Potential conflict of interest: Nothing to report.
The upper limit of normal for ALT activity has been recommended to be lowered to ≤30 U/L in men and ≤19 U/L in women. These changes have been suggested to be diagnostically useful in subjects with nonalcoholic fatty liver disease (NAFLD). Our aim was to investigate the prevalence and spectrum of NAFLD with regard to the new ALT guidelines in 233 women with class II/III obesity. We compared our prior reference range for ALT (ULN ≤ 30 U/L in women) with the new standard. Our study demonstrates that only 86 patients (36.9%) would be classified as having normal ALT levels compared with 169 patients (72.5%) by the new and old standards, respectively. In patients with normal ALT activity (new vs. old standard), the prevalence of fatty liver (FL: 39.5% vs 40.2%), portal fibrosis, and steatosis (IPF: 37.2% vs. 33.7%) and nonalcoholic steatohepatitis (NASH: 23.3% vs. 26%) were similar. In comparison, newly defined patients with elevated ALT levels (>19 U/L) demonstrated an increased prevalence of FL (36%) and IPF (11.6%) but a 23.8% decrease in the prevalence of NASH as compared with the old standard. The sensitivity and specificity for NASH were 42% and 80% (ALT > 30 U/L) compared with 74% and 42% (ALT > 19 U/L). In conclusion, a significant increase in the prevalence of FL and IPF is detected in subjects with elevated ALT levels with the application of the new standard. However, the diagnostic utility for ALT to identify NASH or IPF remains poor, and significant healthcare expenditures may be incurred if this standard is adopted. (HEPATOLOGY 2005.)
Obesity is a major health issue in the United States that is associated with increased morbidity and mortality.1 The prevalence of obesity (body mass index [BMI] ≥ 30 kg/m2) in the United States is 31%, and the prevalence of morbid obesity (BMI ≥ 40 kg/m2) is even growing at a greater rate than obesity.2 The degree of obesity is further subdivided into class I (BMI ≥ 30-34.9 kg/m2), class II (BMI 35-39.9 kg/m2), and class III (BMI ≥40 kg/m2).3 Individuals with class II obesity that is associated with medical comorbidities or class III obesity alone are considered suitable candidates for gastric bypass surgery.4
Nonalcoholic fatty liver disease (NAFLD) constitutes a spectrum of liver injuries associated with obesity and insulin resistance.5, 6 It encompasses steatosis (fatty liver [FL]) with and without inflammation; steatosis and portal fibrosis without zone 3 liver injuries (isolated portal fibrosis [IPF]), and nonalcoholic steatohepatitis (NASH). NASH is defined by zone 3 hepatocellular injury (ballooning degeneration, apoptosis, or Mallory bodies) or perivenular/pericellular fibrosis.7 NASH is an aggressive form of NAFLD and can progress to cirrhosis and hepatocellular carcinoma.8
Hepatic aminotransaminases, alanine (ALT) and aspartate (AST), are both markers for hepatocellular injury, although ALT is considered more specific. Normal levels of liver transaminases have been demonstrated in subjects with the entire spectrum of NAFLD, and therefore ALT and AST activity has not been very useful in predicting NASH or advanced fibrosis.9–11 Current standards of normal liver enzymes were defined 2 decades ago by using general populations that may have included covert liver pathological conditions such as nonalcoholic liver disease. Prati and colleagues12 have recently proposed a new ALT upper limits of normal for healthy males and females to be ≤30 U/L and ≤19 U/L, respectively. Our laboratory reference range for ALT has been ≤30 U/L for men and women. Therefore, we were particularly interested in investigating the application of the new guidelines in women. The aims of this study were to apply the new ALT standard to our database of class II/III obese women undergoing gastric bypass surgery to determine the prevalence of an elevated ALT level and to assess the diagnostic implications for the spectrum of NAFLD.
BMI, body mass index; NAFLD, nonalcoholic fatty liver disease; FL, fatty liver; IPF, isolated portal fibrosis; NASH, nonalcoholic steatohepatitis; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATP III, Third Report of the National Cholesterol Education Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults Treatment Panel.
