Liver disease in adolescents: A cohort study of high-risk individuals


  • The views expressed in this article do not necessarily reflect those of the industry partners.

  • Potential conflict of interest: Nothing to report.


Little is known about the health and behavior of adolescent offenders as they relate to abnormalities of liver biochemistry and hepatitis C virus (HCV) infection. A large study of male juvenile offenders was undertaken that allowed a re-evaluation of the normal limits of alanine aminotransferase (ALT), associations with elevated ALT, and HCV antibody positivity. Young offenders (age 12–19 years) serving community orders participated in a wide-ranging health survey and laboratory assessment between October 2003 and December 2005. Normal ranges for liver biochemistry were calculated from the 95th percentile of males at the lowest risk for liver disease. The final sample comprised 682 males, of whom 439 (64%) gave blood. The calculated upper limit of normal for ALT was 28 IU/L. Seventeen percent of adolescents had an elevated ALT. Strong associations with elevated ALT included HCV antibody positivity [odds ratio (OR) 14.6], overweight and obesity (OR 6.9), and elevated total cholesterol (OR 3.6). More than 90% of adolescents with elevated ALT levels had 1 or more features of the metabolic syndrome. HCV antibody was positive in approximately 3% of the cohort, with the most significant risk factor being injecting drug use (OR 7.8; P < 0.01). The new infection rate was 3.7% per year. Conclusion: New upper limits for ALT provide greater sensitivity for the early diagnosis of liver disease in adolescents. High rates of HCV infection and obesity-related liver disease exist in this group, and targeted interventions are needed to reduce future health-related morbidity. (HEPATOLOGY 2007.)

There is a dearth of information about the health-related outcomes of adolescents, and even less about adolescent offenders. The rates of and associations with hepatic pathology in this population have also not been clarified. Nonalcoholic fatty liver disease (NAFLD) and chronic hepatitis C (CHC) are 2 of the most common liver diseases in affluent nations and are a major cause of liver-related morbidity and mortality.1, 2 Reports on the epidemiology of overweight and obesity3 and case series of type 2 diabetes and cirrhosis in pediatric populations4–6 suggest that the pathophysiological correlate of NAFLD, insulin resistance, has its genesis in childhood and adolescence. With regard to CHC, up to 70% of infections in Western countries are acquired through injecting drug use.7, 8 In addition, although there is a low overall prevalence of CHC in countries such as the United States and Australia (≈1%), injecting drug use, the principal high-risk behavior, often commences during adolescence.9 Therefore, defining the prevalence of CHC in at-risk adolescent populations is crucial, because it will contribute significantly to the adult pool of chronic infection, serve as a reservoir for transmission, and provide a valuable opportunity to prevent transmission.

To our knowledge, only 3 publications in the English literature have previously examined the prevalence of CHC exposure and viremia in adolescent offenders.10–12 The associations with liver biochemistry were not reported in any of these cohorts. The prevalence of abnormal liver tests and their associations with determinants of metabolic liver disease has been reported on 1 previous occasion in adolescents13; however, in that study, an abnormal alanine aminotransferase (ALT) level was based on a historical definition. A definition of the normal ranges for liver biochemistry excluding all those with subclinical liver disease has never been examined in adolescents. This study, part of a comprehensive survey of the physical and mental health needs of a large cohort of young offenders serving community orders14 in whom testing for hepatitis C virus (HCV) and hepatitis B virus (HBV) infection was undertaken, therefore provided us with a unique opportunity to determine the clinical correlates of abnormal liver biochemistry among adolescents.

Our analysis focused on liver-related morbidity and subclinical liver disease in a large cohort of adolescents. The aims were (1) to create a new definition for the upper limit of normal for ALT, aspartate aminotransferase (AST), and γ-glutamyltranspeptidase (GGT); (2) to define the associations and implications of elevated ALT in adolescents; and (3) to define the risk factors and associations with HCV antibody positivity. The analysis was confined to males who comprised the majority of subjects in the sampling frame.


ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CHC, chronic hepatitis C; CI, confidence interval; GGT, γ-glutamyltranspeptidase; HBV, hepatitis B virus; HCV, hepatitis C virus; NAFLD, nonalcoholic fatty liver disease; OR, odds ratio; ULN, upper limit of normal.

