• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services


  • Print
  • Share
  • E-mail

Section Contents Menu

Drug Safety and Availability

Postmarket Reviews - Volume 2, Number 1, 2009

Back to Table of Contents


The Smoking Cessation Aids Varenicline (Marketed as Chantix) and Bupropion (Marketed as Zyban and Generics): Suicidal Ideation and Behavior

Abstract: An analysis of Adverse Event Reporting System (AERS) reports identified cases of suicidality associated with the use of varenicline and bupropion, two products used as smoking cessation aids. This report provides FDA’s summary analysis of these cases.

Keywords: varenicline, bupropion, nicotine, suicide

A postmarket safety review of varenicline (marketed as Chantix) and bupropion (marketed as Zyban and generics), two products used for smoking cessation, has identified cases where suicidal ideation and behaviors followed the initiation of treatment with these drugs. Although this review also evaluated transdermal nicotine products (marketed as Habitrol, Nicoderm, Nicoderm CQ, and Prostep), no clear association was identified between the use of these products and suicidal ideation and/or behavior (these data have not been included in this report).

For both varenicline and bupropion, cases of suicidality were reported among patients with and without a prior psychiatric history.  Healthcare professionals should continue to closely monitor patients treated with these drugs for neuropsychiatric symptoms (e.g., changes in behavior, agitation, depressed mood, and suicidal thoughts and behavior), as directed in the Warnings section of varenicline label and the Boxed Warning, Warnings and Precautions section of the bupropion label.1,2

In 2006, varenicline (0.5 and 1 mg tablets) was approved for use as a pharmacological aid for smoking cessation. Varenicline is a high affinity, partial agonist at the α4β2 nicotinic acetylcholine receptor subtype. Its competitive binding with nicotine and partial agonist activity at the α4β2 receptor are believed to be responsible for this drug’s therapeutic effect.3,4

Bupropion was first approved in 1985 for the treatment of Major Depressive Disorder. Buproprion 150 mg extended release tablet (Zyban) was approved for the smoking cessation indication in 1997. In the following years, generic forms of this drug became available for smoking cessation treatment. Bupropion is a monocyclic aminoketone antidepressant chemically unrelated to the more traditional tricyclic, tetracyclic or SSRI antidepressants.  It is a relatively weak inhibitor of the neuronal norepinephrine and dopamine, and does not inhibit monoamine oxidase or alter the re-uptake of serotonin. Although the mechanism by which bupropion produces its therapeutic effect is unknown, it is generally assumed that its action on noradrenergic and/or dopaminergic activity contributes to the drug’s efficacy.5

FDA reviewed AERS cases describing suicidal ideation and behavior†† reported for these compounds from the date of each compound’s approval (from approval of the smoking cessation indication for bupropion) to November 27, 2007. FDA identified 153 reports of suicidal adverse events for varenicline (suicidal ideation-116, suicide-37) and 75 reports for bupropion (suicidal ideation-46, suicide-29). These cases likely represent a fraction of those that occurred during this time period due to underreporting to FDA’s Adverse Event Reporting System (AERS).  The total yearly prescriptions and units sold for these products is estimated in the millions.  The data summarized below represent these AERS cases.

Reported Cases of Suicidal Ideation and Behavior

The demographic data of these AERS cases, if provided, are listed in Table 1. When evaluating the temporal relationship between the initiation of treatment and the time-to-onset for these adverse suicidal events, a large percentage of AERS reports indicated that suicidal ideation and behavior occurred while the individual was undergoing treatment for smoking cessation. For varenicline, out of the 128 cases that reported time-to-onset, 86% reported that symptoms began while taking the drug. Of the 72 bupropion cases reporting time-to-onset, 93% noted an onset of suicidal ideation and behavior while the patient was taking the drug. For both drugs, the median time to event onset was less than two weeks. A similar proportion of cases for varenicline (35%) and bupropion (33%) reported symptom abatement after discontinuation of the drug.

Table 1.  Selected demographics of AERS reports of suicidal ideation (SI) and suicidal behavior (SB) with varenicline and bupropion.

