Vancomycin-Induced Liver Injury, DRESS, and HLA-A*32:01

Abstract

Background:

Intravenous vancomycin therapy can cause liver injury as well as “drug-reaction with eosinophilia and systemic symptoms” (DRESS) syndrome. This study aimed to better define the clinical features and human leukocyte antigen (HLA) associations of vancomycin induced liver injury.

Objective:

To describe clinical, biochemical, and temporal characteristics of vancomycin-induced liver injury

Methods:

Cases of liver injury with recent exposure to vancomycin who were enrolled in the US Drug-induced Liver Injury Network between 2004 and 2020 were assessed. Sequencing of HLA alleles was performed on stored blood samples.

Results:

Among 1697 cases of drug induced liver injury identified between 2004 and 2021, 9 (0.5%) were attributed to intravenous vancomycin. The 9 cases included 6 men, median age 60 years (range: 23–85), and treatment for 26 days (range 1 to 34). The clinical presentation was DRESS syndrome in 8 patients, of whom 6 received corticosteroids. Liver injury varied from hepatocellular to cholestatic and from mild (n=5) to fatal (n=1). In survivors, liver injury and DRESS ultimately resolved. HLA typing demonstrated the HLA-A*32:01 allele in 7 vancomycin cases (78%, all with DRESS), versus 1 of 81 cases (1.2%) exposed but not attributed to vancomycin, and 113 of 1708 cases (6.6%) without vancomycin exposure. The allele frequency in vancomycin cases was 0.44 compared to less than 0.04 in US populations.

Conclusions:

Vancomycin-induced liver injury is commonly associated with DRESS and linked to HLA-A*32:01. HLA-A*32:01 testing could be considered early to risk-stratify patients using long term intravenous vancomycin therapy.

Summary:

Vancomycin is commonly used to treat multidrug-resistant bacterial infections and rarely has been associated with drug hypersensitivity syndromes. In our cohort, vancomycin-induced liver injury was associated with HLA-A*32:01, and frequently associated with drug reaction with eosinophilia and systemic symptoms.

Introduction

Vancomycin is a macromolecular glycopeptide antibiotic derived from Streptomyces orientalis which has potent bactericidal activity against most gram positive organisms[1]. Its antimicrobial activity is predicated upon inhibition of peptidoglycan synthesis, an essential structural component of the bacterial cell wall[2]. Initial widespread use of vancomycin was stymied by concerns of ototoxicity and nephrotoxicity, occurring typically when combined with aminoglycosides. Its potent activity against multi-drug resistant organisms such as methicillin-resistant Staphylococci has subsequently made vancomycin an invaluable agent in therapy of multidrug resistant endocarditis, osteomyelitis, and surgical site infections[3]⁴. Vancomycin requires parenteral administration; the absorption of oral formulations (used primarily for Clostridioides difficile colitis) being minimal except in rare instances of abnormal intestinal permeability, as in inflammatory bowel disease.

Intravenous (iv) vancomycin is generally well tolerated, but when given for more than a few days, it can cause transient, mild aminotransferase elevations in up to 5% of patients.^5,6,7,8 More rarely, extended therapy with vancomycin has been linked to instances of acute liver injury including “drug-reaction with eosinophilia and systemic symptoms” or DRESS syndrome.[4,5] Indeed, vancomycin is now reported to be the most common cause of DRESS syndrome in hospitalized patients in the United States and the most common cause of antibiotic-induced DRESS overall.[6]^,11 DRESS linked to vancomycin therapy is most commonly associated with renal injury but can result in accompanying serum aminotransferase elevations and in rare instances acute, symptomatic liver injury with jaundice. In contrast, oral vancomycin, which is used largely as therapy of Clostridioides difficile diarrhea, has not been linked convincingly to clinically apparent liver injury.

Recent studies have demonstrated a strong association with vancomycin-induced DRESS syndrome and HLA-A*32:01.[7] Whether there is a similar association with vancomycin-induced liver injury with or without DRESS is not well defined. In the U.S Drug Induced Liver Injury Network (DILIN) prospective study, vancomycin liver injury was uncommon, representing less than 1% of cases.¹³ The aims of the current analyses were to better define the clinical and biochemical features, temporal characteristics, and HLA associations of vancomycin-induced liver injury in the United States.

