11/12/2005  2pm to 8pm 


Abstract ID: 64659

Category: IO2: Liver Imaging Modalities

Validation of an MRI Technique to Measure Liver Fat Content.

S. J. Cotler, University of Illinois at Chicago, Chicago, IL, G. Guzman, University of

Illinois at Chicago, Chicago, IL, J. E. Layden-Almer, University of Illinois at Chicago,

Chicago, IL, J. Berkes, University of Illinois at Chicago, Chicago, IL, K. Schwartz,

University of Illinois at Chicago, Chicago, IL, T. Mazzone, University of Illinois at

Chicago, Chicago, IL, X. Zhou, University of Illinois at Chicago, Chicago, IL




Magnetic resonance imaging (MRI) has the potential to provide an accurate, noninvasive method to quantify liver fat, although MRI techniques have not been systematically validated using liver tissue with known fat content. The aim of this study was to validate an MRI technique for measuring liver fat.



A water phantom and liver phantoms were scanned before scanning each of 10 human subjects to generate a standard curve.

Patient Characteristics:


50 ± 11

Gender (F/M)



86.2 ± 11.7

BMI (Kg/m2)

31.9 ± 3.9


The liver phantoms consisted of calf liver homogenized with 0%, 5%, 10%, 20% and 30% corn oil added by weight. The MRI protocol used a 3.0 Tesla magnet and included fat imaging using a radio-frequency pulse to saturate the water signal and water and fat imaging without using any saturation pulses. A fast gradient echo pulse sequence was used to increase the data acquisition efficiency. The signal intensity from the water phantom was used as a normalization factor to account for the signal variations between different study dates, while the signal from the liver phantoms served as a calibration reference to derive the fat content from the patients. For human subjects, two regions-of-interest (ROI) were selected in the right lobe of the liver for calculation of liver fat content and the results were averaged. Each of the 10 human subjects underwent a liver biopsy. Image analysis was performed on histologic sections. A computer program was developed to fit ellipses to fat globules and to calculate the percent of the area comprised of pixels consistent with fat in a histologic section. A liver pathologist estimated fat content based on the appearance of the histologic section used for image analysis. Pearson correlation coefficients were used to assess the relationships between MRI and image analysis measurements and MRI and the pathologist’s estimate of liver fat content.



MRI measurements of the liver phantoms were highly reproducible with

standard deviations of 0.38% (0% fat), 0.88% (5% fat), 2.12% (10% fat), 2.97% (20% fat), 5.36% (30% fat), and 0.32% (100% fat), and were directly related to fat concentration. MRI measurements of liver fat content in human subjects were strongly correlated with image analysis of histologic sections (r=0.96, p<0.001) and with a pathologist’s estimate of liver fat (r=0.93, p<0.001). The pathologist’s estimates of liver fat content tended to be higher than measurements made by MRI (median 8.5%, range - 4% to 25%).



·       The MRI technique validated in this study provides a non-invasive means to measure liver fat and could be used to follow changes in fat content over time in response to a therapeutic intervention.


·       The technique was confirmed over a range of fat content in human subjects and is sufficiently sensitive to distinguish patients with NAFLD

·       Standard light microscopy provided higher estimates of liver fat content than measurements made by MRI or image analysis. 


Abstract ID: 67487

Category: IO2: Liver Imaging Modalities

Amount and type of liver fibrosis detected by stiffness measurements.

M. Ziol, Jean Verdier Hospital, AP-HP, Bondy, France, N. Barget, Jean Verdier

Hospital, Pathology Department, Bondy, France, V. Bourcier, Jean Verdier Hospital,

Liver Unit, Bondy, France, T. Garnier, Jean Verdier Hospital, Liver Unit, Bondy,

France, M. Beaugrand, Jean Verdier hospital, Liver Unit, Bondy, France




Liver stiffness measurement (LSM) is a new, non invasive method that has been demonstrated to detect significant liver fibrosis in patients with chronic hepatitis C.



We aimed to investigate in a large cohort of patients with various chronic and acute liver diseases the relation between LSM using elastography and collagen deposition quantified on liver biopsies.



Two hundred and seventy patients referred to our institution for liver biopsy whatever the suspected cause of the disease were included. LSM was performed in all patients within less than 3 months before or after biopsy. To avoid sampling error, less than 20mm long liver biopsies were excluded. We quantified collagen deposition using computer assisted image analysis on 5 micrometers thick picrosirius red stained sections. The fibrosis area fraction (FAF) was assessed using interactive thresholding of greyscale converted digital images of 10 consecutive fields (X100) for each liver biopsy. Taking in account clinical data and histological analysis, patient’s liver biopsies were classified as follows: no architectural abnormalities, chronic hepatitis (portal fibrosis), biliary type fibrosis, alcoholic liver disease or non alcoholic steatohepatitis and abnormalities of liver-cell plates without fibrosis.

Both FAF (%) and LSM (kilopascals) were available in 211 patients. LSM values ranged from 2.8 to 75.4 kPa (mean +/-SD:9+/-13) and fraction of fibrosis area from 0.13 to 49% (3.3+/-7). Fraction of fibrosis area was significantly correlated to LSM (Spearman correlation coefficient Rho=0.608; p<0.0001).



All except one patient in the group without architectural abnormalities (acute hepatitis, granulomas, steatosis without fibrosis; n=24) had low LSM values (5.9+/-3.2kPa). When patients were divided into groups according to the etiology and related architectural abnormalities, significant correlation was found between LSM and FAF in chronic hepatitis (n=121, Rho=0.66, p<0.0001) and alcoholic liver disease or non-alcoholic steatohepatitis (NASH) (n=37, Rho=0.64, p=0.0001). In patients with abnormalities of liver cell plates without fibrosis on liver biopsies (nodular regenerative hyperplasia, incomplete or macronodular cirrhosis; n=25), LSM values were

highly scattered, varying from 3 to 46kPa.