Suicidal Ideation
Suicidal Behavior
Suicidal Ideation
Suicidal Behavior
Age in YearsCases Reportingn = 106n = 27n = 33n = 24
Range20 - 8322 - 7226 - 7015 - 70
SexCases Reportingn = 115n = 34n = 43n = 29

Complementing these data, 41% of varenicline cases (63/153 cases reporting) reported that the change in thinking or behavior was a significant change from the past. In 37 of these 63 cases (59%), this was a first time experience, while the remaining 26 cases noted there was a worsening of pre-existing psychiatric disease (which the reporter had previously felt was stable).  A similar pattern was seen with bupropion. Twenty-three percent of cases (17/75 cases reporting) stated a significant change from the past in thinking and/or behavior occurred after initiating treatment. In 11 of these 17 cases (65%), this was a first time experience. The remaining six cases noted there was a worsening of pre-existing, yet stable, psychiatric condition.

For those reports indicating drug dose, most individuals, independent of the type of treatment or whether they experienced suicidal ideation or behavior, were taking the drug at the recommended dose. Specifically, 95% of varenicline reports indicating suicidal ideation (78/82) and 93% of varenicline reports indicating suicidal behavior (15/16) noted that the patient was taking the recommended dose. For bupropion, 96% of reports indicating suicidal ideation (25/26) and 100% of bupropion reports indicating suicidal behavior (14/14) noted that the patient was taking the recommended dose.

The psychiatric variables associated with the varenicline and bupropion reports of suicidal ideation and behavior are listed in Table 2. For varenicline cases, approximately half indicated that the suicidal events occurred in patients with psychiatric histories. Forty-two percent of cases also noted that the patient was currently taking psychotropic medications in addition to the smoking cessation aid at the time of the event. A smaller percentage of cases reported these concurrent factors in bupropion-treated patients. It is important to note that many of the patients who reported suicidality did not have a psychiatric history and/or were not taking psychotropic medications. In addition, there were many cases that did not provide information on psychiatric history or concomitant psychotropic medications.

Table 2.  Psychiatric variables noted in AERS reports of suicidal ideation (SI) and suicidal behavior (SB) with varenicline and bupropion. The variables of Psychiatric History and Concomitant Psychiatric Medications are not mutually exclusive.

Suicidal Ideation
Suicidal Behavior
TotalSuicidal Ideation
Suicidal Behavior
Total Casesn = 116n = 37n = 153n = 46n = 29n = 75
Psych History54%38%50%26%21%24%
No Psych History25%30%26%35%27%32%
Unknown Psych History  21%32%24%39%52%44%
Concomitant Psych Meds42%30%39%20%17%19%
No Concomitant Psych Meds20%24%21%41%21%33%
Unknown Concomitant Psych Meds38%46%40%39%62%48%

For both compounds, a psychiatric history of depression was frequently reported. Specifically, of the 77 varenicline reports indicating the patient had a psychiatric history, 58% (45/77) reported a history of depression. Of the 18 bupropion reports indicating the patient had a psychiatric history, 66% (12/18) reported a history of depression. Depression was also frequently reported as a co-occurring psychiatric event at the time of reporting. Specifically, 50%, and 70% of varenicline and bupropion reports, respectively, had depression listed as a co-occurring psychiatric event.

Table 3 lists the outcomes associated with the varenicline and bupropion suicidal ideation and behavior reports. The majority of outcomes were serious and included, although not mutually exclusively from, death and hospitalization. Four cases describing an adverse suicidal event while patients were taking varenicline or bupropion are presented in Box 1.

Table 3.  Outcomes noted in AERS reports of suicidal ideation (SI) and suicidal behavior (SB) with varenicline and bupropion. The outcomes listed below are not mutually exclusive.


Suicidal Ideation
Suicidal Behavior
TotalSuicidal Ideation
Suicidal Behavior
Serious (cases reporting)

n = 110

n = 37

n = 147

n = 30

n = 29

n = 59





























Required Intervention














Non-Serious (cases reporting)

n = 6

n = 0

n = 6

n = 16

n = 0

n = 16


Box 1


Case 1

A 45 year-old female nurse reported she was taking varenicline, as directed, to aid in the cessation of smoking. She had been a smoker (1-1.5 packs/day) for approximately 4 years. After 10 weeks of varenicline treatment, she reported she was experiencing suicidal ideation, aggression, erratic behavior, uncontrollable emotions, and decreased clarity. One week after discontinuing varenicline, she reported that [her symptoms] were "not at the levels listed above," although she still experienced some "residuals." She described her alcohol use as rare and reported treatment for hypothyroidism with Synthroid.