Methods

The US Drug-induced Liver Injury Network (DILIN) is an NIH-funded prospective study of clinically apparent liver injury suspected to be caused by prescription and over-the-counter medications as well as herbal and dietary supplements (HDS). The design, methods, terminology, and definitions used in the Network have been described in detail.¹³ Briefly, all patients with acute liver injury suspected to be caused by drugs or HDS products seen at 5 to 8 U.S. medical centers were prospectively enrolled in a database maintained by a central data coordinating center (the current and previous sites are provided Table E1). At the time of enrollment, patients underwent medical history, physical examination and laboratory assessments, and all relevant data and blood test results were extracted from the medical record. Patients were scheduled to be seen again 6 months after onset and at 12 and 24 months if there was evidence of persistent liver test abnormalities. All patients provided written informed consent for enrollment, and all details of the study protocol were reviewed and approved by institutional review boards at each site as well as by a Data Safety and Monitoring Board appointed by the NIDDK.

Causality assessment was done by committee based upon a written clinical narrative and case report forms that summarized test results for up to 6 months after enrollment. Each case was scored independently by 3 DILIN investigators on whether injury was drug-induced and which of the implicated agents was responsible, with likelihood scores of definite (>95%), highly likely (75%-95%), probable (50%-74%), possible (25% to 49%), or unlikely (<25%). Discrepancies among the 3 reviewers were then resolved by consensus on conference calls. For the current analysis, all cases for which vancomycin was one of the implicated drugs (n=12) as well as all cases that had received vancomycin within 2 months of onset but listed only as a “concomitant” medication (n=78) were re-reviewed by 3 of the authors (BAA, CK, and JHH) and scored similarly for likelihood, with discrepancies being resolved in a group conference call. All causality scores were made without knowledge of the HLA testing results.

DRESS syndrome was diagnosed using a modified DRESS scoring system using elements of the RegiSCAR system (Table E2).^14,15 The modification was required because the DILIN database did not have specific data related to the nature and extent of rash, specifics of peak temperature, and details of other extrahepatic manifestations. Furthermore, patients were enrolled up to 6 months after clinical onset, often after resolution of the clinical features. Scores included one point for the following features: liver involvement, renal dysfunction, fever, rash (1 point if compatible and another if biopsy proven), facial edema, lymphadenopathy, eosinophilia (>500/μL and another point if >1500μL), atypical lymphocytosis or other hematologic abnormalities (thrombocytopenia, leukocytosis, leukopenia, anemia) and thus ranged from 1 to 10. Fever and rash, either with a skin biopsy or a clinical statement that the rash was typical of DRESS syndrome, were required. Scores of 6 or above were considered definite, 5 highly likely, 4 probable, 3 possible and 2 or less unlikely.

HLA typing was done on all enrolled patients who gave informed consent for genetic testing and from whom adequate DNA samples were available. High resolution class I and II HLA sequencing was done on DNA extracted from whole blood using the Illumina MiSeq platform at the Vanderbilt University Medical Center’s Immunogenomics, Microbial Genetics and Single Cell Technologies core.¹² Two control groups were used: (1) DILIN patients who were exposed to vancomycin and who developed acute liver injury that was attributed to other agents (vancomycin likelihood scores of possible or unlikely) and (2) DILIN patients with liver injury attributed to another drug or HDS products and who had not been exposed to vancomycin in the previous two months. Allele frequency rates were also compared to U.S. population controls.¹³

Descriptive statistics were computed to describe the study cohorts, using median and range for continuous variables and frequency and percentage for categorical variables. Comparisons of the allele or carriage frequency of HLA-A*32:01 between groups were assessed by Fisher’s exact test. All analyses were carried out in SAS version 9.4. All authors had full access to the study data and reviewed and approved the final manuscript.

Results

Among 2272 cases of suspected drug-induced liver injury enrolled in DILIN between September 2004 and January 2021, 2150 cases had a full 6 months of follow up and had undergone underwent formal adjudication by the DILIN Causality Committee. Of the 2150, 1697 (79%) were scored overall as definite, highly likely or probable drug-induced liver injury (high confidence cases). Vancomycin was listed as a potential cause in 12 cases, but in only 6 was vancomycin considered the primary cause (highly likely in 2 and probable in 4); in the other 6, another drug was considered a more likely cause and vancomycin was considered only possible (n=2) or unlikely (n=4) the cause. In addition, there were another 78 high confidence cases in DILIN in which vancomycin was not initially implicated but was listed as a concomitant agent that had been taken within two months of onset. Re-review of these cases independently by 3 authors (BA, CK, JHH) raised the vancomycin causality score to highly likely in one and probable in two cases, with 8 being scored as possible and the remaining 67 as still unlikely. Thus, for this analysis, there were 90 patients with drug-induced liver injury who were exposed to vancomycin, of which 9 were considered highly likely (n=3) or probably (n=6) due to vancomycin. In the other 81 cases, vancomycin was considered only a possible cause in 10 and an unlikely cause in 71 (Figure E1).

