Monitoring liver transplant rates in persons diagnosed with hepatitis C: a data linkage study, England 2008 to 2017

Introduction Liver transplantation is an important measure of burden from hepatitis C virus (HCV)-associated liver disease. Aims To describe transplant rates and survival in individuals with HCV infection from 2008 to 2017 in England through data linkage. Methods This is a retrospective observational cohort study. Laboratory reports of HCV infection were linked to the Liver Transplant Registry for individuals aged 15 years and over, first diagnosed between 1998 and 2017. We estimated age-sex standardised incidence rates and used Poisson regression to investigate predictors of liver transplantation and test for a change in incidence after introduction of direct-acting antivirals (DAAs) in 2014. Kaplan-Meier survival analysis was used to calculate post-transplant survival rates. Results Of 124,238 individuals diagnosed with HCV infection, 1,480 were registered and 1,217 received a liver transplant. Of individuals registered, 1,395 had post-HCV cirrhosis and 636 had hepatocellular carcinoma (618 also had post-HCV cirrhosis). Median time from HCV diagnosis to transplant was 3.4 years (interquartile range: 1.3–6.8 years). Liver transplant rates were lower 2014–17 compared with 2011–13 (incidence rate ratio: 0.64; 95% confidence interval: 0.55–0.76). Survival rates were 93.4%, 79.9% and 67.9% at 1, 5 and 10 years, respectively. Data linkage showed minimal under-reporting of HCV in the transplant registry. Conclusion In the post-DAA era, liver transplant rates have fallen in individuals with HCV infection, showing early impact of HCV treatment scale-up; but the short time from HCV diagnosis to liver transplant suggests late diagnosis is a problem.


Introduction
In 2018, around 113,000 individuals were estimated to be chronically infected with hepatitis C virus (HCV) in England [1]. These individuals are at increased risk of cirrhotic end-stage liver disease (ESLD) and hepatocellular carcinoma (HCC), for which mortality rates had been increasing -doubling over 10 years from 2005 to 2014 -until recent years [2]. Liver transplantation may be indicated if liver decompensation deteriorates. Between 2008 and 2014 in England, an average of 134 individuals with post-HCV cirrhosis were registered for a liver transplant and 108 received a liver transplant each year, accounting for 17-21% of all liver transplants [2].
HCV-related ESLD is preventable with early diagnosis and treatment of HCV infection, but historically estimated treatment levels in the United Kingdom (UK) have been low (ca 3% annually) [3]. Low treatment rates have been attributed to: (i) interferon-based regimens which required injections, (ii) long treatment durations, (iii) poor tolerability and (iv) moderate efficacy, as defined by a sustained virological response (SVR). In comparison, the newly introduced directacting antivirals (DAAs) are taken orally, have short regimens and have SVRs that are above 95% [4]. Since 2015, National Health Service (NHS) England has rolled out DAAs in a managed care programme with annual scale-up in treatment slots; an estimated 24,592 people had been treated between the financial years of 2015/16 and 2017/18 [5]. Widespread treatment with DAAs not only prevents the development of severe liver disease in individuals with HCV but can also improve liver function in individuals with advanced liver disease, which should contribute to a fall in the burden of HCV-associated disease and the need for liver transplantation [4,6,7].
Since the expansion of treatment with DAAs in England, a 43% drop in the number of liver transplant registrations has been recorded for individuals with post-HCV cirrhosis [2]. Vaziri et al. reported that the proportion of liver transplants attributed to HCV-associated cirrhosis fell from 10.5% to 4.7% between 2013 and 2016 and the proportion of liver transplants for cancer-associated HCV fell from 46.4% to 33.7% over the same period [8]. Similar results have been found in Italy, the United States (US) and Argentina [9][10][11]. A declining contribution of HCV-associated cirrhosis and cancer to transplants is helpful but is not a substitute for monitoring rates of HCV-associated transplants and may be limited by under-reporting of HCV coding in the liver transplant dataset, as has been observed in death registry data [12].
Through data linkage of routine laboratory reports of HCV and the NHS Blood and Transplant Service (NHSBT) liver transplant registry, we estimate liver transplant incidence rates and survival in the pre-(2008-13) and post-DAA (2014-17) eras, describe the characteristics of individuals who are registered for and underwent a liver transplant (2008-17) and explore any under-reporting of HCV in the transplant registry. We hypothesise that a reduction in transplant rates post introduction of DAAs will be observed, indicative of the early impact of DAA on HCV burden.

Methods
Using data linkage, we conducted a retrospective observational cohort study to describe the characteristics of individuals diagnosed with HCV infection in England between 1998 and 2017, who were registered for and had received a liver transplant, and estimate liver transplant rates between 2008 and 2017.