Using both quantitative morphometric evaluation of liver fibrosis and qualitative assessment of architectural abnormalities as a gold standard, our results suggest that LSM accurately reflects liver fibrosis observed in chronic hepatitis and alcoholic liver disease or NASH.
Abstract ID: 63689

Category: IO2: Liver Imaging Modalities

Assessment of Hepatic Fibrosis in Chronic Liver Disease — Usefulness of the StrainRate Imaging Method

N. Hotta, Aichi Medical University, Nagakute, Japan, A. Okumura, Aichi Medical

University, Nagakute, Japan, T. Ishikawa, 21 Karimata, Nagakute, Japan, S. Kakumu,

Aichi Medical University, Nagakute, Japan




The development of hepatic fibrosis in patients with chronic liver disease

increases the risk of liver cancer. Assessing the degree of hepatic fibrosis is therefore one of the most important factors in therapeutic planning. Conventionally, liver biopsy is performed to assess hepatic fibrosis, but this method is associated with complications such as hemorrhage. The present study was conducted to determine whether an easily performed myocardial examination technique can be applied to the assessment of hepatic fibrosis. Strain Rate Imaging is a new method based on Tissue Doppler Imaging (TDI). This method is useful for eliminating the effects of translational motion or tethering of the myocardium and for evaluating the local contraction and expansion of the myocardium. The strain value is calculated by dividing the change in the length of an object before and after motion by the length before motion. The usefulness of Strain Rate Imaging in assessing the degree of hepatic fibrosis was evaluated.



Strain Rate Imaging was performed using a diagnostic ultrasound system

(AplioTM, Toshiba Medical Systems Corporation, Tochigi, Japan) in a total of 30 subjects: 17 in the chronic hepatitis group, 8 in the cirrhosis group, and 5 in the normal control group. Dynamic images were acquired as raw data in Harmonic TDI mode, in which the harmonic components are used for data transmission/reception. Similar to the Tissue Harmonic method used in B-mode imaging, Harmonic TDI improves the accuracy of TDI velocity values and suppresses artifacts. TDI-Q, the Tissue Doppler analysis software installed in the Aplio system, was used for analysis. Measurement was

performed five times from the epigastrium, with the ROI size set to 10 mm and the derivative pitch to 3 mm. The mean of the five measurements was used as the strain value.



The mean strain value was 0.204 in the chronic hepatitis group, 0.078 in the

cirrhosis group, and 0.32 in the normal control group.



The results of the present study suggest that this noninvasive method permits

quantitative assessment of the degree of hepatic fibrosis to be performed easily and in a short time. It is expected that the accuracy of the Strain Rate Imaging method in determining the degree of hepatic fibrosis will be improved when it is used in combination with histological examination.

Abstract ID: 65302

Category: IO2: Liver Imaging Modalities

Prospective randomised comparison of mini-laparoscopy and percutaneous liver biopsy. Analysis of 1000 examinations. Diagnosis of cirrhosis and complications.

U. Denzer, Medical department I, university medical center Hamburg Eppendorf,

Hamburg, Germany, A. Arnoldy, I Medical department Johannes Gutenberg University

Mainz, Germany, Mainz, Germany, Germany, U. Weickert, Medical department C,

Medical center Ludwigshafen, Germany, Ludwigshafen, Germany, Germany, S.

Kanzler, I Medical Department, University Mainz, Germany, Mainz,Germany, Germany,

P. R. Galle, I Medical Department, University Mainz, Germany, Mainz, Germany, J. F.

Riemann, medical Department C, Medical Center Ludwigshafen, Germany,

Ludwigshafen, Germany, H. P. Dienes, Institution for Pathology, University Medical

Center Köln, Germany, Koeln, Germany, Germany, A. W. Lohse, Medical center I,

University Medical Center Hamburg Eppendorf, Hamburg, Germany, Germany




Percutaneous liver biopsy is the gold standard for gaining liver histology in

chronic liver disease (pLB). Nevertheless false negative biopsy rates in cirrhosis range from 10 - 50 %. Diagnostic mini-laparoscopy (optic 1.9 mm) (ML) is an alternative procedure with the advantage of additional macroscopic evaluation. This study compared both methods prospectively and randomized.



Patients gave written informed consent. Exclusion of severe coagulation

disorders. Liver biopsy: Vim Silverman needle (ML); Menghini needle (pLB) (biopsy diameter 1.4 mm). Macroscopic liver evaluation documented as normal, fibrosis/irregular cirrhosis and cirrhosis. Formalin fixed paraffin embedded sections were stained with H&E and IVG. The slides were blindly coded and graded (HAI score). Controls day 1 after biopsy: Blood count and ultrasound.



Patients: 1000 patients were randomized (9/00 - 7/04), study drop out: 75. Biopsy groups showed comparable demographic and clinical characteristics of patients. Indication for liver biopsy: cHCV: 48,8 % (ML) 41.8 % (pLB); Hepatopathy of unknown origin: 29 % 33,3 %; cHBV: 10 %, 29,8 %; AIH / PBC/ PSC: 11,4 %, 10,7%; others: 1,3 %, 0,6 %.  Diagnosis of cirrhosis: Criteria for cirrhosis: PLB: Modified Ishaak score 5 or 6; ML: Combination of macroscopic (irregular cirrhosis scored as fibrosis) and histologic scoring. Available for blinded histologic scoring pLB in 410 cases, ML in 401 cases. No significant difference in length of biopsy specimen, marginally difference in count of portal fields per specimen. Comparison of blindly coded liver histology resulted in no significant differences in histologic grading (No liver fibrosis: score 0 vs. liver fibrosis: score 1-4 vs. liver cirrhosis: score 5-6; p = 0.34;). Histologic grading revealed the diagnosis liver cirrhosis in pLB in 23.0 % (n = 94) compared to ML in 25.3 % (n = 103). Comparison of histologic staging in pLB and combined macroscopic /histologic staging in ML resulted in a significant higher rate of 10 % more frequently diagnosed liver cirrhoses with ML procedure. pLB diagnosed liver cirrhosis in 22.9 % (n = 94) compared to ML in 33.2 % (n = 133) (p = 0.001). Complications: PLB: Severe (requiring hospital stay longer than one day): 0.9 % (Hemobilia: 1; post biopsy bleeding: 3). ML: 0.2 % (Hemobilia: 1) p = 0.188 ns. (table 3). Ultrasound control day 1: Intraabdominal fluid (3.4 % vs. 1.4 %, p = 0.068) and intrahepatic hematoma (2.0 % vs. 2.1 %, p = 1.00).



Mini-laparoscopic evaluation diagnoses of about 10 % more liver cirrhoses.

Mini-laparoscopic liver biopsy is shown to be safe and superior in staging of liver diseases.


Abstract ID: 65144

Category: IO2: Liver Imaging Modalities

Prevalence and factors associated with failure of liver stiffness measurement using FibroScan® in a prospective study of 2114 examinations.