The 9 high confidence vancomycin cases included 6 men and 3 women, 8 white and 1 black, ages 23 to 85 (median = 60) years (Table 1). The duration of vancomycin therapy ranged from 1 to 37 (median = 26) days, latency to onset of symptoms 6 to 28 (median= 20) days, and latency to first liver test abnormalities 7 to 33 days (median = 24). Therefore, most cases (89%) were still receiving vancomycin when symptoms first arose. Initial symptoms were often suggestive of hypersensitivity and included fever (n=8), rash (n=8), and jaundice (n=4). Seven patients had significant eosinophilia (above 500/μL). A modified DRESS scoring system suggested that 1 had possible, 2 probable, 3 highly likely, and 3 definite DRESS syndrome. Strikingly, all 9 patients were exposed to multiple drugs in the 2 months before onset, ranging in number from 6 to 28 (median = 13). Importantly, 7 patients were taking at least one other medication that has been linked to DRESS syndrome in the literature.^17,18,19

Table 1.

Vancomycin Hepatotoxicity: 9 cases Clinical Features and Outcomes

Feature	Median or No.	Range or Percent
Age (years)	60 years	23–85 years
Male Sex	6	67%
Race: White	8	89%
Black	1	11%
Ethnicity: Hispanic	0	0%
Body Mass Index (kg/m2)	33 kg/m 2	24–45 kg/m 2
Duration of therapy (days)	26 days	1–34 days
Latency to onset (days)	27 days	6–80 days
Symptoms: Any	9	100%
Rash	8	89%
Fever	8	89%
Renal dysfunction	6	75%
Hospitalization	8	89%
Liver biopsy	2	22%
Skin biopsy	1	11%
Corticosteroid therapy	6	67%
DRESS Syndrome	8	89%
Other medications being taken within 2 months	16.4	9–28
Taking another drug implicated in DRESS	7	78%
Chronic injury [n=7]	0	0%
Severity Score
1 (mild)	2	22%
2 (moderate)	1	11%
3 (moderate/hospitalized)	2	22%
4 (severe)	3	33%
5 (fatal)	1	11%
Liver Transplant	0	0
Death	2	22%

Initial and peak laboratory abnormalities are shown in Table 2. The initial median serum bilirubin was 1.0 mg/dL (range 0.4–14.1), and 6 patients had a peak bilirubin of 2.5 mg/dL or higher, the level used to define jaundice. Serum alanine aminotransferase (ALT) and alkaline phosphatase (Alk P) levels also varied greatly such that the initial R value (R = ALT/ULN ÷ Alk P/ULN) ranged from 0.2 to 28.8 (median = 2.1); two cases were in the range of hepatocellular injury (R ≥ 5), four cholestatic (R ≤ 2), and three mixed (R >2 and <5). The peak international normalized ratio (INR) ranged from 1.1 to 5.4 (median 1.5). The liver injury in the nine vancomycin cases was considered anicteric and mild (bilirubin <2.5 mg/dL and INR normal) in two, moderate (bilirubin ≥2.5, normal INR, or bilirubin <2.5 and abnormal INR) in three, severe (jaundice, raised INR) in three, and fatal in one patient who died of multi-organ failure approximately 2 months after onset. Another patient died of recurrent renal cell carcinoma approximately 3 months after onset and after resolution of the liver injury. Six patients, all with DRESS syndrome, were treated with corticosteroids.

Table 2.

