The present and future disease burden of hepatitis C virus infections with today's treatment paradigm: Volume 4

Factors influencing the morbidity and mortality associated with viremic hepatitis C virus (HCV) infection change over time and place, making it difficult to compare reported estimates. Models were developed for 17 countries (Bahrain, Bulgaria, Cameroon, Colombia, Croatia, Dominican Republic, Ethiopia, Ghana, Hong Kong, Jordan, Kazakhstan, Malaysia, Morocco, Nigeria, Qatar and Taiwan) to quantify and characterize the viremic population as well as forecast the changes in the infected population and the corresponding disease burden from 2015 to 2030. Model inputs were agreed upon through expert consensus, and a standardized methodology was followed to allow for comparison across countries. The viremic prevalence is expected to remain constant or decline in all but four countries (Ethiopia, Ghana, Jordan and Oman); however, HCV‐related morbidity and mortality will increase in all countries except Qatar and Taiwan. In Qatar, the high‐treatment rate will contribute to a reduction in total cases and HCV‐related morbidity by 2030. In the remaining countries, however, the current treatment paradigm will be insufficient to achieve large reductions in HCV‐related morbidity and mortality.


Hepatitis C virus (HCV) infection incurs a substantial global disease
burden with dire implications for morbidity and mortality. 1 Analyses have shown significant associations between infection with HCV and outcomes such as cirrhosis, hepatocellular carcinoma (HCC) and liver transplants as well as a corresponding increased association with mortality. [2][3][4] With the emergence of effective direct-acting antiviral treatments for HCV, policies to address the HCV disease burden at the national and global levels will require the use of more accurate epidemiological data. 5 There remains, however, a lack of robust data on HCV-and liver-related mortality at the country level, as well as significant heterogeneity in pooled global estimates. 6 It was recently estimated that globally 71 million people have active viremic HCV infections. 7 This study aimed to provide estimates of incidence, prevalence, diagnosis, treatment and mortality in 2015 for multiple countries

| Inputs
The historical epidemiology of HCV was gathered through a literature search, analysis of unpublished data and discussion with expert panels. 8 When no input data were available, analogues (data from countries with a similar health care and/or risk profiles) or expert inputs were used. Ranges were used to capture uncertainty in inputs with wider ranges implying greater uncertainty.

| Model
A disease progression model was constructed in Microsoft Excel ® (Microsoft Corp., Redmond, WA, USA) to quantify the size of the HCV-infected population, by the liver disease stages, from 1950 to 2030. The model has been previously described in detail. 9 Microsoft Excel was selected as a platform due to its transparency, availability and minimal need for operator training. The model was set-up for sensitivity and Monte Carlo analysis using Crystal Ball ® , an Excel ® add-in by Oracle ® . Beta-PERT distributions were used for all uncertain inputs.
The model starts with the annual number of acute infections that progressed to chronic HCV infection after accounting for spontaneous clearance of the virus (Figure 1). The progression of these new cases was followed along with all chronic infections from prior years, after accounting for mortality and cure. Unless specified, the scope of the model was limited to HCV-RNA-positive cases. Nonviremic cases (those who spontaneously cleared the virus or were treated and cured) were not considered even though they would test positive to HCV antibodies (anti-HCV) and may still progress to more advanced stages of liver disease despite viral clearance. 3 The total number of cases, at each stage of the disease, was tracked by age and gender.
The historical number of HCV infections, and the age and gender distributions, was gathered through a literature search and discussions and a standardized methodology was followed to allow for comparison across countries.
The viremic prevalence is expected to remain constant or decline in all but four countries (Ethiopia, Ghana, Jordan and Oman); however, HCV-related morbidity and mortality will increase in all countries except Qatar and Taiwan. In Qatar, the high-treatment rate will contribute to a reduction in total cases and HCV-related morbidity by 2030. In the remaining countries, however, the current treatment paradigm will be insufficient to achieve large reductions in HCV-related morbidity and mortality.

