Loss to follow‐up in the hepatitis C care cascade: A substantial problem but opportunity for micro‐elimination

Abstract Since the advent of direct‐acting antivirals, elimination of hepatitis C viral (HCV) infections seems within reach. However, studies on the HCV cascade of care show suboptimal progression through each step for all patient groups. Loss to follow‐up (LTFU) is a major issue and is a barrier to HCV elimination. This review summarizes the scale of the LTFU problem and proposes a micro‐elimination approach. Retrieving LTFU patients and re‐engaging them with care again has shown to be feasible in the Netherlands. Micro‐elimination through retrieval can contribute to reaching the World Health Organization's viral hepatitis elimination targets by 2030.


| BACKG ROU N D
The global hepatitis C virus (HCV) epidemic stimulated the World Health Organization (WHO) to develop viral hepatitis elimination targets in 2016. 1 An estimated 71 million people worldwide were infected by HCV in 2015. 2 Thus, the WHO set the target of a 90% reduction in new infections and a 65% reduction in viral hepatitis-related mortality by 2030 as compared to 2015. These are ambitious but feasible goals, since we have ample tools at hand to curtail the current HCV epidemic. The diagnosis of active HCV can be readily made, by means of sample analysis in a central facility or through point-of-care testing. Direct-acting antivirals (DAAs) cure the infection in ≥95% of cases. 3 Pangenotypic DAAs can be used in all patients with only a few barriers such as potential drug-drug interactions or presence of (decompensated) cirrhosis. 4,5 Most countries have assessed their specific HCV population and the availability of tools in their countries and subsequently developed national hepatitis plans in line with the WHO elimination targets. 6 HCV elimination according to the WHO goals can be achieved in various ways, which ideally should be incorporated in a multifaceted approach. We can focus on prevention, by developing a vaccine or by increasing awareness and educating groups at risk of transmission of the virus. Secondly, we can develop or augment existing screening strategies, in order to diagnose more patients.
Lastly, we can treat as many infected patients as possible. Since the development of highly effective and tolerable DAAs, HCV elimination projects have primarily focused on prevention and screening, since treatment was not seen as a problem anymore. However, ensuring treatment for all diagnosed patients remains a problem to this day.
Loss to follow-up (LTFU) prevents patients from receiving the care they need to be cured of their infection. The extent of this problem remains unclear, especially in the DAA era. In order to grasp the scope of the LTFU problem, one needs to understand the HCV care cascade and how patients move through its phases. This review aims to assess published literature on LTFU in the HCV cascade of care during the DAA era and will provide an overview of issues and possible solutions.

| CON CEP T OF LOSS TO FOLLOW-UP
Different definitions for LTFU are used in the literature, since patients become lost to follow-up for various reasons. For example, they may have moved house, emigrated, died or been imprisoned. Many times, the reason for LTFU cannot be ascertained as contact with the patient cannot be established. Retrospective observational studies often do not provide a specific definition or use nonattendance to any appointment as a definition. Suppl file 1-11 Some of these studies mention death separately and do not include this as a reason for LTFU. Suppl file [12][13][14][15][16][17] Other studies that reviewed ever-diagnosed patients defined LTFU as patients who never or not recently had an appointment with an HCV specialist. Suppl file 18,19 Interventional studies aiming to improve the cascade of care also do not give a definition or define LTFU as nonattendance anywhere in the care cascade, Suppl file 20-49 often separating death from the LTFU group. Suppl file 50-58 There are some studies that include multiple LTFU definitions and report data on all of them, such as nonattendance, nonresponse to invitation, moved, incarcerated, no insurance and comorbidities. Suppl file 59-62 There may be a lesson to be learned on defining LTFU from studies in other fields of medicine. Previously mentioned HCV studies did not take time into account when defining LTFU. Prospective studies defined LTFU as nonattendance at the end of their study period, which varied greatly among studies. Retrospective studies defined LTFU as nonattendance since their last visit up to study initiation. HIV studies have investigated LTFU extensively and showed that the way you define LTFU greatly influences your LTFU outcomes. 7 In addition, these studies have demonstrated different ways to determine the ideal timeframe to classify someone as LTFU, that is x days after last clinic visit. 8,9 When different studies use different definitions, it is virtually impossible to compare care cascades and combine results. However, since this is the case for the HCV studies assessed in this review, we chose a pragmatic approach that suits the illustrative purpose of this review. We define LTFU as nonattendance to any appointment in the care cascade at any time since their last visit. Patients who had died were not included in the definition of LTFU.

