- Top of page
- MATERIALS AND METHODS
Background: The aim of the study was to evaluate the incidence of abnormality of liver tests (LTs) or hepatotoxicity in a large group of inflammatory bowel disease (IBD) patients and, specifically, to assess the incidence of azathioprine (AZA) / mercaptopurine (MP)-induced liver injury in a long-term follow-up study.
Methods: All consecutive IBD patients followed for at least 5 years were included in this retrospective study. LTs including alanine transaminase, aspartate transaminase, alkaline phosphatase, γ-glutamyl transferase, and bilirubin were periodically monitored. “Abnormality-of-LTs” was defined as LTs between N (upper limit of the normal range) and 2 N, and “liver injury/hepatotoxicity” as LTs >2 N.
Results: A total of 786 patients were included, and 138 received AZA/MP; 120 patients (15%) and 39 (5%) presented abnormality of LTs or hepatotoxicity, respectively, during follow-up. The most frequent explanations were AZA/MP treatment and fatty liver disease. Among AZA/MP-treated patients (690 patient-years follow-up) the incidence of abnormal LTs and hepatotoxicity was, respectively, 7.1% and 2.6% per patient-year. Most patients spontaneously normalized LTs despite maintaining AZA/MP. These drugs were withdrawn due to hepatotoxicity (LTs >5 N and lack of decrease despite 50% dose reduction) in 3.6% of the patients and all of them normalized LTs.
Conclusions: In IBD patients, AZA or MP treatment induces abnormality of LTs in a relatively high proportion of the cases, but the development of true hepatotoxicity/liver injury is exceptional. Moreover, most of the cases of thiopurine-induced hepatotoxicity in IBD patients are mild, and the abnormalities in LTs spontaneously return to normal values despite AZA/MP being maintained, therapy withdrawal being necessary in only ≈4% of the patients.
It has been previously reported that inflammatory bowel diseases (IBDs) are frequently associated with pathologic findings in the liver and biliary tract, ranging from minor alterations, such as liver fatty changes, to severe conditions, like primary sclerosing cholangitis.1–3
The thiopurine drugs azathioprine (AZA) and 6-mercaptopurine (MP) have been shown to be effective at inducing and maintaining remission in IBD.4–6 The occurrence of side effects, however, is a major drawback in the use of these drugs, and concerns regarding short- and long-term toxicity have restricted their use.6–8 Both acute hepatocellular hepatitis and cholestatic disease have been described during AZA/MP therapy.8 The frequency of hepatotoxicity was initially described as low, but this incidence markedly varies among different series and data are sometimes contradictory.9 Furthermore, these data are based on a relatively low number of patients and with a limited follow-up.9 Consequently, the actual incidence of hepatotoxicity in IBD patients, mainly in those receiving AZA or MP therapy, has not been determined, and its real impact and management in clinical practice has not been fully characterized. Thus, despite increasing recognition of the potential for this serious adverse event, homogeneous guidelines for monitoring liver function in patients prescribed AZA/MP are not widely employed.9
Therefore, our aim was to evaluate the incidence of abnormality of liver tests (LTs) or hepatotoxicity in a large group of ≈800 patients with IBD and, specifically, to assess the incidence of thiopurine-induced liver injury in a long-term (5-year) follow-up study.
- Top of page
- MATERIALS AND METHODS
IBD is frequently associated with pathologic findings in the liver and biliary tract. Several series evaluating LTs in these patients reported abnormal findings both in CD and in UC ranging from 3% to almost 50%.2, 3, 13–19 More recently, a multicenter study described a 12% prevalence of hepatobiliary alterations among patients with IBD.20 In our study, 15% of the patients presented some type of abnormality of LTs at some time during the 5 years of the study, although only 5% had true hepatotoxicity (e.g., an increase >2 N in LTs).
The lack of universally standardized definitions of adverse drug reactions represents an important limitation of studies evaluating drug-induced liver toxicity.12 Furthermore, assessment of the risk/benefit ratio of a drug takes into account not only the incidence but also the severity of its adverse effects. Thus, lack of a definition of the reaction and of measuring its severity may result in the use of inaccurate or misleading terms, induce bias in evaluating safety, and lead to unjustified actions.12 Therefore, we classified the grade of hepatic analytical alterations using the proposed definitions of the International Consensus Meeting recommendations organized under the auspices of the Council for International Organizations of Medical Sciences (CIOMS)12: “Abnormality of LTs” or “liver injury”/“hepatotoxicity.”
When abnormal LTs are detected in a patient with IBD, several causes need to be ruled out: the patient should be interrogated as to drug intake (mainly thiopurines), alcohol abuse, serological test to rule out hepatitis A, B, and C infection, and laboratory tests including autoantibodies should be considered, and patients must undergo an abdominal ultrasound with the aim of excluding biliary tract abnormalities, focal lesions, portal hypertension, or hepatic steatosis.12 In this respect, nonalcoholic fatty liver disease, due to overweight (sometimes as a consequence of improvement of IBD activity) or to previous steroid treatment, may be responsible for some cases of abnormal LTs.