K E Y W O R D S
diagnosis, disease burden, epidemiology, HCV, hepatitis C, incidence, mortality, prevalence, treatment with an expert panel. 8 These data were used to estimate the historical number of new HCV infections, as described below.

| New HCV infections & re-infection
When available, reported or calculated annual estimates of new infections were used. In most countries, the number of new HCV infections was not available and was, therefore, back-calculated using a two-step process that first calculated the annual number of new cases and then calculated the age and gender distribution of these cases.
The annual number of new cases was calculated using the known number of total HCV infections in a given year. The model calculated the annual number of all-cause mortality, liver-related deaths and cured cases, for that year as will be described below. The Excel ® optimization add-in, Solver, was then used to determine an average number of new infections per year going back to 1950. The size and impact of the HCV-infected population prior to 1950 were considered negligible for the purposes of this analysis. An annual relative incidence value was used to describe the change in the number of new infections from 1950 to the year of known HCV prevalence. This was done to account for the fact that annual number of new cases did not remain flat since 1950. Literature reviews and discussions with the expert panel were used to identify the years when new infections peaked using the risk factors common in the country (transfusion-related infections, injection drug use [IDU], etc.). When immigration from endemic high-risk countries was highlighted as an important source of new infections, the annual number of new cases due to immigration was calculated by gathering net annual immigration, by country of origin and from national databases regarding the anti-HCV prevalence in the country of origin.
In the second step, the age and gender of the acute infections were calculated using the known age and gender distribution of the total infected population in a given year. The age and gender distributions of the new infections in 1966 and every 5 years thereafter were modified to match the known distribution of the prevalent population, and trended linearly between the 5-year increments. The age and gender distributions in years 1950-1965 were set to equal 1966.
It was assumed that in the absence of better information, future HCV infection and re-infection will remain the same as they are today. This is a more conservative approach than a dynamic model, which would show a reduction in HCV incidence with treatment of high-risk populations (treatment as prevention). This conservative approach was deemed appropriate given the uncertainties present for HCV epidemiology and lack of detailed data on infection and reinfection rates.

| Progression rates
Disease progression was simulated by multiplying the total number of cases at a particular stage of the disease by a progression rate to the next stage. The rates were gathered from previous studies or calculated using known number of HCC cases/mortality, as explained previously. 9 The number of new cases at a stage of the disease was calculated by multiplying the progression rate and the total number of cases at the previous stage of the disease in the previous year. The total number of cases was adjusted for ageing, all-cause mortality and cured in any given year.

| All-cause mortality
The all-cause mortality rates by age and gender were gathered from the United Nations Population Database 10 unless stated otherwise.
The rates were adjusted for incremental increase in mortality due to IDU and transfusion, as described previously. 11 Unless specified, a standard mortality ratio (SMR) of 10 (95% uncertainty interval 9.5-29.9) was used for the portion of the HCV-infected population between ages of 15-44 years who were active injection drug users (active IDU). [12][13][14][15][16][17] An SMR of 2.1 (1.3-17.6) was applied to all ages for the portion of the population infected due to transfusion. 18 While blood transfusion still posed a risk in some of the low-income countries in this analysis, general consensus was that new HCV infections due to transfusion are on the decline as safer transfusion practices are implemented. A linear declining rate was applied to get the per cent of total infections attributed to transfusion to zero by 2030.
Country-specific adjustments to all-cause mortality for active IDU and transfusion were made using the following assumptions:

| Bahrain
In 2015, experts estimate that 30% of the HCV-infected populations were active IDU. According to a 2008 study of medical records from 183 HCV-infected patients, 35% had acquired their infection through past transfusion. 19

| Bulgaria
In 2015, approximately 11% of the HCV-infected population was active IDU. This percentage was back-calculated using the estimates of

| Cameroon
Based on expert consensus, in 2015, <1% of the infected population was assumed to be active IDU, and 5% of the infected population was assumed to have been infected through blood transfusion.
Infection due to traditional practices (circumcision, ritual cutting, etc.) poses a significant but as of yet unquantified risk for HCV infection in Cameroon.

