A continuous 13C methacetin breath test for noninvasive assessment of intrahepatic inflammation and fibrosis in patients with chronic HCV infection and normal ALT

Up to 30% of patients with hepatitis C virus (HCV) infection and normal serum alanine aminotransferase (NALT) have significant liver disease. Currently, many of these patients undergo a liver biopsy to guide therapeutic decisions. The BreathID® continuous online 13C-methacetin breath test (MBT) reflects hepatic microsomal function and correlates with hepatic fibrosis. To assess its role in identifying intrahepatic inflammation and fibrosis in NALT patients, we tested 100 patients with untreated chronic HCV infection, and 100 age- and sex-matched healthy volunteers using 13C MBT following ingestion of 75 mg methacetin. All HCV patients had undergone a liver biopsy within 12 months of performing the MBT. Patients with a necroinflammatory grade ≤4 or >4, based on Ishak (modified HAI) score, HAIa + HAIb + HAIc + HAId, were defined as having low or high inflammation, respectively. Patients with a histological activity fibrosis stage ≤2 or >2, were defined as having nonsignificant or significant fibrosis, respectively. A proprietary algorithm to differentiate intrahepatic inflammation within chronic HCV patients with NALT achieved an area under the curve (AUC) of 0.90. Setting a threshold on the point of best agreement (at 83%) results in 82% sensitivity and 84% specificity. With application of another proprietary algorithm to differentiate patients with nonsignificant or significant fibrosis, 67% of liver biopsies performed in the patient group could have been avoided. This algorithm achieved an AUC of 0.92, with a sensitivity of 91% and a specificity of 88%. There was no correlation between body mass index (BMI) and MBT scores for patients with the same histological score. The continuous BreathID®13C MBT is an accurate tool for measuring the degree of inflammation and fibrosis in patients with chronic HCV infection and NALT. As such, it may prove to be a powerful, noninvasive alternative to liver biopsy in the management of this patient population.

somal function and correlates with hepatic fibrosis. To assess its role in identifying intrahepatic inflammation and fibrosis in NALT patients, we tested 100 patients with untreated chronic HCV infection, and 100 age-and sexmatched healthy volunteers using 13 C MBT following ingestion of 75 mg methacetin. All HCV patients had undergone a liver biopsy within 12 months of performing the MBT. Patients with a necroinflammatory grade £4 or >4, based on Ishak (modified HAI) score, HAIa + HAIb + HAIc + HAId, were defined as having low or high inflammation, respectively. Patients with a histological activity fibrosis stage £2 or >2, were defined as having nonsignificant or significant fibrosis, respectively. A proprietary algorithm to differentiate intrahepatic inflammation within chronic HCV patients with NALT achieved an area under the curve (AUC) of 0.90. Setting a threshold on the point of best agreement (at 83%) results in 82% sensitivity and 84% specificity. With application of another proprietary algorithm to differentiate patients with nonsignificant or significant fibrosis, 67% of liver biopsies performed in the patient group could have been avoided. This algorithm achieved an AUC of 0.92, with a sensitivity of 91% and a specificity of 88%. There was no correlation between body mass index (BMI) and MBT scores for patients with the same histological score. The continuous BreathID Ò 13 C MBT is an accurate tool for measuring the degree of inflammation and fibrosis in patients with chronic HCV infection and NALT. As such, it may prove to