Vancomycin Hepatotoxicity: 9 cases Laboratory Test Results

Feature	Initial		Peak
	Mean	Range	Mean	Range
ALT (U/L)	292	51–2018	820	117–2018
AST (U/L)	255	73–1232	644	73–1494
Alk P (U/L)	302	61–597	395	265–968
Total Bilirubin (mg/dL)	1.0	0.4–14.1	3.3	0.9–19.4
INR	1.1	1.1–1.6	1.6	1.1–5.4
R ratio	1.6	1.1–5.4
Creatinine (mg/dL)	1.3	0.7–8.9	2.5	0.6–9.8
Hemoglobin (g/dL)	10.5	8.4–12.6
White cell count (/μL)	8.6	1.8–13.3
Eosinophil count (/μL)	208	70.2–1734	952	0.6–9.8
Platelet count (/μL)	181	116–450
	Number	Percent
Enzyme Pattern
Hepatocellular (R ≥ 5)	2	22%
Mixed (R > 2 <5)	3	33%
Cholestatic (R ≤ 2)	4	44%
ANA Positive	1	11%
SMA Positive	1	11%
HLA A*32:01	7	78%

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; Alk, alkaline phosphatase; R = (ALT/ULN ÷ Alk P/ULN); ANA, anti-nuclear antibody; SMA, smooth muscle antibody; HLA, human leukocyte antigen.

Seven patients were seen in follow-up 6 or more months after onset, and all had recovered, had normal liver tests, and were without residual features of DRESS syndrome. Overall, the clinical and biochemical pattern of injury varied considerably. Hepatocellular cases tended to have no or only mild jaundice and responded rapidly to corticosteroid therapy (Figure 1A); whereas cholestatic cases were more likely to be jaundiced and responded more slowly to immunosuppression (Figure 1B).

Figure 1:

A Clinical course of a 60-year-old man with hepatocellular liver injury after vancomycin therapy of aortic valve endocarditis (Case #1). He developed rash, facial edema, and eosinophilia (16% eosinophils) 20 days after starting iv vancomycin. Liver tests results showed marked elevations in serum alanine aminotransferase (ALT: peak value 2018 U/L) but modest increases in alkaline phosphatase levels (Alk P: peak 300 U/L) and total bilirubin (peak 3.0 mg/dL). He was started on prednisone 1 week after liver injury onset and all clinical symptoms and laboratory abnormalities resolved within the next few weeks.

B Clinical course of a 57-year-old male with cholestatic liver injury after treatment with vancomycin, ceftriaxone and metronidazole for osteomyelitis (Case #5). After 27 days of iv vancomycin therapy, he developed rash, fever, and jaundice. Liver test results 7 days later showed marked elevations in alkaline phosphatase levels (Alk P: peak 968 U/L) and total bilirubin (peak 19.4 mg/dL), with moderate elevations in alanine aminotransferase levels (ALT: peak 561 U/L). Liver biopsy showed portal, lobular, and granulomatous hepatitis with evidence of bile duct injury and portal fibrosis. All antimicrobials were stopped, and he was treated with prednisone. At follow up 3 and 8 months later, all clinical symptoms and liver test abnormalities had resolved.

HLA testing demonstrated a close association of the vancomycin cases with HLA-A*32:01 which has been previously linked to vancomycin-induced DRESS syndrome irrespective of liver disease (Table 3)¹². All 9 had HLA testing, 7 of whom carried at least one HLA-A*32:01 allele (carriage frequency [CF] = 78%), and one of whom was homozygous (thus, allele frequency [AF] = 0.44). This allele was present in only 1 of 81 cases with a history of vancomycin exposure but in which liver injury was considered due to another agent and only possibly or unlikely due to vancomycin (CF= 1.2% AF= 0.006: [p<0.001 compared to vancomycin high confidence cases]). This allele was found among DILIN cases due to other agents and not exposed to vancomycin at rates similar to that in the general population (CF=6.6%, AF=0.034 [p<0.001 compared to vancomycin cases]). Thus, the AF of HLA-A*32:01 among vancomycin cases (0.44) was far higher than rates from large cohorts of U.S. Caucasians (AF=0.036), Blacks (AF=0.015), and Asians (AF=0.006) (all comparisons p<0.001).¹³

Table 3:

HLA-A^*32:01 Allele and Carrier Frequency

Population	No.	AF	CF	p value
Vancomycin (High Confidence)	9	0.444	78%	ref
Vancomycin (Possible & Unlikely)	81	0.006	1.2%	<0.001
All other Drug-induced liver injury	1708	0.034	6.6%	<0.001
Caucasian (European) *	1,242,090	0.035	~6.8%	<0.001
African Americans *	416,581	0.015	~2.5%	<0.001
Asian Americans (Chinese) *	99,672	0.006	~1%	<0.001
Hispanic (Mexican-Americans) *	261,235	0.027	~5%	<0.001

