Disparate patterns of disease time burden in patients with HCC on immunotherapy or tyrosine kinase inhibitors.

Introduction
Hepatocellular carcinoma (HCC) is a prevalent malignancy, consistently ranking as one of the leading causes of cancer mortality worldwide.1 Surveillance and early diagnosis are paramount for improving HCC outcomes, because curative therapies, including resection, ablation, and liver transplant, are associated with the optimal outcomes.[2], [3], [4] Given suboptimal adherence to HCC surveillance programs, 30–50% of patients with HCC still present at advanced disease stages.5,6 Tumor size, major vessel invasion, extrahepatic metastases, and concomitant cirrhosis can preclude patients from receiving curative or locoregional therapies. Thus, systemic therapy might be the only treatment option for patients with advanced disease.7,8
Broadly classified into immunotherapies and tyrosine kinase inhibitors (TKIs), systemic therapy for HCC has developed rapidly in recent years. Systemic therapy significantly prolongs overall survival and has become the standard of care for advanced HCC.[9], [10], [11] Although cancer survivorship in HCC has improved, health-related quality-of-life (HRQoL) and patient-reported outcomes remain frequently overlooked in patients with HCC.12,13 Even among studies that reported HRQoL, comparisons were mainly between systemic therapy agents and placebo, with head-to-head comparisons between systemic therapy agents lacking. Increased efforts are warranted to study patient-reported outcomes in HCC systemic therapy, especially in the real-world setting. Indeed, the AASLD highlighted the importance of using patient-reported outcomes to guide treatment decisions in HCC.14
HRQoL is difficult to measure in patients with HCC taking systemic therapy, because it is influenced by multiple factors, including treatment adverse effects, complications of cirrhosis, and HCC-related symptoms. Days at home (DAH) is a patient-centered metric that can encapsulate the complexity of HRQoL. Defined as days alive and not requiring healthcare utilization,15 DAH captures multiple factors, including survival, inpatient stay, investigations, and medical procedures. It has been studied in patients with cirrhosis16,17 and in patients with cancer,18,19 yet it has not been specifically studied in HCC. Thus, we performed a population-based cohort study of patients with HCC receiving systemic therapy. We compared DAH in patients taking immunotherapy and TKIs, providing an in-depth analysis of disease time burden and healthcare utilization. Ultimately, we aimed to generate real-world data to guide treatment selection in patients with advanced HCC.

Patients and methods

Patient population
The data from this study were obtained from the Clinical Data Analysis and Reporting System (CDARS), a territory-wide electronic database managed by the Hospital Authority of Hong Kong. The Hospital Authority is the sole provider of public healthcare services in Hong Kong, covering ∼90% of healthcare services in the region. CDARS has detailed data covering patient demographics, disease diagnoses, drug prescriptions, laboratory tests, hospital admissions, emergency department attendance, and inpatient procedures.
We included patients with HCC who received systemic therapy between 1 January 2008 and 31 December 2023. Patients were identified based on the International Classification of Disease, 9th Revision, Clinical Modification (ICD-9-CM) diagnosis code 155.0. Systemic therapy agents were classified into immunotherapy (nivolumab, pembrolizumab, atezolizumab, durvalumab, tremelimumab, and ipilimumab; including monotherapies or combinations) or TKIs (sorafenib, lenvatinib, cabozantinib, and regorafenib; including monotherapies or combinations). Bevacizumab was predominantly used in combination with atezolizumab, and other combinations were also recorded if present. Patients were categorized based on the first systemic therapy agent prescribed (immunotherapy or TKIs), with the index date defined as the first date of receiving treatment. Given that our primary aim was to compare immunotherapy vs. TKIs, analysis was performed based on these distinct subgroups. Patients who received both immunotherapy and TKIs during the observation period (e.g. switching of drug class or use of combination regimens) were categorized and analyzed separately. Patients who were found to have other types of cancer before HCC were excluded from the analysis.
This study was performed in accordance with the Declarations of Helsinki and Istanbul. The study protocol was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (UW 23-102 and UW 25-073). Informed patient consent was waived, because data were collected for routine clinical service, and the study involved the use of archived data without identifiable information.