Possible explanations for the abnormal LTs in our patients are summarized in Table 2. As shown, the most frequent causes were AZA and MP treatment (38 and 11 patients, respectively), and fatty liver disease at abdominal ultrasound (49 patients). A recent study evaluated the prevalence of ultrasonography hepatobiliary changes in a large IBD population and found these abnormalities in more than 50% of the cases.1 Liver enlargement and steatosis were found in >35% of both CD and UC patients, a higher prevalence than among healthy controls.1 Other studies have also confirmed that more patients with IBD than controls have liver steatosis, with percentages ranging from 13% to almost 100%.20, 21 However, ultrasound changes including hepatomegaly or a dysechoic liver echo pattern have also been observed in the absence of any LT abnormalities in about 60% of IBD patients.19
The incidence of AZA/MP-associated hepatotoxicity markedly varies among different series and the data are sometimes contradictory.9 Furthermore, these data are based on a relatively low number of patients with a limited follow-up.9 Our study includes a large sample of patients (N = 786) with IBD in general, and also a high number of AZA/MP-treated patients (N = 138). Furthermore, our patients received thiopurine drugs for 5 years, giving a total of 690 patient-years of follow-up. Previous studies evaluating the incidence of AZA/MP-induced liver injury in patients with IBD are summarized in Table 4.7, 22–54 When all the studies were considered, including a total of 3485 patients, the mean overall prevalence of AZA/MP-induced liver disorders (abnormal LTs, without taking into account the follow-up period) calculated from these studies was 3.4%.9 However, when only studies where the follow-up time was specifically stated were considered, the mean annual drug-induced liver disorder rate (abnormal LTs per patient-year) was only 1.4%.9 In summary, the incidence of hepatotoxicity in IBD patients receiving AZA/MP, calculated from previously (mainly retrospective) published studies, seems to be only ≈1%–2% per patient and year of treatment.
Table 4. Studies Evaluating Incidence of Thiopurine-induced Liver Disorders in Patients with Inflammatory Bowel Disease
|Author||No. Patients||Thiopurine Drug||Mean Time Follow-up (Months)||Drug-induced Liver Disorder (%)||Definition of Liver Disordera|
|Ansari et al22||106||AZA||6||0/106 (0%)||NA|
|Bastida et al23||161||AZA/MP||9||21/161 (13%)||Abnormality of LTs|
| || || || ||16/161 (10%)||Liver injury|
|Bouhnik et al24||157||AZA/MP||14||4/157 (2.5%)||NAb|
|Candy et al25||31||AZA||15||0/31 (0%)||NA|
|Caprilli et al26||10||AZA||3||0/10 (0%)||NA|
|Cuffari et al27||25||MP||17||1/25 (4%)||NAd|
|De Jong et al28||50||AZA||28||0/50 (0%)||NA|
|Dubinsky et al29||92||AZA/MP||>4||16/92 (17%)||NAb|
|Ewe et al30||21||AZA||4||0/21 (0%)||NA|
|Fraser et al31||622||AZA||53||17/622 (2.7%)||NAb|
|George et al32||105||MP||3.5||5/105 (5%)||NAb|
|Goldenberg et al33||74||AZA/MP||24||0/74 (0%)||NA|
|Goldstein et al34||24||MP||7.4||0/24 (0%)||NA|
|Hanauer et al35||131||MP||24||0/131 (0%)||NA|
|Hawthorne et al36||33||AZA||12||0/33 (0%)||NA|
|Hindorf et al37||364||AZA/MP||18||19/364 (5.2%)||Liver injury|
|Jewell et al38||80||AZA||1||0/80 (0%)||NA|
|Kirk et al39||24||AZA||6||0/24 (0%)|| |
|L. Sanroman et al40||34||AZA||25||5/34 (15%)||NA|
|Mantzaris et al41||70||AZA||24||1/70 (1.4%)||NA|
|Markowitz et al42||27||MP||18||4/27 (15%)||NA|
|O'Donoghue et al43||51||AZA||12||0/51 (0%)||NA|
|O'Brien et al44||78||AZA/MP||19||1/78 (1.3%)||NA|
|Oren et al45||32||MP||9||0/32 (0%)||NA|
|Present et al7||396||MP||60||11/396 (2.8%)e||NAb|
|Present et al46||39||MP||24||0/39 (0%)||NA|
|Qasim et al47||110||AZA||3||4/110 (3.6%)||NAb|
|Regueiro et al48||58||AZA||3||1/58 (1.7%)||NAc|
|Rhodes et al49||15||AZA||2||0/15 (0%)||NA|
|Rosenberg et al50||10||AZA||6||0/10 (0%)||NA|
|Rosenberg et al51||15||AZA||6||0/15 (0%)||NA|
|Schawab et al52||83||AZA||—||3/83 (3.6%)||Abnormality of LTs|
| || || || ||2/83 (2.4%)||Liver injury|
|Summers et al53||39||AZA||9||0/39 (0%)||NA|
|Weersma et al54||318||AZA||—||7/318 (2.2%)||NAc|
Nevertheless, the progressive increasing number of reports of AZA/MP-induced liver disorders suggests that this adverse drug reaction may be underrecognized and underreported.9 In this respect, it is worth mentioning that almost all studies evaluating the hepatotoxicity potential of thiopurines are retrospective, and that until very recently no prospective studies had been published. Thus, Bastida et al23 in 2005 performed the first prospective epidemiological study that specifically addressed the incidence of thiopurine-induced hepatotoxicity in IBD patients. A cohort of 161 patients was prospectively followed for a median of 271 days. Abnormal LTs and hepatotoxicity was detected in 13% and 10% of the patients, respectively, a figure considerably higher than that reported in retrospective studies. These figures are closer to those reported in our study, where the incidence of abnormal LTs and hepatotoxicity in patients treated with AZA/MP was, respectively, 7.1% and 2.6% per patient and year of follow-up. Despite the retrospective nature of our study, the management scheme (indications, dose, criteria of adverse effects and definition of liver injury, and decision to pursue or discontinue treatment) was homogeneous since the same clinicians provided care from the onset of AZA/MP treatment in this single-center study.
