Background: Individuals with Gilbert's syndrome present with mild, unconjugated hyperbilirubinaemia, resulting from impaired glucuronidation by reduced uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) expression. The A(TA)7TAA polymorphism responsible for the syndrome has been associated with nilotinib-induced hyperbilirubinaemia in patients with chronic myeloid leukaemia (CML). Our study extends UGT1A1 molecular analysis to a larger cohort of CML patients receiving other tyrosine kinase inhibitors (TKIs) and explores the relationship with other abnormalities of liver function.

Methods: We interrogated our database of 832 patients treated with a TKI for CML at our center and identified 467 individuals who presented in first chronic phase, who had received only a TKI (prior interferon, combination experimental therapy, autologous and allogeneic transplant patients were all excluded) and for whom serial liver function results were available. We then performed PCR to identify variations in dinucleotide repeats in the UGT1A1 promoter region. Genotypes were assigned as follows: 6/6 (homozygous for (TA)6 allele; wild-type), 7/7 (homozygous for (TA)7allele) and 6/7 (heterozygous). Because individual patients may have received more than one TKI, we defined the period on treatment with each TKI as an 'episode' in order to be able to attribute abnormalities of liver function to a specific drug. Hyperbilirubinaemia was graded according to the Common Terminology Criteria for Adverse Events (CTCAE), version 4.0. Both ALT and ALP were defined as increased if they were above the upper limit of normal. To date we have completed the analysis for 340 patients comprising a total of 568 episodes of TKI therapy (imatinib 313, dasatinib 125, nilotinib 130).

Results: The UGT1A1 genotype analysis showed six variants. The majority of patients displayed one of 6/6 (149 patients), 6/7 (138 patients) or 7/7 (48 patients) genotypes. Small numbers of patients with the 5/6, 5/7 and 6/8 genotypes were excluded from further analysis. Hyperbilirubinaemia was seen in patients with all three genotypes, but was more frequent in patients on nilotinib (44%), compared to imatinib (14%) and dasatinib (8%). The incidence of hyperbilirubinaemia in individuals with the 6/6 genotype was 6%, 0% and 22% for imatinib, dasatinib and nilotinib respectively, and 10%, 6% and 56% for patients with the 6/7 genotype. The frequency of hyperbilirubinaemia in those with the 7/7 genotype was significantly higher compared to the occurrence of such events in any other genotype, irrespective of treatment. However, it occurred more often in patients on nilotinib (80%), followed by imatinib (61%) and dasatinib (44%). A significantly higher number of patients with the 7/7 genotype on nilotinib reported Grade 3 hyperbilirubinaemia (25%), with no grade 3 events on imatinib and only one on dasatinib. Hyperbilirubinaemia on any single drug did not predict for this event on any subsequent drug.

Further analysis of liver function tests (LFTs) and bilirubin levels divided episodes into four groups: 1. Episodes with normal bilirubin and normal LFTs (n = 273), 2. Episodes with isolated raised LFTs but normal bilirubin (n = 179), 3. Episodes with isolated hyperbilirubinaemia (n = 27) but normal LFTs and 4. Episodes with raised bilirubin and raised LFTs (n = 84). Abnormal LFTs with or without hyperbilirubinaemia were most commonly seen on nilotinib (65%) compared to imatinib (40%) and dasatinib (46%), confirming previous reports of the frequencies of transaminitis on the various TKIs. Isolated hyperbilirubnaemia was uncommon (<6%) on any of the drugs, but was more frequent in the 7/7 than the 6/6 or 6/7 genotypes for imatinib or dasatinib. Interestingly it was not seen in patients with the 7/7 genotype on nilotinib, as 80% of the 7/7 patients with hyperblirubinaemia on nilotinib had other abnormalities of liver function.

Conclusion. We confirmed the increased risk of hyperbilirubinaemia in individuals with the A(TA)7TAA polymorphism when treated with nilotinib. In contrast to other reports, we found that all patients with the 7/7 genotype are susceptible to increases in bilirubin when treated with any of imatinib, dasatinib or nilotinib. Hyperbilirubinaemia most commonly occurs alongside elevations in LFTs on nilotinib, indicating a nilotinib-induced hepatotoxicity that is compounded by the additional UGT1A1 inhibition.


Milojkovic:ARIAD Pharmaceuticals Inc.: Honoraria; Pfizer: Honoraria; BMS: Honoraria; Novartis: Honoraria. Apperley:Pfizer: Honoraria; Bristol Myers Squibb: Honoraria; Ariad: Honoraria; Novartis: Honoraria. Foroni:Bristol Myers Squibb: Research Funding; Novartis: Research Funding; Ariad: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.