Hepatotoxicity in Children Receiving Isoniazid Therapy for Latent Tuberculosis Infection

Transcription

1 Original Article Hepatotoxicity in Children Receiving Isoniazid Therapy for Latent Tuberculosis Infection Shiow-Huey Chang, 1 Payam Nahid 2 and Sarah R. Eitzman 3 1 Public Health Department, Santa Clara County, San Jose, California; 2 Division of Pulmonary and Critical Care, University of California San Francisco, San Francisco General Hospital, California; and 3 Department of Pediatrics, Santa Clara Valley Medical Center, San Jose, California Corresponding Author: Shiow-Huey Chang, RN, PhD, 660 South Fair Oaks Avenue, Sunnyvale, CA Shiow-huey. Chang@ucsf.edu. Received August 4, 2013; accepted November 20, 2013; electronically published January 15, Background. The frequency of isoniazid hepatotoxicity is low in children receiving isoniazid therapy for latent tuberculosis infection. However, isoniazid hepatotoxicity may cause liver failure and death. We evaluated children who developed isoniazid hepatotoxicity to determine demographic and clinical characteristics. Methods. A retrospective review was performed of medical records of 1582 patients aged <18 years who were evaluated for isoniazid therapy at a public health department and clinic in California. Results. There were 13 patients who had latent tuberculosis infection and who developed isoniazid hepatotoxicity (0.8% of all 1582 patients who started isoniazid; 1.1% of 1235 patients who completed the 9-month isoniazid therapy). There were 8 girls (62%) and 9 Hispanic children (69%) who had hepatotoxicity. Sex, age, and race were not independently associated with the development of isoniazid hepatotoxicity. Symptoms and signs of hepatotoxicity were present in 11 of the 13 patients, and 2 other patients had alanine aminotransferase >5 times the upper limit of normal and no signs of hepatotoxicity. The most common symptoms included abdominal pain, anorexia, vomiting, and nausea. Most patients developed hepatotoxicity within 6 months of starting isoniazid, but 3 patients developed hepatotoxicity 6 months after starting isoniazid. After stopping isoniazid, the alanine aminotransferase levels decreased to normal in all patients. Conclusions. In children who have latent tuberculosis infection, isoniazid hepatotoxicity has low frequency and typically is reversible when isoniazid is stopped. Evidence of late drug-induced liver injury indicates the importance of monitoring symptoms and serum transaminases throughout isoniazid therapy. Key words. liver; muscular dystrophy; rifampin; transaminase. Latent tuberculosis infection (LTBI) in children and adolescents often is treated with a 9-month course of isoniazid [1]. Isoniazid therapy may be complicated by hepatotoxicity, and children who develop drug-induced liver injury may require liver transplant [2]. Isoniazid hepatotoxicity is defined in the guidelines of the American Thoracic Society as an increase in serum alanine aminotransferase (ALT) > 3 times the upper limit of normal (ULN) with symptoms of hepatitis and/or jaundice, or ALT > 5 times ULN without symptoms [3]. The frequency of hepatotoxicity is low in children receiving isoniazid therapy. In a national sample of 2473 patients aged < 20 years, hepatitis associated with isoniazid was noted in 1 per 1000 patients [4]. In a Seattle study of 11,141 patients, hepatitis was defined by serum aspartate aminotransferase (AST) level >5 times ULN and symptoms of hepatitis; there were no cases of hepatotoxicity in 1468 patients aged <14 years [5]. A study from San Diego County that used the current American Thoracic Society definition of hepatotoxicity showed no cases of liver injury associated with isoniazid in 1277 patients aged <14 years, and there were 6 patients who had hepatotoxicity in 1939 patients aged years (0.3%) [6]. Although the frequency of hepatotoxicity from isoniazid is low in children, there are 2 reports of liver transplant and mortality caused by isoniazid. In 20 children who had hepatic failure caused by isoniazid, 10 children had liver transplant, 6 children died before transplant, and only 4 children recovered without transplant [2]. In another report of 2 children who had severe adverse events from Journal of the Pediatric Infectious Diseases Society, Vol. 3, No. 3, pp , DOI: /jpids/pit089 The Author Published by Oxford University Press on behalf of the Pediatric Infectious Diseases Society. All rights reserved. For Permissions, please journals.permissions@oup.com.

