SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Nonalcoholic steatohepatitis (NASH) is common in morbidly obese persons. Liver biopsy is diagnostic but technically challenging in such individuals. This study was undertaken to develop a clinically useful scoring system to predict the probability of NASH in morbidly obese persons, thus assisting in the decision to perform liver biopsy. Consecutive subjects undergoing bariatric surgery without evidence of other liver disease underwent intraoperative liver biopsy. The outcome was pathologic diagnosis of NASH. Predictors evaluated were demographic, clinical, and laboratory variables. A clinical scoring system was constructed by rounding the estimated regression coefficients for the independent predictors in a multivariate logistic model for the diagnosis of NASH. Of 200 subjects studied, 64 (32%) had NASH. Median body mass index was 48 kg/m2 (interquartile range, 43-55). Multivariate analysis identified six predictive factors for NASH: the diagnosis of hypertension (odds ratio [OR], 2.4; 95% confidence interval [CI], 1-5.6), type 2 diabetes (OR, 2.6; 95% CI, 1.1-6.3), sleep apnea (OR, 4.0; 95% CI, 1.3-12.2), AST > 27 IU/L (OR, 2.9; 95% CI, 1.2-7.0), alanine aminotransferase (ALT) > 27 IU/L (OR, 3.3; 95% CI, 1.4-8.0), and non-Black race (OR, 8.4; 95% CI, 1.9-37.1). A NASH Clinical Scoring System for Morbid Obesity was derived to predict the probability of NASH in four categories (low, intermediate, high, and very high). Conclusion: The proposed clinical scoring can predict NASH in morbidly obese persons with sufficient accuracy to be considered for clinical use, identifying a very high-risk group in whom liver biopsy would be very likely to detect NASH, as well as a low-risk group in whom biopsy can be safely delayed or avoided. (HEPATOLOGY 2008.)

Non-alcoholic steatohepatitis (NASH) is a stage of nonalcoholic fatty liver disease (NAFLD) characterized by the presence on liver biopsy of steatosis and of necroinflammation with variable amounts of fibrosis in individuals who do not consume alcohol to excess.1–3

NASH may progress to cirrhosis in 15%-25% of patients, and NASH-related cirrhosis is now considered the major cause of cryptogenic cirrhosis.4–7 Once cirrhosis develops, 30%-40% of these patients succumb to a liver-related death over a 10-year period, the mortality rate being similar to or worse than cirrhosis associated with other forms of hepatitis.8, 9

The prevalence of NASH in morbidly obese persons ranges from 20%-35%, whereas simple steatosis (a benign condition) is present in approximately 70%.10–12 Obese patients are at particularly high risk for NASH in view of the frequent co-existence of other features of the metabolic syndrome. Metabolic syndrome and its physiologic correlate, insulin resistance, are fundamental to the pathogenesis of NASH.13–15 Biopsy series in morbidly obese persons undergoing bariatric surgery have shown rates of cirrhosis, often called cryptogenic, of approximately 2%-4%.3, 12, 16–18 Histologic analysis also provides unique prognostic information indicative of patients more likely to progress from NASH to cirrhosis.2, 19 However, liver biopsy is invasive, associated with complications, and physically challenging in those who are obese.20–23

An additional concern is the rising prevalence of morbid obesity (defined as a body mass index [BMI] > 35 kg/m2 together with obesity-associated diseases or a BMI > than 40 with or without obesity-associated diseases) in the United States, where approximately 14 million people are currently affected.24–26 NASH is usually asymptomatic, and in obese persons, often is associated with aminotransferases that are within the normal range.5, 27 Radiologic studies are frequently impractical, often not technically feasible because of scanner weight limitations, and at best only determine the extent of steatosis without any estimation of the more important features—necroinflammation or fibrosis.28, 29 In the absence of other reliable biochemical or genetic markers of disease presence or severity, practical clinical scoring systems are urgently required that would predict which morbidly obese persons are most likely to have NASH, rather than simple steatosis. Such a scoring system would then better identify those at higher risk who might benefit most from pursuing a liver biopsy. In addition, in patients undergoing bariatric surgery, a scoring system could identify those at low risk of NASH, thereby avoiding the risk and expense of liver sampling. The aim of this study therefore was to develop a practical and clinically useful scoring system to predict the probability of having NASH in morbidly obese patients.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Study Population

The study population consisted of 200 consecutive morbidly obese adult (at least 18 years of age) patients without known liver disease who underwent elective bariatric surgery at the University of California San Francisco between August 2001 and January 2006. Subjects with a history of excessive alcohol use (greater than 20 g/day if male or 10 g/day if female) were excluded. All subjects had standard serologic testing to exclude chronic hepatitis C or B and iron overload. Subjects with histopathological findings suggestive of liver disease other than NAFLD were also excluded. The population basis used to select the 200 morbidly obese patients studied was a sample of individuals identified from the University of California San Francisco bariatric surgery center who had been scheduled for bariatric surgery. The consent rate was over 90%, with only 14 refusals among the 214 subjects approached during the recruitment period. The study was approved by the University of California San Francisco Institutional Committee on Human Research, and all subjects provided specific written informed consent to undergo liver biopsy as part of their bariatric operation.