| Colombia
Using Argentina as an analogue, it was estimated that approximately 9% of the infected population in 2001 had been infected via IDU. 20,21 Anti-HCV prevalence among active IDU in Colombia ranges from 1.70% to 20.9%, but the size of PWID population has not been well defined. 23

| Dominican Republic
The rate of IDU is low in the Dominican Republic. According to expert input, 0.5% of HCV cases in 2015 were due to IDU. Transfusion accounted for an estimated 40% of cases in 2015, also according to expert input.

| Ethiopia
IDU is very uncommon in the Ethiopian population, and thus, none of the HCV-infected population was assumed to have received their infection through this means (expert consensus). Expert consensus estimates that 10% of the HCV population was infected through blood transfusion in 2014.

| Ghana
No formal study of IDU-associated HCV has been carried out in Ghana, but rates are believed to be very low. Expert consensus estimated that <1% of the 2015 infected population was infected through this means. Meanwhile, between 2008 and 2013, 1.20% of HCV cases were assumed to have occurred as a result of contaminated blood transfusions. 28

| Hong Kong
In 2015, an estimated 16.5% of the infected population was active IDU. This percentage was calculated using estimates of 6000 active IDU in Hong Kong in 2011 and an anti-HCV prevalence of 56% among IDU. 29 In a study of 273 HCV-positive Chinese blood donors in Hong Kong, 37.9% received their infection through transfusion.

| Kazakhstan
There have been no published studies regarding the prevalence of HCV among active injection drug users in Kazakhstan, and thus, Uzbekistan was used as an analogue. An estimated 7% of the infected population was active IDU. 33 This corresponds to about 2000 individuals. Uzbekistan was also used as an analogue source for transfusionrelated HCV rates, with an estimated 41% of the infected population infected via transfusion in the year 2000. 33

| Malaysia
Based on expert input, about 65% of the viremic population in 2009 was infected via IDU. In 2005, an estimated 42% of the HCV-infected population had been infected through a past blood transfusion, with the majority infected prior to 1994. 34

| Morocco
The per cent of HCV-infected individuals who are active IDU was estimated to be 7.6% based on Libyan analogue data from 2008. 35,36 In a nationwide study carried out from 2005 to 2011, 16.1% of HCV-infected individuals in Morocco had experienced a blood transfusion. 37

| Nigeria
In 2013, expert consensus estimated that 1% of the infected population was assumed to be active IDU, and 10% of the infected population was assumed to have been infected through blood transfusion.

| Oman
In 2010, clinical data showed that 22.6% of the infected population was active IDU and that 21.1% had become infected through past transfusion. 38

| Qatar
Data on risk factors in Qatar are scarce. According to expert input, blood transfusion and dialysis are the most common risk factors among Qatari nationals, while active injection drug users are uncommon and make up <5% of the HCV-infected population in 2015. Analogue data from Saudi Arabia were used to estimate that 15% of the infected population has a history of blood transfusion. 39

| Taiwan
Experts estimated that there are 60 000 active IDU in Taiwan in 2015.

| Diagnosed
The total number of diagnosed cases was reported previously. 8 To estimate future total diagnosed cases, it was assumed that the number of newly diagnosed cases stayed the same as the last reported year.

