Mortality among Persons in Care with Hepatitis C Virus Infection -- The Chronic Hepatitis Cohort Study (CHeCS), 2006-2010: liver disease deaths 12 times higher

Provided by NATAP

Download the PDF here

"The age-adjusted mortality rate for liver disease in CHeCS was twelve times higher than the MCOD rate......mean age of death was 59 years, 15 years younger than MCOD deaths......only 19% of all CHeCS decedents and only 30% of those with recorded liver disease had HCV listed on their death certificates.....This effect was seen with extra-hepatic causes as well"

Reena Mahajan1, Jian Xing1, Stephen J. Liu1, Kathleen N. Ly1, Anne C. Moorman1, Loralee Rupp2, Fujie Xu1, Scott D. Holmberg1, for the Chronic Hepatitis Cohort Study (CHeCS) Investigators*
1Centers for Disease Control and Prevention, Atlanta, Georgia
2Henry Ford Health System, Detroit, Michigan, USA

Summary: Using US health system data from an observational cohort study, HCV is under-documented on death certificates. Only 19% of those with known HCV infection had HCV listed on their death certificate although two-thirds had pre-mortem indications of chronic liver disease......Disease-specific, liver- and non-liver-related, mortality for HCV-infected patients in an observational cohort study, the Chronic Hepatitis Cohort Study (CHeCS) at four US health care systems, were compared with Multiple Cause of Death (MCOD) data in 12 million death certificates in 2006-2010......The age-adjusted mortality rate for liver disease in CHeCS was twelve times higher than the MCOD rate......only 19% of all CHeCS decedents and only 30% of those with recorded liver disease had HCV listed on their death certificates.

"In Table 4 just below you can see a 15 year difference in age at death between those in this study with HCV & the general population 59 vs 74 yrs old......The mortality rate estimated from this analysis was twelve times higher than the general population; this is much higher than the two-to-five times higher mortality rates in HCV-infected vs uninfected persons seen in other studies...... This effect was seen with extra-hepatic causes as well.....In summary, our analysis of a known HCV-infected cohort demonstrates that less than one-fifth of deaths in HCV-infected persons are coded as having HCV; this indicates a significant underestimation of the number of deaths among people with HCV and the true medical and public health impact of HCV. In this analysis, we have tried to be clear about the difference between dying with HCV and dying from HCV, but both represent a substantial public health burden. For purposes of public health, policy planning, disease modeling, and medical care, this is a huge burden that should be reported and hopefully spur public health action as curative, all-oral therapies are becoming available to treat HCV. Addressing the true impact of HCV, including of those chronically infected with HCV who are not utilizing health services, will be essential to appropriately respond to this epidemic. 35"

"whether the death was considered HCV or non-HCV-related, mean age of death in HCV-infected persons (59 years) was 15 years younger than for all cause mortality in the general population (74 years), a finding similar to previous research. 28......Despite confirmed chronic HCV infection, only 19% (306/1,590) had HCV infection listed on the death certificate."

---------------------------------------
Mortality among Persons in Care with Hepatitis C Virus Infection--The Chronic Hepatitis Cohort Study (CHeCS), 2006-2010

Abstract

Background. Numbers of deaths in hepatitis C virus (HCV)-infected persons recorded on US death certificates have been increasing, but actual rates and causes of death in them have not been well elucidated.

Methods. Disease-specific, liver- and non-liver-related, mortality for HCV-infected patients in an observational cohort study, the Chronic Hepatitis Cohort Study (CHeCS) at four US health care systems, were compared with Multiple Cause of Death (MCOD) data in 12 million death certificates in 2006-2010. Pre-mortem diagnoses, liver biopsies, and FIB-4 scores (a non-invasive measure of liver damage) were examined.

Results. Of 2,143,369 adult patients seen at CHeCS sites in 2006-2010, 11,703 (0.5%) had diagnosed chronic HCV infection, and 1,590 (14%) died. CHeCS decedents were born from 1945-1965 (75%), white (50%), and male (68%); mean age of death was 59 years, 15 years younger than MCOD deaths. The age-adjusted mortality rate for liver disease in CHeCS was twelve times higher than the MCOD rate. Before death, 63% had medical record evidence of chronic liver disease, 76% had elevated FIB-4 scores, and of those biopsied 70% had moderate or worse liver fibrosis. However, only 19% of all CHeCS decedents and only 30% of those with recorded liver disease had HCV listed on their death certificates. Conclusions. HCV infection is greatly under-documented on death certificates. The 16,622 persons with HCV listed in 2010 may represent only one-fifth of about 80,000 HCV-infected persons dying that year, at least two-thirds of whom (53,000 patients) would have pre-mortem indications of chronic liver disease.

Excerpts

Of deaths from 1999-2007 with HCV infection listed as a primary or underlying cause of death, 57% had chronic liver disease as a cause 2 but many also had extra-hepatic manifestations.4 Presence of HCV may potentially accelerate the disease process in heart disease, 5-7 diabetes, 8,9 various malignancies, 10,11 and genitourinary conditions. 12,13

In this study, we looked at hepatic and extra-hepatic causes of death for a cohort of hepatitis C-infected patients in care in the United States.

The age-adjusted CHeCS mortality rate in these HCV-infected patients was 12,854 per 100,000 persons compared to the MCOD mortality rate of 1,046 per 100,000 persons.\

The age-adjusted CHeCS mortality rate in these HCV-infected patients was 12,854 per 100,000 persons compared to the MCOD mortality rate of 1,046 per 100,000 persons. The age-adjusted mortality rate for all liver disease categories in CHeCS was higher than the national MCOD rate; for example, alcohol-related liver disease was 6 times greater while other hepatitis was 86 times higher than the national rate (Table 2). The most frequently cited cause of death was non-alcohol related liver disease; it was listed for 32% of all deaths (513 /1,590). The CHeCS mortality rate among persons with non-alcohol related liver disease was 669 vs. 51 per 100,000 for alcohol-related liver disease. In addition, all age adjusted death rates from extra-hepatic causes were higher than the national MCOD rates; the highest rates were seen with genitourinary causes, mental/behavioral disorders, and diabetes (Table 2). Of 1,590 deaths, only 12 (1%) were due to intentional self-harm, 93 (6%) due to sepsis due to any cause, and none were due to overdose/poisonings.

Despite confirmed chronic HCV infection, only 19% (306/1,590) had HCV infection listed on the death certificate. HCV was not listed for the majority of deaths across disease categories whether liver or non-liver related. The total number of deaths listed as liver-related was 47% (752/1,590), but of these, only 41% (306/752) had HCV listed as a cause of death. By type of liver-related death, HCV was listed for 36% (183/513) of those with non-alcohol related liver disease, 27% (13/49) of those with alcohol-related liver disease, 31% (53/169) of those with liver cancer, and only 9% (6/70) of those with ÒotherÓ hepatitis (Figure; Table 3). Even among 156 (10%) of decedents who had a liver transplant before death, HCV was only listed on 46 (29%) of death certificates.

Discussion:

Data from this study suggests a much greater role of HCV on mortality in the United States than has been previously understood based on analyses of death certificate data. The data in this paper document and contradict prevalent views that, perhaps because of its long incubation period (30 years), HCV infection is an indolent infection that is not of urgent concern. Originally intended as a study of causes of death in approximately 1,600 well characterized decedent HCV patients in the CHeCS, we found that only 19% had HCV listed on their death certificates, and only in 30% of death certificates in which liver disease had been noted. Even among the 156 HCV-infected CHeCS patients who had a liver transplant before death, only 46 (29%) had had HCV noted on their death certificate. As there were 16,622 death certificates in the US listing HCV as an underlying or contributing cause of death in 2010, we extrapolate that only one-fifth of those with HCV who die are having HCV recorded on their death certificates. Thus, our analysis suggests that at least 80,000 persons with HCV may have actually died in 2010. Given that 63% of the well-characterized CHeCS patients had medical records (ICD 9 codes) indicating pre-mortem liver disease--and 76% had FIB4 scores indicative of substantial or more liver damage-this suggests that total US deaths contributed to by HCV total at least 53,000.

Our results may be a conservative estimate as recent studies indicate that only about half of all HCV-infected persons have been diagnosed with the infection. 20-22 Further, approximately 50% of all deaths in those with known HCV had liver disease listed on their death certificates. Thus, even if we exclude other diseases associated with HCV infection such as diabetes and non-Hodgkin lymphoma, 23-26 it appears that most are dying not just with HCV but in possibly from HCV. These considerations are especially important because identifying and treating HCV patients in an era of rapidly evolving and effective, curative therapies could have a major public health impact.

