Telephone: 919-966-2516; FAX: 919-966-1700
Liver transplantation for acute liver failure from drug induced liver injury in the United States
Article first published online: 23 JUL 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 10, Issue 8, pages 1018–1023, August 2004
How to Cite
Russo, M. W., Galanko, J. A., Shrestha, R., Fried, M. W. and Watkins, P. (2004), Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl, 10: 1018–1023. doi: 10.1002/lt.20204
- Issue published online: 23 JUL 2004
- Article first published online: 23 JUL 2004
- Mentored Career Development award. Grant Number: 5-K12-RR017667
- NIH. Grant Number: 1 U01 DK65201-01
Studies of acute liver failure from drugs have included cases mostly attributed to acetaminophen (APAP) but have reported limited data on other drugs. We used the United Network for Organ Sharing (UNOS) liver transplant database from 1990 to 2002 to identify recipients and estimate a U.S. population-based rate of liver transplantation due to acute liver failure from drugs. Patients were identified if their diagnosis was acute hepatic necrosis from an implicated drug at the time of transplant. Liver transplantation for drug hepatotoxicity accounted for 15% of liver transplants for acute liver failure over the study period. In our cohort (n = 270), 206 (76%) recipients were female. APAP alone, or in combination with another drug, accounted for 133 (49%) cases. In the non-acetaminophen (non-APAP) group (n = 137), the most frequently implicated drugs were: isoniazid, n = 24 (17.5%); propylthiouracil, n = 13 (9.5%); and phenytoin and valproate in 10 (7.3%) cases each. One-year patient and graft survival for the entire cohort was 77 and 71%, respectively. Among Caucasians (n = 206) and African-Americans (n = 48), APAP only was implicated in 110 (53%) patients and 12 (25%) patients, respectively, and non-APAP drugs were implicated in 96 (47%) patients and 36 (75%) patients, respectively (P = .0004). Among African-Americans in the non-APAP group, 28 (78%) were women. In conclusion four drugs were implicated in 42% of patients undergoing liver transplantation for acute liver failure due to drugs other than APAP. The increased frequency of African-American women undergoing liver transplantation for non-APAP drug induced liver injury warrants further study. (Liver Transpl 2004;10:1018–1023.)
Drug induced liver injury is an important medical problem in the United States. Serious drug induced liver injury may lead to hospitalization, and it is the most common identifiable cause of acute liver failure in the United States.1, 2 Drug induced liver injury has also been the most common single adverse event causing major regulatory problems during drug development and marketing, including failure to obtain FDA approval and the withdrawal of drugs from the market place.3
In recognition of the need for more investigation of drug induced liver injury, the National Institutes of Diabetes, Digestive and Kidney Diseases recently established the Drug Induced Liver Injury Network. Prior population-based studies from Europe suggest that severe drug induced liver injury is underestimated in single or multicenter studies.4, 5 Population based data on drug induced liver injury in the United States are limited. Reporting of adverse drug reactions through MedWatch is voluntary, and cases of serious drug induced liver injury may not be captured. A study that was conducted in Boston area hospitals in 1974 reported that 3 to 5% of patients are hospitalized for drug induced liver injury.6 More recently, in a study of acute liver failure from 17 centers in the United States, 308 cases of acute liver failure were captured of which 160 (52%) were due to drugs.2 Acetaminophen (APAP) accounted for 75% of the drug induced liver injury cases. Details of the other drugs implicated in causing acute liver failure were not reported.
Studies of drug induced liver failure in the United States have been small single center studies or large multicenter studies that have not provided details on the types of drugs, other than APAP, that led to acute liver failure and liver transplantation.1, 2, 8–10 All liver transplants performed in the United States should be entered into the United Network for Organ Sharing (UNOS) liver transplant database. This database, therefore, provides the opportunity to estimate a population-based incidence of liver transplantation for acute liver failure from drugs in the United States, including drugs other than APAP, and to identify characteristics of liver transplant recipients who developed acute liver failure from drugs.
We analyzed the UNOS liver transplant database from January 1, 1990 through October 31, 2002. The database consists of 51,741 liver transplant recipients of whom 492 adult and pediatric liver transplant recipients (<18 years old) indication for liver transplant was acute hepatic necrosis from a drug. The UNOS database consists of information from the transplant candidate registration and transplant recipient registration forms that collect demographic, clinical, and laboratory data.