Data Collection

Data were prospectively collected and consisted of demographic information (age, sex, and self-reported race), medications used, coexisting medical conditions including type 2 diabetes mellitus (patients with a clinical diagnosis of adult-onset diabetes mellitus and receiving oral hypoglycemic agents with or without added insulin replacement), hypertension (patients with a clinical diagnosis of essential hypertension and receiving antihypertensive medications), hyperlipidemia (patients with a clinical diagnosis of hyperlipidemia or abnormal serum lipid levels or receiving lipid-lowering medications), arthritis or degenerative joint disease (patients with chronic joint pain and using prescription non-narcotic or narcotic medications for more than 6 months, or a history of joint replacement surgery), and obstructive sleep apnea (patients with a clinical diagnosis of obstructive sleep apnea confirmed by polysomnography and using continuous positive airway pressure therapy), anthropomorphic measurements (BMI), laboratory data including aspartate aminotransferase (AST), alanine aminotransferase (ALT), total bilirubin, prothrombin time, high-density lipoprotein, low-density lipoprotein, total cholesterol, triglycerides, albumin, serum creatinine and platelet count; and imaging data (preoperative ultrasound reading, categorized as “normal” versus “consistent with steatosis”). During the bariatric surgery, a 16 gauge Tru-cut (Cardinal Health, McGaw Park, IL) liver biopsy of the left hepatic lobe, without targeting a specific area in the lobe, was performed under direct vision.

Diagnosis of NASH

All patients included in the study underwent serologic evaluation for known underlying causes of chronic liver disease, including testing for hepatitis C and hepatitis B, and evaluation of iron studies. Each liver biopsy was individually coded and stained with hematoxylin-eosin and Masson trichrome for standard histopathological interpretation. Biopsies were interpreted by a single experienced liver histopathologist (L.F.), who was blinded to the patients' clinical data. A diagnosis of NASH was made using the NAFLD Activity Score, which has been recently published by Kleiner et al.30 The NAFLD Activity Score is an assessment of the qualitative and quantitative presence of steatosis, inflammation, hepatocyte ballooning and fibrosis, which are all histologic components of NASH. The histologic diagnosis of NASH was made in the context of also conducting a thorough histologic evaluation of each patient's liver biopsy for evidence of other liver disorders, such as autoimmune hepatitis, hepatitis C, hepatitis B, iron overload, and primary biliary cirrhosis. With respect to the diagnoses of hepatitis C and autoimmune hepatitis, the following histologic findings were exclusion criteria for this NASH study: (1) dense or diffuse mononuclear portal infiltrates, which may be seen in hepatitis C or autoimmune hepatitis; (2) prominent plasma cell infiltrates, either portal or lobular, which may be seen in autoimmune hepatitis; (3) predominance of portal-based liver injury, as may be seen with hepatitis C; (4) predominant portal-based scarring, which is not characteristic of NASH; (5) presence of portal-based granulomas, which is suggestive of primary biliary cirrhosis; and (6) presence of ductular reaction or significant cholestasis, suggestive of a biliary obstructive process.

Definition of Study Variables

Dependent/Outcome Variables.

The outcome of interest in this analysis was a final pathologic diagnosis of NASH or NASH-related cirrhosis. Consequently, two comparison groups were used: patients without NASH (normal liver histology or simple steatosis) and patients with NASH (either NASH or NASH-related cirrhosis).

Independent/Potentially Predictive Variables.

The independent variables considered for inclusion in the scoring system for prediction of NASH included demographics (age, sex, and race); clinical characteristics (BMI and co-existing medical condition including type 2 diabetes mellitus, hypertension, hyperlipidemia, cholelithiasis, arthritis or degenerative joint disease, and obstructive sleep apnea); laboratory data (AST, ALT, prothrombin time, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, albumin, creatinine, and platelet count); and results of preoperative abdominal ultrasound diagnosis (normal versus “consistent with steatosis”).

Statistical Analysis

We first compared the demographic, clinical, laboratory, and ultrasound-based characteristics of patients with NASH and without NASH using chi-square and Mann-Whitney U-tests as appropriate. To derive the binary or categorical predictors required for a clinically useful scoring system in morbid obesity, we categorized continuous variables, including age and BMI, by quartile, whereas for laboratory values, we used accepted cut points defining abnormal values. Because they were hypothesized a priori to be of primary importance, better-performing cut points for ALT and AST were found by using the search algorithm used in classification trees; this procedure empirically determines the cut point for each of the measures that best distinguishes patients with NASH from those without NASH. Non-black race groups were combined because NASH prevalence was similar in whites and Hispanics, and because the two other non-black groups were very small. Unadjusted single-predictor logistic models were then used to select a subset of predictors associated with NASH at P < 0.1 to be considered for inclusion in the multivariable scoring model, which was then obtained using forward selection with an inclusion criterion of P < 0.1. The estimated regression coefficients from a final model were then rounded to the nearest whole number to obtain the points assigned to each predictive variable included in the final model. Finally, the summed points were grouped into predefined categories for the sample prevalence of NASH: low (<15%), intermediate (15%-49%), high (50%-79%), and very high (≥80%).