Often, the high mortality and burden from HCV infection are minimized because other non-HCVrelated causes of death are considered to be more proximal or immediate reasons. For example, in a recent survey of New York resident physicians, over-documentation of cardiopulmonary causes of death and other inaccuracies-- both knowing and unavoidable--were reported; those surveyed believe that the current cause-of-death reporting system is generally inaccurate.27 This study also indicates that in the HCV-infected population over 70% had pre-mortem liver disease by ICD-9-CM electronic hospital record coding, liver biopsy, or FIB-4 score. So, in addition to under-recording HCV infection, even verified pre-mortem liver disease is also under-recorded, Further, whether the death was considered HCV or non-HCV-related, mean age of death in HCV-infected persons (59 years) was 15 years younger than for all cause mortality in the general population (74 years), a finding similar to previous research. 28

The mortality rate estimated from this analysis was twelve times higher than the general population; this is much higher than the two-to-five times higher mortality rates in HCV-infected vs uninfected persons seen in other studies. 29-32 Even with significant underreporting, persons who died in our cohort with non-alcohol related liver disease had 24 times the risk of death and those with liver cancer had almost 29 times the risk of death compared to over 12 million deaths in the age-matched general population. This effect was seen with extra-hepatic causes as well: compared to the general population, cases had three times the rate of injuries and genitourinary causes of death, ten times the rate of HIV, and twice the rate of mental/behavioral disorders. Other researchers have attributed higher rates of injuries/trauma as well as mental/behavioral disorders to lifestyle factors, including a previous history of substance abuse. 7,33 However, results from our death certificate data show that only 1% of CHeCS patients had suicide listed, 6% had sepsis, and none died of overdose or poisoning.

Our data represent findings from four health care systems in the United States and thus have a number of limitations. While two sites have transplant centers associated with them, we cannot measure how many patients are ÒattractedÓ to these medical centers because they have tertiary care facilities vs the fact that they are in the catchment area of these large integrated health systems. However, as only a minority (10%) of the CHECS decedents were seen at the transplant centers, these patients do not affect the overall picture. Level of care provided at a particular site should not affect the low rate of death certificate recordings of HCV (29%). Due to the variability in the definition of ICD-10 mortality codes used for chronic liver disease, we compared the codes that we used for our definition of chronic liver disease with a previously established definition. We found that for both definitions, 46% of cases were identified as having liver-related causes of death.34 This concordance further substantiates our findings.

An additional limitation is the use of the FIB-4 index as a measure of chronic liver disease; although validated, changes to liver enzymes and platelet count may be affected by non-liver related conditions such as infection or malignancy. However, the overlap between liver disease defined by ICD-9-CM codes and FIB-4 scores correlates with the findings using the ICD-10 mortality codes indicating that there is truly underlying liver disease in patients dying with and from HCV.

In summary, our analysis of a known HCV-infected cohort demonstrates that less than one-fifth of deaths in HCV-infected persons are coded as having HCV; this indicates a significant underestimation of the number of deaths among people with HCV and the true medical and public health impact of HCV. In this analysis, we have tried to be clear about the difference between dying with HCV and dying from HCV, but both represent a substantial public health burden. For purposes of public health, policy planning, disease modeling, and medical care, this is a huge burden that should be reported and hopefully spur public health action as curative, all-oral therapies are becoming available to treat HCV. Addressing the true impact of HCV, including of those chronically infected with HCV who are not utilizing health services, will be essential to appropriately respond to this epidemic. 35

Source

Determining the Effect of Hepatitis C on Mortality: Sorting the Signal From the Noise - Editorial

Provided by NATAP

Download the PDF here

Clinical Infectious Diseases Advance Access published February 12, 2014 Keith M. Rose Mount Sinai Health System St. Luke's Roosevelt Hospital

Hepatitis C is a major cause of morbidity and mortality estimated to affect over 150 to 200 million people worldwide. 1,2,3 Infection with hepatitis C virus (HCV) carries a large clinical impact and high cost burden to health care systems, and is one of the leading contributing causes of end stage liver disease requiring liver transplantation in the US. 4 Advancements have been made in areas of both diagnosis and treatment, improving our ability to detect and treat the disease earlier. This carries the potential benefit of decreasing the morbidity and mortality caused by this disease.
This month's issue of Clinical infectious Diseases features two articles which address the impact of hepatitis C, highlighting both the significant mortality it brings as well as the potential underreporting of the disease. Both of these studies utilize data collected from death certificates and either disease reporting/ surveillance systems, or electronic medical records.

The study by Pinchoff et al entitled "Death among people with Hepatitis C in New York City, 2000-2011'"examined surveillance data for Hepatitis C reporting, and compared it to cause of death data obtained from death certificates from 2000 to 2011 in New York City (NYC). They evaluated the effect of Hepatitis C on age of death, cause of death (COD), as well as the effect of co-infection with HIV compared to the population without these diseases. This was a well designed study taking advantage of New York City having several robust disease surveillance registries that were able to be cross-matched and then compared to mortality data. By doing the study in NYC, they were able to evaluate a large, well defined, and diverse population with a particularly high incidence of this disease.5 This study adds to the literature as it further helps to delineate the natural history of Hepatitis C in the real world.The authors were able to convincingly demonstrate an increased risk of premature mortality (age <65) in patients infected with hepatitis C, stressing the importance again of early identification and potential treatment of this disease.The study also attempted to further evaluate the cause of death in this population from a review of death certificate data. It is very important to understand when and how people are being diagnosed with the disease and ultimately what they are specifically dying from.

There are however some weaknesses in the study that need to be discussed. Unlike a true cohort, this study only captured people who died in NYC. Patients who were diagnosed and treated in NYC would not be included in the COD analysis if they were to die outside of the city limits. Although the study implies that earlier diagnosis and treatment would likely decrease premature mortality, the study was not able to evaluate the subset of people that were treated for hepatitis C, and with what regimen. Overall the study does an excellent job demonstrating the associated mortality with HCV, but causation is much more difficult to prove. Lastly, the utilization of death certificate data may lead to a classification bias with overrepresentation of certain CODs ie HIV/AIDS and cardiovascular causes. 6,7,8

Also in this issue of Clinical Infectious Diseases, Mahajan et al publish their findings from The Chronic Hepatitis Cohort Study (CHeCS). This large cohort of 11,703 patients was formed from a review of electronic medical records from four large healthcare systems from 2006-2010, extracting demographics and data from patients diagnosed with HCV. They found a significantly higher than expected effect of hepatitis C on mortality, with a mortality rate twelve times higher in their cohort compared to the general population. An interesting finding of this study, was the paucity of HCV being listed on death certificates of these individuals. In fact, the majority of deaths, whether liver or non-liver related, did not have hepatitis C listed as a COD. Although this finding was specifically for hepatitis C, it reiterates the importance of proper death certificate completion and accuracy. It also sheds light on the potential impact on other studies such as the article by Pinchoff et al above that utilize death certificate data, to determine the impact on mortality of a certain disease. Potentially the Pinchoff study may be missing large numbers of patient s that died from both hepatic and non-hepatic causes related to Hepatitis C. Clearly those that had hepatitis C listed on their death certificates tended to died younger with a higher incidence of premature death, but the full impact is uncertain if we don't have an accurate representation of the deaths that HCV may have contributed to.

This paper by Mahajan et al also has some limitations, and opens the door for many further questions. This study was done with data from four large academic centers and may not necessarily be extrapolated to all hospital systems. Furthermore it would be interesting to know who completed the death certificates, ie attending versus resident or intern, how often it was completed by a covering physician who may not have known the patient as well, and what specific training, if any the completing physician had received in proper death certificate completion. Lastly it would be useful to know what percentage of patients were being treated for Hepatitis C and the impact this may have had on this subset of the cohort.

It's important to note that both of the above studies are only able to take into account HCV cases that were picked up by surveillance. HCV is often under diagnosed, so we don't know if the findings from these studies would apply to all subjects infected with HCV. We can only address the mortality risk, and associations in those with an established diagnosis of hepatitis C. Many of these cases may have been detected because of liver function test abnormalities or exam findings consistent with liver disease. It's important to also consider that hepatitis C infection is associated with several social factors and behaviors that can increase the risk of death. This comes back to an important question of whether people are dying with hepatitis C or from hepatitis C. To this point, it is also possible that some of the controls who died in both studies may have had HCV, as not everyone who dies is checked for the disease.

As more studies are done to look at the effect of disease, and ultimately determine what Americans or other populations are dying from, the importance of accurate information on death certificates becomes paramount. Based on their data, Mahajan et al propose that 80,000 Americans died with Hepatitis C in 2010 instead of the reported 16,622 secondary to gross underreporting on death certificates. This huge disparity is unlikely unique to Hepatitis C. Wexelman et al in a survey of residents in NYC found that the majority of residents in NYC felt the death certificate reporting system was inaccurate and often knowingly listed inaccurate CODinformation on death certificates.9 Other studies have also documented the general inaccuracies of death certificates, particularly noting cardiovascular disease being overrepresented. 6,7,8 As less autopsies are being performed, it is becoming more and more important to strive to improve the accuracy and consistency of cause of death reporting.