Patients and Methods

The study formed part of a comprehensive analysis of the physical and mental health needs of young offenders serving community orders with the New South Wales Department of Juvenile Justice, Australia. Young offenders under the supervision of one of the 36 Juvenile Justice Community Service Centers throughout the state of New South Wales were approached to participate in a wide-ranging health survey between October 2003 and December 2005. Six centers were not sampled due to geographical distance or low numbers of eligible participants. Adolescents were excluded if they had serious mental health problems, had undergone substance withdrawal, were considered too violent or disruptive by Department of Juvenile Justice staff, or had court appearances or admission to custody on the day of the survey. Participants were interviewed in locations across New South Wales, stratified into 3 main areas: Sydney, “other metropolitan,” and “regional.” Sydney included the Greater Sydney metropolitan area (population 4.2 million), “other metropolitan” included cities with populations of more than 100,000, and “regional” comprised cities and towns with populations of less than 100,000.

The survey investigated several health-related behaviors, including alcohol and drug use, exercise, diet, sexual practice, and body piercing. Standard demographic data were collected, as were medical history, anthropomorphic data, and blood pressure. All participants were asked to provide blood samples for biochemistry, liver tests, lipid analysis, blood-borne viruses, and sexually transmitted infections. Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). BMI ranges for participants were calculated based on age according to Cole et al.'s data from 97,876 males between the ages of 2 and 18.15 Adolescent hypertension was defined according to age and height as per recently published pediatric guidelines.16 Normal ranges for adolescent lipids were taken from The Cholesterol in Childhood Guidelines.17 This suggested that there was an increased risk of cardiovascular disease with total cholesterol ≥5.18 mmol/L, low-density lipoprotein cholesterol ≥3.4 mmol/L, triglycerides ≥2.25 mmol/L, and high-density lipoprotein cholesterol ≤1.03 mmol/L.

Study approval was independently granted by the University of Sydney Human Research Ethics Committee, the Research Applications Subcommittee of the Department of Juvenile Justice Collaborative Research Unit, the Justice Health Human Research & Ethics Committee (formerly Corrections Health at the time of study commencement), and the Aboriginal Health and Medical Research Council. Written informed consent was required as a condition for participation. Parental consent was required for participants under the age of 14 years.

Liver tests and lipid studies were analyzed using a Beckman Synchron LX 20 and LXi machine respectively (Beckman Coulter Inc., Fullerton, CA) using Beckman Coulter reagents. HCV antibody was measured by an enzyme immunoassay on an Abbott Axsym platform (Abbott Diagnostics, Abbott Park, IL). HCV polymerase chain reaction results were obtained using a Versant HCV RNA qualitative assay kit (Bayer HealthCare, Germany).

Statistical analyses were performed using SPSS version 13.0 (SPSS Inc., Chicago, IL). Results are reported as means ± standard deviation and medians (range) as appropriate. Continuous variables were compared using nonparametric tests (Mann-Whitney) or 2-tailed Student t tests as appropriate. Categorical data were compared using the chi-square test. Factors with a P value of 0.1 or less on univariate analysis were further analyzed using a multiple logistic regression model in order to identify the independent determinants of outcome variables.


Baseline Characteristics.

Approximately 3412 young male offenders were eligible for inclusion in the survey. Our stratified sampling frame comprised 2388 (70%) of the eligible group. Of these, 341 refused to participate, 597 could not be seen at an appropriate time/location, 682 were uncontactable or did not respond to repeated efforts to contact them, and 90 were excluded because of serious mental health problems, substance withdrawal, or excessively disruptive behavior on the day of testing. These exclusions may have resulted in an underestimation of some of the parameters tested in this study. The final sample comprised 682 males, of whom 439 (64.4%) provided blood samples.

Table 1 describes the distributions of the participants by ethnicity, region, and age. The median age was 17 years (range, 12-19 years). The majority of subjects were born in Australia (82.9%), New Zealand (7.2%), Asia (2.7%), or the Middle East (2.3%).

Table 1. Baseline Characteristics of the Cohort Segregated by Ethnicity, Region, and Age
CharacteristicMales (n = 439)
Culturally and linguistically diverse7517.0
Other metropolitan6815.5
Age <16 years9120.7
Age >16 years34879.3

Normal Ranges for Liver Biochemistry in Adolescent Males.