J. Foucher, Hopital haut-Leveque, Pessac, France, L. Castera, Hopital Haut-Leveque,

Pessac, France, D. Laharie, Hopital Haut Leveque, Pessac, France, N. Le Provost,

Hopital Haut Leveque, Pessac, France, P. Couzigou, Hopital Haut Leveque, Pessac,

France, P. Bernard, Hopital Saint-Andre, Bordeaux, France, X. Adhoute, Hopital Haut

Leveque, Pessac, France, V. de Ledinghen, Hopital Haut-Leveque, Pessac, France



Background & aims:

Recently, we have shown that liver stiffness measurement (FibroScan®, Echosens, France) was a reliable non-invasive method for the assessment

of liver fibrosis. However, in around 5% of cases, no value could be obtained. The aim of this prospective study was to assess the prevalence and factors associated with failure of liver stiffness measurement.



All patients with chronic liver disease and liver stiffness measurement between May 2003 and April 2005 were included. Failure of liver stiffness measurement was defined as no value obtained after 10 measurements. Clinical and biological factors associated with failure were analyzed using Chi-square, t-Student test, and logistic regression.



2114 liver stiffness measurements were analyzed. Characteristics of patients

were: 1168 males, mean age 52 ± 13 years, BMI 24.4 ± 4.4 (range: 12-49). Indications for FibroScan® were HCV (55.5%), HBV (5.5%), alcohol (9%), NASH (3.6%), hemochromatosis (2.4%), miscellaneous (24%). Median value of FS was 6.8 kPa (range: 1.5 – 75 kPa). Failure was observed in 4.5% of cases. By univariate analysis, factors associated with failure were BMI > 28 (OR 9.1, 95%CI 5.8-14.0, p<0.001), diabetes (OR 2.1, 95%CI 1.2-3.7, p=0.01), age > 50 years (OR 4.0, 95%CI 2.7-6.3, p<0.001), NASH (OR 3.4, 95%CI 1.7-6.9, p=0.001), and gGT level (OR 2.0, 95%CI 1.3-3.1, p=0.002).

Failure was not associated with operator, gender, or transaminases level. By multivariate analysis, the only factor aasociated with failure was BMI > 28 (OR 10.0, 95%CI 5.7- 17.9, p=0.001).



·       Liver stiffness measurement was successful in 95.5% of cases

·       The only factor associated with failure is obesity.  However, FibroScan® is feasible even in severely obese patients (BMI=49)

·       Other non-invasive methods of liver biopsy could be used only in patients with failure of FibronScan® examination. 


Abstract ID: 62900

Category: IO2: Liver Imaging Modalities

Non-invasive detection of cellular proliferation in the woodchuck model of hepatocellular carcinoma.

E. McKenzie, National Research Council of Canada - Institute for Biodiagnostics,

Winnipeg, Canada, J. Sun, National Research Council - Institute for Biodiagnostics,

Winnipeg, Canada, M. Jackson, National Research Council - Institute for Biodiagnostics,

Winnipeg, Canada, M. L. Gruwel, National Research Council - Institute for

Biodiagnostics, Winnipeg, Canada




Carcinogenesis can be measured in the livers of woodchucks chronically

infected with woodchuck hepatitis virus (WHV) through repeated magnetic resonance imaging (MRI) and localized phosphorus spectroscopy (31P-MRS), and the results of the MRS can be correlated with the degree of liver injury detected by serum gamma glutamyltransferase.



Prior to imaging, serum samples were collected from the hind limb of each

woodchuck for analysis of gamma glutamyltransferase concentration, viral load and expression of woodchuck hepatitis surface antigen (WHsAg). A custom-built doubly tunable quadrature volume coil was tuned and matched to 300MHz and 121.5MHz for proton imaging (1H-MRI) and phosphorus spectroscopy (31P-MRS) respectively to image age- and sex-matched control (n=5) and chronically infected (n=5) adult woodchucks in a 7T horizontal bore magnet (Magnex, UK). Sagittal and axial scout FLASH images (128x128, slice thickness = 5mm, TR/TE=1000/4.1ms, 8 averages) were acquired to locate the largest portion of liver with the least amount of signal contamination from surrounding muscle and adipose tissues. Two-dimensional chemical shift imaging (2D-CSI, 16x16 data matrix, 24x24x2cm, 1024 data points, TR=1000ms, 16 averages) was acquired for all woodchucks each month for a six months period.



Cellular proliferation within the tumours of infected woodchucks was detected as an increase in the ratio of the amounts of phosphomonoesters relative to the b phosphate group of ATP (p=0.01). Significantly elevated levels of serum gamma glutamyltransferase indicated hepatocyte injury in chronically infected woodchucks (p=0.02). An active chronic infection was confirmed by viral load and the expression of WHsAg in infected animals.



Repeated imaging of chronically infected woodchucks can detect changes

in liver tissue composition using proton imaging and phosphorus spectroscopy. Liver injury can be correlated to the expression of gamma glutamyltransferase.

Abstract ID: 65541

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Administration of Erythropoietin for Hepatitis C Treatment-Related Anemia Is Not Cost Effective.

M. S. Campbell, University of Pennsylvania, Philadelphia, PA, J. C. Sun, University of

Pennsylvania, Philadelphia, PA, B. Y. Lee, University of Pennsylvania, Philadelphia,

PA, D. E. Kaplan, University of Pennsylvania, Philadelphia, PA, K. Reddy, University

of Pennsylvania, Philadelphia, PA




To determine the cost-effectiveness of erythropoietin administration to genotype 1 hepatitis C patients with advanced fibrosis who develop treatment-related anemia.



We constructed a 7-state lifetime Markov cost-effectiveness model with 1

year cycle length. Patients cycled between bridging fibrosis, compensated cirrhosis, hepatocellular carcinoma, decompensated cirrhosis, liver transplantation, and death. Competing age-adjusted mortality was included. Transition probabilities, quality of life measurements based on standard gamble data, and costs were obtained from systematic literature review. We used wholesale drug costs and performed the analysis from a Medicare payer perspective (2003 US dollars). Base case was a 50 year old anemic patient (hemoglobin < 12 mg/dL) with at least bridging fibrosis. We assumed that 80% of patients who receive erythropoietin and 46% of those who do not receive erythropoietin were able to maintain recommended weight-based pegylated interferon and ribavirin dosages (Afdhal NH et al Gastroenterology 2004). Using data examining effect of adherence on response rate, we modeled that 46.4 % of patients given

erythropoietin and 42.6% who are not given erythropoietin achieved sustained virologic response (Reddy KR et al J Hepatol abstract 2005). We accounted for treatment discontinuation due to lack of early virologic response. We performed Monte Carlo microsimulation and sensitivity analyses.