Increasing education initiatives seem to help, 10 and perhaps with more of these programs we may be moving toward a more reliable system, but this is not the only obstacle. There are system based issues which also need to be further addressed.9 For instance, in the Wexelman et al study, residents reported not being able to enter into the Electronic Death Registration System (EDRS) system what they believed was the true COD for a variety of reasons. These issues seem to be multi-factorial and will ultimately need to be addressed on several levels. However, perhaps it is only when you look at studies like these two well designed, interesting papers on hepatitis C, do you really appreciate the importance of getting this right.

The author has no reported conflicts of interest.

Source

The Risk of Long-term Morbidity and Mortality in Patients With Chronic Hepatitis C

JAMA Internal Medicine

February 2014, Vol 174, No. 2

Original Investigation | February 2014

Results From an Analysis of Data From a Department of Veterans Affairs Clinical Registry

Jeffrey McCombs, PhD¹; Tara Matsuda, BA^1,2; Ivy Tonnu-Mihara, PharmD²; Sammy Saab, MD³; Patricia Hines, BA⁴; Gilbert L’Italien, PhD⁴; Timothy Juday, PhD⁴; Yong Yuan, PhD⁴

[+] Author Affiliations

JAMA Intern Med. 2014;174(2):204-212. doi:10.1001/jamainternmed.2013.12505.

Abstract

Importance The impact of viral load suppression, genotype, race, and other factors on the risk of late-stage liver-related events in patients with hepatitis C (HCV) has been assessed previously using data from small observational cohorts or clinical trials. Data from large real-world practice samples are needed to improve risk factor estimates for late-stage liver events and death in HCV.

Objective To describe the natural history of HCV in real-world clinical practice.

Design, Setting, and Participants Observational cohort study. Patients with a detectable viral load (>25 IU/mL) and a recorded baseline genotype were selected from the Veterans Affairs (VA) HCV clinical registry (CCR), which compiles electronic medical records data from 1999 to present.

Exposures Risk factors included genotype, race, age, sex, and time to achieving an observed undetected viral load.

Main Outcomes and Measures The primary outcomes were time to death and time to a composite of liver-related clinical events. Secondary outcomes included the components of the composite clinical outcome. Outcomes were measured using a time-to-event format and were analyzed using Cox proportional hazards models.

Results A total of 28 769 of 360 857 unique HCV CCR patients met all study criteria. Only 24.3% of patients received treatment, and 16.4% of treated patients (4.0% of all patients) achieved an undetectable viral load. The unadjusted death rates were 6.8 (95% CI, 6.0-7.7) per 1000 person-years for patients who achieved viral load suppression vs 21.8 (95% CI, 21.5-22.2) deaths per 1000 person-years in patients who did not achieve this goal. Cox model results found that achieving viral suppression reduced risk of the composite clinical end point by 27% (hazard ratio [HR], 0.73 [95% CI, 0.66-0.82]) and the risk of death by 45% (HR, 0.55 [95% CI, 0.47-0.64]). Genotype 2 patients were at significantly lower risk, and genotype 3 patients were at higher risk for all study outcomes relative to genotype 1. Black patients were at lower risk for all liver events than white patients.

Conclusion and Relevance Achieving an undetectable viral load was associated with decreased hepatic morbidity and mortality. It remains to be determined whether newer treatment regimens can offer higher response rates with fewer adverse effects in real-world settings.

Source

A MELD-Based Model to Determine Risk of Mortality Among Patients With Acute Variceal Bleeding

Gastroenterology. 2014 Feb;146(2):412-419.e3. doi: 10.1053/j.gastro.2013.10.018. Epub 2013 Oct 19.

Reverter E¹, Tandon P², Augustin S³, Turon F⁴, Casu S⁴, Bastiampillai R², Keough A², Llop E⁴, González A³, Seijo S⁴, Berzigotti A⁴, Ma M², Genescà J⁵, Bosch J¹, García-Pagán JC¹, Abraldes JG⁶.

Abstract

BACKGROUND & AIMS: Patients with cirrhosis with acute variceal bleeding (AVB) have high mortality rates (15%-20%). Previously described models are seldom used to determine prognoses of these patients, partially because they have not been validated externally and because they include subjective variables, such as bleeding during endoscopy and Child-Pugh score, which are evaluated inconsistently. We aimed to improve determination of risk for patients with AVB.

METHODS: We analyzed data collected from 178 patients with cirrhosis (Child-Pugh scores of A, B, and C: 15%, 57%, and 28%, respectively) and esophageal AVB who received standard therapy from 2007 through 2010. We tested the performance (discrimination and calibration) of previously described models, including the model for end-stage liver disease (MELD), and developed a new MELD calibration to predict the mortality of patients within 6 weeks of presentation with AVB. MELD-based predictions were validated in cohorts of patients from Canada (n = 240) and Spain (n = 221).

RESULTS: Among study subjects, the 6-week mortality rate was 16%. MELD was the best model in terms of discrimination; it was recalibrated to predict the 6-week mortality rate with logistic regression (logit, -5.312 + 0.207 • MELD; bootstrapped R(2), 0.3295). MELD values of 19 or greater predicted 20% or greater mortality, whereas MELD scores less than 11 predicted less than 5% mortality. The model performed well for patients from Canada at all risk levels. In the Spanish validation set, in which all patients were treated with banding ligation, MELD predictions were accurate up to the 20% risk threshold.

CONCLUSIONS: We developed a MELD-based model that accurately predicts mortality among patients with AVB, based on objective variables available at admission. This model could be useful to evaluate the efficacy of new therapies and stratify patients in randomized trials.

Copyright © 2014 AGA Institute. Published by Elsevier Inc. All rights reserved.

KEYWORDS: AUROC, AVB, Cirrhosis, HCC, HVPG, Logistic Regression, MELD, Prognostic Model, ROC, TIPS, acute variceal bleeding, area under receiver operating characteristic curve, hepatic venous pressure gradient, hepatocellular carcinoma, model for end-stage liver disease, receiver operating characteristic, transjugular intrahepatic portosystemic shunt

PMID: 24148622 [PubMed - in process]

Source

HCV+ Died 23 Yrs Earlier Than Persons without HCV. Blacks Have Double the Rate vs Whites-CDC Study

Provided by NATAP

Download the PDF here

"....Our most significant finding, however, was that among decedents with the same cause of death, persons with hepatitis B or C died about 2 decades younger than persons without these infections. Prevention and early treatment of hepatitis B and C will help prevent these early deaths. Evidence suggests that early therapeutic intervention improves all-cause hepatitis B and C mortality; prevention efforts should be expanded to further (1) promote hepatitis A and B vaccination among recommended target groups, (2) increase hepatitis B and C screening to get more people into care and earlier treatment, and (3) treat alcohol- and drug-related disorders......Our study demonstrated racial/ethnic disparities in deaths with hepatitis C, specifically among American Indians/Alaskan Natives, NH blacks, and Hispanics." (from jules: this study was presented this Fall at conference, I think it was IDSA.

"Among decedents with hepatitis C, the highest mortality rates were observed among persons aged 55-64 years; the American Indian/Alaskan Native, NH black, and Hispanic race/ethnic groups; and males. In comparison, among decedents without viral hepatitis, the highest mortality rates were observed among persons aged ≥75 years, NH blacks, and males."

YOU CAN SEE in Table 1 the rates of mortality in HCV among Blacks (7.87), whites (3.97), hispanics (6.9), American Indian/aAlaska Native 10.07).

(Click on table to enlarge)

Among the 10 most frequently cited causes, deaths listing hepatitis B and C occurred at an average median age of 22-23 years younger than deaths not listing hepatitis B and C......Our analysis of 2010 MCOD data identified 18 473 deaths reported with viral hepatitis. If ranked as a leading cause of death using NCHS's list [25], viral hepatitis would rank as the 15th leading cause of death......Hepatitis C alone was identified as a cause of nearly 90% of these deaths; the majority of those occurred in persons aged 45-64 years.....The disproportionate burden in this age group is consistent with other studies [4, 7, 18]. As a result, in 2012, the CDC recommended 1-time hepatitis C testing for persons born during 1945-1965 (aged 45-65 years in 2010) [26], especially because birth-cohort screening in primary care settings is cost effective [27]. Our study provides evidence that strengthens this national recommendation. Moreover, the well-demonstrated increased mortality of both liver-associated and nonliver-associated conditions in HBV- and HCV-infected decedents provides key evidence to get more people treated before they develop serious illness.......we did find a significant elevated risk of dying with substance-related mental disorders in HCV-infected decedents aged 0-44 years (data not shown).

The most frequently listed category among deaths with either hepatitis B or hepatitis C was fibrosis, cirrhosis, and other liver diseases (45.3% and 48.4%, respectively; Tables 2 and 3)......The most frequently reported nonliver-associated conditions among decedents with either hepatitis B or C were cardiac arrest and ventricular fibrillation, substance-related mental disorders, diabetes mellitus without complication, essential hypertension, and adult respiratory failure insufficiency arrest (8.9%-14.0%). Septicemia (except in labor) was among the top 10 most frequently reported conditions for hepatitis C but not for hepatitis B.