Patients and Methods
Patient and methods were similar to our previous publication.13 A retrospective analytical study was carried out in 237 consecutive women with class II/III obesity undergoing laparoscopic Roux-en-Y gastric bypass surgery between January 2002 and September 2004. During surgery, all patients had routine liver biopsies, and the specimens were analyzed by an experienced pathologist. Four subjects had isolated elevated AST levels and were excluded. Therefore, 233 women constituted our study cohort. All patients were negative for hepatitis B and C serology. Patients were excluded if either positive titers for auto-antibodies or elevated iron studies and liver biopsies were consistent with autoimmune hepatitis or iron overload. Subjects consumed less than 1 alcoholic beverage daily, and more than 90% consumed less than 1 drink weekly (by history). None of the subjects were taking medications such as methotrexate, amiodarone, tetracycline, high doses of estrogen, tamoxifen, steroids, or calcium channel blockers. The protocol was approved by the Human Investigative Committee of the University of Alabama at Birmingham.
Preoperative serum samples were analyzed for ALT and AST using the alpha-ketoglutarate reaction. Total bilirubin, alkaline phosphatase, total protein, albumin, fasting blood sugar, cholesterol, triglycerides high-density lipoprotein, and low-density lipoprotein were also measured. ALT activity was classified as normal and elevated as defined by our prior reference laboratory cutoff value of ≤30 U/L and compared with the new proposal of 19 U/L. In addition, the spectrum of NAFLD was noted in the subgroup of patients between these two criteria (19 U/L < ALT < 31 U/L).
Metabolic syndrome was defined in all the subjects defined by the Third Report of the National Cholesterol Education Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel—ATP III).14 Subjects with an elevated blood pressure (≥130/≥85 mm Hg) documented on at least 2 of 3 clinic visits before surgery was considered as hypertensive as per ATP III criteria. Oral glucose tolerance tests were done in all patients except for patients with known diabetes. The American Diabetes Association criteria for diabetes was used: fasting blood sugar > 126 mg/dL or 2 hours oral glucose tolerance tests > 200 mg/dL and patients with 2 hours oral glucose tolerance tests between 140 and 200 mg/dL were considered to have impaired glucose tolerance.15
Liver biopsies were performed on the left lobe of liver at the beginning of the operation using a Tru-cut needle (Bard Max-Core, Covington, GA). The liver biopsy specimen size was calculated for a minimal and maximal average amount of tissue as previously described.6 The minimal sample size is biased toward underestimating the amount of tissue, and the maximal sample size overestimated the specimen sample.
The liver biopsy was interpreted by a single pathologist blinded to all clinical data. Steatosis was graded as minimal (1%-5%), mild (6%-33%), moderate (34%-66%), and severe (67%-100%). Hepatic inflammation was identified as panlobular, portal, or both and subjectively assessed as mild, moderate, or severe. The pathological diagnosis of NASH was in accordance with Kleiner et al.,16 and staging of fibrosis was in accordance with the National Institutes of Health–sponsored NASH Clinical Research Network guidelines: stage 1a/b—mild/moderate zone 3 fibrosis; 1c—portal/periportal fibrosis only; stage 2—zone 3 and portal fibrosis; stage 3—bridging; and stage 4—cirrhosis.
Statistical analysis was performed by using SPSS (SPSS Inc., Chicago, IL; V11.5 for windows). Data were analyzed as both continuous and categorical variables. Continuous variables (such as age, liver enzymes) were expressed as mean ± standard deviation (SD). Categorical variables (such as NASH, stages of fibrosis) were analyzed using chi squared. Odds ratios, 95% CI, and P values were used to analyze the association between liver enzymes, metabolic syndrome, and histopathological injuries. The 95% CI were calculated on the “absolute difference between prevalence estimates,” to analyze the difference between NAFLD spectrum and elevated ALT by old and new criteria. Independent t tests were used when required. A P value of less than .05 was considered statistically significant.
The cohort consisted of 233 women with a mean ± SD age of 41.2 ± 9 years (Table 1). The BMI was 48.2 ± 7.3 kg/m2 (range, 35-76.6 kg/m2) and weight of 300 ± 55.3 pounds (range, 191-465 pounds). Two hundred two patients (86.7%) were non-Hispanic whites.