For a 50 year old with metavir 2 fibrosis, our Markov model yielded median survivals of 31 and 25 years for patients who did not respond to HCV therapy, respectively.  Ten year survival for compensated cirrhotics not responding to therapy was 71%.


Incremental cost effective ratio (ICER) for administration of erythropoietin during HCV therapy were $77,313, $61,886, and $66,674 for patients with metavir stage 2 fibrosis, bridging fibrosis, and compensated cirrhosis, respectively .  One way sensitivity analysis showed that age, effect of erythropoietin in sustained virological response and on quality of life (utility) during treatment, cost of treatment with and without the use erythropoietin, transition possibilities from compensated to decompensated cirrhosis to hepatocellular carcinoma (HCC), and discount were associated with the largest ranges in ICER.


For erythropoietin administration to be cost effective (<$50,000), the incremental improvement in sustained virological response must be 5.7% or greater.  The base case assumes the erythropoietin improves sustained virological response by 3.8%.



Our comprehensive Markov model suggests that administration of

erythropoietin is not cost-effective for hepatitis C patients with advanced fibrosis who develop treatment-related anemia. The small gains in response rate are eclipsed by the large increase in cost of therapy due to the addition of erythropoietin.




Abstract ID: 64853

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

HIV Testing and HIV Seropositivity Among Patients with Chronic Hepatitis C Virus Infection: Missed Opportunities for Early Diagnosis.

G. Villanueva, Bellevue Hospital and NYU School of Medicine, New York, NY, N.

Shukla, NYU School of Medicine, New York, NY, C. T. Tenner, VA New York Harbor

Healthcare System and NYU School of Medicine, New York, NY, A. Aytaman, VA

New York Harbor Healthcare System, Brooklyn, NY, E. J. Bini, VA New York Harbor

Healthcare System and NYU School of Medicine, New York, NY



Coinfection with hepatitis C virus (HCV) and human immunodeficiency virus (HIV) is a major public health problem worldwide. The 2002 National Institutes of Health HCV Consensus Conference recommended that patients with HCV infection who are at risk for HIV should be offered HIV testing. To date, however, little is known about HIV testing in this population. The aims of this study were to determine the proportion of patients with chronic HCV infection who were tested for HIV and to evaluate the frequency of HIV seropositivity in this population.



569 patients with chronic HCV infection and 670 HCV antibody negative controls completed a survey at the time of their outpatient clinic visit at 3 study sites. Data collected included patient demographics, risk factors for HCV and HIV infection, and information regarding prior HIV testing.



Overall, the mean age of the 1,239 patients was 55.2 ± 10.3 years, 80.8% were men, and the population was ethnically diverse with 36.5% non-Hispanic whites, 37.4% non-Hispanic blacks, 19.6% Hispanics, and 6.5% who reported their ethnicity as “other”. Of the 569 HCV(+) patients, 62.6% were tested for HIV, 35.7% were never tested, and 1.8% did not know if they were tested and this differed significantly (p <0.001) from HCV(-) patients (43.9%, 49.4%, and 6.7%, respectively). Among HCV(+) patients, the proportion tested for HIV decreased with age (p <0.001), was more common in men than in women (70.0% vs. 38.4%, p <0.001), and differed according to ethnicity (p = 0.001). Importantly, risk factors for HIV infection were common among the 203 HCV(+) patients who were never HIV tested, including injection drug use (68.0%), ³10 lifetime sexual partners (58.1%), sex with a prostitute (58.1%), and sex with a same-sex partner (13.3%). Of the HCV(+) patients who were tested for HIV, 23.9% were positive, 74.2% were negative, and 2.0% did not know the results of their HIV test; this differed significantly (p = 0.04) from HCV(-) patients (16.0%, 81.3%, and 2.7%, respectively). Among HCV(+) patients, the proportion that tested positive for HIV did not differ according to age (p = 0.27), but was higher in women than in men (41.9% vs. 21.9%, p = 0.004) and differed according to ethnicity (p = 0.049).



Although HCV(+) patients were more likely to be tested for HIV than HCV(-) subjects, missed opportunities for early diagnosis of HIV infection exist. The high prevalence of HIV seropositivity and risk factors for HIV infection in patients with HCV suggests that the current testing guidelines should be revised to recommend universal HIV testing for all HCV infected persons.



Abstract ID: 64395

Category: LO1: Public Policy, Epidemiology, and Decision Analysis


Quality-of-life tradoffs for hepatitis C treatment: do patients and health professionals agree?


B. R. Schackman, Weill Medical College of Cornell University, New York, NY, P. A.

Teixeira, Weill Medical College of Cornell University, New York, NY, G. Weitzman,

Weill Medical College of Cornell University, New York, NY, A. I. Mushlin , Weill

Medical College of Cornell University, New York, NY, I. M. Jacobson, Weill Medical

College of Cornell University, New York, NY



To investigate differences between how HCV patients and health professionals in an urban setting evaluate the quality-of-life tradeoffs associated with HCV treatment.



Hypothetical descriptions of HCV disease symptoms and of treatment side

effects were presented to 1) 92 treatment-naïve HCV patients recruited at gastroenterology, methadone maintenance, and HIV clinics at 3 urban hospitals; and 2) 23 health care professionals experienced in treating HCV patients. Respondents performed rating scale (RS) and standard gamble (SG) evaluations of the health states on a scale from 0 (death or worse than death) to 100 (best possible health). Evaluations were in English and Spanish. 


78% of patients and 83% of professionals were able to complete all the evaluations.  Treatment side effects were rated worse by patients than professionals using both evaluation methods (p<0.02). RS responses indicated that a year of severe treatment side effects was equivalent to 4.3 years of mild chronic HCV symptoms for patients versus 1.9 years for professionals; a year of moderate treatment side effects was equivalent to 1.5 years of moderate chronic HCV symptoms for patients versus 0.7 years for professionals. SG responses were similar. For the 44 patients with depression (10-item CES-D score ³ 10), the implied value of HCV treatment was low unless it would also relieve their depression.