Normal ranges for liver biochemistry in adolescents were derived from the 95th percentile value for the population cohort of males with the lowest risk of liver disease. Subjects had to meet the following criteria: normal BMI, normal low-density lipoprotein cholesterol, and normal triglycerides. Subjects with high systolic blood pressure, excessive alcohol consumption and hepatitis B or C infection were excluded as per Prati et al.18

For the 206 males that met these criteria, the median ALT was 15 IU/L (range, 4–61 IU/L), with values of 9, 12, 19, and 28 IU/L for the 5th, 25th, 75th, and 95th percentiles, respectively (Table 2). The median value for AST was 16 IU/L (range, 4–44 IU/L); the median value for GGT was 16 IU/L (range, 6–53 IU/L). For AST, the values of the 5th, 25th, 75th, and 95th percentiles were 7, 13, 21, and 32 IU/L, respectively; the corresponding values for GGT were 10, 14, 19, and 29 IU/L, respectively. The upper limit of normal, therefore, was 28 IU/L for ALT, 32 IU/L for AST, and 29 IU/L for GGT. Histograms for each of these values are presented in Fig. 1.

Table 2. Normal Ranges for Liver Biochemistry
Figure 1.

Histogram defining normal values for ALT, AST, and GGT in 206 adolescent males.

Liver Biochemistry and Metabolic Profile.

The baseline metabolic profile and liver biochemistry of the total cohort is presented in Table 3. Of note, 33% were either overweight or obese, 10% had elevated low-density lipoprotein cholesterol, and 40% had low levels of high-density lipoprotein cholesterol. Abnormal liver biochemistry was present in a large proportion of adolescents, with 17% having elevated ALT and 14% elevated GGT. Exposure to both HBV and HCV were more prevalent in this cohort (4.4% and 3.2%, respectively) compared with the general population. The majority of adolescents remained susceptible to HBV (70%). There were no cases of human immunodeficiency virus.

Table 3. Baseline Characteristics of 439 Adolescent Males with Blood Results
CharacteristicMales with Blood Results (n = 439)
  • Abbreviations: Anti-HBc, antibody to hepatitis B core antigen; Anti-HBs, antibody to hepatitis B surface antigen; Anti-HCV, antibody to hepatitis C virus; HBsAg, hepatitis B surface antigen; HDL, high-density lipoprotein; HIV, human immunodeficiency virus; LDL, low-density lipoprotein; SD, standard deviation.

  • *

    Analyzed according to The Cholesterol in Childhood Guidelines (American Academy of Pediatrics, 1992) in which LDL ≥3.4 mmol/L, triglycerides ≥2.25 mmol/L, or HDL ≤1.03 mmol/L were considered to convey an excess cardiovascular risk.

  • According to analysis in this article.

Age (years), mean (SD)16.6 (1.3)
BMI, mean (SD)23.69 (4.9)
 Underweight, n (%)15 (3.5)
 Normal weight, n (%)266 (62.1)
 Overweight, n (%)90 (21.0)
 Obese, n (%)57 (13.3)
Blood sugar level, mmol/L (SD)5.8 (1.1)
Diabetic, n (%)5 (1.2)
Abnormal lipid analysis* 
 LDL cholesterol, n (%)44 (10.4)
 HDL cholesterol, n (%)176 (41.4)
 Triglycerides, n (%)34 (8.0)
HBsAg-positive, n (%)3 (0.7)
 Not immune, n (%)301 (70.0)
 Immune: vaccinated, n (%)110 (25.6)
Anti-HBc, n (%)19 (4.4)
Anti-HCV, n (%)14 (3.2)
HIV-positive, n (%)0
Elevated liver biochemistry 
 ALT, n (%)76 (17.5)
 GGT, n (%)62 (14.3)
 AST, n (%)30 (6.9)

A comparison of the cohort with elevated ALT levels compared with those with normal ALT values is presented in Table 4. There was a strong association for overweight or obesity (72% versus 25.1%; P < 0.001) to predict elevated ALT, and on multivariate analysis these factors conveyed a risk [odds ratio (OR)] of 6.9 [confidence interval (CI) 3.7–13.1; P < 0.001] (Table 5). Likewise, low-density lipoprotein cholesterol, triglycerides, and total cholesterol levels were all higher in subjects with elevated ALT (P <0.01), with total cholesterol remaining significant when controlled for BMI (Table 5). There were no significant differences in high-density lipoprotein cholesterol levels for the 2 groups.

