the following formula (11, 12): ALBI score=[0.66 × log10 birirubin (μmol/l)] + [−0.085 × albumin (g/dl)], with the result defined by the following score: ≤−2.60=grade 1; >−2.60 to ≤−2.27=grade 2a; >−2.27 to ≤−1.39=grade 2b; and >−1.39=grade 3.
The starting dose of lenvatinib was determined according to the drug information: Patients weighing <60 kg with Child–Pugh class A received 8 mg lenvatinib once daily, whereas those weighing ≥60 kg received an initial dose 12 mg lenvatinib once daily, and those with Child–Pugh class B received an initial dose of 8 or 4 mg of lenvatinib once daily. The administration of lenvatinib was performed 30 min after breakfast.
Serum examination for liver function and measurement of portal vein velocity (PVV) were performed just before treatment and on the second day after the initiation of the therapy. The clinical stage of HCC was determined by Barcelona Clinic Liver Cancer (BCLC) staging (13), and the effect of lenvatinib was judged as a partial response (PR), stable disease (SD), or progressive disease (PD) based on the modified Response Evaluation Criteria in Solid Tumors (14). When adverse events were found, they were classified according to the Common Terminology Criteria for Adverse Events (15).
All patients were required to provide written informed consent before study participation according to institutional guidelines. The trial was reviewed and approved by our Institutional Review Board (no. 214).
Measurements of portal flow. PVV was measured by Doppler ultrasonography (Aplio i800; Canon Medical Systems, Tochigi, Japan) with a 3.75-MHz convex probe. The examination was performed 2 h after lunch.
Portal hemodynamics were evaluated using both pulsed and color Doppler ultrasonography. Diameter, flow direction, and mean velocity were measured at the right or left portal vein trunk. Measurements of the PVV were taken five times and averaged. Only one physician performed all measurements to reduce the intra-observer measurement error variability.
Statistical analysis. Factors that might influence PVV or serum ammonia concentration were evaluated using simple/multiple linear regression analysis. Spurious correlations were recognized by examination of every partial correlation. Significance was established at p<0.05. All statistical analyses were performed using JMP software (Version 15.1.0, SAS Institute Inc., Cary, NC, USA).
Results
Characteristics of the patients. The enrolled patients included nine diagnosed with Child–Pugh class A and two with class B. Regarding mALBI grading, patients with Child–Pugh class A consisted of three each with grade 1, 2a, and grade 2b, while patients with Child–Pugh class B were classified as mALBI grade 2b. In the classification of clinical stage, six were BCLC stage B and five were stage C (Table I).
Response and adverse effects of lenvatinib. The mean period of lenvatinib administration was 214.8 days, ranging from 2 to 531 days. Of the 11 patients, PR and SD was observed in three and five, respectively (Table II).
Clinical stage and baseline hepatic functional reserve of the patients.
On the second day of administration, adverse events of greater than grade 2 were observed in three patients; two showed grade 2 hypertension, and one had grade 3 hypertension and grade 3 HE, for whom treatment with lenvatinib was abandoned that day. One patient (case no. 11) experienced fatigue on the third day of treatment, and continued administration after the dose was reduced from 12 mg to 8 mg per day.
Factors influencing PVV. All patients showed forward direction of portal flow, which did not change after 2-day administration of lenvatinib. Although earlier reports revealed that PVV reduced as liver fibrosis progressed (16), baseline PVV had no significant correlation with ALBI score (R2=0.0405, p=0.5529) or with Child–Pugh score (R2=0.1781, p=0.1961). In the examination of Doppler ultrasonography on the second day, 10 out of 11 patients showed reduced PVV compared with the value before treatment.