Case 2

A 30 year-old female reported that she was taking varenicline 0.5 mg/day to help her quit smoking. She indicated that she had taken the pill for 5 days as directed. In her report, she stated, "I became very depressed for no apparent reason and started having suicidal thoughts. I worried that I was going to hurt myself or someone else. I was also in a state of panic and unable to eat. I thought that I might go crazy and felt completely out of sorts. I quit taking the pill. About 36 hours later, I felt like myself again."


Case 3

A 27 year-old female with a history of smoking 1 pack/day for the past 11 years was taking bupropion 150 mg BID as a smoking cessation aid. Fifteen days into bupropion treatment, the patient reported feeling “emotional” and was having regular crying fits. The patient threatened to kill herself and stated, "I didn’t care if I lived." The patient had no documented history of depression.

Case 4

A 46 year old male smoker was taking bupropion 150 mg BID for four days when he experienced depression, panic, paranoia, sensory changes and thoughts of suicide.  The patient discontinued his medication at that time. No other information on the resolution of this event was provided.  The patient reported that he had a family history of chronic depression.

Together, the AERS data suggest a possible association between suicidal events and the use of varenicline and bupropion. Healthcare providers are reminded to closely monitor patients for neuropsychiatric symptoms (e.g., changes in behavior, agitation, depressed mood, and suicidal thoughts and behavior) while they are using varenicline and bupropion as smoking cessation aids. Healthcare providers should report any cases of suicidal ideation and/or behavior in patients taking these drugs to FDA's Medwatch Program.  Given the well-established health risks of smoking, healthcare providers should continue to work closely with patients to assist them in quitting smoking.

Relevant Web Sites

National Institute on Drug Abuse website on nicotine
Centers for Disease Control and Prevention website on smoking and tobacco use
Public Health Advisory on varenicline


† Suicidal ideation included the search terms 'suicidal ideation',
'self-injurious ideation', and 'suicidal depression'.

†† Suicidal behavior included the search terms 'completed suicide', 'attempted suicide', 'intentional self-injury', 'self-injurious behavior', 'suicidal behavior', 'multiple drug overdose', 'gun shot wound', 'intentional misuse', and 'overdose'.


  1. Varenicline (Chantix) Product Labeling
  2. Bupropion (Zyban) Product Labeling
  3. Stack NM. Smoking cessation: an overview of treatment options with a focus on varenicline. Pharmacotherapy. 2007;27(11):1550-7.
  4. Potts LA, Garwood CL. Varenicline: the newest agent for smoking cessation. Am J Health Syst Pharm. 2007;64(13):1381-4.
  5. Jorenby D. Clinical efficacy of bupropion in the management of smoking cessation. Drugs. 2002;62(Suppl 2):25-35.

Quitting Smoking Reduces the Risk of Cardiovascular Disease, Pulmonary Disease, and Cancer

Back to top


Abacavir (Marketed as Ziagen) and Abacavir-Combination Products (Marketed as Trizivir and Epzicom): Hypersensitivity Reaction, HLA-B*5701, and Skin Patch Testing

Abstract:  As of July 2008, the Boxed Warning of the abacavir product labeling recommends that patients be screened for the human leukocyte allele, HLA-B*5701, prior to initiating abacavir treatment. Screening patients for HLA-B*5701 reduces the risk for an abacavir (ABC) hypersensitivity reaction (HSR). If an ABC HSR is suspected, healthcare professionals are directed to discontinue ABC immediately and not rechallenge patients with this drug.  Although the abacavir skin patch test (research use) may appear to be a simple solution to aid in the diagnosis of ABC HSR, a clinical syndrome with a degree of diagnostic uncertainty, the predictive value of skin patch testing is not fully understood. The use of skin patch testing is difficult to justify as the basis for rechallenging a patient with ABC, given the risk for a fatal systemic reaction.