INTRODUCTION
The natural history of chronic hepatitis C virus (HCV) infection is variable, and some carriers have an indolent disease course with no complications even after decades of follow-up [1,2]. The decision to start treatment for chronic HCV infection depends on several factors, including: alanine aminotransferase (ALT) [3] levels, viral load, liver biopsy parameters of fibrosis and inflammation, patient determination and expected compliance [4]. The protracted course of the illness coupled with the intention of treating those who are most likely to benefit makes liver biopsy an important decision-making tool.
Serum ALT concentration, the most widely used indirect marker for liver disease activity, remains within the normal range in 25-30% of chronic HCV carriers, and an additional 40% have ALT levels less than twice the upper limit of normal [3,5]. It is generally accepted that the natural history of the subgroup of HCV carriers with persistently normal or minimally elevated ALT levels (NALT) is characterized by a slower progression rate [6][7][8]. Accordingly, follow-up and deferring therapy has been suggested in this patient group. A recent review of three large randomized trials has shown that patients with NALT have significantly lower inflammation and fibrosis scores on liver biopsy than patients with elevated ALT [9]. Nevertheless, these patients can have progressive liver disease and develop advanced fibrosis or cirrhosis [3,10]. These studies suggest that patients with chronic HCV with NALT should be evaluated similarly to patients with elevated ALT levels because they are at risk for developing significant liver disease [9].
Using periodic liver biopsies to determine if and when to use antiviral treatment is unlikely to elicit a favourable patient response and can lead to higher costs, increased cumulative cirrhosis incidence and decreased survival rates in comparison with Ôempirically basedÕ treatments [11,12]. Although considered the gold standard for assessment of liver fibrosis, liver biopsies have limitations, including inter-observer variability, sampling error and risks for complications. Reliable and inexpensive noninvasive methods to assess disease progression are a necessity in this setting [12].
Breath testing is based on the principal that an ingested substrate is metabolized, and a measurable metabolite is then expelled by the respiratory system. An ideal compound for this purpose is metabolized solely by the liver and therefore reflects liver function. Breath testing has been used experimentally and clinically for several decades [10], including for follow-up on patients with chronic liver disorders. The major drawbacks of these tests are the need for traditional, cumbersome isotopic ratio mass spectrometry methods, a prolonged testing time and patient inconvenience.
The BreathID Ò continuous online 13 C-methacetin breath test (MBT), which reflects hepatic microsomal function (CYP1A2), is a laser-based technology that creates an infrared emission precisely matching the absorption spectrum of CO 2 and can detect variations of less than 1/1000 in the 13 CO 2 / 12 CO 2 ratio measurement. The system is based on the measurement of CO 2 by molecular correlation spectroscopy. This test offers several advantages: It is an office-based, noninvasive tool for the assessment of both liver inflammation and fibrosis does not involve a blood test and can provide an immediate result at the point-of-care. The aim of the present study was to determine its accuracy in assessing the degree of liver fibrosis and inflammation in patients with chronic HCV infection and NALT.

Study population
Patients. From 1 March 2006 to 31 May 2006, we enrolled 100 consecutive, unselected, patients with previously untreated, chronic HCV. All were anti-HCV and HCV RNA positive, with a normal serum ALT level (£·2 ULN) on two separate tests during the preceding 6 months. All patients underwent a thorough physical examination and liver ultrasonography. Patients with ALT >·2 times the upper limit of the reference range were excluded. Patients were enrolled if they fulfilled the above criteria and had undergone a liver biopsy within 12 months of the breath test, as described below. Patients with other concomitant causes of liver disease such as hepatitis B virus (HBV), HIV, autoimmune hepatitis, alcohol abuse (excess of 40 g/day) and hepatocellular carcinoma were excluded from the study. Ultrasonographic evaluation of the abdomen was performed in all patients, and those with vessel occlusion were excluded.
Healthy volunteers. A group of 100 healthy volunteers (57 males and 43 females) were enrolled as controls in the study. They were screened by medical history, physical examination, liver ultrasound and routine liver function tests. All healthy volunteers had blood test results within normal limits. None had a history of active or previous liver disease or alcohol or drug abuse, and none were taking medications.
All participants gave written informed consent to their participation in the study, which was conducted in strict adherence to the principles of the Declaration of Helsinki. All experiments were approved by the Institutional Review Board committees and the Israel Ministry of Health Committee for Human Clinical Trials. Tables 1 & 2 show the main clinical, laboratory and histological characteristics of the patients and healthy volunteers at the time of liver biopsy, when applicable. The average age and body mass index (BMI) of the patients (36 females and 64 males) were 46 (SD 13.6; range 19-76) and 25.2 (SD 3.9; range 17.5-34.6), respectively. Difference in gender distribution between patients and healthy volunteers (chisquare test) was not significant. Comparing age and BMI between patients and healthy volunteers (t-tests) yields a significant difference (P = 0.0047) for age and a nonsignificant difference (P = 0.306) for BMI. For healthy volunteers, average age and BMI were 40.7 (SD 12.6, range 18-75) and 24.6 (SD 3.9, range 18-37), respectively. For HCV patients,