Among decedents aged 45-64 years, the relative risks for conditions associated with a significantly increased risk of dying with hepatitis B ranged from 14.0 (cancer of liver and intrahepatic bile duct) to 2.6 (acute and unspecified renal failure; Table 4). Coronary atherosclerosis and other heart disease and substance-related mental disorders (relative risk, 0.5 and 0.7, respectively) were the 2 conditions associated with a significantly decreased risk of dying with hepatitis B. .....Among the same age group, the relative risks for conditions associated with a significantly increased risk of dying with hepatitis C ranged from 12.0 (cancer of liver and intrahepatic bile duct) to 1.5 (septicemia, except in labor; Table 5).

Causes of Death and Characteristics of Decedents With Viral Hepatitis, United States, 2010

Abstract

Background. Previous research indicates that the mortality burden from viral hepatitis is growing, particularly among middle-aged persons. To monitor progress toward prevention goals, it is important to continue to document characteristics and comortalities of these deaths. This study sought to examine demographic characteristics and the most frequent causes of death among decedents with a viral hepatitis-related death.

Methods. A cross-sectional study was performed on approximately 2.4 million death records from 2010. We calculated mortality rates for decedents with and without hepatitis A, B, and C virus (HAV, HBV, and HCV) and relative risks for the most frequently cited conditions in decedents with and without HBV and HCV.

Results. In 2010, there were 18 473 (0.7%) deaths with HAV, HBV, and HCV listed among causes of death, disproportionately in those aged 45-64 years. Among the 10 frequent causes of death, decedents listing HBV or HCV died, on average, 22-23 years earlier than decedents not listing these infections. HBV- and HCV-infected decedents aged 45-64 years had an increased risk of having the following conditions reported than decedents without these infections: cancer of liver and intrahepatic bile duct; fibrosis, cirrhosis, and other liver diseases; alcohol-related liver disease; gastrointestinal hemorrhage; human immunodeficiency infection; acute and unspecified renal failure; and septicemia (HCV only).

Conclusions. Decedents with other causes of death that include HBV or HCV died 22-23 years earlier than decedents not listing these infections. These data suggest and support the need for prevention, early identification, and treatment of HBV and HCV.

In the United States, hepatitis A virus (HAV), hepatitis B virus (HBV), and hepatitis C virus (HCV) are nationally notifiable infectious conditions and are routinely reported and monitored through national surveillance [1]. Hepatitis A and acute hepatitis B and C are reportable by law in all states and the District of Columbia; chronic hepatitis B and C are reportable by law in 43 states and the District of Columbia. Reporting of these conditions is based upon standard case definitions established in collaboration with the Centers for Disease Control and Prevention (CDC) and the Council of State and Territorial Epidemiologists [2].

While the incidence of hepatitis A, B, and C is at an all-time low [1]-creating the misconception that further prevention efforts are unnecessary-foodborne outbreaks for hepatitis A and healthcare-associated outbreaks for hepatitis B and C continue to occur each year [1]. For hepatitis A, mortality occurs most frequently among persons aged >45 years [1] and persons with underlying chronic liver disease [3]. For hepatitis B and C, mortality occurs most frequently among persons aged 45-64 years [4], and chronic infection accounts for the majority of the total burden. One reason why mortality occurs disproportionately in this relatively younger age group is because up to 65% [5] of the estimated 730 000 [6] US residents with hepatitis B and up to 75% [5] of the estimated 3.2 million [7] US residents with hepatitis C are asymptomatic; many are unaware of their infection [8]. Still, these prevalence estimates, which were obtained from the National Health and Nutrition Examination Survey, are considered conservative because high-risk groups, specifically homeless and institutionalized persons, were not included. Although important for estimating and describing the total burden of disease and for tracking and targeting prevention activities, an analysis describing hepatitis A, B, and C mortality rates by detailed demographic characteristics has not been performed for the United States. Additionally, while hepatitis B and C are well-recognized causes of liver-related disease [9], understanding the effect of these viruses on other frequently occurring comorbidities that lead to death is becoming appreciated [10-13]. Specifically, studies have concluded that HCV infection significantly increased the risk of dying from all causes and nonliver-related causes [10, 11]. Additionally, all-cause mortality among HCV-infected patients in 4 US healthcare networks was nearly 3 times higher than all-cause mortality among HCV-uninfected persons in the noninstitutionalized US population [13]. However, similar studies for hepatitis B have not been performed in a representative US population. Additionally, to our knowledge, this is the first study to document the most common or frequent causes of death among decedents with hepatitis B and C.

Our goal was to use US multiple-cause-of-death (MCOD) data, mainly for hepatitis B and C, from 2010 to (1) characterize the national burden of mortality associated with and without hepatitis A, B, and C by describing incidence rates for select detailed demographic characteristics; (2) compare the most frequently listed causes of death among persons with and without a death associated with hepatitis B and C; and (3) calculate the risk of dying with the most frequently cited conditions among persons aged 45-64 years with a death associated with hepatitis B and C.

METHODS

MCOD Data

This study used the public-use 2010 US MCOD data file, which contains information on all registered deaths occurring within the calendar year [14]. State vital registration offices house these death certificates and, through a cooperative agreement, compile and share this information with the National Center for Health Statistics (NCHS). NCHS then uses this information to generate the annual national multiple-cause mortality datasets.

The conditions on the cause of death section of the death certificate are reported by the decedent's physician, hospital residents, medical examiner, or coroner. The types of causes of death are the underlying, immediate, intermediate, and contributing causes of death that, together, are called the multiple causes of death. In MCOD files, these conditions are translated into International Classification of Diseases, Tenth Revision (ICD-10) [15], codes by highly skilled nosologists at NCHS using 2 schemes: entity axis and record axis. The entity axis represents a direct transcription of each disease entity listed on the cause of death section of the death certificate. The record axis represents a modified version of the entity axis in which repetitive conditions and inconsistencies are removed, related conditions are joined, and coding rules are followed [16]. An example of how conditions are coded is as follows: a death certificate with cirrhosis of liver and alcoholism as causes of death would be directly transcribed to ICD-10 codes K74 (cirrhosis of liver without mention of alcohol) and F10 (alcohol dependence syndrome) as entity axis conditions. These conditions represent separate entities for the same death record. Searching for death records with ICD-10 code K74 would, on the surface, seem that such records had no mention of alcohol. Therefore, a preferable record axis code would be K70.3 (alcoholic cirrhosis of liver), which would encompass both ICD-10 codes K74 and F10.

Definitions

For the purpose of this study, hepatitis A-, B-, and C-related deaths were defined using 2 definitions. First, deaths citing HAV (ICD-10: B15), HBV (ICD-10: B16, B17.0, B18.0, and B18.1), or HCV (ICD-10: B17.1 and B18.2) as the underlying cause or associated cause of death in the record axis was counted. Second, any death with HIV (ICD-10: B20-B24) as the underlying cause and HAV, HBV, or HCV as an associated cause of death in either the record or entity axis was counted. The second definition was implemented to ensure that deaths where coinfection with HIV and viral hepatitis occurred were not excluded due to the frequent reassignment of HIV as the underlying cause of death and the tendency for viral hepatitis to be excluded when translation from entity to record axis occurs [4, 18, 19]. For hepatitis B and C, the decision to combine acute and chronic ICD-10 codes was based on a study that found that chronic hepatitis B and C deaths were often incorrectly coded as acute [20]. The term "viral hepatitis" refers to hepatitis A, B, and C, collectively. The terms "with hepatitis" and "without hepatitis" are used for decedents who had and did not have hepatitis listed as a cause of death, respectively.

Statistical Analyses

Demographic information on age, race/ethnicity, and sex were examined, and mortality rates were calculated from this information. For this analysis (Table 1), deaths listing more than 1 hepatitis infection were assigned a single infection based on a mutually exclusive hierarchy: hepatitis A > hepatitis B > hepatitis C. The hierarchy was based on the need to fully describe characteristics of deaths with hepatitis A and B as hepatitis C carried the highest mortality burden and the degree of co-hepatitis infection was small. Age was divided into the following categories: 0-34, 35-44, 45-54, 65-74, and ≥75 years. Race/ethnicity was classified as Asian/Pacific Islander, American Indian/Alaska native, non-Hispanic (NH) white, NH black, and Hispanic. Mortality rates were calculated using the 2010 US bridged-race approximations [21] and were standardized to the age distribution of the 2000 US standard population [22]. The Poisson distribution was used to estimate the variance for rates and to calculate 95% confidence intervals (CIs) [23].