Cross-validation was then used to assess the predictiveness of our scoring system that might be expected in actual clinical practice. Specifically, we estimated the risk of having NASH for the patients in each of 10 randomly selected and mutually exclusive subsets of the study population using a scoring system based on data for the other 90% of the study population and using exactly the same methods used to derive the primary scoring system. These cross-validated risk classifications were then compared with the true prevalence of NASH in each of the four risk categories. Because the cross-validation assessment of accuracy uses the “new” observations for each excluded 10% subset to assess performance of the cross-validation scoring system based on the remaining 90%, it provides a better estimate to the expected performance of the scoring system derived from the complete data than a naïve assessment based on the complete data. This procedure is similar to but more efficient than learning set/test set methods in which only some of the data are used to derive a prediction rule, and the remainder to evaluate it.31 Analyses were performed using SPSS Version 10.0.1 (SPSS Inc., Chicago, IL: Tables 1 and 2); R (R Foundation for Statistical Computing, Vienna, Austria: tree algorithm for AST and ALT); and STATA version 9.1 (STATA Corp., College Station, TX: derivation and cross-validation of clinical scoring system for risk classification).

Table 1. Demographic and Clinical Features in Patients with NASH and in Patients Without NASH
FeatureALLNormal Liver/Simple SteatosisNASHP value
n = 200n = 136n = 64
  • *

    Race is self-reported and was available in 197 subjects. Abbreviation: IQR, interquartile range.

Sex: Female n(%)168 (84)118 (86.8)50 (78.1)0.12
Age: years, median (IQR)43 (37–51)42 (36–50)44 (38–54)0.096
Race*:    
 White, n(%)131 (66.5)84 (61.8)47 (73.4) 
 Black, n(%)27 (13.7)24 (17.6)3 (4.7)0.01 White versus black
 Non-black, n(%)170 (86.3)110 (80.9)60 (93.8)0.01 Non-black versus black
Body mass index: kg/m2, median (IQR)48 (43–55)48 (43–54)50 (44–57)0.462
Comorbidity prevalence:    
 Diabetes mellitus, n(%)52 (26)27 (19.9)25 (39.1)0.004
 Hypertension, n(%)112 (56)65 (47.8)47 (73.4)0.001
 Hyperlipidemia, n(%)100 (50)61 (44.9)39 (60.9)0.034
 Cholelithiasis, n(%)79 (39.5)52 (38.2)27 (42.2)0.594
 Arthritis or degenerative joint disease n(%)62 (31)44 (32.4)18 (28.1)0.546
 Obstructive sleep apnea n(%)27 (13.5)12 (8.8)15 (23.4)0.005
Table 2. Laboratory and Ultrasound Data in Patients with NASH and Patients Without NASH
ParameterAllNormal Liver/Simple SteatosisNASHP Value*P Value**
n = 200n = 136n = 64
Median (IQR)% AbnormalMedian (IQR)% AbnormalMedian (IQR)% Abnormal
  • *

    Comparison of distribution of continuous variables between groups Normal/Steatosis versus NASH/Cirrhosis (Mann-Whitney U test).

  • **

    Comparison of proportions between groups Normal/Steatosis versus NASH/Cirrhosis - Chi-squared Test.

Platelet count303 (268–349)1.5310 (271–353)1.5290 (257–334)1.60.031
Creatinine0.8 (0.7–0.9)5.60.8 (0.7–0.9)7.40.8 (0.7–0.9)1.60.110.2
Total bilirubin0.6 (0.5–0.8)2.10.6 (0.5–0.7)0.80.7 (0.6–0.9)5.10.020.9
AST22 (19–28)5.720 (18–25)3.028 (23–35)11.7<0.010.04
ALT24 (19–36)6.322 (18–30)4.634 (26–42)10<0.010.2
Alkaline phosphatase71 (58–86)6.468 (56–85)6.980 (65–90)5.30.021
Prothrombin time11.9 (11.6–12.4)1.611.9 (11.6–12.3)1.511.9 (11.5–12.5)1.60.91
Albumin3.9 (3.6–4.1)10.73.9 (3.6–4.1)8.63.9 (3.6–4.2)15.30.90.4
Triglycerides140 (102–178)38.6138 (102–167)35.6148 (106–195)45.80.30.5
Total cholesterol192 (174–224)45.8200 (179–226)50.8185 (158–214)33.30.10.2
HDL45 (40–53)20.546 (40–53)16.943 (36–52)29.20.30.2
LDL119 (99–143)41121 (104–146)45.8106 (92–140)29.20.10.2
Liver ultrasound 44.5 40.6 53.70.9 