Based on the results that adverse events were more frequently observed in patients who were administered higher doses of lenvatinib or who were in a more progressive stage of liver fibrosis in the phase I and II studies of lenvatinib for HCC (5, 17), the influence of the dosage per body surface area (d/BSA) and the hepatic functional reserve to the PVV was examined. The change in PVV was expressed as the difference between PVV before and that 2 days after the introduction of lenvatinib (dPVV). Regarding the presenting factor of hepatic functional reserve, we adopted the ALBI score because it was superior to the Child–Pugh score in predicting the prognosis of patients treated with lenvatinib (17). Simple regression analysis showed that both d/BSA and ALBI score were significantly correlated with dPVV (Figure 1). The dosage of lenvatinib was determined partially according to Child–Pugh staging. Therefore, the possibility that either factor might be a spurious correlation should be confirmed. To determine the relation, the multiple regression analysis was performed. The dPVV was mainly affected by d/BSA (Table III).
Response and adverse events of lenvatinib treatment.
Factors influencing the changes in serum NH3 level. Of the enrolled patients, HE was found in only one patient, whereas eight out of 11 showed increased serum ammonia concentration on the second day. We set change in serum ammonia concentration as a factor indicating the risk of HE (dNH3), and examined the relationship between dNH3, d/BSA, ALBI score, and dPVV (Table IV). Excluding dPVV, which showed no correlation with dNH3, multiple regression analysis was performed using the other three factors, which indicated that d/BSA and the ALBI score had a significant independent correlation with dNH3 (Table V). The multiple regression model formula predicting dNH3 using d/BSA and the ALBI score was as follows: dNH3=118.8 + 94.9×ALBI score + 23.4×d/BSA (R2=0.5429, p=0.0437)
A recent study of patients with HCC treated with lenvatinib including not only Child–Pugh class A but also class B revealed that the rate of worsened Child–Pugh score was significantly higher in patients with a baseline Child–Pugh score of ≥6 in than those with a score of 5 (18). We also investigated another model that replaced the ALBI score with a variable that distinguished patients with a Child–Pugh score of 5 and others (Table V). The formula was as follows: dNH3=−198.58+98.91×α+32.32×d/BSA (R2=0.6399, p=0.0113) (Child–Pugh score=5; α=0, Child–Pugh score ≥6; α=98.91)
Comparing Child–Pugh score with mALBI grading, the threshold between Child–Pugh score 5 and ≥6 was equivalent to that between mALBI grade 2a and 2b. Therefore, the formula above can be described with grade 2a and 2b mALBI instead of Child–Pugh scores 5 and ≥6.
Discussion
TKIs are accepted as first- or second-line treatment for unresectable HCC. Currently, sorafenib and lenvatinib are available as first-line therapies, while regorafenib is placed as second-line treatment in Japan. Although HE was considered to be an adverse effect in the clinical trials of these drugs, it has not always been recognized as a specific event of TKIs. This is probably because most of the patients with HCC originally had liver fibrosis of various degrees and tended to develop HE with common triggers such as dehydration, constipation, and infection, which meant that it was difficult to distinguish whether the cause of HE observed during the treatment was from TKIs or merely an event of the clinical course (5, 19). Several authors, however, have reported cases of TKI-induced HE based on the lack of ordinary triggers. To minimize this ambiguity, we set a very short observation interval (2 days), which would reduce other incidents triggering HE and exclude the risk of deterioration of the hepatic functional reserve.
Before initiating this research, we predicted that portal flow would decrease with administration of lenvatinib through its effects on vasoconstriction and capillary rarefaction. In this study, we found a decrease in PVV in proportion to the dosage of lenvatinib, indicating that lenvatinib might have directly influenced the portal vein flow. However, it should be emphasized that portal flow volume can be affected by various factors such as oral intake and exercise (20, 21). Considering the high frequency of general fatigue and appetite loss during treatment with lenvatinib, it is difficult to assert the direct effect of lenvatinib on PVV with the small number in our analysis.
Simple linear regression analysis showed that the change in mean portal vein velocity (dPVV; PVV after 2-day administration of lenvatinib–PVV before treatment) was significantly correlated with albumin-bilirubin (ALBI) score (A) as well as with dosage per body surface area (d/BSA) (B).
