Low hepatitis C prevalence in Belgium: implications for treatment reimbursement and scale up

The newly developed direct-acting antivirals (DAAs) provoked a revolution in the treatment of chronic hepatitis C virus (HCV) infection. With a 95% success rate, minimal side effects and shorter treatment duration, the burden of the disease can drastically be reduced. Elimination of hepatitis C could now be within reach [1, 2] and scaling up of treatment of chronic HCV infection is crucial in the recently set hepatitis C elimination targets for 2030 of the World Health Organization (WHO) [3, 4]. Several studies have indicated that treatment of all those with chronic HCV infection is cost-effective [5‐7]. However, cost-effectiveness does not imply affordability and the high cost of DAAs has resulted in reimbursement restrictions in many countries. Reimbursement criteria often target more severe liver fibrosis stages and sometimes include restrictions concerning drugs or alcohol use [8, 9]. In Western Europe, Belgium remains one of the few countries restricting DAA reimbursement to fibrosis stage F2 or higher [9].

In order to estimate the cost of different reimbursement strategies, data on the prevalence of chronic HCV infection are needed. In Belgium, no nationwide HCV prevalence data exist and the last regional study on serum samples was conducted in 1993–1994 without distinction between resolved and current infection [10, 11]. The Belgian restricted DAA reimbursement criteria are therefore mainly guided by the 1993–1994 seroprevalence estimate of 0.87%, the lack of recent prevalence data and caution concerning the budgetary impact of different reimbursement strategies.

Here, we set out to update the HCV seroprevalence and prevalence of chronic HCV infection estimates in the Belgian general population in order to guide further decisions on DAA reimbursement criteria.

In total 3209 specimens, collected by 28 laboratories, were tested. Specimens were representative of the Belgian population in terms of sex and fairly representative for population per region, although the Walloon region was slightly undersampled. Individuals under 20 years of age were oversampled, as aimed for in the ESEN sampling strategy (Table 1). Of the 22 laboratories that shared this information, nine reported to receive samples from correctional facilities, and one from needle exchange programs.

Of all specimens, 14 (0.44%) tested ELISA-positive and were further tested in LIA. Six of 14 (43%) were found to be seronegative. Of the eight samples tested in RT-qPCR, two were found seropositive with resolved infection, four were found to be indicative of chronic HCV infection, and of two the HCV status remained undetermined (Fig. 1). The seropositive samples were found in a 12, 34, 40, 50, 53 and 67 year old. The two latter had a resolved infection.

After weighting for sex, age and population per district and adjusting for clustered sampling, HCV seropositivity was estimated to be 0.22% (95% CI: 0.09–0.54%), and prevalence of chronic HCV infection 0.12% (95% CI: 0.03–0.41%). In those of 20 years and older, these estimates were 0.26 (95% CI: 0.10–0.64) and 0.13 (95% CI: 0.04–0.43), respectively. No statistically significant differences were detected by sex, region or age group. We estimated there to be 24,420 (95% CI: 9990–59,940) HCV seropositive and 13,320 (95% CI: 3770–45,510) chronically infected individuals in the Belgian general population. The estimated clearance rate was 45%. Of the total estimated number of HCV seropositive individuals in Belgium, 66% were between 50 and 69 years old (Table 2, Table 3).

In this first Belgian nationwide HCV prevalence study, we estimated the HCV seroprevalence to be 0.22% which is considerably lower than the 0.87% found in the Flanders region in 1993–1994 by Beutels et al. [10]. The exact reasons for this difference are difficult to determine. We believe that it reflects, at least partially, a true decrease in HCV prevalence, due to both a decline in incidence and a natural turnover of HCV infected individuals, as the interval between the collection of both serum banks was exactly 20 years. The decline in incidence is achieved through efforts made to eliminate nosocomial transmission by screening of blood donors [13] and haemophilia patients, through programs such as needle exchanges programs targeting IDU, and through treatment breakthroughs in the last decades. Previously, Jadoul et al. reported a two-fold decrease in HCV prevalence between 1991 and 2000 in Belgian hemodialysis patients [14], hereby showing a considerable impact of precautionary measures on the HCV prevalence in certain risk groups. However, we acknowledge that differences in study design may have contributed as well to the different estimations. Although Beutels et al. also used residual sera and sampled similar hospital wards, they slightly oversampled migrants, without correcting for this afterwards, and did not ask for exclusion of known immunocompromised patients or patients with evidence of multiple blood transfusions. A different representation of these populations in both studies may have contributed to the different estimates.

On the other hand, our estimate of HCV seroprevalence is slightly higher than the 0.12% found in a saliva-based study in the Flanders region in 2003 [11]. In 2003, participants were recruited into the study and saliva tests, with a slightly lower sensitivity than serum tests [11], were used to test for the presence of HCV antibodies. Both these factors could explain our, slightly higher, estimate.

Within Europe, nationwide seroprevalence estimates range from 0.22 to 3.5%, with countries in Southern and Eastern Europe generally being the most affected [15‐17]. Our estimate is comparable to the 2012 Dutch serum bank-based estimate of 0.3% [16], and is exactly the same as the Dutch model-based estimate of 0.22% in individuals between 15 and 79 years of age [17]. The first of these studies recruited participants into the study, which could lead to an underestimation, however, the second one used a model to account for underrepresentation of specific risk groups and estimated the seroprevalence to be slightly lower.