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

Two hundred subjects were studied. The median age was 43 years (interquartile range, 37-51) and median BMI was 48 kg/m2 (interquartile range, 43-55); 168 (84%) were women. Data regarding race were available in 197 patients; 131 (66.5%) were Caucasian, 27 (13.7%) were African American, and 39 (19.8%) were “other.” Three patients declined to disclose their race. Histologic findings from liver biopsies were available for all 200 patients. The mean biopsy length was 11.7 mm (range, 4-31 mm), and the mean number of portal tracts per liver biopsy was 7.8 (range, 6-30). Sixty-four subjects (32%) had NASH, 61 (30.5%) had simple steatosis, and 75 subjects (37.5%) had normal histologic findings. Three of the 64 patients of the NASH group had the diagnosis of NASH-related cirrhosis. Bariatric surgery was completed in two patients with this diagnosis and deferred in one.

Univariate Analysis

Univariate analysis showed that the proportion of patients with a diagnosis of type 2 diabetes, hypertension, hyperlipidemia, or obstructive sleep apnea was higher in patients with NASH than in patients without NASH (Table 1). Additionally, the proportion of white and Hispanic patients with NASH was significantly higher than that in blacks. There were no significant differences between the groups in sex distribution, age, BMI, or other comorbidities.

A higher proportion of patients with NASH had abnormal AST values (11.7% versus 3% in non-NASH patients) based on the standard laboratory cutoff values. Patients with NASH had higher total bilirubin, AST, ALT and alkaline phosphatase values, and lower platelet counts (Table 2).

Identification of Optimal Cutoff Points for AST and ALT

The cutoff point selected by the tree algorithm for both AST and ALT was 27 U/L. Thirty-four of 58 patients (58%) with an AST ≥ 27 U/L had NASH, as compared with 26 of 134 (19%) with an AST below this cutoff point; similarly, 45 of 83 of patients (54%) with ALT ≥ 27 U/L had NASH, as compared with 15 of 108 (14%) with ALT < 27 U/L.

Multivariate Analysis and NASH Clinical Scoring System for Morbid Obesity

A total of 186 patients, including 58 (31.2%) with NASH, were used for the multivariate analysis and creation of the Scoring System. Fourteen patients were omitted from the analysis because one or more of the independent predictive variables was not available. Multivariate analysis identified six independent predictive factors for the presence of NASH in the morbidly obese: hypertension, diabetes, AST ≥ 27 IU/L, ALT ≥ 27 IU/L, obstructive sleep apnea, and non-black race (Table 3). These variables were then used to develop a NASH Clinical Scoring System for Morbid Obesity (Table 4), with each predictive factor assigned a score equal to the integer nearest its regression coefficient (Tables 3 and 4). Table 5 shows the naïve estimate of the performance of the Scoring System, in which the data for the full study population was used both to derive and evaluate the prediction rule. The prevalence of NASH in each risk level is interpretable as the positive predictive value (PPV) of that risk classification, whereas its complement represents the negative predictive value (NPV). Thus, both the PPV of the very high-risk classification and the NPV of the low-risk classification are 93%. The performance estimate based on cross-validation is shown in Table 5; as compared with the classification using the full-study population (Table 4), the PPV of the cross-validated very high-risk classification is lower than the PPV of the full-cohort very high-risk classification (93% versus 80%, respectively). Similarly, the NPV of the cross-validated low-risk classification is lower than the NPV of the full-cohort low-risk classification (87% versus 93%, respectively). Table 6 illustrates that for 85% of patients (159/186), the NASH risk classification based on the full cohort is concordant with the NASH risk classification based on cross-validation. Of the 27 patients for whom the NASH risk classification was discordant, the discrepancy spans only a single NASH risk class in 26, with the majority (17/27, 63%) reclassified from intermediate to low risk in the cross-validation. The one patient moving more than a single category was reclassified from intermediate to very high risk. The area under the receiver operating characteristic curve for the risk classification was 0.80 based on the complete data and 0.75 based on the cross-validation.

Table 3. NASH Clinical Scoring System for Morbid Obesity, Part 1. Results of Multivariate Analysis: Independent Predictors of NASH and Assigned Score Values
Independent PredictorOdds Ratio (95% CI)P ValueRegression CoefficientScore Value
Hypertension2.4 (1.0–5.6)0.0380.8851
Type 2 diabetes2.6 (1.1–6.3)0.0290.9731
AST ≥ 27 IU/L2.9 (1.2–7.0)0.0211.0521
ALT ≥ 27 IU/L3.3 (1.4–8.0)0.0081.1901
Sleep apnea4.0 (1.3–12.2)0.0141.3931
Non-black8.4 (1.9–37.1)0.0052.1262
Table 4. NASH Clinical Scoring System for Morbid Obesity, Part 2. Prevalence of NASH in Study Population According to Summed Scoring System Points
Summed PointsNASH Risk Classificationn/N*Prevalence of NASH (95% CI)
  1. n = number of patients with NASH; N = total number of patients in each risk classification. A total of 186 patients were used for the multivariate analysis and creation of the Scoring System, because in 14 patients, one or more of the independent predictive variables was not available.