Data sources
Routine laboratory reports of HCV infection HCV diagnoses were obtained from routine laboratory reports of HCV, defined as the detection of HCV antibody (anti-HCV) or HCV RNA in blood, submitted by English virology laboratories to Public Health England (PHE). The laboratory surveillance system does not distinguish between anti-HCV or HCV RNA positive individuals, so laboratory 'confirmed' cases are a mix of current and ever infected individuals. Laboratory HCV reports have been submitted to PHE (and its predecessor organisations) through paper forms or electronically since 1990 but laboratory reporting of notifiable organisms became mandatory in 2010. Reports include basic demographics (name, date of birth, sex and NHS number (a unique identifier for all individuals registered with the NHS within England)) and variable amounts of risk factor information (e.g. History of injecting drug use, testing within prison services, infection as a result of tattoos, piercings or blood products). Ethnicity is poorly recorded.
For this study, the routine laboratory reporting dataset was enhanced with information from additional sources. As the date of the first HCV diagnosis is critical, this was updated with information from the Sentinel Surveillance of Blood Borne Virus Testing (SSBBV) and Hospital Episode Statistics (HES) when possible. Established in 2002, SSBBV collects information on hepatitis A-E, HIV and HTLV tests (regardless of the result) from 23 participating sentinel laboratories in England [13,14]. Earlier diagnosis dates were identified in HES where an earlier inpatient stay was recorded with HCV ICD-10 diagnosis codes (B17.1 and B18.2). HCV treatment with DAAs (date and outcome) was obtained through linkage with the national HCV treatment monitoring and outcomes dataset, established in 2015. Date of death was obtained from the NHS Spine Patient Demographic Service.

Data linkage
The two datasets were linked deterministically using a combination of name, date of birth, sex and NHS number.

Analysis
Age-sex standardised incidence rates (ASRs) of first liver transplant and first HCC-associated liver transplant were calculated for individuals aged 15 years and older who were diagnosed with HCV infection between 1998 and 2017 and are presented for 2008 to 2017. ASRs for 1998-2007 are not presented due to unstable rates and wide 95% confidence intervals (CI) (calculated using Poisson distribution) resulting from the smaller cohort size and limited follow-up time. Individuals who received a liver transplant before or within 6 months of HCV diagnosis were excluded from the analysis, as there was potential bias towards higher rates of diagnosis in individuals with major morbidity and in individuals who received a transplant before their reported HCV diagnosis were likely to have had a diagnosis prior to transplant which had not been reported to PHE. As a result, follow-up began 6 months after diagnosis and ended when the individual had a liver transplant, died or 31 December 2017 whichever occurred first. ASRs were also calculated for all liver transplants in England using ONS mid-year population estimates [15] and NHSBT data on all first liver transplants in England.

Ethical statement
The PHE Caldicott Guardian approved the collection and processing of confidential patient data, without patient consent, from routine laboratory reports of HCV infection, SSBBV and for linkage of these laboratory data to the liver transplant registry under Regulation 3 of the Health Service (Control of Patient Information) Regulations 2002. Regulation 3 allows the processing of patient information for recognition, control and prevention of communicable disease and other risks to public health.

Results
Of 161,820 individuals aged 15 years and older diagnosed with HCV in England between 1998 and 2017, 124,238 (76.8%) had sufficient identifiers for linkage

Received a liver transplant
Of 1,217 individuals who received a transplant, the median age at transplant was 53 years (IQR 48-58) and individuals with a record of alcoholic cirrhosis were transplanted at a younger median age than those with no record of alcoholic cirrhosis (51 vs 54 years respectively, p < 0.0001 Age-sex standardised liver transplant incidence and HCC-associated liver transplant incidence between 2008 and 2017 are presented in Figure 1.  Figure 1).  Figure 2).  Limitations There are several limitations of this study. First, it is possible that individuals were diagnosed with HCV before PHE was notified or their diagnosis was not reported to PHE at all, thus impacting follow-up time. This would be more of an issue for individuals diagnosed before 2010, as since then, laboratories are legally required to notify PHE of all positive HCV tests. However, individuals, who were diagnosed before 2010 (pre-legislation) may also have been tested with their HCV diagnosis reported post-2010 (post-legislation) and therefore captured in later years. Linkage with other surveillance systems can be used to update diagnoses dates before 2010. Second, we did not have any information on whether individuals were acutely or chronically infected with HCV as this information cannot be derived from the laboratory reports. However, as individuals with chronic infection are more likely to progress to advanced liver disease, we can assume that most individuals who were registered for a liver transplant had chronic hepatitis rather than acute fulminant hepatitis type presentation. Third, we had no information on the stage of liver disease at diagnosis in the HCV laboratory reports. This information would have been useful, as we found short time lag between HCV diagnosis and transplant registration, which we can only speculate is due to late diagnosis of liver disease and/or HCV when disease stage is already advanced. If the disease stage was available, we could have investigated liver transplant rates by disease stage at presentation.