Biochemical analysis
All patients underwent biochemical work-up, including a complete blood count, aspartate aminotransferase (AST), ALT(3), alkaline phosphatase, c-glutamyltranspeptidase, lactate dehydrogenase, albumin, total bilirubin and prothrombin activity. Routine biochemical tests were performed using commercially available kits. The AST/ALT ratio and AST/platelet ratio index were calculated. For ALT measurements, an upper normal limit of 53 U/L was used ( Table 2).

Liver histology
Following assessment of prothrombin time and platelet count, patients underwent a percutaneous, ultrasound-guided liver biopsy under local anaesthesia (lignocaine 1%). Specimens obtained by means of Menghini needles, diameter 1.6 mm, had an average length of 20 ± 5 mm (range, 15-25 mm), and representative according to accepted standards. Biopsy specimens were fixed with formalin, embedded in paraffin and stained with haematoxylin and eosin. All sections were reviewed by an expert pathologist blinded to patient clinical data and breath-test results. Necroinflammatory score was graded using the HAI score based on periportal or periseptal interface hepatitis (piecemeal necrosis) (0-4), confluent necrosis (0-6), focal (spotty) lytic necrosis, apoptosis, and focal inflammation (0-4) and portal inflammation (0-4) [13]. Fibrosis was staged using the Ishak (modified HAI) fibrosis score on a scale from 0 to 6 [13]. Table 3 shows selected patient data grouped by fibrosis score.

Noninvasive breath testing
Following an overnight (>8 h) fast, patients and healthy volunteers were connected to the breath-testing unitÕs Bre-athID Ò system (BreathID Ltd, Jerusalem, Israel) via nasal cannula (IDcircuit TM ), and received 75 mg of N-(4-methoxy-13 C-phenyl)acetamide (methacetin, Isotec) dissolved in 150 mL of water. Breath samples were collected using an automatic breath sampling unit under continuous capnographic control, before and for 60 min after the labelled substrate was administered to the patient. The 13 CO 2 / 12 CO 2 ratios in the breath samples were determined and mapped on the screen at a high frequency (once every 2-3 min). During the test period, all patients and healthy volunteers continued fasting and were at rest to eliminate any variability in CO 2 excretion due to the ingestion of food or physical activity.

Analysis of breath-test data
Results obtained from the device were expressed as percentage of administered dose of 13 C per cent dose recovered (PDR) and the cumulative PDR (CPDR) percentage of 13 C recovered over time at 10, 15, 20, 30 and 60 min after ingestion of methacetin, respectively, as well as the PDR peak and peak time. PDR refers to the rate at which the 13 C substrate is metabolized and is expressed in %/h. PDR expresses the rate of substrate metabolization derived from the change in the 13 C/ 12 C ratio, in which the specific test details are taken into account, thereby normalizing the results and making them independent of differences in weight, height, dose or substrate type and purity [14,15]. CPDR is the numeric integral of PDR and describes the total amount of substrate metabolized at any given accumulated time. Data are expressed in units of %/h for PDR and per cent for CPDR. The BreathID Ò device plots the PDR and CPDR in real-time and provides PDR peak value and peak time.