Keywords:  abacavir, hypersensitivity reaction, HLA-B*5701, skin patch testing

Abacavir (ABC) sulfate (Ziagen; also in the combination products Trizivir and Epzicom), a nucleoside reverse transcriptase inhibitor indicated for treatment of HIV-1 infection, has been associated with a unique, serious, and sometimes fatal hypersensitivity reaction (HSR). ABC HSR is a multi-organ syndrome which typically occurs within six weeks of initiating ABC treatment.1 The most common symptoms of an ABC HSR include fever (~80% of cases), rash (~70% of cases), malaise/fatigue, nausea, and vomiting. Other signs and symptoms of an ABC HSR may include myalgia or arthralgia, headache, diarrhea, pruritis, hypotension, and various respiratory symptoms. Development of an ABC HSR requires immediate and permanent discontinuation of the drug, as rechallenge can be fatal.2

Correctly diagnosing an ABC HSR is critical given the severe consequences of this reaction to the patient and the utility of this drug in managing HIV. Unfortunately, the diagnosis of an ABC HSR can be difficult. An ABC HSR may mimic common infections, or may be confused with adverse events seen with other concomitant or prophylactic medications used for the treatment of HIV, or with opportunistic infections commonly seen in patients with HIV.  A review of literature indicates that misdiagnosis of an ABC HSR can occur in approximately two to seven percent of patients, independent of whether patients are even taking ABC.1,3

One risk factor that has been shown to be associated with the development of an ABC HSR is the presence of the  human leukocyte allele, HLA-B*5701.4,5 In one prospective, randomized clinical trial (PREDICT-1), not treating patients positive for HLA-B*5701 with abacavir significantly reduced the incidence of clinically suspected cases of an ABC HSR from 7.8% to 3.4%.6   Data from this study suggest that 61% of HLA-B*5701 positive subjects may develop an abacavir HSR during the course of therapy compared with 4.5% of HLA-B*5701 negative subjects. As of July 2008, the Boxed Warning on the abacavir product labeling recommends that patients be screened for HLA-B*5701 prior to initiating abacavir treatment.6

Even with genetic screening for HLA-B*5701, and refraining from initiating abacavir treatment in those patients positive for this allele, some patients may develop an ABC HSR. As noted above, HLA-B*5701 is not 100% predictive of an ABC HSR. Some patients negative for this allele may still develop signs and symptoms suggestive of an ABC HSR. For this reason, healthcare professionals should closely monitor patients taking abacavir for potential signs and symptoms of this reaction (see Table 1).

If an ABC HSR is suspected, healthcare professionals are directed to discontinue ABC immediately and not rechallenge patients with the drug or by any other means.

Table 1. This table lists the systems and signs and symptoms involved in the "> 2 signs and symptoms from 2 systems" rule used in the clinical diagnosis of ABC HSR.

SystemSigns and Symptoms
Group 1Fever
Group 2Rash
Group 3Nausea, Vomiting, Diarrhea, Abdominal Pain
Group 4Malaise, Tiredness, Myalgias, or Arthralgias
Group 5Shortness of Breath, Cough, Sore Throat

Several research reports have emerged that have used skin patch testing to immunologically confirm suspected cases of ABC HSR.7,8  Specifically, in an analysis of race differences in ABC HSRs among black and white subjects positive for HLA-B*5701, white HLA-B*5701 positive subjects with a HSR were more likely to be skin patch test positive than black HLA-B*5701 positive subjects with a HSR.8  Among the 10 HLA-B*5701 positive black subjects, 5 were skin patch test positive and 5 were skin patch test negative. These data suggest that skin patch testing was not predictive for an ABC HSR in patients positive for the HLA-B*5701 allele.  Conversely, only 32.3% (42/130) of white and 7.2% (5/69) of black patients who were skin patch tested and who met criteria for clinically suspected ABC HSR had a positive skin patch test for an ABC HSR. Healthcare professionals must be aware that the use of skin patch testing to immunologically confirm cases of an ABC HSR remains in the realm of research and that skin patch testing has not been clinically validated. Skin patch testing may miss cases of true ABC HSR or provide false positive results. The accuracy of skin patch testing results is unknown in a broad population.

Given the risks associated with rechallenging patients with ABC when a case of HSR is suspected, that is, an increased risk of mortality, the diagnosis of an ABC HSR must remain a clinical one in all patients (see Table 1). Although skin patch testing appears to be a simple solution to aid in the diagnosis of a clinical syndrome with diagnostic uncertainty, the positive and negative predictive value of the abacavir skin patch test are not fully understood, making its use to justify a potentially fatal systemic abacavir rechallenge a risky endeavor.