To determine the most frequently reported causes of death (Tables 2 and 3), ICD-10 codes among deaths with and without hepatitis B and C were isolated from the record axis and classified according to the Clinical Classifications Software (CCS) for ICD-10; this is a well-developed categorization scheme that collapsed approximately 32 000 ICD-10 codes into 260 clinically meaningful and manageable categories [24]. For the CCS category labeled "Other Liver Diseases" in decedents with hepatitis B or hepatitis C, liver fibrosis and cirrhosis (ICD-10, K74) was the most frequently listed; therefore, we renamed this category with the more descriptive label of "Fibrosis, Cirrhosis, and Other Liver Diseases." Liver-associated conditions included the following CCS categories: fibrosis, cirrhosis, and other liver diseases; other hepatitis infections; cancer of the liver and intrahepatic bile duct; alcohol-related liver disease; and gastrointestinal hemorrhage. The median age at death for each of the top 10 CCS categories was examined. Then, the difference was determined by calculating the average of the median ages for the 10 most frequently cited causes among decedents with and without hepatitis B and C and subtracting these 2 averages.

Relative risks were calculated to quantify the risk of dying with the 15 most frequently cited conditions among decedents aged 45-64 years who had hepatitis B and C listed among causes of death. The comparison group was decedents belonging to the same age group who did not have hepatitis B and C listed. For this analysis, we included deaths with either hepatitis B alone or hepatitis C alone (but not those who were coinfected) to remove the effect of potential confounders on the results of deaths with hepatitis B, which accounted for 30% of hepatitis B deaths. We separately compared the relative risks of all deaths with hepatitis B together (hepatitis B alone plus hepatitis B/C coinfected) with hepatitis B alone and found that the magnitude of association of liver-associated and HIV-associated conditions with hepatitis B would be more similar to those of hepatitis C had hepatitis B/C coinfected deaths been included, therefore justifying the removal of the coinfected.

The 95% CIs were used to determine the variance and statistical significance of each relative risk estimate. Data were analyzed using SAS software, version 9.2 (SAS Institute, Cary, NC).

RESULTS

In 2010, 2 472 542 deaths were registered in the United States, of which 18 473 (0.7%) listed hepatitis A, B, or C among causes of death. This implied an age-adjusted mortality rate of 5.2 deaths per 100 000 population (Table 1). Hepatitis A was listed as the underlying cause of 30 deaths (<0.01%) and as any cause of 96 deaths (<0.01%; Figure 1). Hepatitis B was listed as the underlying cause of 589 deaths (0.02%) and as any cause of 1875 deaths (0.08%). Hepatitis C was listed as the underlying cause of 6857 deaths (0.28%) and as any cause of 17 113 deaths (0.69%). Among deaths with hepatitis A, B, or C, 92.6% were with hepatitis C, 10.1% were with hepatitis B, and 0.5% was with hepatitis A. Among deaths with hepatitis C, 3.4% had hepatitis A and/or B. Among deaths with hepatitis B, 1.7% had hepatitis A.

Among decedents with hepatitis A, the highest mortality rates were observed among persons aged ≥45 years and accounted for 89.6% of hepatitis A-related deaths (Table 1). Among decedents with hepatitis B, the highest mortality rates were observed among persons aged 55-64 years, Asians/Pacific Islanders, and males. Among decedents with hepatitis C, the highest mortality rates were observed among persons aged 55-64 years; the American Indian/Alaskan Native, NH black, and Hispanic race/ethnic groups; and males. In comparison, among decedents without viral hepatitis, the highest mortality rates were observed among persons aged ≥75 years, NH blacks, and males.

The most frequently listed category among deaths with either hepatitis B or hepatitis C was fibrosis, cirrhosis, and other liver diseases (45.3% and 48.4%, respectively; Tables 2 and 3). Cancer of the liver, including hepatocellular carcinoma, and intrahepatic bile duct and alcohol-related liver disease were also frequently reported in deaths with hepatitis B and C (11.6%-22.8%). The most frequently reported nonliver-associated conditions among decedents with either hepatitis B or C were cardiac arrest and ventricular fibrillation, substance-related mental disorders, diabetes mellitus without complication, essential hypertension, and adult respiratory failure insufficiency arrest (8.9%-14.0%). Septicemia (except in labor) was among the top 10 most frequently reported conditions for hepatitis C but not for hepatitis B.

For deaths without hepatitis B or C, liver-associated conditions were not among the top 10 categories. Further, the top causes occurred at a lower frequency for any 1 category (7.2%-15.1%) than the top causes among deaths with hepatitis B or C (8.9%-48.4%; Tables 2 and 3). Among the 10 most frequently cited causes, deaths listing hepatitis B and C occurred at an average median age of 22-23 years younger than deaths not listing hepatitis B and C.

Among decedents aged 45-64 years, the relative risks for conditions associated with a significantly increased risk of dying with hepatitis B ranged from 14.0 (cancer of liver and intrahepatic bile duct) to 2.6 (acute and unspecified renal failure; Table 4). Coronary atherosclerosis and other heart disease and substance-related mental disorders (relative risk, 0.5 and 0.7, respectively) were the 2 conditions associated with a significantly decreased risk of dying with hepatitis B.

Among the same age group, the relative risks for conditions associated with a significantly increased risk of dying with hepatitis C ranged from 12.0 (cancer of liver and intrahepatic bile duct) to 1.5 (septicemia, except in labor; Table 5). The relative risks for conditions associated with a significantly decreased risk of dying with hepatitis C ranged from 0.6 (all external causes of injury and poisoning) to 0.9 (respiratory failure, insufficiency, arrest-adult).

DISCUSSION

Our analysis of 2010 MCOD data identified 18 473 deaths reported with viral hepatitis. If ranked as a leading cause of death using NCHS's list [25], viral hepatitis would rank as the 15th leading cause of death.

Hepatitis C alone was identified as a cause of nearly 90% of these deaths; the majority of those occurred in persons aged 45-64 years. The disproportionate burden in this age group is consistent with other studies [4, 7, 18]. As a result, in 2012, the CDC recommended 1-time hepatitis C testing for persons born during 1945-1965 (aged 45-65 years in 2010) [26], especially because birth-cohort screening in primary care settings is cost effective [27]. Our study provides evidence that strengthens this national recommendation. Moreover, the well-demonstrated increased mortality of both liver-associated and nonliver-associated conditions in HBV- and HCV-infected decedents provides key evidence to get more people treated before they develop serious illness. In our comparative cause-of-death analysis, hepatitis B- or C-related deaths most frequently also had liver-associated conditions, which supports the established literature on outcomes of chronic hepatitis infection. Our most significant finding, however, was that among decedents with the same cause of death, persons with hepatitis B or C died about 2 decades younger than persons without these infections. Prevention and early treatment of hepatitis B and C will help prevent these early deaths. Evidence suggests that early therapeutic intervention improves all-cause hepatitis B and C mortality [28-30].

Illicit drug use is associated with an increased likelihood of HBV and HCV infection [7]. In this study, while we showed that substance-related mental disorders were reported frequently among causes of death in persons aged 45-64 years, regardless if HBV and HCV infections were present, the relative risk of dying with a substance-related mental disorder, however, was less likely in deaths with hepatitis B and was not significant in deaths with hepatitis C. For hepatitis B, the decreased likelihood of dying may be explained by the fact that the majority of HBV-infected decedents in the 45-64 year age group were Asians/Pacific Islanders, who most likely acquired their infection during birth or early childhood. For hepatitis C, the insignificant relative risk of dying may be explained by the fact that most persons aged 45-64 years may have used illicit drugs infrequently during their youth, and the behavior is more common among adolescents and young adults. To support this assumption, we did find a significant elevated risk of dying with substance-related mental disorders in HCV-infected decedents aged 0-44 years (data not shown).

Our study demonstrated racial/ethnic disparities in deaths with hepatitis C, specifically among American Indians/Alaskan Natives, NH blacks, and Hispanics. While health disparities in minorities with chronic HCV infection have been documented since as early as 1999 [4, 31, 32], we and others [4] showed that this trend, unfortunately, has not improved.

Although we could not obtain vaccination status from death certificates, our data showed that 3.4% of decedents with hepatitis C had hepatitis A and/or hepatitis B and 1.7% of decedents with hepatitis B had hepatitis A, indicating vaccination was probably not received. Even though hepatitis A and B vaccination is recommended for HCV-infected patients [33], this recommendation had the lowest quality-of-care indicator score in an evaluation study of HCV-infected patients-only 22% received hepatitis A vaccination and only 26% received hepatitis B vaccination or had documented immunity [34].