Statistical analysis
Using SpearmanÕs nonparametric Rho correlation, the correlation between the different breath-test parameters and modified Ishak HAI inflammation and fibrosis scores, gender, BMI and age were assessed. Patients were grouped according to fibrosis scores, using breath-test parameters to compare between HAI fibrosis scores of £2 vs >2, and HAI inflammation scores (HAIa + HAIb + HAIc + HAId) £4 and >4, respectively. Mann-WhitneyÕs two-samples test and logistic regression with receiver operating characteristic (ROC) curve analysis were used to evaluate the ability of different breath-test parameters and their combination to predict the severity of fibrosis and inflammation. Finally, the repeatability of the test was determined by assessing several participants more than once during a period of <2 weeks. Two algorithms which include several breath-test parameters and patient data were developed to allow differentiation of high vs low inflammation, and significant vs nonsignificant fibrosis, with high sensitivities and specificities while maximizing the number of liver biopsies identified as avoidable.

Breath-test parameters significantly differentiate grade of intrahepatic necroinflammation in chronic HCV patients with NALT
The Mann-Whitney Ôtwo-samples testÕ, used to compare inflammation groups (HAIa + HAIb + HAIc + HAId £ 4 vs Breath-test parameters significantly differentiate degree of fibrosis on liver histology in chronic HCV patients with NALT Most of breath-test parameters evaluated showed a statistically significant (P < 0.005) difference between the

Breath-test parameters significantly differentiate between chronic HCV patients with NALT and healthy volunteer groups
A binary logistic regression model using PDR20 and age (P < 0.001 for each of the two parameters) showed that the MBT can differentiate patients and healthy volunteers with an area under the curve (AUC) of 0.67 (95% CI, 0.59-0.74), sensitivity of 56% and specificity of 86% (Table 6, Fig. 1).

Assessment of serum blood test parameters
Laboratory parameters such as ALT, albumin, prothrombin time, international normalized ratio levels and the AST⁄platelet ratio index score were analysed for dependence with the modified Ishak HAI fibrosis groups. Results are summarized in Table 7. Since most models using blood-test parameters to assess liver disease are often based on elevated ALT values, they failed in the patient population studied in the present study.

Correlation with patient characteristics
No significant correlation was shown between gender and modified Ishak HAI fibrosis stages (chi-square analysis, P = 0.145). When compared by two-tailed t-tests, BMI and age differed significantly (P = 0.004 and P < 0.001, respectively) between significant and nonsignificant fibrosis (HAIf < 3 or HAIf ‡ 3). SpearmanÕs Rho correlations showed that some breath-test parameters were correlated with age (PDR peak, PDR20, CDPR30, CPDR60) or BMI (PDR peak, CPDR60).