Health professionals have fewer concerns about HCV treatment side effects than patients. Peer-led counseling and pre-treatment of depression may improve patient acceptance of quality-of-life tradeoffs associated with initiating HCV treatment. Conversely, helping physicians to recognize patients’ perceptions may also lead to improved communication and higher treatment acceptance rates.  In addition, patients with symptoms of depression represented 60% of the sample and their responses indicating the importance of relieving depressive symptoms to their consideration of HCV treatment.




Abstract ID: 62351

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Coffee Drinking Decreases the Risk of Chronic Liver Disease in the United States Population.

C. E. Ruhl, Social & Scientific Systems, Inc., Silver Spring, MD, J. E. Everhart,

National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD




Coffee drinking has been suggested to protect against liver injury, but it is

uncertain whether this is of clinical significance. We examined the relationship of coffee and tea consumption with the incidence of hospitalization or death from chronic liver disease (CLD).



Participants in the population-based first National Health and Nutrition

Examination Survey, 1971-75, were asked about their typical total coffee and tea consumption. Daily coffee and tea intake was categorized as <1 cup (mean 0.2 cups), 1-2 cups, and >2 cups (mean 4.0 cups). A second analysis included follow-up of persons free of liver disease in 1982-84 who were asked more detailed questions on coffee and tea drinking. Persons with evidence of liver disease at baseline were excluded. Participants were followed through 1992-93 for CLD as identified by hospital or death certificate diagnosis of CLD or cirrhosis (ICD-9-CM 571). Ninety-six percent of the baseline cohort was recontacted. Hazard rate ratios for CLD according to coffee and tea intake level were calculated using Cox proportional hazards analysis.



9,849 persons were followed for a median of 19.0 years (range 0.02-22.1). Cumulative incidence of CLD was 1.4% at 20 years. Participants who consumed more coffee and tea had a lower incidence of CLD (test for trend p=0.002) (Figure). In multivariate analysis, participants who drank >2 cups per day had less than half the rate of CLD as those who drank <1 cup per day (hazard ratio = 0.43, 95% confidence interval = 0.24-0.78). Protection by coffee and tea was limited to persons at higher risk for liver diseases from heavier alcohol intake, overweight, diabetes, or high iron saturation. Among 9,650 participants who provided detailed drink information in 1982-84, intake of regular ground coffee and of caffeine was associated with lower incidence of CLD.



Coffee and tea drinking decreases the risk of clinically significant CLD.  Further investigation into the mechanism and potential clinical benefit of this relationship should be pursued. 

Abstract ID: 62378

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

The Cost of Variceal Bleeding in Cirrhosis - A Comparison of a Decision Based Economic Model and Patient Data.

L. M. Cheshire, University College London, UK, London, AK, United Kingdom (Great

Britain), R. Jalan, University College London, London, AK, United Kingdom (Great





A bottom-up cost analysis of direct health care resource utilization was

carried out in a sample of patients presenting with oesophageal variceal bleeding and compared with the predicted cost of managing a bleed from a decision analysis model.



The decision analysis model was developed in Excel®. Treatment strategies

for the model were based on the ‘UK guidelines on the management of variceal haemorrhage in cirrhotic patients’ (Jalan & Hayes 2000). Clinical outcomes and probabilities were derived from the literature and physician interviews. Sensitivity analysis was performed to evaluate the effect of changes in resource use or unit costs. For the patient analysis, data was collected from 31 patients case notes, each having experienced an acute episode of variceal bleeding requiring endoscopic therapy. Resource data on: length of stay; blood product use; medication; consultations; and tests/procedures were collected. Data sources were medical/nursing notes, charts and electronic records. The main unit of analysis was an ‘episode of care’, defined as starting with index bleed and continuing for a 42-day follow-up period, or death.



In the patient analysis, mean survival at 42 days was 55% vs. 54.8% in the model. Out of the 31 initial bleeds, there were a further 11 re-bleed/new bleeds; 8 (26%) patients required transjugular intrahepatic portosystemic stent shunt (TIPSS) insertions for uncontrolled bleeding. Mean cost of an episode of care was £17,738 in the patient analysis vs £15,595 in the model; the actual cost of a life saved is £32,345. Non-survivors cost significantly

more on a mean daily cost basis (p=0.0007). Disease severity impacted on cost, although this was not incorporated into the model. Major cost drivers in both model and patient data were length of stay, especially intensive care, use of blood products and number of procedures/investigations carried out. Patient analysis reflected these estimates and findings in the decision model.




There was good concordance between the economic model and actual

patient data, making it an extremely powerful tool to evaluate the cost of different clinical scenarios. Variceal bleeding carries a burden of disease to both individuals and health service providers. To bring down costs, management strategies need to be found to reduce length of stay and improve control of initial bleeding, whilst preventing further rebleeding.


Abstract ID: 64466

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Impact of Patient Race and Gender on the Care of a Large Cohort of Community Based Hepatitis C Patients.

T. M. Shehab, Huron Gastro, Ypsilanti, MI, R. R. Randhawa, St. Joseph Mercy

Hospital, Ypsilanti, MI, A. M. Jankowski, Huron Gastro, Ypsilanti, MI, R. L. Stoler,

Huron Gastro, Ypsilanti, MI




Previous research has demonstrated that patient race and gender affect the

management of a number of diseases. While race affects the efficacy of hepatitis C therapy, the impact of race and gender on the care of hepatitis C patients has not been assessed. The aim of this study is to determine if race and gender influence the provision of care (diagnostic evaluation/treatment) of a large cohort of patients seeking consultative care for hepatitis C in a community based, private practice gastroenterology group.



Retrospective chart review was performed for all patients referred for evaluation of hepatitis C between 1997 and 2002. The review assessed patient demographics, laboratory evaluation, disease severity and treatment response (Interferon/Ribavirin or Pegylated Interferon/Ribavirin).



670 patients ( male=396; female = 274)were seen for an initial consultation for hepatitis C during the study period. 59% of patients were male, 538 patients or 80% were Caucasian (Cauc), 101 patients or 15% African-American (AA) and mean age was 47 ± 10.0 (17-90) years. A statistically significant difference (p<0.05) was seen between Cauc and AA patients in the following categories: BMI, proportion of patients who returned for a

second office visit, compliance with office visits, genotype and sustained virologic response (SVR). The mean number of office visits during the study based on race/gender was: 5 for Cauc men and women, 4 for AA women and 2 for AA men. There was no significant difference in indication for referral, rates of viremia, diagnostic evaluation, proportion of patients offered therapy or proportion of patients with elevated liver enzymes based on race or gender. The patient’s gender did not affect the proportion of patients with advanced fibrosis, elevated liver enzymes or likelihood of

receiving hepatitis C therapy.