Definition of covariates
Baseline characteristics, including age, sex, pre-existing comorbidities (defined by the Charlson Comorbidity Index (CCI)), liver disease etiology (viral hepatitis: HBV or HCV; and non-viral hepatitis), cirrhosis, albumin–bilirubin (ALBI) score, model for end-stage liver disease (MELD) score, alpha-fetoprotein (AFP) levels, and previous HCC treatment (including curative treatment: surgical resection, liver transplant, and ablative therapies; and non-curative treatment: locoregional therapy, including transarterial chemoembolization [TACE] and stereotactic body radiation therapy) were determined based on ICD-9-CM diagnosis codes, treatment records, and clinical parameters. The grading for the ALBI score was defined as: grade 1, ALBI score ≤ -2.60; grade 2, ALBI score between -2.60 and -1.39; and grade 3, ALBI score > -1.39. ALBI grade 1 was defined as a low ALBI score, and ALBI grade 2–3 was defined as a high ALBI score.20 Among patients who received immunotherapy, the occurrence of immune-related adverse events (irAEs) was also documented. The detailed definitions and coding are depicted in Table S1.

Study outcomes
The primary outcome of this study was DAH within the first year of systemic therapy, which was defined as days alive and not requiring healthcare utilization.15 DAH was calculated by 365 days minus days utilized on inpatient stay, emergency department attendance, day procedures, blood tests, and imaging. For patients who passed away within the first year of systemic therapy, mortality days were also subtracted from the DAH (Table S2).16 As a secondary outcome, we assessed factors associated with DAH180, defined as patients who had ≥180 DAH within the first year of systemic therapy. The cutoff of 180 days was selected because it represents patients who spent more than half the year free from healthcare utilization. The cutoff of 180 days has also been used in previous reports.15,16

Statistical analysis
Categorical variables are presented as number (percentage) and were compared by Chi-squared test. Continuous variables were presented as median (IQR) or mean (SD), and were compared by Mann-Whitney U test or t test as appropriate. Differences with 95% CIs were estimated using linear regression models.16
Subgroup analyses were performed with stratification by age (<60 vs. ≥60-years old), etiology of HCC (viral vs. non-viral), ALBI grades (low [grade 1] vs. high [grade 2–3]), baseline AFP level (<400 vs. ≥400 ng/ml), MELD score (<10 vs. ≥10), and history of previous treatment for HCC. Analysis according to treatment regimen (first-line vs. non-first-line regimens by the American Society of Clinical Oncology [ASCO] 2024 HCC guidelines21) was performed. The first-line immunotherapy agents recommended by ASCO 2024 are atezolizumab plus bevacizumab or durvalumab plus tremelimumab, whereas the first-line TKIs are lenvatinib or sorafenib. Multivariate logistic regression was performed to determine independent factors associated with DAH180. Multivariate logistic regression was also performed in the immunotherapy group to assess for potential associations between irAEs and DAH180. Analyses were performed using Stata, version 17 (StataCorp LLC, College Station, TX, USA). Two-tailed p <0.05 was considered statistically significant.

Results

Patient characteristics
We included 4,677 patients with HCC who received systemic therapy between 1 January 2008 and 31 December 2023 (Fig. 1). Of these patients, 578 were in the immunotherapy group, 3,410 were in the TKI group, and 689 received both immunotherapy and TKIs within the study period. Their baseline characteristics are presented in Table 1. The mean age was similar between the immunotherapy and TKI groups (63.2 years vs. 62.7 years, p = 0.111). Most patients were men (86.0% male on immunotherapy and 85.5% male taking TKIs, p = 0.779), had viral hepatitis (86.0% in the immunotherapy group and 84.7% in the TKI group, p = 0.422), and had cirrhosis (73.7% in the immunotherapy group and 68.2% in the TKI group, p = 0.009).
In the immunotherapy group, the most commonly prescribed regimens were nivolumab monotherapy (35.8%), atezolizumab + bevacizumab (20.6%), and pembrolizumab monotherapy (14.4%) (Table S3). In the TKI group, sorafenib and lenvatinib was used in 77.3% and 21.9% of patients, respectively (Table S4). In patients who had received both immunotherapy and TKIs, 56.3% were initially taking TKIs with a subsequent switch to immunotherapy, 35.1% were initially taking immunotherapy with a subsequent switch to TKIs, and 8.6% were started on a combination of immunotherapy + TKIs (Table S5).