| THE H C V C ARE C A SC ADE
In order to grasp the magnitude of the LTFU problem in chronic HCV patients, we must first understand the HCV care cascade.
Reviewing published literature on this subject shows that definitions of the care cascade vary with each paper. However, efforts to come up with an unambiguous description of the HCV care cascade have been made. In 2018, the WHO established a monitoring framework that includes 10 core indicators addressing prevention, diagnosis, treatment and mortality. 10 The WHO states that four of the 10 core indicators should be used for cascade of care reporting: the number of patients infected, diagnosed, treated and cured. 11 Recently, a study group comprised of clinical, epidemiological and public health experts from Australia, Europe and North America have proposed a clarified and slightly extended care continuum. 12  Step 1: HCV prevalence; step 2: diagnosed with chronic HCV; step 3: linked to care; step 4: liver disease assessed; step 5: started on treatment; step 6: achieved SVR; step 7: accessed chronic post-SVR care. Figure freely adapted with permission from Safreed-Harmon et al. 12 HCV, hepatitis C virus; LTFU, lost to follow-up; SVR, sustained virological response chronic HCV; (c) linked to HCV care; (d) liver disease assessed; (e) started on treatment in (year); (f) achieved sustained virological response (SVR) in (year); and (g) accessed chronic post-SVR care. The authors provided pragmatic definitions for the four key steps, which stakeholders can use to report on elimination progress. Understandably, by increasing the number of steps in the care cascade, the chances of being lost from care also increase. LTFU is seen as a major problem, because it remains unsure whether the patient is cured or not. Liver disease in these patients may progress, and they may even contribute to HCV transmission if they still exhibit certain risk behaviour. When reviewing the literature published on HCV care cascades in the DAA era and their LTFU rates, we used the CHCoC to report our findings (see Figure 1). And overview and characteristics of the included studies in this review can be found in Table 1.

| LTFU DURING D IAG NOS TI C A SS E SS MENT (CH CO C S TEP 2)
The first step in diagnosing chronic HCV is the determination of presence of HCV antibodies. However, the key step in confirming the diagnosis of chronic HCV is determining HCV RNA (or HCV core antigen when RNA assays are not available or not affordable). 13

| LTFU DURING LIVER D IS E A S E A SS E SS MENT (CH CO C S TEP 4)
Several diagnostic procedures are available to grade and stage liver disease. Where liver biopsy was standard of care in the past, nowadays noninvasive methods are largely preferred. Liver fibrosis may be quantified by using serological panels, such as the widely used FIB-4 (using the patient's age, platelet count, AST and ALT levels) or APRI score (using AST levels, the AST upper limit of normal and platelet count), or by using transient elastography.

| LTFU B EFORE INITIATING TRE ATMENT (CH CO C S TEP 5)
Even in the era of highly effective DAAs, treatment initiation rates are low. LTFU proves to be a large contributor to this problem.
Retrospective studies have shown that only 12%-77% (median 29%) of patients diagnosed or engaged in care during the DAA era initiated treatment after being diagnosed with chronic HCV. Suppl file [1][2][3][4][5]13,34 Interventional studies aimed to improve the care cascade show that this rate can increase to 16%-100% (median 73%). Suppl file 18-28,33,35-39,50-53,59-62 Studies in the HIV field show similar results, with 36%-91% (median 90%) initiating treatment in retrospective studies Suppl file 9-11 and 25%-100% (median 80%) in interventional studies. Suppl file 47-49 However, the treatment rate remains suboptimal in PWID with only 20%-90% (median 53%) initiating treatment. Suppl file 29-32,40,41 Generally, treatment initiation rates are higher in decentralized settings, both in PWID and non-PWID populations. Reasons for poor treatment initiation rates vary. Unfortunately, many countries still experience restrictions in who can and cannot be treated with DAAs. 14,15 This problem may especially apply to studies from the first stages of the DAA era. 16 Other reasons for poor treatment initiation rates may be comorbidities or perceived lack of compliance. However, LTFU contributes to a large extent to these poor rates. Studies showed that LTFU is the reason for nontreatment in 0%-67% (median 33%) of cases. Suppl