Relevant Web Sites

Information sheet on Abacavir and Abacavir-containing medications

Abacavir (Ziagen) product labeling

Abacavir sulfate; lamivudine (Epzicom) product labeling

Abacavir sulfate; lamivudine ; zidovudine (Trizivir) product labeling

Medication Guide for Abacavir products

Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents


  1. Shear NH, Milpied B, Bruynzeel DP, Phillips EJ. A review of drug patch testing and implications for HIV clinicians. AIDS. 2008;22(9):999-1007.
  2. Escaut L, Liotier JY, Albengres E, et al. Abacavir rechallenge has to be avoided in case of hypersensitivity reaction. AIDS. 1999;13(11):1419-20.
  3. Hernandez J, Cutrell A, Bonny T, et al. Diagnosis of abacavir hypersensitivity reactions among patients not receiving abacavir in two blinded studies. Antivir Ther. 2003;8:L88.
  4. Mallal S, Nolan D, Witt C, et al. Association between presence of HLA-B*5701, HLA-DR7, and HLA-DQ3 and hypersensitivity to HIV-1 reverse-transcriptase inhibitor abacavir. Lancet. 2002;359(9308):727-32.
  5. Hetherington S, Hughes AR, Mosteller M, et al. Genetic variations in HLA-B region and hypersensitivity reactions to abacavir. Lancet. 2002;359(9312):1121-2.
  6. Abacavir product labeling.
  7. Mallal S, Phillips E, Carosi G, et al. HLA-B*5701 screening for hypersensitivity to abacavir. N Engl J Med. 2008; 358:568–79.
  8. Saag M, Balu R, Phillips E, et al. High sensitivity of Human Leukocyte Antigen-B*5701 as a marker for immunologically confirmed abacavir hypersensitivity in white and black patients. Clin Infect Dis. 2008;46:1111-18.

Back to top


Atomoxetine (Marketed as Strattera): Serious Liver Injury

Abstract:  In 2004, the atomoxetine label was updated to include information about cases of serious liver injury. From January 2005 to March 2008, six postmarket cases of serious liver injury with atomoxetine use were reported to FDA. None of the patients that developed serious liver injury had liver transplants. One patient with serious liver injury died, although it is unclear if atomoxetine caused or contributed to this event. Patients and caregivers should be alert to the signs and symptoms of liver injury throughout atomoxetine treatment. Atomoxetine should be discontinued and not resumed if a patient presents with jaundice or laboratory evidence of hepatotoxicity.

Keywords:  atomoxetine, hepatotoxicity, ADHD

FDA continues to receive reports of serious liver injury in patients given atomoxetine. Atomoxetine received FDA approval on November 26, 2002 as the first non-stimulant medication used for the treatment of attention deficit hyperactivity disorder (ADHD) in children (ages 6 years and above) and adults.1  Atomoxetine's therapeutic action is believed to be due to its selective inhibition of norepinephrine reuptake. From the year 2002 to 2007, approximately 3.3 million patients received a prescription for atomoxetine in the United States.  Of those, approximately 2.1 million patients (64%) were children ages 17 years and younger.2

While a signal for serious liver injury was not detected during premarket clinical trials of atomoxetine, two published postmarket reports did identify instances of atomoxetine-induced hepatitis.3,4 In one of these reports, there was a positive rechallenge with atomoxetine. Subsequent to these reports, a bolded warning was added in 2004 to the atomoxetine label indicating an increased risk for severe liver injury.

Since the 2004 labeling change, FDA received six additional reports of serious liver injury in patients taking atomoxetine. Following an evaluation of the information from the drug sponsor and additional literature articles submitted to FDA, the atomoxetine product label was again revised in 2007. The Warnings and Precautions section of the drug label advises prescribers about the risk for severe liver injury with this drug.1 Healthcare professionals and patients should be watchful for serious liver injury associated with the use of atomoxetine and report cases to FDA's MedWatch.

Box 1

What is drug-induced liver injury?

The liver is the main organ for metabolizing, activating and/or deactivating drugs prior to excretion via the bile or urine. Sometimes the drug (or its metabolites) can cause chemical injury to liver cells. This injury can vary in severity from asymptomatic elevations of blood enzyme activities to liver failure with a need for transplantation. This injury can also result in death. Drug-induced liver injury (DILI) is now the most frequent cause for acute liver failure in the United States, exceeding all other causes (e.g., viral, alcoholic, autoimmune, ischemic) combined. It is also one of the main reasons for not approving new drugs, or for removing drugs from the market after they were approved for clinical use.