Table 1. General and Clinical Characteristics of the Cohort
All Patients (N = 233)
ALT: Old Criteria
ALT: New Criteria
≤30 U/L (n = 169)
>30 U/L (n = 64)
≤19 U/L (n = 86)
>19 U/L (n = 147)
41.2 ± 9
41.4 ± 8.9
40.4 ± 9.3
41.7 ± 9.4
40.9 ± 8.8
Race (W %)
48.2 ± 7.3
48.1 ± 7.1
48.4 ± 7.9
48.1 ± 6.6
48.3 ± 7.7
26.8 ± 18
19.7 ± 4.9
45.5 ± 25.1
15.6 ± 2.6
33.3 ± 19.8
24.4 ± 14.8
19.5 ± 4.1
37.5 ± 22.8
17.2 ± 3.4
28.6 ± 17.1
Diabetes mellitus (%)
↑ Glucose %
↑ Triglycerides (%)
↓ High-density lipoprotein (%)
Metabolic syndrome (%)
The ALT activity range was 10 to 216 U/L (mean, 26.8 ± 18 U/L); 64 of 233 patients (27.5%) had elevated ALT levels by our old reference standard. The AST activity range was 10 to 186 IU/L (mean, 24.4 ± 14.8 IU/L), and 33 of 233 patients (14.2%) had elevated levels (>30 IU/L), all of whom had elevated ALT. Applying the new ALT standard resulted in a significant prevalence increase of 147 patients (63.1%, 95% CI 35.6 ± 8.43%; P < .0001) with elevated levels. Therefore, 169 patients (72.5%) had normal ALT (old reference cutoff) compared with 86 patients (36.9%) after applying the new guidelines. No significant differences in the prevalence of hypertension, diabetes, dyslipidemia, and the metabolic syndrome were noted when comparing old and new ALT criteria.
Liver Biopsy Specimen Size.
Liver biopsy specimens were 1.0 cm or larger in 225 of 233 (96.5%), 1.5 cm or larger in 205 of 233 (88%), and 2.0 cm or larger in 152 of 233 (65.2%) subjects. The average minimum biopsy size was 2.4 ± 1 cm, and the average maximum biopsy size was 3 ± 1.5 cm for the entire cohort. In patients with elevated ALT levels, no differences were noted in the minimum or maximum biopsy sample (old criteria: 2.38 cm and 2.96 cm, compared with new criteria: 2.38 cm and 3.0 cm), eliminating sampling variation when applying the new ALT standard to the histological results.
Liver biopsy results for the 233 subjects were categorized into FL, IPF, and NASH. The prevalence of liver injury among the 3 groups is: FL (36.1%), IPF (31.3%), and NASH (32.6%). Table 2 demonstrates the characteristic histopathology and spectrum of NAFLD, comparing old and new ALT criteria.
Table 2. Prevalence of NAFLD Histological Findings Among the Patients With Normal and Elevated ALT by Old and New Criteria
NAFLD Histological Findings (%)
All Patients (N = 233)
ALT: Old Criteria
ALT: New Criteria
Subgroup of Patients Between Two Criteria (19 U/L < ALT < 31 U/L) (n = 83)
≤30 U/L (n = 169)
>30 U/L (n = 64)
≤19 U/L (n = 86)
>19 U/L (n = 147)
Zone 3 ballooning
Zone 3 fibrosis
NASH (Zone 3 ballooning and/or fibrosis)
Isolated Portal Fibrosis (IPF)
All the patients had some amount of steatosis; minimal 23.6% (55/233), mild 34.2% (80/233), moderate 21.9% (51/233), and severe 20.2% (47/233). The grades of steatosis were similar in patients with normal ALT levels (old or new criteria). However, patients with elevated ALT levels by the new criteria had an increased prevalence of minimal/mild steatosis (34.4%-51%) compared with a slightly lower prevalence of moderate/severe steatosis (55.6%-49%) when compared with the old standard. Most patients 63.8% (53/83) with newly characterized elevated ALT levels were noted to have minimal/mild steatosis.