Table 4. Characteristics of the Cohort of Adolescent Males with Elevated ALT
 Elevated ALT (≥28 IU/L), n = 76Normal ALT (<28 IU/L), n = 359P Value
  1. Abbreviations: Anti-HBc, antibody to hepatitis B core antigen; Anti-HBs, antibody to hepatitis B surface antigen; Anti-HCV, antibody to hepatitis C virus; HBsAg, hepatitis B surface antigen; HDL, high-density lipoprotein; HIV, human immunodeficiency virus; LDL, low-density lipoprotein; SD, standard deviation.

Age (years), mean (SD)16.8 (1.4)16.5 (1.3)0.173
BMI, mean (SD)28.1 (6.3)22.6 (3.8)<0.001
Overweight or obese54 (72)87 (25.1)< 0.001
Systolic blood pressure elevated10 (13.2)24 (6.7)0.06
Blood sugar level, mmol/L (SD)5.8 (1.8)5.8 (0.9)0.790
Diabetic, n (%)1 (1.3)4 (1.1)0.657
Never exercise or play sports, n (%)13 (20)36 (12.6)0.120
Frequent exercise (>2 times/week), n (%)52 (69.3)253 (70.9)0.828
Current smoker, n (%)54 (72)301 (84.3)0.014
Alcohol in last 12 months   
 None, n (%)7 (9.5)22 (6.2)0.468
 Unsafe, n (%)61 (82.5)309 (87.5)0.266
HBsAg-positive, n (%)1 (1.3)2 (0.6)0.96
 Not immune, n (%)51 (68.9)248 (70.7)0.815
 Immune: vaccinated, n (%)20 (27)88 (25.1)0.757
Anti-HBc, n (%)3 (4.0)14 (4.3)0.756
Anti-HCV, n (%)8 (10.5)6 (1.7)< 0.001
HIV-positive, n (%)00
Elevated liver biochemistry   
 GGT, n (%)40 (52.6)22 (6.1)< 0.001
 AST, n (%)23 (30.3)7 (1.9)< 0.001
Abnormal lipid analysis   
 Total cholesterol, n (%)22 (29.7)26 (7.4)< 0.001
 LDL cholesterol, n (%)15 (20.5)29 (8.3)0.002
 HDL cholesterol, n (%)35 (47.3)140 (40.1)0.26
 Triglycerides, n (%)16 (21.6)18 (5.2)< 0.001
Table 5. Multiple Logistic Regression Model for Factors Associated with Elevated ALT
FactorUnadjusted OR (95% CI)Adjusted OR (95% CI)P value
Anti-HCV6.9 (2.3–20.4)14.6 (3.7–57.6)<0.001
BMI: overweight or obese7.7 (4.4–13.5)6.9 (3.7–13.1)<0.001
Total cholesterol5.3 (2.8–9.9)3.6 (1.7–7.7)0.01
Current smoker0.48 (0.27–0.85)0.49 (0.24–0.99)0.05
Triglycerides elevated5.0 (2.4–10.5)2.1 (0.9–5.0)0.08
LDL cholesterol elevated2.5 (1.5–4.2)

Blood glucose levels, the number of diabetics, physical activity, and alcohol consumption were also not important markers for elevated ALT. HCV antibody positivity conveyed a very high risk for elevated ALT values (adjusted OR 14.6; CI 3.7–57.6), although the wide CI is a reflection of the small patient numbers. Patients with HBV (hepatitis B surface antigen–positive) were too few to have any statistical impact (n = 3). When adolescents with viral hepatitis were excluded, 76% of those remaining with abnormal ALT values were either overweight or obese, and 92% had 1 or more features of the metabolic syndrome according to Adult Treatment Panel III criteria.19 Current smokers overall had lower ALTs and this remained a modestly protective factor on multivariate analysis (OR 0.49; CI 0.24–0.99; P = 0.05), possibly a reflection of the anorectic properties of nicotine.