^*data from www.alleleFrequencies.net, assessed 12.06.2021

Abbreviations: AF, allele frequency; CF, carrier frequency

A listing of the 9 high confidence vancomycin cases, their HLA-A results, and selected clinical features are given in Table 4. The two cases without A*32:01 had latency periods of 6 and 10 days, considerably shorter than the median time to onset of vancomycin associated DRESS of 21 days.¹² Both cases were judged as only probable vancomycin-induced and DRESS scores were 1 (not DRESS) and 4 (probable DRESS). Both patients not carrying HLA-A*32:01 were also exposed to multiple other agents capable of causing liver injury including piperacillin/tazobactam, ciprofloxacin, metronidazole, olmesartan, carvedilol and sertraline.^15,17,18,19

Table 4.

Vancomycin Hepatotoxicity: 9 Cases

Case	HLA-A (1)	HLA-A (2)	Rash	Fever	Eos	Renal	DRESS	DRESS score	Duration Therapy (days)	Time to Onset (days)	R value	Pattern	Peak Bilirubin	Severity Score	Causality Score
1	A*32:01	A*32:01	1	1	1	1	Yes	8	37	20	28.2	HC	3.0	4	3
2	A*25:01	A*32:01	1	1	1	1	Yes	5	22	22	3.7	Mix	0.9	1	2
3	A*02:01	A*03:01	1	1	1	1	Yes	4	15	10	9.9	HC	1.9	3	3
4	A*24:02	A*24:02	0	0	0	0	No	1	1	6*	0.6	Chol	14.3	2	3
5	A*26:01	A*32:01	1	1	Unk	1	Yes	3	26	27	3.4	Mix	19.4	4	3
6	A*01:01	A*32:01	1	1	1	1	Yes	5	31	20	0.8	Chol	0.7	1	2
7†	A*03:01	A*32:01	1	1	1	U	Yes	5	18	16	0.2	Chol	3.3	3	3
8†**	A*32:01	A*33:03	1	1	1	1	Yes	6	30	16	2.1	Mix	38.6	5	3
9†	A*01:01	A*32:01	1	1	1	1	Yes	6	30	19	1.7	Chol	13.0	4	2

^*History of jaundice after previous exposure to vancomycin.

^†Initially concomitant medications thought to be an unlikely cause which on re-review was considered at least probable

^**African American, fatal outcome after re-exposure after onset of DRESS and acute exacerbation of liver and renal injury

Abbreviations: HLA, human leukocyte antigen; DRESS, drug-reaction eosinophilia and systemic symptoms; R = (ALT/ULN ÷ Alk P/ULN); ULN, upper limit of normal; HC, hepatocellular; Mix, mixed; Chol, cholestatic; Unk, unknown

Discussion

Among 1697 cases of confirmed drug-induced liver injury seen at U.S. medical centers participating in DILIN between 2004 and 2020, 9 (0.5%) were attributed to intravenous vancomycin. Most cases presented with DRESS syndrome, and in several patients the liver injury arose after onset of allergic symptoms of fever and rash. Vancomycin-induced liver injury was often associated with carriage of HLA-A*32:01 (78%), which is found in less than 7% of US Caucasian and Black populations. One patient who was homozygous for HLA-A*32:01 had the highest DRESS score but did not appear to have more severe liver injury than the 6 who were heterozygous for the allele.

The results from this case series support previous findings suggesting a strong association between HLA-A*32:01 and vancomycin DRESS. Furthermore, these findings indicate that testing for HLA-A*32:01 may help in assigning causality of vancomycin-induced DRESS syndrome. This is particularly important because most patients on prolonged intravenous vancomycin treatment are exposed to multiple medications, in particular other antibiotics, which have also been implicated in causing DRESS syndrome. In this study, all 3 cases judged to be highly likely due to vancomycin carried HLA-A*32:01 and 4 of the 6 cases (67%) judged to be probably due to vancomycin had one HLA-A*32:01 allele. This distribution matches the definitions of these causality scores, probable cases being 50% to 75% likely. Indeed, the two subjects without HLA-A*32:01 both had atypical clinical features: one of the cases (#4) was the only patient without DRESS syndrome, had the shortest duration of therapy (1 day) and latency to onset (6 days). Both individuals were exposed to other agents implicated in causing liver injury and DRESS, which raises the possibility that the liver injury was incorrectly attributed to vancomycin.