| Treated & cured
The total number of treated HCV patients was similarly reported previously. 8 It was assumed that the number of treated patients stayed constant after the last reported year. It was also assumed that the number of treated patients for each genotype was proportional to the genotype distribution of the HCV-infected population. 8 The annual number of cured patients was estimated using the average sustained viral response (SVR) rate in a given year. A separate SVR was used for each of the major genotypes, as shown in Table 1.
Different methods were used to estimate the average SVR. All countries took into consideration a weighted average of different treatment options in a given year; pegylated-interferon (Peg-IFN)-based therapy in combination with RBV (dual therapy); Peg-IFN with RBV and a protease inhibitor (PI) (triple therapy); or direct-acting antiviral therapies (DAAs) with or without RBV. Some also took into consideration the percentage of the population who were treatment experienced and treatment naïve on each treatment option, while other countries took into account the disease stages of the patients being treated (eg F1, F2, F3 and F4).

| Treatment protocols
The pool of patients who could be treated was impacted by explicit or implicit treatment protocols. Explicit protocols were determined by national or international guidelines, whereas implicit protocols were determined by actual practice in the country revealed by consultations with experts. Prior to 2015, decompensated cirrhotic patients were considered ineligible for treatment in all countries.
According to the literature, approximately 40%-60% of HCV patients are eligible for Peg-IFN/RBV. 42,43 The definition of eligibility included contraindications to the drugs (eg psychiatric conditions) as well as patient's preference. In this analysis, 50%-100% of the patients were considered treatment eligible for the standard of care (Table 1).
In each country, the expert panel provided the most common stages of fibrosis considered for treatment (Table 1). Many countries use, or are starting to use, noninvasive testing methods to determine the level of fibrosis in patients. However, the METAVIR scale was used in this model to represent the severity/stage of liver fibrosis. The age of the patients was also considered. Table 1 outlines the most common age ranges considered for treatment. The data presented here do not imply that patients with lower METAVIR score or older/younger patients were not treated in each country. Instead, the data provided a range for the majority of treated patients.

| Future treatment protocols
In this analysis, it was assumed that the future treatment paradigm will remain the same as today. Thus, all assumptions (the number of acute cases, treated patients, per cent of patients eligible for treatment, treatment restrictions, the number of newly diagnosed annually and the average SVR by genotype) were kept constant in future years.

| RESULTS
The results of the analysis for 2015 are shown in Table 1. Figure 2 shows the age distribution of the HCV-infected population by country. Table 2

| Bahrain
Annual incidence was modelled with expert input to peak in 1991,

| Bulgaria
Expert consensus was used to estimate annual incidence.

| Colombia
Prevalence data, historical trends and expert input informed the estimated annual number of new cases. Incidence is assumed to have

| Croatia
Expert consensus was used to estimate annual incidence. In 2015, it was estimated that there were 190 new cases in Croatia (4.5 per 100 000).

| Dominican Republic
Expert consensus was used to estimate annual incidence.

| Ghana
Annual incidence was modelled with expert input to increase during the 1980s and 1990s, peak in the late 1990s and decrease thereafter cases, respectively, in 2015. In the same time period, the number of cases of HCC and liver-related death is also expected to increase by 70% and 65% from 1300 and 1400 cases, respectively, in 2015.

| Jordan
Expert consensus was used to estimate annual incidence.

| Kazakhstan
The number of new cases was estimated through a combination of expert consensus, prevalence data and historical trends. Incidence is assumed to have increased from the 1960's before stabilizing in the late  In 2015, there were an estimated 331 000 (256 000-398 000) viremic individuals in Kazakhstan. Viremic cases are expected to peak at 332 000 cases in 2018 and then decrease to 320 000 cases in 2030, a 3% reduction from 2015. Compensated and decompensated cirrhosis cases will increase 60% and 65% from a 2015 base of 18 000 and 1800, respectively. The number of HCC cases and liver-related deaths will increase by 65%, from a base in 2015 of 830 and 920, respectively.

| Malaysia
Annual incidence was estimated with expert input. In 2015, it was Similarly, liver-related deaths and HCC are both projected to rise by 100% from 2015 bases of 1400 and 1200 cases, respectively.

| Morocco
Expert consensus was used to estimate annual incidence.