Risk Factors for HCV Exposure.

Fourteen males (3.2%) were found to be HCV antibody–positive. A comparison of this group with subjects who were HCV antibody–negative is given in Tables 6 and 7. Overall, subjects positive for the HCV antibody were significantly more likely to have injected drugs in the previous 12 months (OR 16.1; CI 5.0–51.5) and to have been exposed to HBV (OR 6.0; CI 1.2–29.4). Use of heroin was highly significant on univariate analysis (P < 0.001), but clearly followed injecting drug use and lost significance on multivariate testing (Table 7). Amphetamine (OR 3.4) and benzodiazepine use (OR 1.6) appeared more common in the HCV antibody–positive cohort; however, this was not significant on multivariate regression, possibly because of the small sample size.

Table 6. Risk Factors Associated with Hepatitis C Antibody Positivity in Adolescent Males
FactorHCV-Positive (n = 14)HCV-Negative (n = 425)P Value
  1. Abbreviations: Anti-HBc, hepatitis B core antigen antibody; Anti-HBs, hepatitis B surface antigen antibody; HBsAg, hepatitis B surface antigen; HIV, human immunodeficiency virus; LDL, low-density lipoprotein; SD, standard deviation.

Age (years), mean (SD)17 (0.7)16.6 (1.3)0.06
BMI, mean (SD)24.2 (6.2)23.7 (4.9)0.72
Overweight or obese, n (%)4 (30.8)143 (34.5)0.78
Blood sugar level, mmol/L (SD)5.8 (0.7)5.8 (1.1)0.42
Never exercise or play sports, n (%)6 (42.9)43 (10.2)<0.001
Frequent exercise (>2 times/week), n (%)7 (50)301 (71.3)0.115
Current smoker, n (%)12 (85.7)344 (81.5)0.69
Alcohol in last 12 months, n (%)12 (85.7)388 (93.2)0.464
 None, n (%)2 (14.3)29 (6.8)0.613
 Unsafe but not often, n (%)12 (85.7)339 (79.8)0.989
Use of illicit drugs, n (%)14 (100)380 (90.3)0.554
 Heroin, n (%)9 (64.3)45 (10.7)<0.001
 Marijuana, n (%)14 (100)379 (90)0.21
 Benzodiazepines, n (%)5 (35.7)42 (10)0.009
 Amphetamines, n (%)12 (85.7)196 (46.6)0.004
 Cocaine, n (%)4 (28.6)71 (16.9)0.28
 Injected drugs last 12 months, n (%)7 (50)23 (5.5)<0.001
 Unsafe injection last 12 months, n (%)5 (71.4)6 (26.1)0.068
Lifetime sexual partners <11, n (%)6 (42.8)102 (24.7)0.221
Sexually transmitted infections, n (%)7 (50.0)110 (25.9)0.064
Sex worker, n (%)1 (7.1)2 (0.5)0.191
Condom use with casual sex   
 Never/not always, n (%)8 (57.1)145 (40.3)0.21
Body piercing or tattoos, n (%)9 (69.1)218 (51.9)0.51
HBsAg, n (%)1 (7.1)2 (0.5)0.09
 Not immune, n (%)6 (42.9)295 (70.9)0.024
 Immune: vaccinated, n (%)5 (35.7)105 (25.2)0.354
Anti-HBc, n (%)3 (21.4)16 (3.8)0.02
HIV-positive, n (%)00
Elevated liver/lipid biochemistry   
 ALT, n (%)8 (57.1)68 (16.3)<0.001
 LDL cholesterol, n (%)044 (10.7)0.20
 Triglycerides, n (%)2 (15.4)32 (7.8)0.32
Table 7. Multiple Logistic Regression Model of Factors Associated with Hepatitis C Antibody Positivity
FactorUnadjusted OR (95% CI)Adjusted OR (95% CI)P Value
  1. Abbreviations: Anti-HBc, antibody to hepatitis B core antigen; Anti-HBs, antibody to hepatitis B surface antigen.