0–2Low4/597% (2–16%)
3–4Intermediate22/8127% (18–38%)
5High19/3259% (41–76%)
6–7Very high13/1493% (66–100%)
Table 5. NASH Clinical Scoring System for Morbid Obesity, Part 3. Prevalence of NASH by Cross-Validation Risk Classification
Summed PointsNASH Risk Classificationn/N*Prevalence of NASH (95% CI)
  1. Abbreviations: n = number of patients with NASH; N = total number of patients in each risk classification.

0–2Low10/7613% (6–23%)
3–4Intermediate20/6829% (19–42%)
5High16/2759% (39–78%)
6–7Very high12/1580% (52–96%)
Table 6. NASH Clinical Scoring System for Morbid Obesity, Part 4. Comparison of Full-Cohort and Cross-Validated Risk Classifications
Full-Cohort Risk ClassificationCross-Validation Risk Classification
LowIntermediateHighVery high
Low59000
Intermediate176211
High06251
Very high00113

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References

The NASH Clinical Scoring System for Morbid Obesity derived in our study is based on data from patients who underwent bariatric surgery, which means it is limited to patients who have a BMI greater than 40 or those with a BMI greater than 35 together with obesity-associated diseases. However, the increasing incidence of morbid obesity in the United States, with an estimated prevalence of 14 million individuals, calls for the development of diagnostic and treatment strategies specific for this growing population. The goal of the NASH Clinical Scoring System for Morbid Obesity is to assist the clinician in predicting the risk for an individual morbidly obese patient of having NASH, and therefore whether a liver biopsy is more likely to yield significant histopathologic findings that will be of clinical and prognostic importance. We suggest a liver biopsy be pursued in patients at high or very high risk for NASH as defined by the NASH Clinical Scoring System. We found cross-validated PPVs of 80% and 59% for patients who fall into the very high and high NASH risk classifications, respectively. Conversely, cross-validated NPVs were 87% and 71% for patients in the low and intermediate NASH risk classifications, respectively. Thus, patients with scores in the high or very high risk classification range would likely benefit from undergoing liver biopsy. For patients with scores that place them in the low or intermediate NASH risk classifications, a physician and the patient may choose not to undertake liver biopsy, therefore avoiding its associated risks and costs, opting instead for clinical follow-up. The NASH Clinical Scoring System in Morbid Obesity can be used in the clinic setting with information that is easily obtained in the standard evaluation of morbidly obese patients.

The NASH Clinical Scoring System included the independent predictors for NASH selected during the multivariate analysis. The results of the multivariate analysis corroborate previous studies in showing a strong correlation between the presence of other systemic manifestations of severe obesity and metabolic syndrome, such as hypertension and type 2 diabetes, and NASH11, 21, 32 and also highlight the usefulness of the concept for a lower limit for AST and ALT to differentiate patients at higher risk for NASH. Hypertension and type 2 diabetes are central components of the metabolic syndrome and have previously been associated with a higher prevalence of NASH in obese and nonobese populations.9, 12, 17, 33–40 It is therefore not surprising to find that these two comorbidities were significant predictors of NASH in our study. The diagnosis of obstructive sleep apnea (OSA) was another independent predictor identified. OSA has recently been linked to a higher prevalence of diabetes, but it is still unknown whether this is solely related to a higher prevalence of OSA in morbidly obese patients with truncal obesity, and therefore a chance association, or whether there is a causal association related to increased nocturnal sympathetic activity and high serum glucagon and corticosteroid levels and increased urinary catecholamines, all of which may be causally related to increased insulin resistance and are found in OSA patients.41–43 Regardless of the reason, OSA is a key risk factor in our scoring model.

In our study, the proportion of white and Hispanic patients with NASH was higher than that of blacks (35% versus 11%; P = 0.01). Ethnic differences in the prevalence of NASH and cryptogenic cirrhosis have been reported before, with blacks having lower prevalence.44, 45 This lower prevalence may suggest a protective effect of black race/ethnicity regarding the development of obesity-associated liver injury, even when other systemic manifestations are present. Alternatively, it may represent variation in referral patterns or genetic differences in body fat distribution or metabolic thermogenesis.36, 46–49 In a subset analysis of our population comparing blacks with whites and Hispanics combined, there were no significant differences in BMI, prevalence of metabolic syndrome, hypertension, or hyperlipidemia. However, nearly twice as many blacks in our study had a diagnosis of type 2 diabetes (44% versus 23% for whites and Hispanics combined; P = 0.03). Taken together, these results support the possibility of an independent “protective” effect of being black in terms of risk of developing NASH.

Among patients with NAFLD, the prevalence of hyperlipidemia is known to vary between 20% and 92%.2, 20 However, in our study, hyperlipidemia, despite having a higher prevalence at the univariate level in patients with NASH, was not selected as an independent predictor on multivariate analysis.