Repeatability of breath testing
A total of 42 healthy volunteers and 11 patients were assessed for test repeatability using the Ôwithin-subject coefficient of variationÕ of MBT parameters ( Table 8). The number of repetitions per person was between two and six. Estimations were based on four BT parameters for healthy volunteers⁄patients separately and for both groups together. Repeating the breath testing in both patients and in healthy volunteers (repeats ‡2) resulted in an inter-test variability of £13% for the PDR peak height (95% CI, 0.11-0.15).
The use of MBT as a tool to avoid the need for liver biopsy Necroinflammation. By using a proprietary algorithm that includes breath-test parameters, age and other patient data to differentiate intrahepatic inflammation (HAIa + HAIb + HAIc + HAId £ 4 vs > 4) for chronic HCV patients with NALT, an AUC of 0.90 was achieved (Fig. 2). Setting a threshold on the point of best agreement (at 83%) results in sensitivity of 82% and specificity of 84%. At the datasetÕs prevalence of 68%, the positive predictive value (PPV) was Ó 2008 The Authors Journal compilation Ó 2008 Blackwell Publishing Ltd 92% and the negative predictive value (NPV) was 69%. Assuming a prevalence of 45.5%, this would lead to a PPV of 82% and an NPV of 85%.
Fibrosis. By using an algorithm that includes breath-test parameters, age and other patient data, 67% of liver biopsies performed in the patient group could have been avoided (Fig. 3). This algorithm achieved an AUC of 0.92, with a sensitivity of 91% and a specificity of 88%, a PPV of 88% and an NPV of 91%. Thirty-four patients were identified as having significant fibrosis, including four false positives: two with a HAI fibrosis score of 2, and an additional two with a score of 1. Thirty-three patients were identified as having nonsignificant fibrosis, including three false negatives: two with a HAI fibrosis score of 3 and one with a score of 5. There was no correlation between age or BMI and MBT scores for patients with the same histological score.
Applying the same proprietary algorithm developed to differentiate significant from nonsignificant fibrosis on the healthy volunteer group combined with the significant fibrosis group (n = 150), 67% of the tested subjects (n = 98) would get an answer (Fig. 4). This algorithm achieved an AUC of 0.92, with a sensitivity of 91% and a specificity of 88%, a PPV of 79% and NPV of 95%. Thirtyeight subjects were identified as having significant fibrosis, including eight false positives. Sixty subjects were identified as having nonsignificant fibrosis including three false negatives; two with a HAI fibrosis score of 3 and one with a score of 5.
Combination of inflammation and fibrosis algorithms. Applying the described inflammation algorithm on the subset of patients analysed by the fibrosis algorithm (67% of the initial population), resulted in an area under the ROC of 0.89. When the same threshold was used, sensitivity and specificity were 83 and 81% respectively, with PPV and NPV of 91 and 68%, respectively (Fig. 5).
Applying the described inflammation algorithm on the subset of patients not analysed by the fibrosis algorithm (33% of the initial population), resulted in an area under the ROC of 0.96. When the same threshold was used, sensitivity and specificity were 82 and 91%, respectively, with PPV and NPV of 95 and 71%, respectively (Fig. 6).