2 222 Chang et al isoniazid from 2004 to 2008, 1 child underwent liver transplant [7]. The clinical characteristics of children who develop isoniazid hepatotoxicity are unknown. The purpose of this retrospective, descriptive study was to evaluate the prevalence of hepatotoxicity caused by isoniazid, based on serum transaminase levels and symptoms, and to evaluate the relation between demographic and clinical characteristics and the development of hepatotoxicity in children and adolescents receiving isoniazid therapy. MATERIALS AND METHODS Subjects Patients aged <18 years (n = 1582 patients) who were referred to a public health department or a pediatric LTBI clinic in a California county and who initiated isoniazid therapy for LTBI between January 2005 and August 2011 were reviewed retrospectively. Patients who restarted isoniazid therapy or became pregnant during treatment were excluded from the study. Consistent with current guidelines of the American Thoracic Society, isoniazid hepatotoxicity was definedby(1)alt>3timesuln(female,alt>93u/l; male, ALT > 120 U/L) and symptoms suggestive of hepatitis or jaundice (nausea, vomiting, abdominal pain, jaundice, or unexplained fatigue), or (2) ALT > 5 times ULN (female, ALT > 155 U/L; male, ALT > 200 U/L) without symptoms [3]. The study was approved by the Committee on Human Research at University of California, San Francisco (IRB# ) and the institutional review board at Santa Clara Valley Medical Center (IRB#11 059). Evaluation Patients aged <2 years received isoniazid therapy (9 months) from their primary care physician and had monthly home visits by public health nurses who evaluated medication adherence ( pill counts) and adverse reactions. Patients aged 2 18 years received isoniazid therapy (9 months) from the LTBI clinic and were monitored monthly at the clinic for medication adherence ( pill counts) and adverse reactions before receiving each 1-month refill of medication. In the LTBI clinic, a nurse practitioner assessed patients for direct contact to active tuberculosis, previous LTBI treatment, history of hepatitis or liver disease, prior hospitalizations or surgery, and history of current medications, including antiseizure, antipsychotic, antifungal medications, hormonal contraceptives, and nonprescription drugs or vitamins. Body weight was measured to determine daily dose by weight categories (daily dose range, mg/kg; maximum daily dose, 300 mg) [8]. Patients who had a history of hepatitis, alcohol abuse, or receiving potentially hepatotoxic drugs had ALT and AST tests before isoniazid treatment (baseline) for LTBI. Children who had abnormal pretreatment transaminase levels were rechecked with ALT and AST tests before starting therapy. Before starting isoniazid for LTBI, patients who had risk factors such as silicosis, gastrectomy, jejunoileal bypass, chronic renal failure, conditions requiring prolonged high-dose corticosteroid or other immunosuppressive therapy, or malignancy were referred to their primary care physician for evaluation. After isoniazid therapy was started, ALT and AST tests were obtained in patients who had nausea, abdominal pain, anorexia, or fatigue. In addition, ALT, AST, total bilirubin, alkaline phosphatase, and an acute hepatitis panel were performed in patients who had icterus, dark urine, or hepatomegaly. Chest radiography was done when activation of tuberculosis was suspected. Patients who developed increased liver enzyme tests > normal and < 2 times ULN were instructed to continue isoniazid and repeat liver enzyme tests after 1 2 weeks. When liver enzyme tests were >2 times ULN, patients were advised to discontinue isoniazid and repeat liver enzyme tests within 24 hours and weekly until liver enzymes decreased. Patients who developed persistent signs of hepatotoxicity or neuropathy that did not resolve with cessation of antituberculosis drug therapy were referred to their primary care physician. Patients who had increased liver enzymes >5 times ULN or jaundice were evaluated with ALT, AST, total bilirubin, alkaline phosphatase, an acute hepatitis panel, prothrombin time, partial thromboplastin time, and occasionally Epstein-Barr virus titers and were referred to a pediatric gastroenterologist for consultation. When the elevation of liver enzymes subsided, rifampin therapy (6 months) was offered to patients in lieu of isoniazid. The practice procedure at the LTBI clinic did not change from 2004 to Serum transaminase testing at the LTBI clinic was performed with a modular analyzer (Roche/Hitachi Modular P800, Roche Diagnostics, Indianapolis, IN). The normal range of laboratory