| LTFU DURING OR AF TER TRE ATMENT (CH CO C S TEP 6)
As we know from registration trials, DAAs are highly effective. Many studies report intention-to-treat SVR percentages, defined as the proportion of patients who reached SVR out of the number of patients that initiated DAA therapy (see Figure 2). In studies including mixed populations, ITT SVR varies from 22% to 98% (median 83%

| LTFU AF TER S VR (CH CO C S TEP 7 )
Guidelines suggest that people with advanced fibrosis (METAVIR score F3) or cirrhosis (F4) who reached SVR should be subjected to surveillance for hepatocellular carcinoma (HCC) every six months by means of ultrasound. 13,17 Furthermore, cirrhotic patients with varices present at pre-treatment endoscopy should be surveyed for oesophageal varices. 13,17 Unfortunately, data on how many cured patients actually receive such surveillance are lacking. Most studies stop reporting on the care cascade at the moment SVR is reached.
A recent review showed that less than 30% of cirrhotic patients are included in surveillance programmes, independent from aetiology. 18 More studies on the surveillance adherence among cured HCV patients with an indication for surveillance are needed.  20 This provides us with an opportunity as the LTFU HCV population may be an excellent candidate for micro-elimination, the process of eliminating HCV in subpopulations. 21 Micro-elimination is the favoured approach in many countries, especially in those with a relatively low national prevalence, but higher prevalence in specific subpopulations.

| FAC TOR S A SSO CIATED WITH LTFU
Lazarus et al have recently described which subpopulations should be considered for micro-elimination, such as aboriginal and indigenous communities, HIV/HCV-coinfected people, migrants from high-prevalence countries, people who inject drugs, people with inherited blood disorders and prisoners. 22 We propose that LTFU patients should be added to this list. As indicated, LTFU is a substantial problem across the entire care cascade. Because this HCV population has already been identified, it is obvious that retrieval of these patients can be considered 'low-hanging fruit'.

| RE TRIE VAL IN THE NE THERL ANDS: THE CELINE PROJEC T
Several regional projects have been executed in the Netherlands focused on the LTFU population. We found that up to 14% of our HCV population diagnosed in the previous 15 years was LTFU before being cured and eligible for retrieval. [23][24][25] Based on best practices from these projects, a nationwide approach was developed. 26,27 The hepatitis C Elimination in the Netherlands (CELINE) project aims to retrieve LTFU chronic HCV patients and re-engage them with care. The protocol is described in detail elsewhere. 27 In short, we identify diagnosed patients based on laboratory data, which we combine with information from their medical records. Patients who were still HCV-positive when they left care are classified as eligible for retrieval if they are alive and currently residing in the Netherlands. They are subsequently invited by letter to an outpatient clinic of their choice after their current address is verified through municipality records or general practitioners. Data will be collected on patient and disease characteristics of patients who sign informed consent.
What we have learned since the start of CELINE in 2018 is that retrieval is feasible when conducted by a dedicated team. The project gives great insight into our care cascade and gives vital information for our hepatitis elimination plan. The nationwide approach ensures that retrieval is done to the same standards in each participating centre.
Identification of LTFU patients and ensuring they adhere to their clinic appointments are the most time-consuming. This is why we advise that a dedicated team, rather than individual clinicians, should execute these tasks.

| IN CLUDE RE TRIE VAL IN S TANDARD C ARE
Ideally, micro-elimination through retrieval should become standard of care. This concept is not (yet) mentioned in any guidelines or elimi-

| CON CLUS ION
LTFU is a problem in each step in the HCV care cascade, even in the current era where highly effective treatments are available and where it has been possible to simplify the cascade of care. HCV can be micro-eliminated in the LTFU population through retrieval.
We present an example of nationwide retrieval in the Netherlands, which shows retrieval is feasible and can contribute to HCV elimination. We propose that micro-elimination through retrieval becomes standard of care on the road to viral hepatitis elimination.
Furthermore, efforts to retain patients in care should be implemented in daily clinical practice.

ACK N OWLED G EM ENTS
The CELINE project is part of the LEGA-C™ programme. LEGA-C is meant to suggest 'Local Elimination Programs leading to Global Action in HCV' , and Gilead is using the LEGA-C programme to help streamline and support multiple HCV elimination initiatives across the world, including the SCALE (SCreening, Access, and Linkage to CarE) programme.