It is unknown whether risk factors exist for thiopurine-induced hepatotoxicity. Although previous reports have shown an increased risk of AZA/MP-induced hepatotoxicity in men,23, 55 we could not confirm this association in our study. At the same time, age was not found to be a risk factor for liver injury in our patients taking AZA/MP.
At present, it is unknown whether there is any difference in the hepatotoxicity potential of AZA and MP, as prospective comparative studies between both drugs have not been performed.9 In our study, both drugs were associated with a similar incidence of hepatotoxicity. From the data included in Table 4, a mean incidence of hepatotoxicity of 2.1% can be calculated in those studies prescribing AZA, and 2.7% in those with MP, suggesting a similar risk with both drugs.9 Thus, although it has been suggested that, compared with MP, hepatotoxicity is less frequent, milder, and less dose-dependent with AZA,56 the hepatotoxicity potential of both drugs in patients with IBD seems to be similar, and, if differences exist, they are probably minor. This observation suggests that the risk with different thiopurine drugs is probably too small to influence the choice of agent.
There are no clear recommendations for the management of thiopurine-induced liver injury. Our proposed algorithm for AZA/MP-induced hepatotoxicity is summarized in Figure 1. A small percentage of AZA/MP-treated patients present a slight elevation of LTs that do not have clinical implications and the abnormalities in LTs return to normal values during the follow-up, indicating that it is not always necessary to adjust the dose of these drugs. In our study, only 18 of the 138 patients taking AZA/MP had LTs >2 N, and in most of these patients the abnormal LTs were classified as mild (<5 N). Furthermore, it should be underlined that most of these patients had spontaneously normalized LTs despite AZA/MP being maintained at the same dose. Similarly, abnormal LTs resolved spontaneously while continued on MP in 4 out of 5 patients in the study by George et al32 and in 3 out of 4 patients in the study by Markowitz et al.42 However, when abnormalities in LTs are more marked (a precise cutoff point has not been established, but we suggest a figure >5 N), but without associated jaundice, the dose of AZA/MP may be reduced by 50%. It is probably not necessary to withdraw AZA or MP, but frequent clinical and analytical controls (e.g., weekly or monthly) should be strictly performed after reducing the dose. With this strategy, AZA/MP treatment had to be withdrawn due to hepatotoxicity (LTs >5 N and lack of decrease despite 50% dose reduction in AZA/MP dose) in only 5/138 (3.6%) of our patients.
A favorable evolution of LTs after withdrawal of AZA/MP, with liver enzymes returning to normal after stopping the medication in all patients, has been reported.24, 31, 57 In our study all 5 patients in whom AZA/MP was withdrawn due to hepatotoxicity had normalized LTs after treatment was discontinued. However, it should be noted that in unusual cases thiopurines may induce a severe cholestatic jaundice that, in contrast with the acute hepatocellular hepatitis generally associated with AZA/MP, may not regress but even progress despite thiopurine withdrawal.58 Therefore, these drugs should be withdrawn, and not only tapered, in those patients presenting a clinically significant jaundice during thiopurine treatment.
In summary, this study shows that in IBD patients AZA or MP treatment induces abnormal LTs in a relatively high proportion of cases, but the development of true hepatotoxicity/liver injury is exceptional. Moreover, most of the cases of thiopurine-induced hepatotoxicity in IBD patients are mild, and the abnormalities in LTs spontaneously return to normal values despite AZA/MP being maintained, therapy withdrawal being necessary in only ≈4% of the patients.