values differed by sex for ALT (female, 0 31 U/L; male, 0 40 U/L) and AST (female, 0 31 U/L; male, 0 37 U/L), and the normal range for bilirubin was mg/dl for both females and males. Laboratory results were unknown for 20 patients (1.3%) who were treated by private primary care physicians or at a county tuberculosis clinic. The public health nurses and nurse practitioner used a checklist of symptoms of adverse reactions to evaluate symptoms monthly from the start to end of isoniazid therapy. Symptoms compatible with hepatotoxicity were documented throughout the course of therapy. The researcher retrieved symptoms or diagnoses associated with isoniazid hepatotoxicity from the medical

3 Isoniazid Hepatotoxicity in Children 223 Table 1. Relation Between Demographic Characteristics and Hepatotoxicity in Children and Adolescents Who Had Isoniazid Therapy for Latent Tuberculosis Infection a Characteristic No. Patients (%) records from the beginning of therapy to 1 month after therapy completion or discontinuation. Statistical Analysis Tabulations and correlations among variables were conducted with statistical software (SPSS version 20.0, IBM Corp., Armonk, NY). We performed multivariate logistic regression to examine potential predictors (sex, age, and race) in relation to the development of isoniazid hepatotoxicity. The problem of predicting a small number of hepatotoxicity cases was addressed by regrouping potential predictors into dichotomous groups for sex (male or female), age (age < 15 years or years), and race (Hispanic or non-hispanic). Reliability of data entry was assessed by inputting a random sample of 10% records twice; patterns of inconsistency in data entry were refined, and variable coding for data entry was revised to reduce errors. The final frequency of errors for the random sample in the final 100 patients was 2.7%. Statistical significance was defined by P.05. RESULTS Most patients who started isoniazid therapy for LTBI were aged < 15 years and were Hispanic (Table 1). There were 125 patients (8%) who had ALT and AST tests before treatment; abnormal results were noted in 32 patients (19 females and 13 males) (26% patients who had pretreatment tests and 2% of all patients who started treatment). These abnormal pretreatment results were within 2 times ULN for both ALT (mean [range]: female, 28 ± 10 [16 52] No. (%) Patients Who Started Therapy and Who Developed Hepatotoxicity Adjusted Odds Ratio (95% Confidence Interval) P Total (0.8 b ) Sex Female 789 (50) 8 (1.0) 1.6 ( ) 0.4 Male 793 (50) 5 (0.6) 1.0 (Reference) Age (y) < (31) 6 (1.2) (26) 3 (0.7) 1.0 (Reference) (26) 2 (0.5) (17) 2 (0.8) 0.9 ( ) 0.9 Race c Hispanic 1234 (79) 9 (0.7) 1.0 (Reference) Non-Hispanic 1.6 ( ) 0.4 Asian or Pacific Islander 196 (12) 2 (1.0) White 30 (2) 1 (3.3) Black 36 (2) 0 (0) Other 73 (5) 1 (1.4) a N = 1582 children who started isoniazid therapy and were included in the analysis; mean age, 9 ± 5 years (range, y). There were 1235 patients (78%) who completed isoniazid therapy. b The overall frequency of isoniazid hepatotoxicity in 1235 patients who completed isoniazid therapy was 1.1%. c Race based on self-reporting. Data for race available for 1569 patients (99%). U/L; male, 51 ± 20 [16 85] U/L) and AST (mean [range]: female, 35 ± 5 [21 44] U/L; male, 39 ± 13 [21 65] U/L). There were 570 patients (36%) who had 1 ALT and AST test after beginning isoniazid therapy. Thirteen of 1582 patients (0.8%) who started therapy for LTBI developed isoniazid hepatotoxicity. The frequency of hepatotoxicity associated with isoniazid was not significantly different between different sex, age, or racial groups (Table 1). There were 208 patients (13%) who did not have hepatotoxicity and who transferred care or were lost to follow-up before completing treatment, and it was unknown whether these patients developed isoniazid hepatotoxicity. Among 1235 patients (78%) who completed the 9-month course of isoniazid therapy, the frequency of hepatotoxicity was 1.1% (13 of 1235 patients) (Table 1). There were 2 patients who developed hepatotoxicity and who were excluded from all analysis of the 13 patients: 1 male child who had Becker muscular dystrophy and 1 male child who had an uncertain number of doses of isoniazid ingested (Table 2). Most patients who had hepatotoxicity were female, Hispanic, and in the healthy weight category (Table 3). There were 11 patients who had symptoms and signs of hepatotoxicity, most commonly abdominal pain, anorexia, vomiting, and nausea (Table 3). All patients had an appropriate dosage of isoniazid, and most patients developed hepatotoxicity within 6 months of starting isoniazid (Table 4). Half of the patients completed therapy with rifampin (Table 4). No patients had known viral hepatitis or human immunodeficiency virus (Table 5). Only 1 patient