Race had a significant impact on a number of factors that are important in the care of hepatitis C patients. In addition to lower rates of virologic response, African-American patients had fewer office visits and were less likely to return for a second office visit. The disparity was even more significant in African American men. Overall, gender had little impact on the evaluation, histologic severity and initiation of treatment. While much attention has been focused on the impact of race on treatment response, further research must clarify barriers that may negatively affect the care of

certain subgroups of hepatitis C patients.




Abstract ID: 64997

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Cost-effectiveness of treatment for chronic hepatitis C infection in patients coinfected with human immunodeficiency virus

J. C. Servoss, GI Unit, Massachusetts General Hospital, Boston, MA, N. G. Campos,

Harvard University Initiative for Global Health, Cambridge, MA, J. A. Salomon,

Harvard University Initiative for Global Health, Cambridge, MA, K. A. Freedberg,

Massachusetts General Hospital, Boston, MA, J. H. Samet, Boston Medical Center,

Boston, MA, S. J. Goldie, Harvard University Initiative for Global Health, Cambridge,

MA, R. T. Chung, GI Unit, Massachusetts General Hospital, Boston, MA




Antiviral therapy with pegylated interferon-alpha and ribavirin has been recently approved for treatment of hepatitis C virus (HCV) in human immunodeficiency virus (HIV)-coinfected persons. However, in view of more limited efficacy rates in coinfection, the cost-effectiveness of this treatment merits investigation. We sought to examine the clinical benefits and cost-effectiveness of treatment for chronic HCV infection in patients co-infected with HIV.



We modified the natural history, HCV treatment efficacy and cost parameters of an existing Markov model of HCV monoinfection to reflect coinfection with HIV and recent randomized clinical trial data. The two treatment strategies examined were 48 weeks of combination therapy with: (A) pegylated interferon-alpha and ribavirin or (B) interferon-alpha and ribavirin. Outcomes included life expectancy and cost-effectiveness

in incremental cost per life-year saved.



Treatment with pegylated interferon alfa 2a and ribavirin was consistently more effective and cost effective than treatment with interferon alfa-2a (standard interferon) and ribavirin.  For a hypothetical cohort of HIV/HCV coinfected patients with CD4 count 200-500 cell/mm3, treatment with pegylated interferon alfa 2a and ribavirin yielded life expectancy gains from 2 months (genotype 1 HCV) to 11 months (non-1 genotype HCV).  Incremental cost effectiveness ratios, stratified by genotype and weighted by sex, ranged from $160,000 Genotype 1 to $35,600 for non-genotype 1 patients.  Because the overall higher efficacy of HCV treatment in those with non-genotype 1 HCV, the incremental cost effectiveness ratios in non-1 genotype coinfected patients were considerably less than $40,000 per life year saved. 



Treatment of chronic HCV with pegylated interferon-alpha and ribavirin is

most cost-effective in HIV-infected patients with CD4>500/mm3 and in those with nongenotype 1 HCV.   For patients with genotype 1 HCV, the cost effectiveness of HCV treatment is highly dependent upon treatment efficacy.  Our data suggest that treatment of genotype 1 HCV in HIV coinfected patients may be reasonable. 


Further clinical studies are needed to assess the effects of treatment regimens in patients with different stages of HIV and liver disease.

Abstract ID: 61982

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

The role of liver biopsy for asymptomatic liver enzyme elevation.

M. Le Mire, Royal Adelaide and Lyell McEwin Hospitals, Adelaide, Australia, K.

Fleming, John Radcliffe Hospital, Oxford, United Kingdom (Great Britain), R.

Chapman, John Radcliffe Hospital, Oxford, United Kingdom (Great Britain)



Liver biopsy findings in patients with asymptomatic liver enzyme elevation vary considerably. In studies of patients without serologic evidence for viral, autoimmune or metabolic liver disease, cirrhosis rates between 3 and 16% are reported and there is considerable variation in etiology. This has created difficulty in defining the role of liver biopsy and evidence-based guidelines are lacking. The study reviewed diagnoses, fibrosis scores and complications from liver biopsy for asymptomatic liver enzyme elevation including selected patients with isolated gamma glutamyltransferase (GGT) elevation.



The records of consecutive patients who underwent liver biopsy to investigate mild transaminase (£5´ULN) or GGT elevation of any level were studied. Patients with suspected alcoholic liver disease (ALD), drug-induced liver disease or risk factors for non-alcoholic fatty liver disease (NAFLD) were included. Patients with symptoms or sera diagnostic of hemochromatosis, a1-antitrypsin-deficiency related, viral or autoimmune liver disease were excluded. The study included 102 patients who were biopsied during the period from 1997-2001. The mean aspartate transaminase was 53.4 +/-36.7 (N<42) and GGT was 227+/- 276 (N<60). ALD was diagnosed in 26, fatty liver in 18, non-alcoholic steatohepatitis (NASH) in 18 and minor or non-specific changes in 20 patients. Treatment or drug withdrawal was possible for 10 patients including 4 with primary biliary cirrhosis, 2 with drug-induced liver disease and 2 with primary sclerosing cholangitis. The pre-biopsy diagnosis was changed in 16 out of 64 patients for whom it was recorded. In 15 patients with isolated GGT elevation 6 had a diagnosis of NAFLD and 6 ALD. Bridging fibrosis was found in 12 and cirrhosis in 9 patients. Milder fibrosis was found in 46 patients (Ishak 1-2) and no fibrosis in 34 patients. Five out of 15 patients with isolated GGT elevation had bridging fibrosis or cirrhosis. Fifty-eight patients underwent biopsy without and 40 with ultrasound guidance, 5 with CT guidance and 3 were by the transjugular route.



Five out of 58 unguided biopsies did not yield an adequate sample compared with 1 out of 40 ultrasound-guided biopsies. Three complications occurred, all with unguided biopsies, including 1 bile leak, 1 hemorrhage and 1 readmission with pain.



·       The incidence of altered diagnosis or advanced fibrosis in this study supports the use of liver biopsy in patients with asymptomatic liver enzyme elevation including selected patients with isolated GGT elevation.

·       The reduced frequency of complications or inadequate samples with ultrasound guidance supports the use of ultrasound to assist liver biopsy.