Although the results of this study are population based, the findings should be interpreted with caution. First, the inaccuracy of cause-of-death coding on death certificates is a significant problem that can lead to underestimates in the viral hepatitis mortality burden [35, 36]; therefore, the hepatitis death estimates in this analysis likely represent only a small fraction of the true burden. Despite having guidelines and training in place, a study at Johns Hopkins Medical Institutions found that more than 40% of causes of death were improperly filled out [36]. In addition to the variability in completeness of recording viral hepatitis deaths, there are data from studies that used medical records to validate hepatitis B and/or C as causes of death in healthcare networks (Mahajan et al, unpublished data) [37]. Investigators found that for patients who had a known HBV [37] or HCV [37, 38] infection associated with chronic liver disease at death, this information was often not reported on death certificates, even when end stage liver disease or hepatocellular carcinoma were listed as the main cause of death. The viral hepatitis mortality burden is even further underestimated by undiagnosed hepatitis infections and deaths unrelated to the decedent's hepatitis infection, such as those that resulted from a suicide or vehicle accident. Because death certificate data imperfectly collect cause of death information, this analysis can only provide data for which viral hepatitis is or is not mentioned on the death certificate. Despite these limitations, MCOD data are invaluable in that they capture all registered deaths in the United States, providing an insightful view into the national burden of viral hepatitis mortality.

In summary, viral hepatitis was listed as a cause of more than 18 000 recorded US deaths in 2010, and there are many who are likely not diagnosed or recorded as having these hepatitides [37]. Because these data demonstrated that death occurred 22-23 years earlier among persons with an HBV- or HCV-related death, prevention efforts should be expanded to further (1) promote hepatitis A and B vaccination among recommended target groups, (2) increase hepatitis B and C screening to get more people into care and earlier treatment, and (3) treat alcohol- and drug-related disorders [26].

Source

\

Predicting Survival after Liver Transplantation Based on Pre-Transplant MELD Score: a Systematic Review of the Literature

PLOS One

RESEARCH ARTICLE

Kristin B. Klein, Taenia D. Stafinski, Devidas Menon

Published: December 12, 2013 DOI: 10.1371/journal.pone.0080661

Abstract

The model for end-stage liver disease (MELD) score is used to stratify candidates for liver transplantation based on objective measures of disease severity. MELD has been validated as a predictor of wait-list mortality in transplantation candidates and has been postulated as a predictor of post-transplant survival. The purpose of this study was to examine the predictive value of the pre-transplantation MELD score on post-transplant survival from relevant existing studies. A systematic review and critical appraisal was performed using Cochrane guidelines. PubMed, the Cochrane Library, Embase, and Web of Science were searched for articles published in the English language since 2005 using a structured search strategy. There were 3058 discrete citations identified and screened for possible inclusion. Any study examining the relationship between pre-transplant MELD and post-transplant survival in the general transplant population was included. Thirty-seven studies met these criteria and were included in the review. Studies were all case series that typically involved stratified analyses of survival by MELD. They represented 15 countries and a total of 53,691 patients. There was significant clinical heterogeneity in patient populations across studies, which precluded performance of a meta-analysis. In 15 studies, no statistically significant association between MELD and post-transplant survival was found. In the remaining 22, some association was found. Eleven studies also measured predictive ability with c-statistics. Values were below 0.7 in all but two studies, suggesting poor predictive value. In summary, while the majority of studies reported an association between pre-transplantation MELD score and post-transplant survival, they represented a low level of evidence. Therefore, their findings should be interpreted conservatively.

Citation: Klein KB, Stafinski TD, Menon D (2013) Predicting Survival after Liver Transplantation Based on Pre-Transplant MELD Score: a Systematic Review of the Literature. PLoS ONE 8(12): e80661. doi:10.1371/journal.pone.0080661

Editor: Evren Alici, Karolinska Institutet, Sweden

Received: May 9, 2013; Accepted: October 5, 2013; Published: December 12, 2013

Copyright: © 2013 Klein et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: Kristin Klein receives funding from the Clinical Investigators program at the University of Alberta. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist

Introduction

The identification of patients who are most likely to benefit from orthotopic liver transplantation (OLT) is a significant challenge in transplantation medicine. Liver transplantation offers the only curative therapy for patients with end-stage liver disease (ESLD). However, the supply of donor livers remains inadequate to meet the demand, necessitating an effective policy for organ allocation. In order to minimize waitlist mortality, a Model for End-Stage Liver Disease (MELD) - based organ allocation was proposed. First adopted by the United States in February 2002, it has become one of the most widely used approaches to prioritizing liver transplant candidates in countries around the world. The MELD score was initially developed by Malinchoc et al [1] to predict mortality in patients undergoing transjugular intrahepatic portosystemic shunts, but has since been validated as predictor of short-term mortality in patients awaiting transplantation. It uses objective variables (creatinine, bilirubin and the international normalized ratio of prothrombin time (PTINR)) to quantify the severity of ESLD, enabling the prioritization of patients in need of liver transplantation by medical urgency. An ideal system would allocate organs to patients not only at the highest risk of dying without transplantation, but also with the highest likelihood of survival following transplantation.

In recent years, the possibility of using pre-transplant MELD to predict post-transplant survival has been explored in many opinion pieces and expert reviews[2]. However, it has yet to be assessed through a systematic review and critical appraisal of published studies that adheres to internationally accepted systematic review guidelines. The need for such a review is heightened by the lack and infeasibility of RCTs on this topic and the fact that considerable debate over the value of MELD in this context remains. Thus, the aim of this study was to assess the association and predictive value of pre-transplantation MELD score on post-transplantation patient survival through a comprehensive, protocol driven systematic review of studies published to date.

Methods

Identification of potentially relevant studies

To identify relevant studies published as of August 2011, a structured search strategy combining relevant controlled vocabulary terms such as Medical Subject Headings (MeSH) and additional non-indexed terms was first developed. Such terms included Model for End-Stage Liver Disease, MELD, liver transplantation, liver failure, and survival. The search strategy was applied to the following electronic bibliographic databases: PubMed (MEDLINE and non-MEDLINE), the Cochrane Library, EMBASE, and Web of Science; and limited to full text, English language studies of adult patients which were published within the past 10 years. For completeness, reference lists of relevant articles were scanned. Also, an internet search for unpublished studies was performed with the Google® search engine. Full search details are provided in Table A in File S1.

Selection of studies for inclusion in the review

Two researchers independently screened the titles and abstracts of citations identified through the literature search using predetermined inclusion criteria (Table 1). The initial search strategy identified studies published in the last ten years, whereas only studies published since 2005 were included in the review.

Parameter	Inclusion Criteria	Exclusion Criteria
General	Full-text articles published in the English language since 2005	Abstracts
Participants	Adults patients with liver failure	Patient populations not representing the general liver transplant population (ex: patients with HCC or HCV only)
Intervention	Patient’s first orthotopic, whole liver, deceased-donor transplant	Multi-organ transplants, non-standard donor or living donor transplants, split liver transplants, sequential transplants
Comparator	Pre-operative Model for end-stage liver disease score	Delta-Meld, MELD-Na, post-operative MELD score
Outcome	Post-transplantation patient survival	Survival rate not reported by MELD score
Study design	Cohort, cross sectional, RCT, quasi-RCT, or controlled studies	Case study or series, commentaries, and opinion pieces without primary data

Table 1. Inclusion/exclusion criteria for review.

Since the purpose of this study was to assess the predictive ability of MELD in the general transplant patient population, studies focussing on specific subgroups of patients were excluded, along with those involving only unique transplant conditions, such as multi-organ transplants, split livers, and non-standard donors. However, studies that considered these conditions within the context of the broader transplantation population were included in the review. Corresponding papers of citations deemed potentially relevant were then retrieved for full review. The level of consensus among reviewers was assessed using the Kappa Statistic. A score of 0.98 was achieved, indicating excellent agreement. Discrepancies among reviewers were resolved through discussion without the need for third party adjudication.

Extraction of data from included studies

Information from included studies was systematically extracted using a pre-tested data abstraction form. The abstraction form contained elements related to study design, patient population, comparators, outcomes measured, and findings. All studies were reviewed by the primary author, with a second reviewer extracting information on 50% of the studies. Reviewers subsequently met to compare results. No discrepancies were found. Therefore, a second, independent review of the remaining studies was deemed unnecessary.

Critical appraisal of included studies

Studies were appraised using the Oxford Center for Evidence-based Medicine Levels and Grades of Recommendation[3].

Data analysis and synthesis of results

Extracted data were tabulated to facilitate a comparison of findings across studies. A meta-analysis of pre-transplant MELD on post-transplant survival using a random effects model was also planned (see Results section). Prior to presenting pooled or summary estimates, clinical heterogeneity and statistical heterogeneity using the I² statistic were assessed.

Results

Results of the literature search are presented in the PRISMA diagram (Figure 1). The search yielded 3058 discrete citations. Forty-eight full-text articles were retrieved for full consideration, of which 37 met the inclusion/exclusion criteria of the review. Among excluded studies, 6 involved inappropriate comparators or outcomes, 4 did not present primary data, and 1 contained data already captured in an included study. The list of excluded studies, along with reasons for exclusion, is presented in Table B in File S1.

Figure 1. Literature search results and study selection for clinical review.