Even though our study was limited to morbidly obese patients and a potential influence of BMI when compared with non–morbidly obese individuals could not be determined, an increasing BMI, per se, was not a risk factor for NASH at the univariate or multivariate level, which suggests that factors other than the total mass of adipose tissue are more relevant in the development of obesity-related complications in this patient population. Higher amounts of the more metabolically active visceral fat deposits (truncal obesity) compared with peripheral fat deposits may be more likely to be associated with systemic complications of obesity. One reason is the important difference between visceral fat and other fat stores regarding the spectrum of adipokines produced and its impact on the individual's general health.50 Although we did not obtain physical surrogate measures of truncal obesity such as abdominal girth or computed abdominal tomography scan, we did collect visceral and peripheral fat samples, and we are currently studying biomarkers in those samples to learn more about the intricate and dysfunctional cross-talk between adipose tissue and liver that is an important factor in the pathophysiology of NAFLD and NASH. Current evidence shows that the increase in fat mass and adipocyte differentiation lead to the production of a variety of cytokines, including leptin, resistin, angiotensinogen, tumor necrosis factor alpha, and free fatty acids, that have a direct impact on hepatocyte function and the development of fibrosis.2, 51–55

This study is one of the largest single-center studies currently available. The outcomes were based on histopathologic examination of liver biopsy tissue, and the NASH scoring system in morbid obesity was validated within the study population. Nevertheless, the Scoring System should be validated in a larger, independent dataset, because our study population comes from a tertiary care medical center, and its demographics and clinical characteristics may differ from those of other centers. Other potential limitations of this study include our inability by design to assess the impact of medications on the severity of NASH. Although this may be of interest for future investigations of NASH diagnostic models, our patients' medical comorbidities and medication regimens are representative of the clinical profile of the majority of morbidly obese patients; thus, our results are relevant and informative. Finally, we acknowledge that the diagnosis of NASH requires a thorough search for, and exclusion of, other causes of chronic liver disease, most notably viral hepatitis and autoimmune hepatitis. With respect to the latter diagnosis, despite the positive autoimmune markers that may occur in patients with NASH, the inflammatory and fibrosis characteristics seen in the histologic findings and interpreted by an experienced liver pathologist are usually sufficient to either make the diagnosis of autoimmune hepatitis or to reasonably exclude it.