·       Liver biopsy for seronegative changed the diagnosis in 25% and defined treatable liver in 10%.

·       Bridging fibrosis or cirrhosis was found in 21% of biopsies.

·       In selected patients with GGT evaluation and alcoholic liver disease, bridging fibrosis or cirrhosis was found.

·       A lower rate of inadequate sampling occurred with ultrasound guided liver biopsy.


Abstract ID: 62181

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Psychosocial Factors in Liver Transplantation: How are U.S. Centers Evaluating Them.

R. C. Anderson, Medical College of Wisconsin, Milwaukee, WI, C. P. Johnson, Medical

College of Wisconsin, Milwaukee, WI, J. Franco, Medical College of Wisconsin,

Milwaukee, WI



The liver transplant waiting list has grown disproportionately to the number of transplants performed at U.S. centers. Minimal listing criteria have been

developed and MELD has prioritized patients based on objective criteria.At this time no standardization of psychosocial factors in those being evaluated for transplant exists. The purpose of this study is to determine if psychosocial factors are considered by transplant centers prior to listing for liver transplantation.



A questionnaire developed by a psychologist experienced in evaluating transplant candidates and addressing 4 major areas (alcohol and drug use, compliance, psychiatric and psychological issues and psychosocial support systems) was sent to 43 U.S. centers performing adult liver transplants. Specific areas addressed included required periods of drug or alcohol abstinence, random drug testing, candidates with non-compliance, illiteracy, decreased mental capacity, psychiatric and psychological issues and incarceration.



Twenty questionnaires were completed and were available for interpretation. ALCOHOL, TOBACCO AND ILLICIT DRUG USE: All 20 centers required a period of alcohol abstinence with 19/20 requiring at least 6 months and 1/20 requiring 12 months. 19/20 centers required participation in drug and alcohol treatment while 1 did not.17/20 centers required marijuana abstinence while 3/20 did not. Random drug testing was performed by 19/20 centers with the frequency ranging from every 6 weeks to on as needed basis. Tobacco cessation was required by 12/20 centers.


COMPLIANCE: All 20 centers excluded patients that did not adhere to medication regimens, medical appointments and the inability to remain drug or alcohol free.


PSYCHIATRIC AND PSYCHOLOGICAL ISSUES: 17/20 centers required a psychiatric or psychological evaluation and all 17 included a mental health professional in the evaluation process. 5/20 centers excluded patients with psychiatric or psychological disease with the most common exclusionary criteria being active psychosis. 6/20 centers excluded patients with illiteracy or decreased mental capacity.


PSYCHOSOCIAL SUPPORT: 15/20 centers excluded candidates without psychosocial support systems with most mandating 1 full time caregiver or support person. 12/20 centers would consider transplantation for incarcerated individuals,7/20 would not and 1/20 would evaluate each case individually.



·       Despite the absence of mandated uniform criterion, transplant centers utilize similar standards when evaluating liver transplant candidates for drug and alcohol abstinence, medical compliance, psychiatric or psychological issues and psychosocial support systems.

·       All centers required at least 6 months abstinence from alcohol.

·       The majority of Centers answering the survey required tobacco cessation.

·       The majority of centers will not consider patients with significant psychiatric or psychological disease for transplant.

·       Illiteracy remains a barrier for transplant in most Centers.



Abstract ID: 66511

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Real time ultrasound guided liver biopsy in real life; a comparison of pre-biopsy versus real time ultrasound.

S. MANOLAKOPOULOS, Gastroenterology Department, ‘Polykliniki’ General

Hospital, ATHENS, Greece, C. TRIANTOS, Gastroenterology Department,

‘Polykliniki’ General Hospital, ATHENS, Greece, S. BETHANIS, Gastroenterology

Department, ‘Polykliniki’ General Hospital, ATHENS, Greece, J. THODOROPOULOS,

2nd Gastroenterology Department, ‘Evangelismos’ General Hospital, ATHENS, Greece,

J. VLACHOGIANNAKOS, 2nd Gastroenterology Department, ‘Evangelismos’ General

Hospital, ATHENS, Greece, E. CHOLONGITAS, Liver Transplantation and

Hepatobiliary Unit, Royal Free Hospital, London, United Kingdom (Great Britain), M.

SIDERIDIS, Radiology Department, 'Polykliniki' General Hospital, ATHENS, Greece,

K. BARBATI, Histopathology Department, ‘Hellenic Red Cross’ General Hospital,

ATHENS, Greece, P. PIPEROPOULOS, Radiology Department, ‘Evangelismos’

General Hospital, ATHENS, Greece, C. SPILIADI, Histopathology Department,

‘Evangelismos' General Hospital, ATHENS, Greece, N. PAPADIMITRIOU, 2nd

Gastroenterology Department, ‘Evangelismos’ General Hospital, ATHENS, Greece, D.

TZOURMAKLIOTIS, Gastroenterology Department, ‘Polykliniki’ General Hospital,

ATHENS, Greece, A. BURROUGHS, Liver Transplantation and Hepatobiliary Unit,

Royal Free Hospital, LONDON, United Kingdom (Great Britain), A. AVGERINOS, 2nd

Gastroenterology Department, ‘Evangelismos’ General Hospital, ATHENS, Greece




Traditionally, liver biopsy has been performed by hepatologists/ gastroenterologists. However, an increasing number are performed by radiologists under ultrasound control. Routine ultrasound assessment of puncture site before performing percutaneous biopsy has been reported to increase diagnostic yield and decrease complication rates. Liver biopsy guided by real time ultrasonography is a safe and effective method for the

diagnosis of liver disease. It is not clear if the latter is superior to marking the puncture site before biopsy as regards biopsy size, fragmentation and complications.



To compare ultrasound assessment of the puncture site before performing percutaneous liver biopsy with real time ultrasound liver biopsy for suspected diffuse liver disease



From January 1995 to September 2004, 558 percutaneous liver biopsies for diffuse liver disease were performed. There were 2 groups: a) group A: real time u/s guided performed by radiologists (213 biopsies, M/F=120/93, median age 50 yrs (range 17 -76), needle 18G) b) group B: u/s assessment by radiologists of puncture site the same day or within previous 24 hours of liver biopsy the (338 biopsies, M/F:246/92, median age 43 yrs (range 15 -75), needle 16G). These biopsies were performed by experienced

gastroenterologists/hepatologists on the ward using the marked site.