Description of included studies

The 37 studies comprised both prospective and retrospective case series, and collectively included a total of 53,691 patients. In most, the main objective was not to assess the relationship between MELD and post-transplant survival. Instead, studies examined a broad range of pre-transplant factors that may or may not influence survival through exploratory analyses. Studies originated from several countries, including: Belgium (2), New Zealand (1), the United Kingdom (3), Brazil (6), Spain (5), the United States (7), Singapore (1), Turkey (1), China (1), Korea (1), Switzerland (1), Poland (1), Italy (3), Canada (1), and Germany (3). The majority were single centered. Sample sizes ranged from 46 to 21,673 patients (mean = 1451, median = 222), most of whom were male. None used a power calculation to determine sample size. Sampling methods comprised consecutive patients who met inclusion/exclusion criteria, which differed considerably across studies. Therefore, patients comprising the “general transplant population” may have varied. The point at which MELD was measured in patients was inconsistent across studies, with some using time of placement on the transplant list and others using time of transplant. In most of the studies, the relationship between MELD and survival was examined through stratified analyses of survival across sub-groups defined by MELD score, where MELD cut-off points for sub-groups were determined post-hoc. Further, the majority(25) measured the association between MELD and survival based on univariate analyses alone, and, therefore, did not control for potential confounders. Eleven of the studies assessed the predictive ability of the MELD score on post-transplantation survival using a receiver operating characteristic (ROC) curve and the c-statistic. In 4 of the studies, there was partial overlap of patient populations since they included data from the Transplant Scientific Registry (Cywinski et al [4], Freeman et al [5], Rana et al [6] and Yoo & Thuluvath [7]). All of these studies were kept in the review as each used different MELD categories and follow-up times in their analysis. A detailed summary of each study is presented in Table C in File S1.

As mentioned above, the majority of studies grouped patients by MELD. Specifically, MELD, which represents a continuous variable, was converted to a categorical variable for the analyses. The cut-off points for such categories varied widely across studies and were typically determined post-hoc. (Refer to relevant outcome measures in Table C in File S1). Effect measures also differed, ranging from proportions to hazards ratios, odds ratios, and relative risks, and follow-up time periods were inconsistent. Lastly, characteristics of the “general transplant population” varied. Therefore, given such clinical heterogeneity across studies, a meta-analysis was deemed inappropriate, and a statistical assessment of heterogeneity was not performed.

Quality of included studies

Based on the Oxford Center for Evidence-based Medicine Levels of Evidence, the quality of all of the included studies was level IV. The studies, which involved a comparison of pre-transplant MELD scores with post-transplant survival, were all case series and predominantly retrospective in design. All studies recruited consecutive patients over a specified time period, thereby reducing the risk of selection bias.

Association between Pre-Transplant MELD Score and Post-Transplantation Survival

Of the 37 studies, 15 found no association between pre-transplant MELD score and post-transplant patient survival, while 22 reported poorer survival with higher MELD. A detailed description of the results of each study is presented in Table C in File S1. Based on qualitative analyses, there were no clear differences in studies with statistically significant findings compared to those with no statistically significant findings. In both groups, sample sizes varied, as did follow-up times. However, findings from the two largest studies (N >15,000) both suggested that survival decreased with increasing MELD. One observed this relationship only when patients with MELD scores under 9 were compared to those with scores of 30 or greater, while the other had treated MELD as a continuous variable. At the same time, of the 7 other studies that analysed MELD as a continuous variable, all but one found no statistically significant association between MELD and survival. In most of the studies, information presented on patient characteristics was limited. Therefore, it was not possible to identify any differences in patient populations that could explain inconsistencies in the findings.

Predictive ability of MELD score for post-transplantation survival

Eleven studies presented a receiver operating characteristic (ROC) curve to determine the predictive ability of pre-transplant MELD score to determine post-transplantation survival. The area under the curve is used to produce a concordance value called the c-statistic. A c-statistic of 0.50 indicates no predictive ability, and is expected if the results are due to chance alone. In contrast, a c-statistic of 1 represents perfect discrimination. Values under 0.7 suggest poor predictive power, while those greater than 0.70 indicate a useful test, and those higher than 0.80 imply excellent predictive accuracy[2]. Among the 11 studies, 10 reported c-statistics less than <0.7, indicating that MELD poorly predicted post-transplant survival. This included the largest study contributing to the review[6]. In 1 study, the c-statistic decreased over time, from 0.711 for 3-month post-transplant survival to 0.679 for 12-month survival[8]. In the single study with a high c-statistic[9], there was no clear difference in sample size, follow-up time, or patient population when compared to the studies with lower values.

Discussion

This review assessed the association and predictive ability of pre-transplantation MELD score on post-transplantation survival in adults with end-stage liver disease. It highlighted discrepancies in findings across studies. Such discrepancies may be related to the nature of the studies, The vast majority were retrospective case series that relied upon exploratory stratified analyses of data to detect a relationship between MELD and post-transplant survival. As such, analytical techniques, rather than study design, were used to control for confounding. In addition, most of the studies were single centered, with each site having its own process for prioritizing patients for transplant. Therefore, patients constituting the general transplant population may have varied across studies. Thus, while using pre-transplant MELD to predict post-transplant survival in transplant candidates may be attractive, there is little evidence to support it. Prospective studies designed specifically to examine this relationship are needed. MELD score does appear to have a greater impact on mortality when observed in combination with other known risk factors for post-transplant mortality, including sub-optimal livers, low graft-to-body ratio and presence of Hepatitis C. Further research in the area of particular patient subgroups (such as those with hepatitis C) may show a stronger association between MELD and post-transplant outcome.

Based on the studies conducted to date, which collectively represent a low level of evidence, MELD could be correlated with survival, but appears to have limited predictive ability. The vast majority of studies presenting concordance statistics found that pre-transplant MELD score offered minimal discriminating power for post-transplantation survival. However, the c-statistic may be of limited value in determining the predictive ability[10]. This is because the c-statistic is intended for diagnostic models, rather than prognostic models. The two types differ in that prognostic models add the element of time. Specifically, diagnostic models are designed to determine the current state of the patient and accurately identify an existing disease state. In contrast, prognostic models are designed to estimate the probability of a future state where the outcome is not yet known and subject to chance.

This review is limited by heterogeneity in key parameters of studies used to date, which precluded performance of a meta-analysis. Studies reported different comparators (in terms of MELD categories) and applied time-points for outcomes. Studies comparing standardized MELD categories would be beneficial in determining whether or not there is, in fact, a threshold level at which liver transplantation does not offer sufficient survival to warrant the use of scarce donor livers. A more accurate assessment of post-transplant survival would also need to look at other factors, such as quality of life. Research examining the combinations of patient factors using more appropriate statistical techniques may also be valuable in improving the predictive ability of pre-transplant elements on post-transplant outcome.

Conclusions

This study provides a comprehensive review of recent articles examining the relationship between pre-operative MELDS score and post-transplantation survival. Based on the results of studies conducted to date, it appears that the use of MELD does not serve as a reliable predictor of post-transplantation survival. This may be a reflection of a reliance on less than ideal analytical measures. However, the use of pre-transplant characteristics may always fall short of ensuring optimal organ allocation due to variability in immeasurable patient factors and the complexity of perioperative and postoperative conditions.

Supporting Information

Table A in File S1. Literature Search. Table B in File S1. Excluded Studies. Table C in File S1. Description of Included Studies [Download File S1, Checklist S1]

Acknowledgments

The authors would like to thank Leigh-Ann Topfer for assistance with the literature search, Mohamed El Shayeb for assistance with manuscript selection, and Andrea Dunn for assistance with data extraction.

Author Contributions

Conceived and designed the experiments: KK TS DM. Performed the experiments: KK TS. Analyzed the data: KK TS. Contributed reagents/materials/analysis tools: N/A. Wrote the manuscript: KK TS.

References

Source

Single measurement of hemoglobin predicts outcome of HCC patients

Med Oncol. 2014 Jan;31(1):806. doi: 10.1007/s12032-013-0806-2. Epub 2013 Dec 11.

Finkelmeier F, Bettinger D, Köberle V, Schultheiß M, Zeuzem S, Kronenberger B, Piiper A, Waidmann O.

Medizinische Klinik 1, Schwerpunkt Gastroenterologie und Hepatologie, Universitätsklinikum Frankfurt, Goethe-Universität, Theodor-Stern-Kai 7, 60590, Frankfurt am Main, Germany.