The mechanisms causing DILI are not well understood. There are several ways the injury can present in patients including jaundice, nausea and vomiting, abdominal pain, bleeding, mental confusion, and/or kidney failure. Although exceedingly rare for most drugs, almost any drug can cause DILI. Still, some drugs are more likely than others to result in this injury. The main determinant for DILI may be individual differences in constitution or manner of responding to the drug. That is, some patients may be more sensitive than others to DILI.  Injuries due to individual differences are generally called "idiosyncratic reactions." Severe idiosyncratic reactions, while generally rare, may occur at much lower doses than those that are well-tolerated and handled by most people.

The FDA, along with partners in academia and industry, is engaged in research into questions about DILI and sponsors an annual public conference on this topic.  The proceedings from the conference can be found at FDA's Drug-Induced Liver Toxicity website.

The following paragraphs summarize FDA's analysis of these six Adverse Event Reporting System (AERS) cases of atomoxetine-associated liver injury. These reports were received between January 2005 and March 19, 2008. Two of the six cases are described in the medical literature.3,4,5

Table 1 summarizes some of the characteristics of these cases. Five of the six cases were among patients 17 years of age and younger.

Table 1. Characteristics of the six cases of atomoxetine-associated severe liver injury

Age (years)Median10.5
SexMalen = 4
Femalen = 2
Dose (n = 4):Median32.5 mg
Range10 – 80 mg
Time to onset (days)Median62.5
Range21 – 730
ReportsU.Sn = 5
Non-U.S.n = 1
Serious OutcomeHospitalizationn = 5
Deathn = 1

In four cases, the patients presented with jaundice and one or more symptoms, including fatigue, lack of appetite, abdominal pain, nausea, or vomiting. Acute impairment of liver function, as manifested by the inability to synthesize sufficient prothrombin (PT) to produce a normal INR or the inability to clear bilirubin, was observed at the time of diagnosis in three of the six patients. Two patients had significant increases in hepatic enzymes.  In one case, profiled in Box 2, peak AST and ALT levels were 6619 IU/L [73 x upper limit of normal (ULN)] and 5182 IU/L (86 x ULN), respectively; peak total bilirubin (TB) was 14.7 mg/dl. Viral titers for hepatitis A, B, and C, Epstein-Barr virus (EBV), and cytomegalovirus (CMV) IgM were negative in two cases with elevated hepatic enzymes. Liver biopsies showed hepatic inflammation/fibrosis in one patient and inflammation with moderate piecemeal necrosis in the other.

Although the majority of patients were hospitalized, none required a liver transplant. Four patients recovered after atomoxetine was discontinued. One adult, male patient died due to hepatic and renal failure. Because there were limited details available about this case, the role of atomoxetine in this patient’s death could not be determined. Over the course of two years, this patient had alternated atomoxetine 80 mg/day with methylphenidate in order to manage his ADHD.

Three patients in this case series reported a history of occasional acetaminophen use, the use of a recent, single dose of mebendazole, or concomitant quetiapine use.  None of these drugs were felt to have contributed to the liver injury.

Two representative cases illustrating the relationship between atomoxetine and serious liver injury are summarized in Box 2. The cases were selected based on a temporal relationship between initiation of atomoxetine and onset of symptoms indicative of a liver injury. Both patients presented with elevated hepatic enzymes and jaundice, and had extensive clinical work-ups, including liver biopsies and tests to rule out non-drug causes of liver injury. Both patients recovered upon atomoxetine discontinuation.  The two cases are also described in the medical literature.3,4,5

Box 2

Case 1

A 12-year-old female diagnosed with ADHD experienced jaundice and elevation of liver enzymes when she restarted atomoxetine 40 mg daily following a 6-week interruption of a year-long therapy. This patient had no history of liver disease, or exposure to hepatitis or any known hepatotoxins. Three weeks after restarting atomoxetine, she was hospitalized for jaundice, abdominal pain, diarrhea, and vomiting. Upon admission, she had stable vital signs, conjunctival icterus, and tenderness in her right upper quadrant. She had no hepatosplenomegaly. She was evaluated for acute hepatitis.