4 224 Chang et al Table 2. Abnormal Transaminase Levels, Sex, and Age in Children and Adolescents Who Had Isoniazid Hepatotoxicity During Treatment for Latent Tuberculosis Infection a ALT or AST Level Descriptor Females (n = 789) Males (n = 793) Total (n = 1582) > 3 times ULN Number (%) 2 (0.3) 1 (0.1) 3 (0.2) Age (y) 9 ± 7 (4 14) 7 ± 6 (4 14) > 5 times ULN Number (%) 6 (0.8) 4 (0.5) 10 (0.6) Age (y) 10 ± 6 (2 17) 7 ± 4 (4 12) 8 ± 5 (2 17) Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ULN, upper limit of normal. a N = 13 patients who had hepatotoxicity. Data reported as number (%) or mean ± SD (range, minimum to maximum). There were 2 patients excluded in addition to these 13 patients: 1 child with Becker muscular dystrophy and 1 child with uncertain number of ingested doses of isoniazid. Elevated ALT or AST level was defined as any elevated level during treatment. Table 3. Demographic and Clinical Characteristics of Children and Adolescents Who Had Isoniazid Hepatotoxicity During Treatment for Latent Tuberculosis Infection a No. (%) Patients or Characteristic Mean ± SD (range) Age (y) b 8±6(2 17) Sex Female 8 (62) Male 5 (38) Race Hispanic 9 (69) Asian or Pacific Islander 2 (15) White 1 (8) Arabic 1 (8) Body mass index c 20 ± 5 (16 30) Weight category c Healthy 8 (62) Overweight 1 (8) Obese 3 (23) Patients taking concurrent medications d 6 (46) Symptoms and signs of hepatotoxicity Abdominal pain 6 (46) Anorexia 5 (38) Vomiting 5 (38) Nausea 4 (31) Jaundice 3 (23) Fever 2 (15) Diarrhea 2 (15) Headache 1 (8) Hepatomegaly 1 (8) Splenomegaly 1 (8) No symptoms or signs 1 (8) a N = 13 patients. Data reported as number (%) patients or mean ± SD (range, minimum to maximum). There were 2 patients excluded in addition to these 13 patients: 1 child with Becker muscular dystrophy and 1 child with uncertain number of ingested doses of isoniazid. b Age at start of isoniazid therapy. c Body mass index (BMI) and weight category not available in 1 patient. BMI values were calculated by the Centers for Disease Control and Prevention BMI percentile calculator for child and teen ( d Concurrent medications: vitamins, 2 patients; diphenhydramine, fluoride, ibuprofen, and iron, 1 patient each. had a pretreatment ALT and AST test, and this test was normal (Table 5). Most patients who had hepatotoxicity had ALT elevation 5 times ULN (Table 5). The mean time from start of isoniazid therapy to onset of hepatotoxicity was 109 days (Table 4). In 3 patients, hepatotoxicity was detected >6 months after the start of isoniazid (Table 4). A 3.9-year-old boy who took isoniazid for 6 months and 4 days (total, 143 pills) developed fever Table 4. Therapy and Hepatotoxicity in Children and Adolescents Who Had Isoniazid Hepatotoxicity During Treatment for Latent Tuberculosis Infection a No. (%) Patients or Characteristic Mean ± SD (range) Isoniazid dosage 300 mg/d 8 (62) 200 mg/d 4 (31) 150 mg/d 1 (8) Average (mg/kg/d) 9 ± 4 (4 14) Time to diagnosis of hepatotoxicity (mo) b <3 6 (46) 3 to <6 4 (31) 6 3 (23) Mean (d) 109 ± 80 (28 257) Therapy completed or not completed Completed Rifampin, 6 months 6 (46) Not completed Isoniazid <9 months 2 (15) Rifampin not started, referred elsewhere 2 (15) Rifampin started, moved or transferred 2 (15) care Rifampin declined 1 (8) a N = 13 patients. Data reported as number (%) patients or mean ± SD (range, minimum to maximum). There were 2 patients excluded in addition to these 13 patients: 1 child with Becker muscular dystrophy and 1 child with uncertain number of ingested doses of isoniazid. b Time from start of isoniazid therapy to diagnosis of hepatotoxicity (elevated ALT >3 times upper limit of normal). and anorexia, and his first transaminases were elevated (ALT, 2690 U/L; AST, 3589 U/L); isoniazid was discontinued the next day, and he declined rifampin therapy. An 8.5-year-old Guatemalan girl took isoniazid for 8.5 months (total, 240 pills) and developed abdominal pain, headache, fever, and elevated transaminases (ALT, 375 U/L; AST, 348 U/L). A 14.1-year-old girl took isoniazid for 8 months (total, 230 pills) and developed abdominal pain, nausea, vomiting, and elevated ALT (ALT, 101 U/L; AST, 26 U/L). Improvement of ALT level occurred in all 13 patients at a mean 69 days after discontinuation of isoniazid, and there were no deaths (Table 5). There was no significant relation between time to improvement of ALT and age or cumulative isoniazid dose.