Abstract

Anemia is a common complication in several types of cancer including hepatocellular carcinoma (HCC). The prognostic potential of hemoglobin (Hb) levels has not yet been investigated in HCC patients. One hundred and ninety-nine patients were prospectively recruited and Hb levels were determined. Hb levels were compared to the stages of liver cirrhosis and HCC stages. The association of the Hb levels and overall survival (OS) was assessed by univariate and multivariate Cox regression models. The relation of Hb levels and OS was further validated in an independent cohort of 87 HCC patients. Hb levels negatively correlated with the stage of liver cirrhosis (model of end stage liver disease score and Child-Pugh stage) and differed between stages of HCC. Low Hb levels (≤13 g/dl) were associated with higher mortality in the test [hazard ratio (HR) 2.422, 95 % confidence interval (CI) 1.357-4.322, P = 0.003] as well in the validation cohort (HR 2.486, 95 % CI 1.097-5.632, P = 0.029) in univariate Cox regression model. Low Hb levels were associated with mortality independently from the tumor stage, age, gender and the C-reactive protein levels in a multivariate Cox regression model. Anemia should be considered as a risk factor for mortality in HCC patients.

PMID: 24326985 [PubMed - in process]

Source

Vitamin D deficiency is associated with mortality in patients with advanced liver cirrhosis

Eur J Clin Invest. 2013 Nov 15. doi: 10.1111/eci.12205. [Epub ahead of print]

Stokes CS, Krawczyk M, Reichel C, Lammert F, Grünhage F.

Department of Medicine II, Saarland University Medical Center, Homburg, Germany.

Abstract

BACKGROUND: Chronic liver disease is the fifth most common cause of mortality in Europe. Recently, vitamin D deficiency has been associated with an increased risk of mortality in the general population. Since patients with advanced liver disease frequently exhibit vitamin D deficiency, we assessed for a possible association of vitamin D deficiency with survival in a cohort of patients with advanced liver disease.

METHODS:  Sixty-five patients with liver cirrhosis (median age, 58 years; range, 19-76 years; 66% male; Child-Pugh stage C, 46%) were included in our prospective single-center survival study. Serum 25-hydroxyvitamin D concentrations were measured by chemiluminescence immunoassay. The optimal cut-off was determined using area under the curve (AUC) and Kaplan-Meier analysis. Chi-square statistics and multivariate binary logistic regression analysis were also conducted.

RESULTS: Median serum vitamin D levels were 8.2 ng/ml (range < 4.0-95.8 ng/ml). Overall, 48% of patients (31/65) died during a 24-month follow-up period. AUC analysis determined a vitamin D level of 6.0 ng/ml as optimal cut-off for discriminating survivors from non-survivors. Kaplan-Meier analysis of survival confirmed low vitamin D levels as significant predictor of death (P = 0.012). Finally, multivariate analysis identified low vitamin D levels (OR = 6.3; 95% CI, 1.2-31.2; P = 0.012) and MELD scores (OR = 1.4; 95% CI, 1.2-1.7; P < 0.001) as independent predictors of survival.

CONCLUSION: Low vitamin D levels are associated with increased mortality in patients with advanced liver disease. Thus, serum levels of vitamin D might represent a critical marker of survival in advanced liver cirrhosis. This article is protected by copyright. All rights reserved.

This article is protected by copyright. All rights reserved.

KEYWORDS: 25-hydroxyvitamin D, cholecalciferol, chronic liver disease, survival analysis

PMID: 24236541 [PubMed - as supplied by publisher]

Source

Hepatitis C Viral Suppression Reduces Liver Morbidity, Death

Medscape Medical News > Conference News

Miriam E. Tucker
November 15, 2013

Washington, DC — Achieving an undetectable hepatitis C viral load is associated with decreased liver morbidity and death, a large observational study using Veterans Affairs (VA) data finds.

"New therapies are needed to increase treatment rates and increase treatment effectiveness," said Jeffrey McCombs, PhD, from the University of Southern California School of Pharmacy in Los Angeles.

Dr. McCoombs presented the results here at The Liver Meeting 2013. The findings were simultaneously published online inJAMA Internal Medicine.

Dr. McCombs and his team used electronic medical records data from the VA National Clinical Case Registry for hepatitis C virus from 1999 through 2010, which includes 360,857 patients.

A total of 128,769 people met the selection criteria with a detectable viral load (>25 IU/mL) at baseline and their viral genotype recorded.

Patients were 97% male, 51% white, and 31% black, with an average age of 52 years. The majority of patients, 79%, were infected with hepatitis C genotype 1, 12% with genotype 2, and 8% with genotype 3.

Twelve percent also had diabetes and 16% had a prior hospital admission.

Primary outcomes were time to death and time to a composite liver event, including compensated and decompensated cirrhosis, hepatocellular carcinoma, or liver-related hospitalization.

Dr. McCombs explained that the time from undetectable viral load to first event was used instead of sustained viral response — the gold standard of treatment success — because it is much easier to obtain. Viral response requires the determination of the time at which the patient maintained consistent viral load suppression for a minimum of 6 months following the termination of treatment.

Only 1 in 4 Treated

In all, just 24% of patients had been treated previously for hepatitis C, and only 16% of those treated achieved an undetectable viral load.

Of the entire study population, only 4% achieved an undetectable viral load.

Overall, viral load suppression reduced the risk for future liver events by 27% and death by 45%.

"This verifies earlier findings that viral load reduction through treatment can significantly reduce the risk of adverse patient outcomes," Dr. McCombs said.

Male sex significantly increased the risk for the composite clinical outcome by 11% and the risk for death by 58%. Both diabetes and prior hospital admissions increased these risks as well.

Compared with whites, black patients had a 35% lower risk for the composite endpoint and a 28% lower risk for death. And compared with patients who had hepatitis C genotype 1, those with genotype 2 were at lower risk: 23% for the composite event and 20% for death.

“This has left me, and many clinicians, in the odd position of feeling that patients are either too healthy or too sick for hepatitis C treatment.” Dr. Mitchell Katz

In an accompanying article, JAMA Internal Medicine deputy editor Mitchell Katz, director of the Los Angeles County Department of Health Services, writes, "The authors demonstrate that patients who do achieve viral suppression, which almost always required  treatment, fared significantly better."

He adds, "The critical issue going forward is whether the new drugs that have been released (eg, hepatitis C protease inhibitors) or are likely to be approved soon (eg, hepatitis C nucleotide polymerase inhibitor) can achieve sustained viral suppression in a high percentage of patients without serious adverse effects."

Can these treatments "be made available without breaking the bank of safety net health systems across the country that care for large numbers of patients with hepatitis C?" Dr. Katz questions. "I certainly hope so," he writes.

Dr. Katz also notes that the low rate of treatment seen in the VA database probably reflects the fact that current treatments aren't very effective in clearing the infection and have serious adverse effects.

He explains that he often feels he is "between a rock and a hard place" when it comes to treating patients with hepatitis C. For patients who are healthy, despite their infection, it may be worthwhile to wait for better treatments, whereas patients with severe liver damage may not be able to tolerate the adverse effects of treatment; for those patients, it is unclear whether suppressing the virus would even improve their outcomes.

"This has left me, and many clinicians, in the odd position of feeling that patients are either too healthy or too sick for hepatitis C treatment," writes Dr. Katz. "It undoubtedly explains why in this VA cohort only 24% of patients had received treatment at any time."

Sustained Viral Response

During the question-and-answer period at the meeting, audience members questioned the study's methods, given its assumption that the patients are all at equal risk going in, which may not be the case.

"Especially when you have relatively low response rates, there's a big difference between those who do and don't respond, before you ever treat them," said Andrea Branch, MD, from the Icahn School of Medicine at Mount Sinai, New York.

She urged Dr. McCombs and his team to compare long-term outcomes in the relapsers with those in patients who actually achieve a sustained viral response.

"That gives you a chance to look at people who have the physiology necessary to respond to treatment with dramatic suppression of virus, but who remain viremic afterward," she explained.

Dr. Branch said other studies have shown that the patients who first suppress the virus, but then relapse and don't achieve a sustained viral response, tend to have long-term outcomes, such as liver cancer and mortality, similar to those in patients who do achieve a sustained response, even though they become hepatitis C load positive again.

"I think it's really important when doing outcomes to tease out patients who never suppress the virus with those who are capable of suppressing it," even if they later relapse, stressed Dr. Branch. "It's a marker of those healthy enough to respond to therapy by becoming viral-load negative. Those people are different from the population who don't become viral-load negative."

But, she added, "as more and more people have a sustained viral response, this will become less of an issue because the more potent the antiviral drugs, the less significant the underlying physiology in achieving viral suppression."

Dr. McCombs replied that the VA dataset would allow for this analysis.

"I would very much encourage you to do that and would be very interested in what you find," Dr. Branch said.

This study was funded by Bristol-Myers Squibb. Dr. McCombs received salary and travel support under the terms of the research grant between Bristol-Myers Squibb and the University of Southern California. Dr. Branch has received research support from Gilead Sciences, the National Institute of Drug Abuse, and the National Institute of Diabetes and Digestive and Kidney Diseases. Dr. Katz has disclosed no relevant financial relationships.

The Liver Meeting 2013: American Association for the Study of Liver Diseases (AASLD). Abstract #246. Presented November 5, 2013.

JAMA Intern Med. Published online November 5, 2013. Abstract Editorial

Source