| Nigeria
Incidence was modelled with expert input to increase exponentially

| Oman
Expert consensus was used to estimate annual incidence. In 2015, it was estimated that there were 310 new cases in Oman (

| Qatar
Expert consensus was used to estimate annual incidence,

| DISCUSSION
The countries in this analysis vary widely in population size, demographics, epidemic history, current risk factors and endemicity, and treatment rates. To account for the heterogeneity of countries considered and the changing disease burden over time, a modelling approach was used to estimate HCV morbidity and mortality in 2015 with forecasts to 2030. The standardized methodology used in this analysis presents an opportunity to compare the HCV disease burden across countries in a variety of contexts.
Globally, the prevalence of viremic HCV is around 1.0%. 7 In Ethiopia, the increase in viremic cases is primarily due to ongoing community and nosocomial transmission.
Viremic cases are forecast to decline most substantially in Taiwan and Qatar, with a 45% and 90% reduction in cases expected by 2030, respectively. In both countries, the average age of the population is >50 years (Figure 2), and in Taiwan, mortality is the primary driver for the declining prevalence. In Qatar, however, a 13% treatment rate has accelerated the reduction in both viremic cases and end-stage liver disease ( Figure 5) (Table 1) largely determined mortality (all-cause and liver-related).
In general, older populations have a higher all-cause mortality rate, 47 and, as HCV disease progression rates increase with age, older individuals were more likely to have HCV-related advanced liver disease and liver-related deaths. As mentioned previously, IDU cases have a higher mortality rate due to associated high-risk behaviour. Table 1 presents the proportion of the infected population who were actively injecting drugs. The all-cause mortality was adjusted accordingly for this portion of the population.

| Limitations
There are several limitations that may influence the outcomes from this study. Empirical incidence data starting in 1950 were impossible to get for all countries, so the distribution of new cases from 1950 to the most recent year of available data was back-calculated using prevalence from a known year, an estimation of the age and gender distribution of new cases and a relative incidence curve. All assumptions were reviewed extensively by the expert panels, and an analysis of key risk factors (eg IDU, nosocomial infection) was conducted in each country to inform incidence estimates after the year of known prevalence. In the absence of better information, the number of new cases was modelled to remain constant after 2015.
A second limitation is the assumption that diagnosis rates in each country will be sufficiently high to provide a pool of patients available and eligible for treatment. In reality, as the diagnosis rate increases, it will become more difficult to find undiagnosed patients. Furthermore, even if diagnosed, not all patients may have easy access to care. Thus, the ability of a country to treat its HCV-prevalent population may be limited by the number of available diagnosed eligible patients. This is particularly relevant to the low-income countries in the analysis where HCV continues to be a neglected disease.
In addition, the model does not consider the potential continued disease progression of cured HCV patients. Previous studies have indicated that it is possible for hepatic inflammation and disease progression to continue to HCC in more advanced patients after achieving SVR, 48,49 even though this happens at a slower rate. 3 As the analysis presented in this study was limited to HCV viremic infections, the outcomes may overestimate the reduction in cases of HCC and decompensated cirrhosis.

| Conclusion
In conclusion, this study demonstrates that overall viremic HCV prevalence is decreasing in most of the analysed countries due to a combination of an aging-infected population, expanding availability of more effective treatments and a concurrent reduction in risk factors, mainly the improvements in the safety of blood products and harm-reduction programs for injection drug users. However, morbidity attributable and mortality attributable to HCV are expected to increase as the current infected population progresses to advanced stages of liver disease. This is especially the case in countries with a large but young infected population, coupled with low diagnosis and treatment rates. A reduced disease burden and eventual elimination are possible if significant improvements are made to the overall cascade of care continuum, especially increases in screening, diagnosis and treatment. Thus, countries will need to evaluate different management strategies to guide decisions on how to best control the expected increase in their HCV-related disease burden. A series of such management strategies are discussed in the next analysis in this volume. 50