We presumed that TKIs including lenvatinib might reduce the hepatic blood flow volume, which can attenuate hepatic metabolism and cause HE. However, the result of the multiple regression analysis for the changes in serum ammonia concentration showed that the dosage of lenvatinib and hepatic functional reserve, instead of the change in portal flow velocity, were significant variables to predict hyperammonemia. Simply accepting this result, lenvatinib seems to directly attenuate hepatic metabolic function in a dose-dependent manner, although the pathway of this is obscure, and HE developed only when a relatively high dosage was administered to a patient with poor hepatic functional reserve. It is noteworthy that there is a threshold affecting serum ammonia concentration between Child–Pugh scores 5 and 6. A recent multicenter study of lenvatinib for HCC including patients with Child–Pugh class A and B showed that the rate of worsened Child–Pugh score was significantly higher in patients with a baseline Child–Pugh score of ≥6 than in those with a score of 5 (11). Hiraoka et al. reported that the prognosis of patients with mALBI 2b was significantly worse as compared to those with mALBI 2a after treatment of HCC with lenvatinib (12). These reports combined with our results suggest that careful judgement of the hepatic functional reserve is required in the use of lenvatinib, even for Child–Pugh class A.
Multiple regression analysis for change in portal vein velocity.
Although the high frequency of hypertension commonly observed following treatment with TKIs encouraged us to investigate the effect on portal flow, we were unable to refer to the relationship between portal flow and hypertension in this study because hypertension of grade 2 or more was found in only three patients, which was a lower incidence than that reported in earlier mass studies. Nonetheless, it might be meaningful that the only patient who developed HE simultaneously experienced grade 3 hypertension.
Several researchers regarded HE during the therapy for HCC with TKIs as a drug class effect but evidence of the existence of a common mechanism has been scarce (7, 9). In contrast, posterior reversible encephalopathy syndrome, which is characterized by the acute onset of headache, nausea, dizziness, changes in consciousness, convulsions and transient visual disturbances (22), and which are generally recognized as class effects of TKIs due to interference with the VEGF pathway, is suspected to be a common theme in the pathophysiology. In the treatment of HCC, it is difficult to distinguish HE caused by some direct drug damage from that due to underlying liver dysfunction. Nevertheless, several case reports of HE with hyperammonemia that developed during the treatment of neoplasms other than HCC with TKIs such as regorafenib, sunitinib, and pazopanib indicate the existence of a common pathway of direct damage to hepatocyte metabolism. Our result showing the changes in serum ammonia that increased in proportion to lenvatinib dosage might support the possibility of HE as a class effect of TKIs.
Correlations between change in serum NH3 (NH3 after 2-day administration of lenvatinib – NH3 before treatment) (dNH3), dosage per body surface area (d/BSA), ALBI score, and change in portal vein velocity (dPVV).
Multiple regression analysis for change in serum ammonia concentration (dNH3).
In this study, we demonstrated that serum ammonia concentration during treatment of HCC with lenvatinib increased according to dosage as well as baseline hepatic liver functional reserve. Concerning the liver fibrosis grade, the threshold that affected the serum ammonia concentration was between Child–Pugh scores 5 and 6. Therefore, even when a patient is classified as Child–Pugh class A, careful ascertainment of their hepatic functional reserve and circumspect decisions regarding the lenvatinib dosage are required.
Acknowledgements
The Authors thank H. Nikki March, PhD, from Edanz Group (www.edanzediting.com/ac) for editing a draft of this article.
Footnotes
Authors' Contributions
Conceptualization: Ryoichi Narita. Methodology: Ryoichi Narita and Kazuhiro Kotoh. Formal analysis and investigation: Ryoichi Narita, Akitoshi Yoneda and Mitsuteru Motomura. Writing - original draft preparation: Ryoichi Narita. Writing – review and Editing: Kazuhiro Kotoh and Masaru Harada. Supervision: Masaru Harada. All Authors approved the final draft of the article.
Conflicts of Interest
All Authors declare no competing interests.
- Received June 6, 2020.
- Revision received June 26, 2020.
- Accepted June 27, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.