Patients with normal ALT by old and new criteria had a similar prevalence of FL (40.2% vs. 39.5%). However, the prevalence of FL was significantly increased in patients with elevated ALT by new criteria compared with old criteria (34% vs. 25%, respectively; 95% CI 9 ± 8.23; P = .03). Figure 1 shows the prevalence of elevated ALT by old and new criteria with respect to different grades of steatosis in subjects with FL. Patients with IPF/NASH were more likely to have elevated ALT as compared with FL when the old criteria are used (OR 2.0, 95% CI 1.1-3.7, P = .04). However, this association was not significant with the new criteria (OR 1.3, 95% CI 0.7-2.2, P = .40) (Table 3).
Table 3. Association Between Histological Spectrum of NAFLD and Elevated ALT by Old and New Criteria
Elevated ALT: Old Criteria
Elevated ALT: New Criteria
NASH/IPF vs. FL
OR 2.0, 95% CI 1.1-3.7, P = .04
OR 1.3, 95% CI 0.7-2.2, P = .40
NASH vs. IPF/FL
OR 3.0, 95% CI 1.6-5.4, P < .0001
OR 2.0, 95% CI 1.1-3.7, P = .02
OR 3.0, 95% CI 1.6-5.5, P = .001
OR 1.7, 95% CI 0.9-3.1, P = .11
OR 2.8, 95% CI 1.5-4.9, P = .001
OR 1.4, 95% CI 0.85-2.4, P = .18
Cytologic ballooning was identified in 66 (28.3%) patients. Patients with normal ALT by old and new criteria showed similar prevalence of hepatocellular ballooning (21.9% vs. 22.1% respectively). However, patients with elevated ALT (new criteria) demonstrated a 29.5% relative decrease in ballooning injury (45.3% vs. 32%). In the 83 newly defined patients with elevated ALT values, ballooning was noted in 21.7%. A positive association was seen between ballooning and elevated ALT levels, but the association was not significant after applying the new ALT standard: (OR 3.0, 95% CI 1.6-5.5, P = .001 by old criteria, compared with OR 1.7, 95% CI 0.9-3.1, P = .11 by new criteria) (Table 3).
The prevalence of NASH was similar in the patients with normal ALT by old and new criteria (26% vs. 23.3%, respectively) but significantly decreased in subjects with elevated ALT using new criteria (50%-38.1%; 95% CI 11.9 ± 8.96; P = .01). Patients with NASH were significantly more likely to have elevated ALT activity as compared with FL or IPF (OR 3.0, 95% CI 1.6-5.4, P < .0001). Using the new criteria, this association remained significant but decreased (OR 2.0, 95% CI 1.1-3.7, P = .02) (Table 3).
Overall, fibrosis was noted in 141 patients (60.5%); mild and moderate fibrosis was noted in 119 (51.1%) and 22 (9.4%) patients, respectively. Of the 22 patients with advanced fibrosis, 2 had cirrhosis and 20 had bridging fibrosis. Ninteen of the 22 patients (86.4%) with advanced fibrosis also had NASH; 3 subjects with bridging fibrosis were classified as IPF. The mean ALT value in patients with advanced fibrosis was significantly different from that of patients with either no or mild fibrosis (no fibrosis: 23 ± 9.3 IU/L, mild fibrosis: 26.6 ± 14.6 IU/L, and advanced fibrosis: 44.5 ± 41 U/L; P < .0001).
The prevalence of elevated ALT levels in patients with no, mild, and advanced fibrosis was statistically significant by the old reference standard (17.4%, 27.7%, 68.2%, respectively; P < .0001); however, it was only trending with the new standard (58.7%, 63%, 81.8%, respectively; P = .06) (Fig. 2). The old ALT standard demonstrated that patients with fibrosis were significantly more likely to have an elevated ALT compared with those without fibrosis (OR 2.8, 95% CI 1.5-4.9, P = .001), whereas the association was not significant when applying the new criteria (OR 1.4, 95% CI 0.85-2.4; P = .18) (Table 3).
Isolated Portal Fibrosis.
The prevalence of IPF is similar irrespective of old and new ALT cutoff criteria (25% vs. 27.9%, respectively; 95% CI 2.9 ± 8; P = .47).
Two hundred twenty-nine liver biopsy specimens (98.3%) showed inflammatory infiltrates that were mild (n = 220), moderate (n = 8), and severe (n = 1). Because of the small number of patients with moderate and severe inflammation, cutoff levels of ALT activity were not analyzed.