DISCUSSION
We assessed the ability of the noninvasive, online, continuous 13 C-MBT in the detection of significant fibrosis and inflammation in patients with HCV. In a cohort of 100 consecutive HCV patients with normal ALT, breath-test parameters correlated with the level of fibrosis and degree of inflammation as indicated by the modified Ishak HAI fibrosis and inflammation scores. The breath test accurately differentiated low and high inflammation (£4 and >4, 83%). The MBT achieved a 90% diagnostic accuracy in differentiating patients with a modified Ishak HAI fibrosis score £2 and >2 while 67% of the biopsies could have been avoided by replacing the assessment with the MBT alone.
Methacetin breath testing has been correlated with fibrosis and overall liver function [16]. Traditionally, this testing was performed using isotopic ratio mass spectrometry, the gold standard for MBT. However, a recent study found that a measurement method with continuous automatic molecular correlation spectroscopy showed a high correlation with mass spectroscopy [17]. In addition to being less cumbersome, the continuous system has an inherent advantage over mass spectrometry in its ability to identify the PDR peak and PDR peak time, which are often missed when noncontinuous measurement is used. Furthermore, being fully automatic and using an internal capnograph, the system mitigates the risk of potential human errors and ensures that the appropriate part of the breath sample is collected. Several noninvasive methods have been explored as tools to assess the degree of liver fibrosis in chronic HCV patients, and some were also evaluated in patients with normal ALT. These include a combination of serum tests such as the AST⁄ALT ratio [18] or the AST⁄platelet ratio index [19]. To date, two methods appear to be the most studied with the goal of using them to supersede liver biopsy in the assessment of liver fibrosis. The first is a patented artificial intelligence algorithm (Fibrotest Ò ; BioPredictive, Paris, France) [20,21]. The second is a technique to measure in vivo liver elasticity, based on one-dimensional transient elastography (Fibroscan Ò , EchoSens, Paris, France) [22,23].
The Fibrotest Ò requires a blood sample and specialized laboratory, which in turn translates into a lag time of several days between test and result. Estimation of liver fibrosis by Fibrotest Ò uses five parameters that were chosen by logistic regression applied to a selection of basic serum biochemical markers, with histological staging as the independent variable [20]. Mean ALT values were three times the upper limit of the reference range for males, and only 13% of the studied patients had ALT within the normal range. Biochemical markers were measured once on the day of biopsy, but because ALT fluctuates widely during the course of chronic  HCV infection, it is likely that only a few, if any, were HCV carriers with NALT. In a subsequent Fibrotest Ò prospective validation study, participants needed documented elevated serum ALT levels (at least 1.5 times the upper limit of normal) on three occasions within 6 months before enrolment. There have been very few independent studies using Fibrotest Ò . In addition to inter-laboratory variations, these studies have shown that in about 15-20% of patients, significant fibrosis could be missed or conversely, significant fibrosis could be diagnosed in the presence of minimal or no fibrosis [24]. In patients with Gilbert syndrome, or any acute inflammation with high haptoglobin values, higher falsepositive and false-negative rates were found. In a recent study of 40 patients with NALT, Fibrotest Ò had an accuracy of only 43%, with a sensitivity and specificity of only 64 and 31%, respectively (11). Both Fibrotest Ò (measuring fibrosis stage in patients with chronic HCV or HBV) and ActiTest Ò (measuring necroinflammatory activity in patients with chronic HCV or HBV) are dependent on inter-laboratory variability of biochemical markers [25].
The Fibroscan Ò provides a noninvasive method for assessing liver fibrosis but does not give information regarding inflammation. In addition, it can be difficult to administer and may produce imprecise results in obese patients. Measurement of liver elasticity so far has been precluded by technical limitations and costs. With Fibroscan Ò , a transmitted elastic wave can be temporally separated from reflected elastic waves, making the technique less sensitive to those boundary conditions (including body fat) that tend to induce artefacts [22].
All the currently used noninvasive methods have a diagnostic accuracy that does not exceed 80-85% [26][27][28][29]. Thus, many patients still require a liver biopsy, and in those classified without one, misdiagnosis is expected in at least 15-20% [30]. The ability of the MBT to accurately assess the degree of intrahepatic inflammation and fibrosis in patients with NALT may add to its value in decision making for these patients.
The clinical management of chronic HCV infection is based on both patient and viral characteristics, and a liver biopsy is often required to guide therapeutic decision making. Paradoxically, patients with NALT, in whom liver biopsy is particularly useful, are more reluctant to undergo one. An attempt to increase the diagnostic performance of noninvasive markers of liver fibrosis by combining them in sequential algorithms was recently suggested. Recently 190 patients with chronic HCV were evaluated for ASTto-platelets ratio, FornsÕ index and Fibrotest Ò results at the time of liver biopsy, and stepwise combination algorithms were developed and validated prospectively in 100 additional patients. The data suggested that a stepwise combination of noninvasive markers of liver fibrosis improves diagnostic performance in chronic HCV, reducing the need for a liver biopsy [31]. The data of the present study show that by using an algorithm that includes breath-test  parameters, age and other patient data, 67% of liver biopsies performed in the patient group could have been avoided. This algorithm achieved an AUC of 0.92, with a sensitivity of 91% and a specificity of 88%. As novel therapies for liver fibrosis evolve, noninvasive measurement of liver fibrosis will be required to help manage patients with chronic liver disease. The BreathID Ò holds several advantages as a noninvasive tool in this setting, including not being limited by patient BMI or other patient characteristics, such as the presence of Gilbert syndrome or acute inflammatory condition. The test provides information on both fibrosis and inflammation. Future studies will determine its correlation with the functional hepatic mass and hepatic reserve along with the clinical course in these patients.
The results of the current study suggest that the continuous BreathID Ò 13 C MBT is an accurate tool for identification of liver inflammation and fibrosis in patients with chronic HCV infection and normal ALT levels, and that its use can avoid the need for a liver biopsy in two-thirds of these patients. As such, it may prove to be a powerful, noninvasive alternative for decision making in the management of this patient population.