5 Isoniazid Hepatotoxicity in Children 225 Table 5. Laboratory Studies in Children and Adolescents Who Had Isoniazid Hepatotoxicity During Treatment for Latent Tuberculosis Infection a No. (%) Patients or Laboratory Study Mean ± SD (range) Viral hepatitis b Negative 11 (85) Not tested 1 (8) HBsAg negative only 1 (8) HIV Negative 3 (23) Not tested 10 (77) ALT and AST before treatment Normal 1 (8) Not tested 12 (92) ALT (U/L) Onset of hepatotoxicity 401 ± 448 ( ) c Peak 559 ± 523 ( ) c Resolution of hepatotoxicity 29 ± 10 (13 44) Amount ALT > ULN at onset of hepatotoxicity < 3-fold 0 3- to <5-fold 5 (38) 5- to <10-fold 3 (23) 10-fold 5 (38) AST (U/L) Onset of hepatotoxicity 361 ± 580 ( ) c Peak 457 ± 663 ( ) c Resolution of hepatotoxicity 33 ± 10 (16 47) Amount AST > ULN at onset of hepatotoxicity < 3-fold 5 (38) 3- to <5-fold 1 (8) 5- to <10-fold 3 (23) 10-fold 4 (31) Time for ALT and/or AST to decrease (d) d 69 ± 63 (21 232) Bilirubin (mg/dl) First or peak 2 ± 4 ( ) Resolution of hepatotoxicity 0.5 ± 0.7 ( ) Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; HBsAg, hepatitis B surface antigen; HIV, human immunodeficiency virus; ULN, upper limit of normal. a N = 13 patients. Data reported as number (%) patients or mean ± SD (range, minimum to maximum). There were 2 patients excluded in addition to these 13 patients: 1 child with Becker muscular dystrophy and 1 child with uncertain number of ingested doses of isoniazid. b Hepatitis A, B, and C. Nonalcoholic steatohepatitis was diagnosed in 1 patient who had a negative test for viral hepatitis; after isoniazid was discontinued, his ALT decreased to 33 U/L and AST decreased to 38 U/L. c Highest value (outlier) not included in mean and SD but shown in range. d Time from onset of hepatotoxicity and discontinuing isoniazid to the resolution to normal ALT and/or AST. DISCUSSION In the present study, the frequency of isoniazid hepatotoxicity was 0.8% in patients who began a 9-month course of isoniazid therapy and 1.1% in patients who completed the therapy. The frequency of hepatotoxicity in patients aged <15 years who started isoniazid was 0.8% (11 of 1318 patients) and the frequency in patients aged years was 0.8% (2 of 264 patients) (Table 1). The present study showed a higher frequency of hepatotoxicity compared with 2 previous studies [5, 6]. In a study conducted with a similar Hispanic population in San Diego, the prevalence of isoniazid hepatotoxicity was 0% in 1277 patients aged <15 years and 0.3% in 1939 patients aged years [6]. Another study that used more stringent criteria for isoniazid hepatotoxicity (AST >5 times ULN and the presence of hepatitis symptoms) showed that the frequency of hepatotoxicity was 0% for patients aged <15 years (n = 1468 patients) and 0.08% (6 of 7449 patients) for patients aged years [5]. Using the criteria of the latter study, the frequency of hepatotoxicity in the present study was 0.5% (6 of 1318 patients) for patients aged <15 years and 0.4% (1 of 264 patients) for patients aged years. The higher frequency in the present study may be attributed to the higher frequency of completion of therapy in the present study (78%) (Table 1) than the previous studies (64% and 74%) [5, 6], which may have facilitated the detection of more patients with isoniazid hepatotoxicity in the present study. In 2 other previous studies, isoniazid hepatotoxicity was defined by a single threshold for serum transaminases (ALT and AST >100 U/L) and not 3-fold >ULN, and this makes it difficult to compare these studies with the present study [9, 10]. In a study of 239 children aged 9 14 years, only 2 patients (0.8%) had elevated liver enzymes (normal range: ALT <30 U/L; AST <40 U/L) above this threshold [9]. The other previous study used similar criteria and reported a frequency of 0.18% isoniazid hepatotoxicity in 564 children [10]. Although the frequency of isoniazid hepatotoxicity in the former study was similar to that observed in the present study, the former