Table 4 details the prevalence of elevated ALT values across the spectrum of NAFLD. The prevalence of an elevated ALT was statistically significant when applying the old standard (P = .002); however, it only trended toward significance applying the new standard (P = .06). The proposed lower ALT cutoff value resulted in a decreased prevalence of normal ALT levels and increased prevalence of elevated ALT levels across the entire spectrum of NAFLD.
Table 4. Prevalence of Normal and Elevated ALT by Old and New Criteria Across the Spectrum of NAFLD
FL (n = 84)
IPF (n = 73)
NASH (n = 76)
Old criteria (%)
New criteria (%)
Diagnostic Utility of ALT Proposed New Guideline.
The diagnostic usefulness of an ALT level using both the old and new reference range for the spectrum of NAFLD is detailed in Table 5. Although the sensitivity for NASH, IPF, and FL increase with the application of the new standard, the negative predictive value does not improve. In addition, the specificity for NASH or IPF decrease with the new ALT cutoff level, thereby decreasing the positive predictive value for the more aggressive liver injury.
Table 5. Diagnostic Utility of Liver Enzymes for Identifying NAFLD
The main findings in this study are that adopting the newly suggested ALT upper range of normal would result in a significant increase in the prevalence of elevated ALT levels (27.5% to 63.1%) in women with class II/III obesity. The complete spectrum of NAFLD is demonstrated in the patients with normal ALT activity even after decreasing the cutoff level to ≤ 19U/L. Importantly, the prevalence of NASH does not change irrespective of the new ALT standard compared with our old reference range (23% vs. 26%, respectively). After applying the new ALT standard, most patients (61.9%) demonstrated either FL or IPF. Despite the significant decrease in the new cutoff level for ALT, 18.2% of patients with advanced fibrosis had normal ALT levels. Although the prevalence of FL and IPF increased and that of NASH decreased with elevated ALT levels as per the new standard, the diagnostic utility of ALT activity for NASH and IPF remains insufficient.
ALT is located in the hepatocellular cytosol, whereas AST is mostly within the mitochondria.17 A major function of ALT is catalyzing the transfer of an amino group from alanine to ketoglutarate to form pyruvic acid and glutamate, respectively. The light absorption that is lost when NADH is oxidized to NAD (during pyruvic acid conversion to lactate) is used to indirectly measure ALT (and AST) activity. Although ALT activity is mostly hepatic in origin, muscle inflammation18 or injury19 can increase serum ALT. Elevated liver enzymes often correlate poorly with the degree of the chronic liver injury,20 steatosis, and fibrosis.21 Nevertheless, a persistently elevated ALT level is often the tipping point for further diagnostic evaluation.
Ruhl and Everhart,22 in a large epidemiological study, noted that 2.8% of the adult population had elevated ALT levels (>43 U/L). Of note, in subjects with class II/III obesity, this prevalence increased to 6.6%. NAFLD was the presumed origin in most individuals with elevated ALT, and this was attributed to an overweight or obese health state. The authors also demonstrated that the application of the new criteria (>19 U/L), would result in a fivefold increase in the prevalence of an elevated ALT level in the general population (men, 12.8%; women, 13.9%). However, the prevalence of using the new ALT cutoff and its association with increasing BMI was not provided. As pointed out by the authors, their study was limited because ALT activity was used as the surrogate diagnosis of NAFLD in the absence of histology. Our study confirms the significant prevalence of elevated ALT levels in class II/III obesity. Using an ALT cutoff (>43U/L) as in the study by Ruhl and Everhart, 8.6% of our cohort had an elevated ALT level (data not shown). All of our subjects had liver biopsies that established the spectrum of NAFLD as well as excluding other causes of liver injury. Of note, an ALT level greater than 19 U/L was present in 63% of women with class II/III obesity in our study, and this represents a 2.3-fold increase over our prior ALT reference range upper limit.
