Accuracy of Liver Fibrosis Degree Based on King ’ s Score to Fibroscan in Chronic Hepatitis B

Background: A great interest has been dedicated to the development of non-invasive predictive models in age (years) x aspartate aminotransferase/AST (IU/L) x [international normalized ratio (INR)/platelets (109/L)]. Hepatitis B. Method: Fibroscan in Division of Gastroenterology and Hepatology at Haji Adam Malik Hospital, Medan. Serum

Accuracy of Liver Fibrosis Degree Based on King's Score to FibroScan in Chronic Hepatitis B

Kata kunci:
INTRODUCTION Chronic liver disease was involving a progressive destructive and regenerative process that begins and hepatocellular carcinoma.Liver fibrosis was caused by chronic damage in liver tissue, related to abundant accumulation of extracellular matrix (ECM) protein. 1,2nfection of Hepatitis B and C virus, alcoholism, autoimmune disease, cholestatic disease, and nonalcoholic steatohepatitis (NASH).Accumulation of ECM protein will damage liver architecture by of liver nodule.A well-formed nodule stage was also known as liver cirrhosis. 2ealing process in response to chronic liver tissue injury.Liver cirrhosis detection and staging were important in chronic hepatitis patient management. 2 Chronic Hepatitis B was the most common infection in chronic hepatitis worldwide.A predictive model patients by Shanghai Liver Fibrosis Group (SLFG), Hui et al, and Mohamadnejad et al. 3,4 However, only several of models above that have been validated and implemented in clinical situation. 3,4oblem that lead to high morbidity and mortality in needed to determine the suitable early treatment. 5Liver biopsy as an invasive method was still become the problem was that clinical presentation of the disease was vary among patients and not always similar to its degree, besides that several patients also refused biopsy procedure. 5Limitation in biopsy procedure was variation among biopsy result of both intra and interobserver, and also a sampling error.There were a lot of research to develop any noninvasive predictive model that strongly correlate to invasively. 4,5,8This technique was known as ultrasound elastography, commercially known as Fibroscan.This new imaging technique has been proven to assess liver fibrosis degree with high accuracy. 4,8routine test in clinical practice.This instrument kilopascal unit (kPa).4,8 stages.This diagnostic accuracy of FibroScan was also higher compared to several biochemistry markers examination. Thevantage of FibroScan was that it is done in a short time periode, without any pain, and less misinterpretation risk compared to liver biopsy.4,7,8,9 There has been previous study to investigate the in compared to Fibroscan in chronic Hepatitis C patients.10   To measure its diagnostic value, receiving operating characteristic (ROC) curve was used to assess its value (PPV), negative predictive value (NPV), diagnostic accuracy (DA), positive likelihood ratio (LR+), and negative likelihood ratio (LR-) based on cut off value in its original journal.Analysis was done using SPSS ver 15 software.

RESULTS
Overall, a total of 62 chronic Hepatitis B patients was involved in this study.Clinical characteristic, Table 1.Mean patient age was 46 years, consist of 39 male patients (62,9%) and 23 female patients (37,1%).All patients was not in decompensated liver cirrhosis condition.Table 1 showed a lowest and highest level of platelet count (58.000/mm 3 and 417.000/mm 3 , respectively), AST level (14 and 124, respectively), and INR (0,64 and 2,62 IU/L, respectively).The FibroScan result showed a lowest and highest score of 3,8 kPa and 67,8 kPa, while King's score showed 0,7 and 88,2, 17,7% from all patients.Based on Kolmogorov-Smirnov normality test, age and platelet count were normally distributed (mean, SD), while other data was nor normally distributed was based on Fibroscan, presented as number and percentage.
Cut off value of King's score and its formula was based on its original investigation (Cross et al). 10 > 16,7 > 12,2.Predictive value of this model in identifying patients was shown in Table 2.Among 27 patients that (48,1%) showed a King's score higher than 12,2.With Cirrhosis cut off value was > 16,7.Five patients (83,3%) that have King's score more than 16,7 from total of six patients was diagnosed to have liver cirrhosis using FibroScan.With King's score more than develop liver cirrhosis.King's score diagnostic value was further evaluated using area under ROC curve (AUROC), LR (+), LR (-), and diagnostic accuracy.In 4).Form this result, although only consist of routine laboratory examination, King's score was accurate and well-predictive in cirrhosis prediction.and accuracy value of this predictive model was shown in Table 2.In Table 2, King's score showed a high predictive value to predict liver cirrhosis, and also a high sensitivity, high NPV, and low LR (-) so that it has lower false negative incidence.Result in King's 76,5% PPV, 68,9% NPV, 0,54 LR (+), 0,53 LR (-), chronic Hepatitis B patients.King's score also showed NPV, 0,98 LR (+), 0,96 LR (-), and 85,48% accuracy to identify cirrhosis in chronic Hepatitis B patients.