study included ALT and AST levels assessed before and at 4, 8, and 12 months after starting isoniazid therapy, which may have enabled the detection of more patients with abnormal ALT and AST levels [9]. The present study did not include routine biochemistry testing, and the detection of affected patients was based on clinical symptoms and signs [9]. Furthermore, compared with the former study, the present study used a definition of isoniazid hepatotoxicity that accounted for the presence of hepatitis symptoms in conjunction with serum transaminase levels >3 times ULN, and the present patients had a wider range of age (0 18 y) [9]. Therefore, comparison between studies must be done with caution [9]. The present study showed that all elevated ALT and AST levels in children on isoniazid therapy were reversible (mean, 69 days) (Table 5). One patient had Becker muscular dystrophy, and in this patient both ALT and AST failed to return to normal despite discontinuing isoniazid. A high serum transaminase level is common in boys who have Becker muscular dystrophy [11, 12]. Several of the present patients developed isoniazid hepatotoxicity 6 months after beginning treatment (Table 4). This emphasizes the

6 226 Chang et al importance of monitoring symptoms and serum transaminase levels throughout treatment to detect drug-induced liver injury. Serum transaminase thresholds may be helpful to monitor isoniazid therapy in children. A previous study showed that children who have isoniazid hepatotoxicity may have high serum transaminase levels (mean ALT, 1030 U/L [range, U/L)]; AST, 1399 U/L [range, U/L]), and 45% of children may have at least 1 risk factor (viral hepatitis; multidrug therapy, and/or acetaminophen) [2]. In the present study, 5 patients (38%) had ALT 10 times ULN (Table 5), but most of the present patients did not have risk factors. If serum transaminase threshold reaches a certain level in children who develop drug-induced hepatitis, it is unknown whether or not liver injury may become reversible. All 13 patients had an appropriate dose of isoniazid within the recommended range of mg/kg/d (Table 4) [8]. Although 1 patient was diagnosed with nonalcoholic steatohepatitis and had a negative test for viral hepatitis, his ALT level stabilized after isoniazid was discontinued (Table 5). Most patients had a healthy body weight (Table 3). Only 1 patient who had hepatotoxicity had pretreatment ALT and AST tests done (Table 5). The frequency of abnormal pretreatment ALT or AST tests in our patients is 2%; therefore, it is unlikely that pretreatment ALT or AST tests would have been abnormal in any of the 13 patients who developed hepatotoxicity. However, nonalcoholic fatty liver disease may occur with drug-induced liver injury, and further study is necessary to evaluate the relation between obesity or nonalcoholic fatty liver disease and drug-induced liver injury [13]. Sex, age, and race were not associated with the development of isoniazid hepatotoxicity (Table 1). Previous studies suggested that age may be a risk factor for of isoniazid hepatotoxicity, but this was not significant in the present study (Table 1) [5, 6, 14]. Females may be at greater risk of developing isoniazid hepatotoxicity, but this was not significant in the present study (Table 1) [5, 15]. Further study with a larger patient sample may determine whether girls before or after puberty are at greater risk of developing isoniazid hepatotoxicity. Limitations of the present study include the inability to determine whether symptoms of hepatitis were directly caused by isoniazid, because other concurrent illnesses may have been present that may have caused symptoms. It is possible that hepatotoxicity unrelated to isoniazid accounted for some cases. However, there were no guidelines that recommended baseline or serial serum transaminases in children who did not have hepatic issues or who were not receiving other potentially hepatotoxic drugs before or during