Injected drugs last 12 months17.3 (5.6–53.6)16.1 (5.0–51.5)<0.001
Anti-HBc6.8 (1.7–26.9)6.0 (1.2–29.4)0.03
Amphetamine use6.8 (1.5–31.2)
Sexually transmitted infections2.8 (0.97–8.2)
Anti-HBs: not immune0.31 (0.1–0.9)
Benzodiazepine use5.0 (1.6–15.8)
Heroin use15.0 (4.8–46.8)
Never exercise or play sports6.0 (1.9–18.6)

HCV antibody–positive subjects did not have an increased prevalence of risk-taking sexual activity when compared with their peers as suggested by similar levels of condom use with casual sex, number of previous sexual partners, and incidence of sexually transmitted infections (Table 6). Whereas metabolic parameters such as BMI and blood sugar level were not statistically different between the 2 groups, the HCV antibody–positive cohort was more likely to have never undertaken exercise or played sports (42.9% versus 10.2%; P < 0.01). When this was controlled for injecting drug use, the association was not apparent, suggesting that drug habit rather than HCV was limiting their activity (Table 7). Only 42% of adolescents with a positive HCV antibody test had a positive HCV polymerase chain reaction (5/12; 2 samples were misplaced).

Informed consent for a second sample was obtained, and blood was collected 12 months later on 81 subjects from the original cohort who remained under community orders. A further 3 subjects had developed HCV antibodies, suggesting a new infection rate of at least 3.7% per year and a cumulative incidence of over 5%. There were also 3 exposures to HBV (new hepatitis B core antigen antibody), but no cases of human immunodeficiency virus.


As a population, juvenile offenders are difficult to study because of their itinerant nature and unwillingness to participate in surveys or research studies. Thus, the data obtained from this cohort provides unique and valuable insight with implications for future screening and targeted health interventions. Our results indicate an alarmingly high level of HCV exposure in male offenders. The antibody positivity rate of approximately 3% is in keeping with the 2 reports in juvenile offenders from the United states (≈2%),10, 12 but lower than a previous Australian study11 in which 21% were anti-HCV positive. The high rate in the latter cohort was likely due to the extraordinarily high rates of injecting drug use (63%) in the small study sample (n = 83). Our results are likely to be more representative given the sampling frame, inclusion criteria, and design, though it may be an underestimation given the possible ascertainment bias (see exclusions).

Overall, HCV antibody was detected at a rate 10-fold to 40-fold higher than in similar-aged adolescents in the United States and Italy.20–22 Of even greater concern was that in the following 12 months, the HCV exposure rate was approximately 4%. It is known that HCV antibody positivity is high in incarcerated adults, with rates between 8%23 and 37%.24 Given that criminal recidivism rates of up to 69% are common in adolescents,25, 26 there is a clear window of opportunity while they are under supervision to try and reduce future HCV transmissions. Murray et al.10 showed that the majority of juvenile offenders have a very poor understanding of HCV transmission. Only 17% of their cohort correctly identified risk behaviors for HCV, and only 5% knew it was a disease affecting the liver. Hence, greater education regarding blood-borne viruses, risk factors for transmission, and the implementation of harm minimization strategies in this population is crucial.

The principle risk factor for HCV exposure in our cohort was injecting drug use. Among the 30 current injectors, 7 (23%) were HCV antibody–positive, consistent with epidemiological surveys in nonincarcerated populations.8, 27 Sexual behavior was not a significant risk factor in this promiscuous cohort, suggesting a very minor role for sexual transmission of HCV. This finding has been widely validated,28–30 but is in contrast to a previous report.7 A high rate of viral clearance was seen, with only 42% of those with anti-HCV having confirmatory HCV polymerase chain reaction. In older adolescents the expected rate of viral clearance should be equivalent to that of adults (≈25%31); however, clearance rates of up to 42% have been seen in cohorts of acute HCV similarly acquired by intravenous drug use.32

Currently, available combination antiviral therapy is associated with viral eradication rates of 54%–63%.33, 34 Reports of treatment in adult prison populations remain very rare, but one such series has shown equivalent sustained virologic response rates to those achieved in community settings.35 Possibly more relevant is a report of 50 current or recently detoxified injecting drug users treated in a community-shared care facility36 that achieved sustained virologic response rates of 62% while maintaining high rates of compliance to treatment. Similar strategies should be advocated in juvenile correction facilities.