DISCUSSION
There were several previous non-invasive diagnostic model in chronic Hepatitis that have been published.Most of those diagnostic model was used in chronic Hepatitis C patients and only a few in chronic Hepatitis B patients. 11Although two last study of FibroTest in chronic Hepatitis B patients showed a 0,77 and 0,78 in AUROC value, those model was using unroutine laboratory examination, such as haptoglobin, A2M, and apolipoprotein A1.Its complexity and cost reduce its utility in daily practice. 12everal predictive model that design for chronic Hepatitis B patients was already proposed, but this study was still have several unique factors.First, SLFG model was designed and validate only for HBeAg positive patients with ALT between 2 and 10 times from upper limit of normal (ULN), while Mohamadnejad et al proposed a formula that only suitable for HBeAg negative patients. 11Hui et al only recruited patients with HBV DNA > 10 5 copy/mL and ALT between 1,5 and 10 from upper limit of normal.In current study, patients recruited was all patients with chronic Hepatitis B without considering its treatment, HBeAg level, ALT, and HBV DNA level.Therefore, this study result was more applicable for chronic Hepatitis B patients with wider coverage. 10,12,13econd, this predictive model of King's score was based in routine laboratory examination.Platelet count, AST, and INR was a routine test that available widely in clinical settings, so there were no need of extra laboratory examination. 10,11,12,13Previous study by Kun Zhou et al showed that accuracy of diagnostic test such as HA and A2M showed that SLFG model and Hepascore was better in identify significant was insignificant for advanced fibrosis and liver cirrhosis. 12,14,15,16,17ocess.Otherwise, unroutine test in daily practice was reduce its utility for standardization, validation, 14 Third, King's score was easy to calculate.Several previous model, except APRI, consist of complex formula that need a calculator in algoritm calculation. 9,12his simplicity of King's score and APRI showed that it is clinically easier to be applied.But, APRI that previously investigated in chronic Hepatitis C patients having one of its indicator, AST, to be proven B patients, according to Kun Zhou et al. 12 This was the explanation why APRI has a lower AUROC compared to King's score. 9,12n the other hand, there were several limitations in this study.In this study, not all patients was undergo on FibroScan (transient elastography), although, not shown in this report, ten patients was undergo liver biopsy with similar result to FibroScan result. 8In Kun Zhou et al, liver biopsy is one of their limitation too by explaining that liver biopsy was not the gold standard because of the possibility in sampling error and observer variability.A prospective analysis result was also claimed that the risk of liver biopsy failure was > 7 times higher than any diagnostic marker. 8,12,18To reduce variability and subjectivity, the use of laparoscopic biopsy, FibroScan, or validating non-invasive test could help the reliability of liver biopsy as gold standard.Other limitation is that this study result was validate in the same population and in small coverage population.
In Asia countries, chronic Hepatitis B was the majority of all chronic liver disease.The main purpose identify eligble patients for antiviral therapy. 19Based Management, liver biopsy was indicated for patients aged > 40 years with ALT < 2x ULN and HBV DNA > 20.000 IU/mL (HBeAg-positive) or > 2000 IU/mL was the main candidate for antiviral therapy.Based on that guideline, if a candidate have been undergo FibroScan, liver biopsy could be skipped.In patients with normal ALT and FibroScan result < 6,0 kPa, no therapy was indicated, < 7,5 kPa was observed, and > 12 kPa should be considered to received therapy. 4,15,19lthough most of non-invasive predictive model could not accurately assess liver fibrosis degree Accuracy of Liver Fibrosis Degree Based on King's Score to FibroScan in Chronic Hepatitis B those predictive model was having a good accurate predictive model was to reduce liver biopsy procedure but not for totally replace liver biopsy. 11,12,13,14Using optimized cut-off core form King's score, it is expected that further livver biopsy needs wa reduced.Furthermore, a combination of several predictive model and invasive diagnostic technique could bring this diagnostic tools into higher level. 10Combination of both FibroScan and King's score was an intersting method in chronic Hepatitis B patients management.
But it has been known that, in daily clinical practice, priority should be given into model that have been validate in large scale study, because of its accuracy could be interfered by different etiologies, populations, and methods.

Figure 4 .
Figure 4. ROC curve of King's score in predicting liver cirrhosis in chronic Hepatitis B patients Medan, and several gastroenterology clinics in Medan.With known proportion of chronic Hepatitis B patients in Indonesia was 0,36 while in this study was approximately 0,11, the sample size of this study was 39 patients.Inclusion criteria was: (1) Male or female aged > 18 years; (2) Chronic Hepatitis B patients with positive viral marker; and (3) Willing to participate by signing informed concent letter.Exclusion criteria was: (1) C virus (HCV) coinfection, alcohol consumption of > 30 g/day; and (2) Other cause of chronic Hepatitis, decompensated liver cirrhosis, and chronic kidney failure patients.Laboratory examination was done in Clinical Pathology Laboratorium in Haji Adam Malik Hospital, Medan.King's score was a non-invasive examination using variable: age, platelet count, aspartate aminotransferase (AST) level, and international normalized ratio (INR).The formula of this score is:

Table 3 . Predictive value of King's score model as predictive model of liver cirrhosis in chronic Hepatitis B patients.
AUROC: area under the ROC curves.
10,12CONCLUSIONKing's score has the ability to predict cirrhosis high accuracy, so that patients with King's score > 16,7 did not need to undergo liver biopsy.Otherwise, for result.King's score as non-invasive predictive model for liver cirrhosis in chronic Hepatitis B patients have a high accuracy.A further study was needed to validate this result in a larger scale with different population.