Subjects were identified if they were diagnosed with acute hepatic necrosis from a drug. The text of each patient diagnosed with acute hepatic necrosis from a drug was reviewed to determine if a specific drug or drugs were implicated. Inclusion criteria for our analysis was: 1) subjects had at least 1 follow-up record after transplant, 2) subjects were UNOS status 1 at the time of transplant, 3) a specific drug was listed as the cause of acute liver failure. Herbal products could be listed as the cause of liver failure without identifying a specific product. Exclusion criteria was: 1) a diagnosis other than acute liver failure was provided in the text, 2) history of a prior organ transplant. Data on patient and graft survival were available for all subjects in our cohort. A total of 222 subjects did not meet the inclusion/exclusion criteria: 135 subjects were excluded because they were not UNOS status 1 at the time of transplant; 69 subjects were excluded because a specific drug was not listed (n = 67) or the text stated a diagnosis other than acute liver failure from drugs (n = 2); and 18 patients were excluded because they had a prior organ transplant. Our cohort consisted of 270 subjects with acute liver failure from an identified drug or drugs (Fig. 1). Nine patients in which APAP was not the sole drug listed were not included in the analysis of the APAP group.
Means were compared using Student's t test or Wilcoxon rank sum test where appropriate, and proportions were compared with Fisher's exact test. Kaplan-Meier survival curves were compared with the log-rank test.
Between 1990 and 2002, 2,291 patients were transplanted for acute hepatic necrosis from any cause of whom 357 (15%) were diagnosed with acute hepatic necrosis from drugs (range 8–20% per year) and UNOS status 1 at the time of transplant. A total of 270 recipients were identified for our analysis who met our inclusion and exclusion criteria, and the name of the implicated drug(s) was provided. A single drug was implicated in 258 (96%) cases, APAP and another drug in 9 (3%) cases, and multiple non-acetaminophen (non-APAP) drugs in 3 (1%) cases. The mean age of the study group was 33 years old, 206 (76%) were female, 206 (76%) were Caucasian, and 48 (18%) were African-American. The characteristics of the 67 recipients whose diagnosis was acute liver failure from a drug, but a specific drug was not listed, were similar to our entire cohort. The mean age of this group was 34 years old, 44 (67%) were female, and 13 (19%) were African-American.
Acute liver failure attributed to APAP hepatotoxicity occurred in 124 (46%) recipients, and other drugs were implicated in 137 (51%) recipients. In subsequent analyses, recipients who experienced drug induced liver injury attributed to APAP were considered separately from those with liver injury attributed to other drugs (referred to as APAP and non-APAP groups, respectively).
Subjects in both the APAP group (n = 124) and the non-APAP group (n = 137) were young and mostly women (Table 1). Mean total bilirubin was higher in the non-APAP group (P < .001). The mean serum creatinine was higher in the APAP group, and a greater proportion had serum creatinine 2 mg/dl or higher, 67 versus 23% in the APAP group, P < .001.
|Acetaminophen* (n = 124)||Non-APAP (n = 137)|
|Age (years)||32 (3–63)||35 (1–72)|
|Female (%)||96 (77)||103 (75)|
|White†||110 (89)||96 (70)|
|African-American||12 (10)||36 (26)|
|Other||2 (1)||5 (4)|
|Labs at time of transplant|
|Total bilirubin (mg/dL)†||9.8||23.1|
|Prothrombin time (seconds)||27||28|
|Hepatitis B surface antigen positive (%)||2 (1.6)||7 (4.8)|
|Hepatitis C antibody positive (%)||5 (4)||1 (.7)|
|O||55 (44)||53 (39)|
|A||48 (39)||62 (45)|
|AB||4 (3)||7 (5)|
|B||17 (14)||15 (11)|
|Days waiting (SD)†||1.6 (2.2)||3.6 (5.5)|
|Mean follow-up (days) (SD)||999 (1,104)||1,113 (1,155)|
Drug Induced Liver Injury in Pediatric Recipients
A total of 41 recipients with acute liver failure due to drug induced liver injury were children (<18 years old). The most common drugs in these children were APAP (n = 15), valproate (n = 8), and propylthiouracil (n = 4). A total of 3 out of 10 total liver transplant recipients (pediatric and adult) with acute liver failure attributed to valproate were 2 years old or younger and 6 cases were 4 years old or younger.