There were no differences in severity of liver disease nor length of specimen

(mean=1.43±0.59 group A vs 1.48±0.38 group B (p=0.39)) nor number of fragments (mean=1.41±0.75group A vs 1.38±0.87.group B (p=0.39)) nor number of passes. There was no difference in establishing diagnosis (p=0.086). No diagnosis was reached in 7 in Group A (4 inadequate sample and 3 no liver tissue) and in 23 in group B (4 inadequate sample and 19 no liver tissue). After using step-wise logistic regression two independent

variables were statistically significant for diagnosis: biopsy length (p=0.004) and fragment number (p<0.001). One patient bled in Group B (required 2 units of blood transfusion and no surgery).



Real time ultrasound did not improve diagnostic yield or result in less complications. Marking the puncture site seems adequate and has practical advantage in employing less of the radiologists time.


Abstract ID: 66908

Category: LO1: Public Policy, Epidemiology, and Decision Analysis

Inform Consent for Clinical Trials- Still some way to go.

a. cheok, National University Hospital, singapore, AL, Singapore




Informed consent is the process of giving participant all of the information

that participant needs to make an decision about involvement in a research study. It is effective only if patients understand the study information presented by the medical team.



We set to determine understanding of the informed consent process in Asian

patients with chronic hepatitis B. Methods 77 Chinese patients with chronic hepatitis B from 12 randomized controlled clinical trials on nucleotide analogues were asked to participate in this survey. All subjects had given informed consent at an earlier date. They were interviewed on a 10-item questionnaire. To avoid bias, the interview was conducted by one administrative assistant, who had not been involved in any clinical trial

or informed consent process. Data were expressed in mean ±S.D. In univariate analysis, Mann-Whitney U test and Kruskal-Wallis test were used.



Age of subjects was 40.32 ±12.07 (range 21-68) years, 66 (85.7%) were male. On a visual analogue scale of 1(minimum) to 5 (maximum), scores for comprehensiveness and complexity of the consent form was 3.67 ± 0.82 and 2.56 ± 1.04 respectively, and score for their understanding of informed consent was 3.84 . On the specific aspects of clinical trials, only 22 (28.6%) subjects thought their trials were randomized . Only 26 (33.8 %) knew placebo was a non-active drug and 20 (26.0%) patients knew that their clinical symptoms might not due to the side effects of the study drug. 76 (98.7%) felt benefited from trials: 65 (84.4%) felt they had better medical attention, 45 (58.4%) patients felt benefited from the new drug, and 51 (66.2%) felt benefited from free treatment and blood tests. The concept of the informed consent were summarized by given scores (0-4). The mean score of the study group is 1.7 ±0.9. In univariate analysis, the differences of the score were seen only in different education level (P=0.01) although different aspects, such as race, religion, income level, marital status were tested.



Despite the fact that all subjects in this study gave an earlier informed consent for the respective clinical trials, many had an incorrect idea and impression about the meaning of placebo, randomization, and the follow up procedures of a clinical trial. More time is needed to explain to patients

with lower education level. Further education and streamlining of the informed consent process may be needed for patients from Asia.

Abstract ID: 68207

Category: LO1: Public Policy, Epidemiology, and Decision Analysis


A. M. Anonychuk, University Health Network, Department of Medicine, Toronto, ON,

Canad, Toronto, Canada, A. Tricco, GlaxoSmithKline Canada, Biomedical Data

Science, Oakville, ON, Canada, Oakville, Canada, C. Bauch, University of Guelph,

Mathematics and Statistics, Guelph, ON, Canada, Guelph, Canada, B. Pham,

GlaxoSmithKline Canada, Biomedical Data Science, Oakville, ON, Canada, Oakville,

Canada, M. Krahn, University Health Network, Department of Medicine, Toronto, ON,

Canad, Toronto, Canada




Current policy regarding hepatitis A (HA) vaccination in high-endemicity

regions and among high-risk groups is influenced by published cost-effectiveness analyses (CEA); however results of these vary widely across studies. A systematic review was conducted to estimate the effects of methodologic quality. Adequate representation of vaccine-induced herd-immunity and accurate estimation of the degree of underreporting were assessed.



CEA studies of HA vaccine were identified (MEDLINE, EMBASE, and two

others; MeSH “cost-effectiveness” AND “hepatitis A”), and included if they were costeffectiveness/ utility studies of HA vaccine. Citations and full-text articles were reviewed independently by two reviewers. Back referencing, author searches, and expert consultation ensured literature saturation. Quality of reporting was assessed using a 21-item quality tool (Neumann 2000). Methodological issues specific to vaccine evaluations were appraised by Beutels 2002 guideline. Key modeling issues were examined using Sculpher’s 2000 framework.



34 CEA studies were included from full-text-article review (n=95) and citation screening (n=700). 9 conducted an additional cost-utility analysis (CUA) and 6 a cost benefit analysis. All were model-based studies, 13 utilizing Markov models. 1 used a dynamic model, capturing effects of herd immunity. Strategies included mass childhood/adult (n=11/n=2) and selective childhood/adult vaccination (n=24/n=8). Populations assessed were healthy participants (n=31), and patients with chronic liver disease (n=3).

The median score on the 21-item tool was 62% (13/21 range[14-90%]) similar to mean scores (58%) of other CEA. CUA (n=9) attained higher quality (median 76%[57-90%]) than CEA (n=25, 52%[14-71%]).

CUA studies assessed using Beutels guideline clearly stated model assumptions (7/9), utilized proper time span (7/9), and captured relevant costs (7/9). Many did not discuss alternative modeling approaches (7/9).

Using Sculpher’s framework, CUA studies specified model assumptions (9/9), and most justified model parameters (7/9). Some did not discuss implications of relaxing model assumptions (3/9). None adjusted for herd immunity, though 4 acknowledged this as a limitation. 1 study adjusted for under-reporting, though 5 studies discussed it.


·       On average, studies that conducted a cost-utility analysis seemed to attain a higher quality of reporting than CEA only studies.

·       Herd immunity effects were not accounted for in most CEA studies.

·       Most CEA studies did not account for under-reporting in HA case-notification data when used as input in the analysis.

·       Large variation was evident in the methods used in these studies. This potentially relates to the large discordant results reported from these studies.

  • The relative efficiency of HA vaccination programs evaluated in these studies were likely to be sensitive to uncertainty in HA incidence, vaccine costs and discount rates.