Over the course of the event, her peak liver biomarkers were AST: 3505 U/L (78 x ULN), ALT: 3264 U/L (65 x ULN), TB:  9.8 mg/dL (~10 x ULN), GGT: 108 U/L (normal reference range: 12-43), and PT: 14.1 seconds (normal reference range: 10.8–13.7). Her serum alkaline phosphatase (AP) was normal. A liver biopsy confirmed hepatic inflammation and fibrosis. Testing for hepatitis A, B, C, CMV, and EBV were negative. An antinuclear antibody (ANA) titer was positive (1:160; normal reference range: < 1:40). Smooth muscle antibodies were negative, ceruloplasmin and alpha-1 antitrypsin were normal. The diagnosis of drug-induced liver injury by atomoxetine was made. Atomoxetine was discontinued. Four weeks later, she had an AST of 800 U/L, ALT of 1438 U/L, and TB of 2.6 mg/dL. Serum transaminase and bilirubin levels returned to normal approximately 6 months after discontinuation of atomoxetine.

Case 2

An 8-year-old female with ADHD was being treated with atomoxetine 25 mg/day and developed vomiting, abdominal pain, jaundice, and very high serum aminotransferases approximately five weeks after starting therapy. She was hospitalized and atomoxetine was discontinued. She had no known history of liver disease. The child occasionally received therapeutic doses of acetaminophen from a parent (~10 times/month) for soccer-related musculoskeletal pains. She also received one dose of mebendazole for a pinworm infection nine days after starting atomoxetine therapy.

Upon admission to the hospital, her physical exam was significant for scleral icterus, jaundice, and hepatomegaly. The peak values of liver markers were AST: 6619 U/L (73 x ULN), ALT: 5182 U/L (86 x ULN), TB: 14.7 mg/dL (~14 x ULN), AP: 528 U/L (3.5 x ULN), and GGPT: 196 U/L (2.5 x ULN). Testing for hepatitis A, B, C, CMV, and EBV were negative. An abdominal ultrasound showed mild hepatomegaly with no gallstones or ductal dilatation. Liver biopsy showed mixed portal inflammation and moderate piecemeal necrosis. Infectious, autoimmune and metabolic causes of liver injury were ruled out.  No serum acetaminophen was detected. She received vitamin K, diphenhydramine, ursodiol, and fresh frozen plasma. The patient's transaminases and bilirubin serum levels began to improve approximately 13 days into her hospitalization. Within two months of her hospital discharge, her AST and ALT serum levels had decreased to 276 U/L (normal reference range: 25-45) and 178 U/L (normal reference range: 0-50), respectively.

Postmarket reports indicate that atomoxetine is associated with serious idiosyncratic liver injury.  Some of the cases reported additional confounding factors, such as occasional acetaminophen use. Some of the cases lacked sufficient clinical detail to convincingly detail the relationship between the use of atomoxetine and liver injury. The mechanism of atomoxetine-induced liver injury remains unknown.  FDA continues to monitor AERS for reports of serious liver injury in association with atomoxetine.

FDA encourages physicians to:

  • Inform patients to immediately contact their physician at the first sign or symptom of fatigue, loss of appetite, nausea, vomiting, pruritis, dark urine, jaundice of the sclerae or skin, right upper quadrant tenderness, or unexplained "flu–like" symptoms
  • Determine liver enzyme levels when a patient presents with signs or symptoms of liver injury
  • Discontinue and not resume atomoxetine treatment if patients present with jaundice or laboratory evidence of liver injury
  • Report cases of serious liver injury to FDA's Medwatch Program


  1. Atomoxetine (Strattera) product labeling
  2. Verispan Vector One®: National, Total Patient Tracker (TPT), 2002-2007, data extracted Sept 2007.
  3. Bangs ME, Ling J, Zhang S, et al. Hepatic events associated with atomoxetine treatment for attention deficit hyperactivity disorder. Drug Safety. 2008;31(4):345-54.
  4. Stojanovski SD, Casavant MJ, Hayat MM, et al. Atomoxetine-induced hepatitis in a child. Clin Toxicol (Phila). 2007;45(1):51-5.
  5. Lim JR, Faught PR, Chalasani NP. Severe liver injury after initiating therapy with atomoxetine in two children. J Pediatr. 2006;148(6):831-4.

Back to top

Back to Table of Contents