The most common drugs reported to lead to acute liver failure in the non-APAP group in descending order of frequency were isoniazid, propythiouracil, phenytoin, and valproate (Table 2). Herbal substances or mushrooms accounted for 16 (11.7%) cases of acute liver failure leading to liver transplantation in the non-APAP group. Herbal substances identified included kava-kava (n = 1) and chaparral tea, and vitamin A was implicated in 1 case.
|n = 137 (%)|
|Amanita mushrooms||9 (6.6)|
|Combination of non-APAP drugs||3 (2.2)|
|Other drugs (1 case each):||20 (14.6)|
|Amoxicillin-clavulinic acid, asparaginase, bromfenac, butorphanol, bupropion, carbemazepine, cocaine, hydrocodone, iron, isoflurane, ibuprofen, itraconazole, lisinopril, 6-mercaptopurine, naproxen paroxetine, pemoline, simvastatin, trimethoprim/sulfasoxazole, zafirkulast||1 (.7) each|
Characteristics of recipients at the time of transplant who developed acute liver failure from the most common drugs in the non-APAP group are shown in Table 3. With all listed drugs, women were more prevalent than men. Most of these injuries were associated with high bilirubin (Table 3). Compared to the APAP group (Table 1), isoniazid, propylthiouracil, and valproate were associated with a lower serum creatinine (P < .01 compared to APAP group) (Table 3).
|Age, Years (Range)||Female (%)||African-American (%)||Total Bilirubin (mg/dl)||Creatinine (mg/dl)|
|Isoniazid (n = 24)||39 (6–69)||16 (67)||8 (33)*||28*||1.2*|
|Propylthiouracil (n = 13)||28 (4–69)||9 (69)||6 (46)*||23.2*||.9*|
|Phenytoin (n = 10)||28 (20–46)||8 (80)||7 (70)*||15.1*||2.4|
|Valproate (n = 10)||10 (1–41)||8 (80)||1 (11)||20.5*||.5*|
|Amanita mushrooms (n = 9)||48 (15–70)||7 (78)||0||12.1||1.6|
|Herbal (n = 7)||41 (33–72)||7 (100)||1 (14)||23.6*||2.3|
|Acetaminophen (n = 124)||32 (3–63)||96 (77)||12 (10)||9.8||3.5|
Data on race were available on 269 recipients. A greater proportion of African-Americans underwent liver transplantation for non-APAP drugs compared to Caucasians, 36 (75%) and 96 (47%), respectively, P = .0004. Among African-Americans in the non-APAP group, 28 (78%) were women. Among African-American women (n = 37), 28 (76%) underwent liver transplantation for non-APAP drugs, while among Caucasian women (n = 158), 73 (46%) underwent liver transplantation for non-APAP drugs, OR = 3.6 (95% CI 1.6–8.0), P = .002 (Table 4). In addition, a greater proportion of African-American women underwent liver transplantation for the most common non-APAP drugs (isoniazid, phenytoin, and propylthiouracil) compared to Caucasian women (Table 4). The greatest difference was seen with phenytoin, 16 versus 1.3%, P = .0007.
|African-American Women (%) n = 37||Caucasian Women (%) n = 158||Odds Ratio [95% CI]||P Value|
|Non-APAP drugs||28 (76)||73 (46)||3.6 [1.6–8.0],||P = .002|
|Isoniazid||6 (16)||7 (4)||4.2 [1.4–12.7]||P = .02|
|Phenytoin||6 (16)||2 (1.3)||15.1 [3.3–∞]||P = .0007|
|Propylthiouracil||4 (11)||5 (3)||3.7 [1.0–13.6]||P = .07|
One-year patient and graft survival were 77 and 71%, respectively. One-year graft survival in the APAP group (n = 124) and non-APAP group (n = 137) were 70 and 73%, respectively, P = .46 (Fig. 2) and 1-year patient survival was 75 and 78%, respectively, P = .50. One-year graft survival for liver transplantation for acute liver failure from isoniazid (n = 24), propylthiouracil (n = 13), and phenytoin (n = 10) was 78, 84, and 60%, respectively, P = .06.
Our analysis of the UNOS liver transplant database is unique because it represents the largest series of patients who developed acute liver failure from any drug and from drugs other than APAP. Because all patients undergoing liver transplantation in the U.S. should be captured in the UNOS database, our study provides a U.S. population-based estimate of liver transplantation for acute liver failure for drug induced liver injury. Prior to this, the largest published series of acute liver failure due to drugs was a prospective study from 17 tertiary care centers that included 120 cases of acute liver failure from APAP and 40 cases of acute liver failure from other drugs.2 The names of implicated drugs other than APAP were not reported.