Other evidence and implications
Through linked data, we identified a decline in HCVassociated liver transplant rates after 2014, consistent with unlinked published data for England [2,8] and internationally [9][10][11]. A separate study within the UK [8], using unlinked data, reported registered transplants for patients with HCV-related cirrhosis fell from 10.5% in 2013 to 4.7% in 2016. In Italy, a decrease was observed between pre-DAA and post-DAA periods for transplant waiting lists overall as well as among patients among those with decompensated cirrhosis, whereas figures associated with HCC remained stable [10]. A downward trend was identified in liver transplants among patients with decompensated cirrhosis in Argentina [11]. A decline has also been reported in the US [9] among patients waiting for a transplant, with a lower likelihood of being on the waiting list in the DAA era compared with the pre-DAA era (Hazard Ratio 0.83). The decrease in liver transplant incidence in England among individuals with HCV within this study contrasts with the trend observed for all liver transplants, where rates have increased in recent years, albeit on a much smaller scale. It is encouraging to see a reduction in HCV-associated liver transplant rates and we hope that the incidence will continue to fall as case-finding activity increases helping to reduce the undiagnosed burden and ensuring that more individuals are treated with DAAs. A similar, although non-significant, decline has also been found in age-sex standardised liver mortality rates among anti-HCV positive individuals [16].
Time from HCV diagnosis to transplant was short.
Mindful of the established natural history of HCV infection causing liver disease progression over several decades [17], this indicates that this linked cohort of individuals with HCV was diagnosed late. As a result, individuals are likely to present with advanced liver disease, which is less amenable to HCV cure and may progress despite HCV cure. Poorer response rates to HCV treatment among patients with cirrhosis have been reported [18][19][20]. In an Italian study, the risk of failing SVR12 was 5.39 times higher among patients with decompensated cirrhosis and 1.56 times with compensated cirrhosis compared with those without cirrhosis [19]. In an American study, the efficacy was found to be significantly reduced to 73% and 91% among those with decompensated and compensated cirrhosis, respectively [20]. Two further studies, one European and Canadian-based and the other an Italian study, observed that the risk of liver disease continued among patients with cirrhosis despite SVR [21,22]. The linkage between SSBBV and the ONS death registry corroborate this finding, as the median time from HCV diagnosis to death (any cause) was 3 years [16]. While there are limited population-level data in England currently on disease stage at time of diagnosis of HCV, data from the national HCV treatment monitoring dataset indicate that around 40% of individuals who have received treatment (2015-18) had moderate fibrosis to decompensated cirrhosis at the time of the decision to proceed with treatment. This is in part due to NHS England's phased implementation of DAA treatments, prioritising patients with advanced cirrhosis and fibrosis before expanding treatment access to patients with milder disease.
Our results from the survival analysis following a transplantation in an individual with an HCV diagnosis (  [23,24]. It is encouraging that survival following liver transplantation in individuals with HCV is similar to all persons receiving a liver transplant and that HCV does not appear to have an additive negative impact on survival rates. It is important to note that under-reporting of HCVassociated transplants is minimal, with only 3.2% of linked individuals having no record of HCV-associated disease or positive HCV tests in the transplant registry. This is much lower than the under-reporting with HCVassociated mortality in England, where 41% of persons who died of liver-related causes linked to an HCV laboratory diagnosis did not have HCV recorded as a contributory cause in the deaths registry [12]. A similar level of under-reporting in mortality was also identified in the US [25] and Scotland [26], with 41% and 52% of individuals who died of liver-related causes, respectively, having HCV indicated on their death certificate. The low level of under-reporting is important to quantify as an accurate recording of HCV in transplant registry data reduces uncertainty in modelling to estimate prevalence and disease burden. These results improve confidence in the use of unlinked HCV-associated liver transplant data as a metric for monitoring progress towards elimination goals in England. Data linkage of HCV diagnoses to a transplant registry also offers an easily reproducible method for other European countries whose surveillance systems are limited by under-reporting of HCV diagnoses in transplant databases and/or lack of information on long-term health outcomes in HCV laboratory diagnoses datasets. Data linkage also allows the burden of disease in the HCV diagnosed population, to be better understood and for an individual in this population to be followed from diagnosis to outcome. This is important when comparing this group to individuals who did not receive a transplant and for understanding these data in the context of a country's HCV epidemic. Liver transplantation, associated with chronic HCV infection, is an important measure of liver disease from HCV infection and countries may, therefore, choose to monitor this metric as an additional indicator (similar to HCV-related mortality and new chronic infections) of the impact of new DAA treatments and their contribution towards the WHO elimination goal [27].

Conclusion
By linking routinely collected datasets, we found 1.2% of persons with HCV were registered for a liver transplant, with higher rates in males. The short median time from diagnosis to transplant suggests missed opportunities for earlier diagnosis before irreversible liver damage has occurred. Although based on early data, our analyses suggest that transplant rates have decreased further since 2014 (the year that DAAs became available in England). Efforts to reduce the undiagnosed HCV burden by active case finding, earlier diagnosis, improved engagement in care and expanded DAA treatment to people with mild or no fibrotic liver disease, could lead to sustained and potentially accelerated reductions in HCV-associated liver transplants as progression to cirrhotic liver disease will become less common. Our study, therefore, provides a baseline from which to benchmark changes in HCV-associated transplant rates as part of evaluating the impact of DAAs and other interventions on eliminating HCV as a global public health threat by 2030.