Another intervention applicable to these adolescents is routine HBV vaccination. Only 30% of our study population had protective antibody levels, with the remainder at high risk of infection, given their risk-taking sexual activity and drug-related behaviors. A 3-dose course of the HBV vaccine can be provided on an accelerated schedule at days 0, 7, and 21 with comparable seroprotection to the traditional 6-month course.37, 38 This strategy is likely to be successful and has been shown to be more acceptable than traditional regimens in similar at-risk populations.39, 40

Defining Normal Liver Biochemistry in Adolescents.

The Department of Juvenile Justice study provided a unique opportunity to re-evaluate the upper limits of normal (ULN) for liver biochemistry in adolescent males. Our sample of 206 adolescent males at the lowest risk for liver disease allowed a 99% CI for determination of this reference interval using the 95th percentile.41 The values obtained for ALT, AST, and GGT were significantly lower than the reference ranges quoted in the recent literature,42, 43 much of which includes populations with a high prevalence of obesity and NAFLD. Prati et al.18 examined healthy ranges for ALT using similar definitions and found that for men, 30 IU/L was at the 95th percentile. They and others44 have shown that sex has a significant influence on ALT levels and suggested that ULN should be assessed separately for males and females. Because age has a significant impact on ALT levels following an inverted U-pattern,44, 45 different ranges are required for adolescents compared with adults. Thus, our new definition of the ULN for liver tests allowed a more robust analysis of the factors responsible for elevated enzymes. Looking to the future, this definition will permit greater sensitivity in diagnosing early liver injury in adolescent male populations and in implementing appropriate intervention.

Associations with Elevated ALT in Adolescents.

Elevated aminotransferases are a biomarker of hepatic necroinflammatory activity. Using our suggested ULN cutoff, 76 patients had abnormal ALT values. This had a good sensitivity for picking up subjects with hepatitis C viremia (80% versus 40% with old cutoff; P value not significant). HBV exposure and HBV surface antigen status were not predictive of elevated ALT, most likely reflecting the fact that HBV infection is usually in the immunotolerant phase in adolescents.

In the remaining subjects with elevated ALT, the majority were either overweight or obese. Moreover, more than 90% of these patients had one or more features of the metabolic syndrome according to Adult Treatment Panel III criteria.19 These individuals were significantly more likely to have elevated total cholesterol or triglycerides, but elevated blood sugar and hypertension were infrequent. Alcohol use and levels of physical activity were also not important risk factors for elevated ALT, possibly a reflection of the overall infrequency of chronic excessive alcohol use and the higher physical activity levels in adolescents when compared with adults. A recent report that examined the chronological development of elevated aminotransferases46 suggests that weight gain occurs prior to hyper-aminotransferasemia; dyslipidemia develops concurrently, and glucose intolerance was the final step in progression of the metabolic syndrome. Our findings are consistent with this chronology of metabolic events.

In patients with unexplained elevated aminotransferases, the vast majority will have NAFLD.47, 48 A recent study showed that obese children with relatively minor ALT elevations have some degree of fibrosis on biopsy in 80%, and 5% had signs of early cirrhosis.49 This figure is likely an underestimation, because the full spectrum of NAFLD from steatosis to nonalcoholic steatohepatitis and cirrhosis has been reported in adults with a “normal” ALT defined using traditional reference ranges rather than the ULN as defined in adults18 or adolescents (this report). Therefore, in clinical practice, the finding of elevated aminotransferases in adolescents is highly significant. Once viral hepatitis is excluded, the most likely diagnosis is NAFLD, and the subject is likely to have elements of the metabolic syndrome. To prevent further hepatic damage and to minimize cardiovascular and diabetes risk, targeted interventions in adolescents at the earliest stages of metabolic dysfunction are a particularly high priority.

In conclusion, this study has provided a wealth of clinical and health-related data relevant to adolescents. The new definition of normal adolescent ALT allows greater sensitivity in diagnosing early liver disease. By identifying those with HBV, HCV, and obesity-related liver disease, targeted interventions can and should be implemented to minimize future health-related morbidity.


We thank Paul Nelson for project and data management and Keith Westbury for assistance with the pathology.