Previous studies have estimated at least 20% individuals with HLA-A*32:01 who are treated with intravenous vancomycin for more than two weeks develop DRESS syndrome.¹² These features indicate that testing for HLA-A*32:01 might be useful in assessing the risk of DRESS syndrome when prolonged therapy is being considered.²⁰ Vancomycin therapy need not be delayed waiting fro the HLA results. Symptoms and signs of DRESS generally arises after 2–3 weeks of therapy, so if blood for HLA testing is obtained when vancomycin is started, test results for HLAT typing as well as bacterial sensitivity should be available well within 2 weeks. DRESS syndrome is a potentially severe adverse reaction that can result in death or permanent disability. If HLA testing is not available, vancomycin should be discontinued immediately if symptoms of hypersensitivity or DRESS arise without waiting for evolution of symptoms and signs.

The mechanism by which vancomycin causes liver injury and DRESS is not well established. Vancomycin is excreted unchanged in the urine without significant metabolites making it an unlikely cause of direct hepatic injury. Prior studies using molecular modeling and auto-dock Vina software have demonstrated strong affinity of vancomycin for the HLA-A*32:01 binding pockets in the absence of peptide¹². In addition, patients with vancomycin related DRESS and RegiSCAR score of 4 or greater (probable DRESS) were significantly more likely to show dose-dependent responses to vancomycin on interferon gamma ELISpot assays suggestive of a T-cell response. On immunohistochemical studies from an HLA-A*32:01 positive vancomycin DRESS case, a mixture of CD8+, CD4+ T cells were present in the acute skin rash as well as in the positive vancomycin skin test, suggesting that recruitment of vancomycin activated CD4+ and CD8+ T-cells mediates the skin and liver injury. FoxP3+ CD4+ regulatory T cells were not seen acutely and only in the vancomycin skin test positive skin following DRESS recovery.

The strengths of the current study were that it was an unbiased sampling of a large number of cases of drug-induced liver injury and employed a prospective, standardized approach to data collection, diagnosis and definitions of severity and outcome. In addition, the study used state-of-the-art HLA sequencing methods and had control groups within the study as well as comparison groups from population-based data from large U.S. cohorts. Furthermore, causality was assessed based upon clinical features alone without knowledge of HLA testing results. Weaknesses of the study included the lack of systematic and prospective collection of data elements of hypersensitivity, skin rash and features of the DRESS syndrome and the need to use a modification of the validated RegiSCAR scoring system for this diagnosis. This study also had no data on the frequency of use of vancomycin and changes of its use over the time of the patient enrollment. Despite these issues, the current findings indicate that vancomycin-induced DRESS syndrome is an uncommon but important and potentially severe cause of drug-induced liver injury.

In conclusion, vancomycin-induced liver injury typically presents or is preceded by features of DRESS syndrome and is closely linked to carriage of HLA-A*32:01. Testing for HLA-A*32:01 can help in diagnosis of vancomycin-induced DRESS syndrome and in assessing risk of this outcome in patients in whom prolonged treatment with vancomycin is planned.

Highlights.

What is already known about this Topic?

Vancomycin is commonly used to treat multidrug-resistant bacterial infections and rarely has been associated with delayed drug hypersensitivity syndromes such as drug reaction with eosinophilia and systemic symptoms (DRESS).

What does this article add to our knowledge?

Vancomycin-induced liver injury was associated with HLA-A*32:01, and manifests as part of DRESS symptoms of fever, rash and eosinophilia which often preceded liver injury.

How does this study impact current management guidelines?

Patients developing fever and rash on vancomycin therapy should stop the drug promptly. HLA-A*32:01 testing can help in the diagnosis in assessing risk of DRESS after initiation of vancomycin therapy.

Abbreviations:

NIDDK

National Institute of Diabetes and Digestive and Kidney Diseases

NIH

National Institutes of Health

iv

intravenous

DRESS

drug reaction with eosinophilia and systemic signs

HLA

human leukocyte antigen

DILIN

Drug Induced Liver Injury Network

HDS

herbal and dietary supplements

ALT

alanine aminotransferase

AST

aspartate aminotransferase

Alk P

alkaline phosphatase

INR

international normalized ratio