In conclusion, the proposed NASH Scoring System in Morbid Obesity can predict presence or absence of NASH in morbidly obese patients with sufficient accuracy to be considered clinically useful. Our cross-validation suggests that it can both identify morbidly obese individuals at high and very high risk for NASH in whom liver biopsy would most likely to be of diagnostic and prognostic utility, and individuals with a low risk of NASH in whom a liver biopsy can be delayed or avoided. The potential risk reductions and cost savings from avoiding unnecessary biopsies in low-risk and intermediate-risk patients would come at relatively little increased risk for missing a patient with NASH, provided they remain in clinical follow-up. Use of a diagnostic tool, such as the NASH Scoring System in Morbid Obesity, to guide decisions regarding liver biopsy might detect more patients with NASH who are currently going undiagnosed, therefore assisting both the clinician and the patient in choosing and tailoring treatment for morbid obesity and its associated complications.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. References
  • 1
    Kowdley KV, Caldwell S. Nonalcoholic steatohepatitis: a twenty-first century epidemic? J Clin Gastroenterol 2006; 40(Suppl 1): S2S4.
  • 2
    McCullough AJ. Pathophysiology of nonalcoholic steatohepatitis. J Clin Gastroenterol 2006; 40(Suppl 1): S17S29.
  • 3
    Abrams GA, Kunde SS, Lazenby AJ, Clements RH. Portal fibrosis and hepatic steatosis in morbidly obese subjects: a spectrum of nonalcoholic fatty liver disease. HEPATOLOGY 2004; 40: 475483.
  • 4
    Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999; 116: 14131419.
  • 5
    Liou I, Kowdley KV. Natural history of nonalcoholic steatohepatitis. J Clin Gastroenterol 2006; 40(Suppl 1): S11S16.
  • 6
    Maheshwari A, Thuluvath PJ. Cryptogenic cirrhosis and NAFLD: are they related? Am J Gastroenterol 2006; 101: 664668.
    Direct Link:
  • 7
    Caldwell SH, Crespo DM. The spectrum expanded: cryptogenic cirrhosis and the natural history of non-alcoholic fatty liver disease. J Hepatol 2004; 40: 578584.
  • 8
    Hui JM, Kench JG, Chitturi S, Sud A, Farrell GC, Byth K, et al. Long-term outcomes of cirrhosis in nonalcoholic steatohepatitis compared with hepatitis C. HEPATOLOGY 2003; 38: 420427.
  • 9
    McCullough AJ. The clinical features, diagnosis and natural history of nonalcoholic fatty liver disease. Clin Liver Dis 2004; 8: 521533, viii.
  • 10
    Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002; 346: 12211231.
  • 11
    Clark JM. The epidemiology of nonalcoholic Fatty liver disease in adults. J Clin Gastroenterol 2006; 40(Suppl 1): S5S10.
  • 12
    Ong JP, Elariny H, Collantes R, Younoszai A, Chandhoke V, Reines HD, et al. Predictors of nonalcoholic steatohepatitis and advanced fibrosis in morbidly obese patients. Obes Surg 2005; 15: 310315.
  • 13
    Bugianesi E, McCullough AJ, Marchesini G. Insulin resistance: a metabolic pathway to chronic liver disease. HEPATOLOGY 2005; 42: 9871000.
  • 14
    Chitturi S, Farrell GC. Etiopathogenesis of nonalcoholic steatohepatitis. Semin Liver Dis 2001; 21: 2741.
  • 15
    Neuschwander-Tetri BA. Evolving pathophysiologic concepts in nonalcoholic steatohepatitis. Curr Gastroenterol Rep 2002; 4: 3136.
  • 16
    Barker KB, Palekar NA, Bowers SP, Goldberg JE, Pulcini JP, Harrison SA. Non-alcoholic steatohepatitis: effect of Roux-en-Y gastric bypass surgery. Am J Gastroenterol 2006; 101: 368373.
    Direct Link:
  • 17
    Beymer C, Kowdley KV, Larson A, Edmonson P, Dellinger EP, Flum DR. Prevalence and predictors of asymptomatic liver disease in patients undergoing gastric bypass surgery. Arch Surg 2003; 138: 12401244.
  • 18
    Spaulding L, Trainer T, Janiec D. Prevalence of non-alcoholic steatohepatitis in morbidly obese subjects undergoing gastric bypass. Obes Surg 2003; 13: 347349.
  • 19
    Marchesini G, Bugianesi E, Forlani G, Cerrelli F, Lenzi M, Manini R, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. HEPATOLOGY 2003; 37: 917923.
  • 20
    Adams LA, Talwalkar JA. Diagnostic evaluation of nonalcoholic fatty liver disease. J Clin Gastroenterol 2006; 40(Suppl 1): S34S38.
  • 21
    Farrell GC, Larter CZ. Nonalcoholic fatty liver disease: from steatosis to cirrhosis. HEPATOLOGY 2006; 43: S99S112.
  • 22
    Janiec DJ, Jacobson ER, Freeth A, Spaulding L, Blaszyk H. Histologic variation of grade and stage of non-alcoholic fatty liver disease in liver biopsies. Obes Surg 2005; 15: 497501.
  • 23
    Maciel AC, Marchiori E, de Barros SG, Cerski CT, Tarasconi DP, Ilha Dde O. Transjugular liver biopsy: histological diagnosis success comparing the trucut to the modified aspiration Ross needle. Arq Gastroenterol 2003; 40: 8084.
  • 24
    Flegal KM. Epidemiologic aspects of overweight and obesity in the United States. Physiol Behav 2005; 86: 599602.
  • 25
    Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999-2000. JAMA 2002; 288: 17231727.
  • 26
    Hedley AA, Ogden CL, Johnson CL, Carroll MD, Curtin LR, Flegal KM. Prevalence of overweight and obesity among US children, adolescents, and adults, 1999-2002. JAMA 2004; 291: 28472850.
  • 27