A novel finding in our study was that 42% of liver transplants for non-APAP liver injury was attributed to just 4 drugs: isoniazid, propylthiouracil, phenytoin, and valproate. Severe liver injury due to these drugs has been well described.10–19 However, prior studies have not been population-based and did not capture a substantial number of patients undergoing liver transplantation. Studies that have included patients who underwent liver transplantation typically include few patients, such as case series.18, 19 In our analysis, post-transplant survival was similar in the APAP and non-APAP groups.
A potentially important observation was the significantly higher proportion of African-Americans among the non-APAP group compared to the APAP group. APAP was the most commonly implicated drug among Caucasians while other drugs were most commonly implicated among African-Americans. This association was almost entirely accounted for by African-American women. The most significant association was seen between African-Americans and acute liver failure from phenytoin, where 60% of these patients were African-American women. This apparent over-representation of African-American women may not be explained by higher rates of epilepsies in African-Americans because investigators have reported similar rates of epilepsy between Caucasians and African-Americans.20 Others have reported that acute liver failure from non-APAP drugs is more common in non-Caucasians compared to Caucasians.2 The proportion of African-American patients was not reported. In our study, the number of African-Americans with acute liver failure from drugs was small, and our results would need to be confirmed in other studies before conclusions could be made about the association between race and acute liver failure from drug hepatotoxicity.
We found that APAP was the most common drug associated with liver transplantation for acute liver failure in children. In 1 series over half of the APAP overdoses in young children were unintentional, and we speculate that a substantial proportion of recipients in our study were accidental overdoses.21 Children also accounted for most cases of valproate hepatotoxicity, and this is consistent with other studies.17
Although there may be a perception that statins are associated with serious hepatotoxicity (with the exception of cerivastatin which has been withdrawn from the market), only 1 of the liver recipients in our cohort had acute liver failure attributed to a statin (simvastatin) over the 12-year study period. Although our study only captures liver transplant recipients and it does not include patients who recovered or died, other analyses support that statins do not lead to life-threatening hepatotoxicity.22, 23 Another interesting observation was the considerable number of cases in which herbal products were identified as the cause of acute liver failure. Herbal products have been reported to be associated with serious hepatotoxicity.24 Unfortunately, in only a few cases was a specific herbal product identified, and other more detailed studies on serious hepatotoxicity with complementary alternative medicine are needed.
There were several limitations to our study, including that we were unable to verify that cases of acute liver failure were due to the identified drug. Although causality assessment was not possible, a transplant physician who evaluated the case provided the data to UNOS and believed the patient developed liver failure from the specified drug. Data on explant pathology and associated symptoms, such as rash, would be important to collect in future studies. Recipients who received a liver transplant for acute liver failure from drug induced liver injury, but a drug was not specified, were not included in our analysis. There may be more cases of acute liver failure from a specific drug because the name of the implicated drug was not provided for all recipients. These recipients had similar characteristics to our study cohort suggesting they were not systematically different. Our analysis included patients who received a liver transplant and did not include patients who may have spontaneously recovered or died. Thus, our analysis should not be considered a comprehensive list of all drugs that may be associated with serious hepatotoxicity. Data such as stage of coma, other medications patients were prescribed, aminotransferases, alcohol intake, dose and duration of medications, and symptoms are either not provided or missing from the UNOS database. Despite these limitations our analysis may be useful to investigators designing trials of drug induced liver injury who may consider targeting patients taking the drugs that have been more commonly associated with acute liver failure from hepatotoxicity.
In conclusion, drugs were the identified cause for acute liver failure in 15% of patients undergoing liver transplantation in the United States between 1990 and 2002. When all patients are considered, APAP is the most frequently implicated drug followed by isoniazid, propylthiouracil, phenytoin, and valproate. A greater proportion of African-Americans were transplanted for acute liver failure from non-APAP drugs compared to Caucasians, with most cases African-American women. The higher than expected rate of African-American women in the group with non-APAP liver injury warrants further study.
The authors thank UNOS for providing the data.
- 23How to use statins in patients with chronic liver disease. Cleve Clin J Med 2003; 17: 58–62., .