As stated by Prati and colleagues,12 the standards for “healthy” liver transaminases were likely defined in patients with subclinical hepatitis C infection and NAFLD. Therefore, the current ALT range of “normal” may underestimate the presence of these common forms of chronic liver disease. Our results support this contention and suggest that lowering the ALT healthy range would significantly increase the detection of NAFLD (Table 2). Importantly, newly identified patients with an elevated ALT value primarily demonstrate minor liver injury on biopsy. Applying the new criteria, the relative increases in individuals with minimal steatosis and mild steatosis on biopsy were 48.1% and 48.4%, respectively. Some experts consider a minimal degree of steatosis to be histologically within normal limits. The prevalence of IPF only increased 11.6%, whereas the prevalence of NASH, considered the more aggressive form of NAFLD, decreased by 23.8%. Taken together, our data confirm that the adoption of the new ALT standard will primarily detect the early stages of NAFLD and thus may be useful in counseling patients.
The increased sensitivity for the entire spectrum of NAFLD is associated with a decrease in the specificity of NASH (Table 3). Additionally, both the negative and positive predictive values for NASH are poor. We also analyzed the diagnostic utility of ALT applying the upper level of 44 U/L in subjects with a BMI >23 kg/m2 (which is obviously all of our cohort), as suggested by Piton et al.23 For NASH, the sensitivity was 18%, specificity 96%, and negative and positive predictive values were 71% and 70%, respectively. Taken together, irrespective of the multiple ALT standards suggested to date, no cutoff value is diagnostically useful to detect the presence of NASH. Of note, predictive values are influenced by disease prevalence; therefore, the diagnostic implications may differ in other populations with a lower prevalence of NAFLD than those found in our study.
To date, a single study has investigated the clinical and histological spectrum of NAFLD in subjects with “normal” ALT activity.10 A comparison to our study is limited by several factors. Most subjects who underwent a liver biopsy had underlying hepatic pathological conditions warranting a liver biopsy, whereas our cohort underwent screening liver biopsies. Our patients were heavier (BMI: 48.2 vs. 29 kg/m2), more women (100% vs. 69%), 10 years younger, and included fewer individuals with hyperglycemia (73.3% compared with 43.5% in our group). Of note, their reference range for ALT activity was considerably higher (>75 U/L). Taken together, this may explain the higher percentage of patients with NASH (72%) among those with normal ALT, in their study compared with our results (23%). A subgroup analysis of individuals (n = 15) with ALT levels less than 31 U/L was too small for comparisons. Despite these differences, the two studies demonstrate that the entire spectrum of NALFD is present in patients with normal ALT levels despite the ALT cutoff level that is applied. In our subjects with normal ALT activity, the prevalence of FL, IPF, and NASH are very similar irrespective of the old or new ALT criteria. Additionally, the relative distribution of NAFLD in patients with normal ALT activity with respect to an elevated ALT level remains the same. In patients with elevated ALT levels, the odds of having NASH were similar (2.1—old criteria, and 1.8—new criteria).
The implications of the new ALT reference range must be considered with regard to healthcare expenditures. Although our data show that most patients with newly defined ALT levels by the proposed criteria have FL or IPF, NASH is present in more than 20% of this cohort, and we speculate that significant increases in diagnostic healthcare expenditures may ensue if the new standard is adopted. For example, evaluating an elevated ALT by ordering only a hepatitis C antibody and an abdominal ultrasound (Medicare charges: $22 and $300, respectively) limited to class II/III obese individuals would cost hundreds of millions of dollars [5% of the adult population (10 million people) is a conservative estimate, and our data suggest 63% of these individuals will have an elevated ALT].
Limitations in this study include fluctuating ALT levels24 that mandate multiple testing that was not done in our study. Also, the cohort investigated included only women who were undergoing gastric bypass surgery (only 35 men had bypass surgery during the study period), and this may not be extrapolated to the general population.
In conclusion, the complete spectrum of NAFLD is demonstrated in women with class II/III obesity with normal ALT levels after the application of the proposed lower ALT standard. Although the new ALT cutoff increases the sensitivity of FL, IPF, and NASH, ALT levels are not useful diagnostically to either include or exclude NASH. The new ALT standard may be helpful for the early recognition of the milder spectrums of NAFLD but may result in significant healthcare expenditure and questionable benefit.
The authors thank Teresa Leath, RN, for her contribution in enrolling subjects and completion of the database.