    Day CP. Non-alcoholic fatty liver disease: current concepts and management strategies. Clin Med 2006; 6: 1925.
  • 28
    Duman DG, Celikel C, Tuney D, Imeryuz N, Avsar E, Tozun N. Computed tomography in nonalcoholic fatty liver disease: a useful tool for hepatosteatosis assessment? Dig Dis Sci 2006; 51: 346351.
  • 29
    Sorrentino P, Tarantino G, Perrella A, Micheli P, Perrella O, Conca P. A clinical-morphological study on cholestatic presentation of nonalcoholic fatty liver disease. Dig Dis Sci 2005; 50: 11301135.
  • 30
    Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW, et al. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. HEPATOLOGY 2005; 41: 13131321.
  • 31
    Efron B. The Jackknife, the Bootstrap and Other Resampling Plans. SIAM (Society for Industrial and Applied Mathematics): Philadelphia, PA; 1982.
  • 32
    Riley TR, 3rd, Kahn A. Risk factors and ultrasound can predict chronic hepatitis caused by nonalcoholic fatty liver disease. Dig Dis Sci 2006; 51: 4144.
  • 33
    Amarapurka DN, Amarapurkar AD, Patel ND, Agal S, Baigal R, Gupte P, et al. Nonalcoholic steatohepatitis (NASH) with diabetes: predictors of liver fibrosis. Ann Hepatol 2006; 5: 3033.
  • 34
    Harnois F, Msika S, Sabate JM, Mechler C, Jouet P, Barge J, et al. Prevalence and predictive factors of non-alcoholic steatohepatitis (NASH) in morbidly obese patients undergoing bariatric surgery. Obes Surg 2006; 16: 183188.
  • 35
    Miele L, Forgione A, Hernandez AP, Gabrieli ML, Vero V, Di Rocco P, et al. The natural history and risk factors for progression of non-alcoholic fatty liver disease and steatohepatitis. Eur Rev Med Pharmacol Sci 2005; 9: 273277.
  • 36
    Agarwal N, Sharma BC. Insulin resistance and clinical aspects of non-alcoholic steatohepatitis (NASH). Hepatol Res 2005; 33: 9296.
  • 37
    Boza C, Riquelme A, Ibanez L, Duarte I, Norero E, Viviani P, et al. Predictors of nonalcoholic steatohepatitis (NASH) in obese patients undergoing gastric bypass. Obes Surg 2005; 15: 11481153.
  • 38
    Friis-Liby I, Aldenborg F, Jerlstad P, Rundstrom K, Bjornsson E. High prevalence of metabolic complications in patients with non-alcoholic fatty liver disease. Scand J Gastroenterol 2004; 39: 864869.
  • 39
    Brunt EM, Neuschwander-Tetri BA, Oliver D, Wehmeier KR, Bacon BR. Nonalcoholic steatohepatitis: histologic features and clinical correlations with 30 blinded biopsy specimens. Hum Pathol 2004; 35: 10701082.
  • 40
    Duseja A, Nanda M, Das A, Das R, Bhansali A, Chawla Y. Prevalence of obesity, diabetes mellitus and hyperlipidaemia in patients with cryptogenic liver cirrhosis. Trop Gastroenterol 2004; 25: 1517.
  • 41
    Singh H, Pollock R, Uhanova J, Kryger M, Hawkins K, Minuk GY. Symptoms of obstructive sleep apnea in patients with nonalcoholic fatty liver disease. Dig Dis Sci 2005; 50: 23382343.
  • 42
    Jouet P, Sabate JM, Maillard D, Msika S, Mechler C, Ledoux S, et al. Relationship between obstructive sleep apnea and liver abnormalities in morbidly obese patients: a prospective study. Obes Surg 2007; 17: 478485.
  • 43
    Savransky V, Nanayakkara A, Vivero A, Li J, Bevans S, Smith PL, et al. Chronic intermittent hypoxia predisposes to liver injury. HEPATOLOGY 2007; 45: 10071013.
  • 44
    Browning JD, Szczepaniak LS, Dobbins R, Nuremberg P, Horton JD, Cohen JC, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. HEPATOLOGY 2004; 40: 13871395.
  • 45
    Weston SR, Leyden W, Murphy R, Bass NM, Bell BP, Manos MM, et al. Racial and ethnic distribution of nonalcoholic fatty liver in persons with newly diagnosed chronic liver disease. HEPATOLOGY 2005; 41: 372379.
  • 46
    Caldwell SH, Harris DM, Patrie JT, Hespenheide EE. Is NASH underdiagnosed among African Americans? Am J Gastroenterol 2002; 97: 14961500.
    Direct Link:
  • 47
    Neuschwander-Tetri BA, Caldwell SH. Nonalcoholic steatohepatitis: summary of an AASLD Single Topic Conference. HEPATOLOGY 2003; 37: 12021219.
  • 48
    Browning JD, Kumar KS, Saboorian MH, Thiele DL. Ethnic differences in the prevalence of cryptogenic cirrhosis. Am J Gastroenterol 2004; 99: 292298.
    Direct Link:
  • 49
    Merriman RB, Aouizerat BE, Bass NM. Genetic influences in nonalcoholic fatty liver disease. J Clin Gastroenterol 2006; 40(Suppl 1): S30S33.
  • 50
    Rocha R, Cotrim HP, Carvalho FM, Siqueira AC, Braga H, Freitas LA. Body mass index and waist circumference in non-alcoholic fatty liver disease. J Hum Nutr Diet 2005; 18: 365370.
  • 51
    McTernan PG, McTernan CL, Chetty R, Jenner K, Fisher FM, Lauer MN, et al. Increased resistin gene and protein expression in human abdominal adipose tissue. J Clin Endocrinol Metab 2002; 87: 2407.
  • 52
    McTernan CL, McTernan PG, Harte AL, Levick PL, Barnett AH, Kumar S. Resistin, central obesity, and type 2 diabetes. Lancet 2002; 359: 4647.
  • 53
    Tilg H, Diehl AM. Cytokines in alcoholic and nonalcoholic steatohepatitis. N Engl J Med 2000; 343: 14671476.
  • 54
    Chalasani N, Crabb DW, Cummings OW, Kwo PY, Asghar A, Pandya PK, et al. Does leptin play a role in the pathogenesis of human nonalcoholic steatohepatitis? Am J Gastroenterol 2003; 98: 27712776.
    Direct Link:
  • 55
    Scheen AJ. Pathophysiology of type 2 diabetes. Acta Clin Belg 2003; 58: 335341.