treatment for LTBI; therefore, data about baseline or serial transaminases were not available for all children. Furthermore, 208 patients (13%) who started isoniazid treatment transferred care or were lost to follow-up, and elevation of serum transaminases in these patients is unknown. In addition, the present patients mostly were Hispanic children, and this may limit generalization of the results to other racial groups or comparison with other studies. In summary, the frequency of isoniazid hepatotoxicity in the present study was similar to previous findings in children (0.8%) [9]. In addition, delayed onset of drug-induced liver injury was noted in the present study, and this suggests that it may be important to monitor symptoms and serum transaminases during the entire treatment. Further study may establish serum transaminase thresholds to guide the monitoring of isoniazid therapy in children. Acknowledgments The author thanks Ms. Cynthia Haines (nurse manager) and her staff at the Tuberculosis Prevention and Control Program of Santa Clara Public Health Department, California for assistance in this project; Kathryn Lee, PhD for reviewing the manuscript; and Bruce Cooper, PhD and Wayne Shoumaker, PhD for statistical consultation. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. REFERENCES 1. Targeted tuberculin testing and treatment of latent tuberculosis infection. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, July This is a Joint Statement of the American Thoracic Society (ATS) and the Centers for Disease Control and Prevention (CDC). This statement was endorsed by the Council of the Infectious Diseases Society of America (IDSA), September 1999, and the sections of this statement. Am J Respir Crit Care Med 2000; 161:S Wu SS, Chao CS, Vargas JH, et al. Isoniazid-related hepatic failure in children: A survey of liver transplantation centers. Transplantation 2007; 84: Saukkonen JJ, Cohn DL, Jasmer RM, et al. An official ATS statement: Hepatotoxicity of antituberculosis therapy. Am J Respir Crit Care Med 2006; 174: Kopanoff DE, Snider DE Jr, Caras GJ. Isoniazid-related hepatitis: A U.S. Public Health Service cooperative surveillance study. Am Rev Respir Dis 1978; 117: Nolan CM, Goldberg SV, Buskin SE. Hepatotoxicity associated with isoniazid preventive therapy: A 7-year survey from a public health tuberculosis clinic. JAMA 1999; 281: LoBue PA, Moser KS. Use of isoniazid for latent tuberculosis infection in a public health clinic. Am J Respir Crit Care Med 2003; 168: Centers for Disease Control and Prevention (CDC). Severe isoniazid-associated liver injuries among persons being treated for latent tuberculosis infection United States, MMWR Morb Mortal Wkly Rep 2010; 59:224 9.

7 Isoniazid Hepatotoxicity in Children Pediatric Tuberculosis Collaborative Group. Targeted tuberculin skin testing and treatment of latent tuberculosis infection in children and adolescents. Pediatrics 2004; 114: Spyridis P, Sinaniotis C, Papadea I, Oreopoulos L, Hadjiyiannis SPapadatos C. Isoniazid liver injury during chemoprophylaxis in children. Arch Dis Child 1979; 54: Nakajo MM, Rao M, Steiner P. Incidence of hepatotoxicity in children receiving isoniazid chemoprophylaxis. Pediatr Infect Dis J 1989; 8: McMillan HJ, Gregas M, Darras BT, Kang PB. Serum transaminase levels in boys with Duchenne and Becker muscular dystrophy. Pediatrics 2011; 127:e Wright MA, Yang ML, Parsons JA, Westfall JM, Yee AS. Consider muscle disease in children with elevated transaminase. J Am Board Fam Med 2012; 25: Yu AS, Keeffe EB. Nonalcoholic fatty liver disease. Rev Gastroenterol Disord 2002; 2: Fountain FF, Tolley E, Chrisman CR, Self TH. Isoniazid hepatotoxicity associated with treatment of latent tuberculosis infection: A 7-year evaluation from a public health tuberculosis clinic. Chest 2005; 128: Franks AL, Binkin NJ, Snider DE Jr, Rokaw WM, Becker S. Isoniazid hepatitis among pregnant and postpartum Hispanic patients. Public Health Rep 1989; 104:151 5.