The estimated Belgian prevalence of chronic HCV infection of 0.12% found in this study is substantially lower than the 2015 model-based viraemic prevalence estimate of 0.6% [18], which is predominantly driven by the seroprevalence estimate of 1993–1994. Our updated estimate is comparable to the Dutch model-based viraemic prevalence estimate of 0.1% and slightly lower than the 0.3% estimates of other surrounding countries like the UK, France and Germany [18]. However, some countries in the region, such as Ireland and Luxemburg, still report higher viraemic prevalences, with 0.6 and 0.9% respectively [18], this could be due to different population demographics and a different contribution of risk groups, such as IDU, to the estimates.

Our clearance rate of 45%, although based on a small number of positive samples, is higher than has been described by Seeff [19]. A similar clearance was observed by Garvey et al. in Ireland [20]. Unfortunately, we do not have details on treatment history of the participants, nor on the number of treated individuals in Belgium, but most likely some of these individuals were known HCV positive and have received treatment.

According to our study, 66% of HCV seropositive individuals in Belgium belong to the so called baby boom generation (born between 1945 and 1965). Due to the facts that this is based on a low number of seropositive individuals, that confidence intervals are wide and that information on risk factors is lacking, additional studies are needed to inform screening strategies. Even more so since screening the baby boom generation has proven to be effective in some countries, such as the United States [21], but other countries such as Germany [22] and the Netherlands [23], with population demographics more comparable to the Belgian one, have not found evidence to support this strategy.

In Belgium, reimbursement of DAAs is being gradually scaled up. Since January 2017, HCV patients with liver fibrosis stage F2 and higher and certain risk groups such as individuals with co-infection with HIV or hepatitis B virus, with blood clotting disorders, hemoglobinopathy, or with significant extrahepatic illness due to hepatitis C, dialysis patients and people who received an organ transplant hematopoietic stem cell transplant or bone marrow transplant (or are on the waiting list) [24], are considered eligible for treatment and reimbursement. However, neighbouring countries do not apply restrictions based on fibrosis stage, a strategy that is partly inspired by their low HCV viraemic prevalence [9, 18]. A 2016 Belgian economic analysis proposed a gradual scaling up of reimbursement of DAAs in order to limit the sudden budgetary impact [25]. The exact impact could not be calculated due to the lack of recent prevalence data of chronic HCV and lack of information on those infected. These new estimates would therefore call for a revision of the Belgian DAA reimbursement guidelines.

One of the main limitations of this study is that we may have underrepresented some specific risk groups, such as people with HIV/HCV coinfection, people with evidence of multiple transfusions, prisoners, migrants and injecting drug users (IDU). Due to exclusion of known immunocompromised individuals, we may have missed those people with HIV/HCV coinfection, which mainly occurs in men who have sex with men (MSM) and IDU. In 2013 there were 5650 diagnosed HIV-positive diagnosed MSM in Belgium [26]. Currently, no Belgian data on HCV seroprevalence in this group are available, but a meta-analysis estimated this to be 8.1% [27]. This would result in 458 HIV/HCV coinfected MSM in Belgium in 2013. In the same year, the total number of diagnosed HIV-positive IDU in Belgium was less than 500 [26]. These numbers show that even a complete exclusion of these groups would have had little impact on the estimates we provided. Unfortunately, the possible impact of excluding people with evidence of multiple transfusions cannot be calculated due to a lack of data. It could be argued that this is an important risk group that we have missed. Nevertheless, since the group was not systematically excluded as the information was not actively retrieved, the impact on our estimates was likely to be limited.

Although other high risk groups like prisoners, migrants and IDUs were not specifically excluded from the study, it is impossible to determine the extent to which they were representatively included in the study sample based on the data we had available. Almost half of the labs reported receiving samples from correctional facilities, but the access to public health care of prisoners is likely to be more restricted than that of the general population. Also IDU and certain migrant populations are less likely to access regular health care. Therefore, we have to acknowledge that our overall estimates might be biased towards medical care seekers and may represent a slight underestimation of the true prevalence. The use of emergency room samples, on the other hand, may have caused an overestimation, as higher prevalences in this population have been observed [28, 29].

Additionally, due to the oversampling of younger age groups, overall estimates and estimates in adults are less precise than they would have been with a population based sampling. For this reason, we had a low power to detect significant differences between age groups, sex or regions.

Finally, we cannot exclude an acute infection was misclassified as chronic which would result in an overestimation of the prevalence of chronicity. This does however not interfere with the conclusions of low seroprevalence and low prevalence of chronicity in the Belgian general population.

In conclusion, prevalence of HCV seropositivity and chronic infection in the Belgian general population is low and comparable to many surrounding countries. The adjusted prevalence can help estimate the cost of reimbursement of DAAs and invites Belgian policy makers to accelerate the scaling up of reimbursement, giving all chronically infected HCV patients a more timely access to treatment. Prompt treatment will not only prevent or limit the liver damage, which paradoxically now has to occur before being considered eligible for treatment reimbursement, but will also stop further transmission, which is key to the elimination of hepatitis C. Further studies are however needed to estimate prevalence in risk groups and to inform a screening strategy.