Days at home
The mean DAH was significantly higher in the immunotherapy group than in the TKI group (223.1 ± 134.7 vs. 183.3 ± 129.4 days, respectively, p <0.001; Table 2). Overall, the DAH within the first year showed a bimodal pattern, with most patients having DAH <90 days or ≥270 days (Fig. 2). In the immunotherapy group, 29.6%, 7.1%, 5.9%, and 57.4% of patients had DAH <90 days, ≥90 and <180 days, ≥180 and <270 days, and ≥270 days, respectively. In patients taking TKIs, 34.2%, 17.8%, 11.1%, and 36.9% had DAH <90 days, ≥90 and <180 days, ≥180 and <270 days, and ≥270 days, respectively. The TKI group had a significantly higher proportion of patients with DAH <90 days (p = 0.030), ≥90 and <180 days (p <0.001), and ≥180 and <270 days (p <0.001). Conversely, the immunotherapy group had a significantly higher proportion of patients with DAH ≥270 days (p <0.001).
The difference in DAH between immunotherapy and TKIs was primarily driven by lower 1-year mortality with immunotherapy (1-year mortality 58.1% in the immunotherapy group vs. 34.2% in the TKI group, improved overall survival by 42 days; 95% CI 30.2–53.8, p <0.001) (Fig. S1). Patients taking immunotherapy averaged 34.2 ± 23.1 days of healthcare utilization within the first year of treatment, whereas patients taking TKIs averaged 31.9 ± 26.7 days (p = 0.059). Immunotherapy was associated with significantly fewer days spent on inpatient stays (3.2 fewer days on average; 95% CI 1.3–5.1, p <0.001) and emergency department attendance (0.3 fewer days on average; 95% CI 0.2–0.4, p <0.001) compared with those taking TKIs. Conversely, immunotherapy was associated with more days spent on day procedures (4.7 more days on average; 95% CI 4.5–4.9, p <0.001; Table S6) and on blood tests/imaging (1.1 more days on average; 95% CI 0.3–1.9, p = 0.009) (Fig. 3). In their time alive within the first year of treatment, patients taking immunotherapy spent 78.1% of their time at home (not requiring healthcare utilization), whereas patients taking TKIs spent 76.4% of their time at home (p = 0.089).
We further analyzed patients based on the whole study period, without restriction to the first year of treatment. In this expanded analysis, patients taking immunotherapy had a median DAH of 185.0 (49.0–687.0), whereas patients taking TKIs had a median DAH of 162.0 (57.0–393.0) (p = 0.187). Immunotherapy remained associated with more day procedures and more time on blood tests/imaging, but was associated with fewer days of inpatient stay and emergency department attendance (all p <0.05) (Table S7).

Comparing immunotherapy and TKIs in subgroup analysis
Among patients with overall survival ≥6 months, the mean DAH was significantly higher in those taking immunotherapy than in those taking TKIs (311.9 ± 52.3 vs. 287.8 ± 71.0 days, respectively, p <0.001). Conversely, among patients with overall survival <6 months, TKIs were associated with a significantly higher mean DAH (57.5 ± 42.9 vs. 44.4 ± 35.6 days, respectively, p <0.001).
After stratification by other patient characteristics, immunotherapy remained associated with higher DAH over TKIs regardless of age, ALBI score, baseline AFP, and history of previous HCC therapies (all p <0.01) (Table S8). The association was also consistent among subgroups on first-line treatment regimens, viral-related HCC, and MELD score <10 (all p <0.001). No significant associations were demonstrable in patients on non-first-line regimens, non-viral HCC, and MELD score ≥10, possibly because of the small sample sizes.
We also performed additional analysis by restricting the study timeframe to 2017–2023, where higher lenvatinib usage in patients taking TKIs was anticipated. Between 2017 and 2023, immunotherapy remained associated with higher DAH compared with TKIs (average 224.2 ± 134.4 vs. 201.9 ± 130.1 days, respectively, p <0.001) (Table S8).

Factors associated with DAH180
In multivariate logistic regression, immunotherapy use (odds ratio (OR) 1.738, 95% CI 1.407–2.147, p <0.001), older age (OR 1.013, 95% CI 1.005–1.020, p = 0.001), and previous HCC therapy (OR 1.722, 95% CI 1.508–2.082, p <0.001) were independently associated with DAH180 (≥180 DAH within first year of therapy). Conversely, higher ALBI grade (OR 0.622, 95% CI 0.508–0.762, p <0.001), MELD score (OR 0.753, 95% CI 0.606–0.935, p = 0.010), AFP (OR 0.522, 95% CI 0.433–0.630, p <0.001), aspartate aminotransferase (AST) (OR 0.646, 95% CI 0.534–0.782, p <0.001), and CCI (OR 0.940, 95% CI 0.915–0.966, p <0.001) were associated with lower odds of DAH180 (Table S9).

Impact of irAEs on DAH
Within the first year of treatment, irAEs occurred in 71 patients (12.3% of patients taking immunotherapy; Table S10). The mean DAH for patients taking immunotherapy with and without irAEs was 194.5 ± 128.1 days and 227.1 ± 135.2 days, respectively, and this difference trended toward significance (p = 0.056). Patients taking immunotherapy without irAEs had significantly higher DAH compared with those taking TKIs (43.8 more days on average; 95% CI 31.6–55.9, p <0.001). Conversely, no significant differences in DAH were observed between patients taking immunotherapy with irAEs and patients taking TKIs (p = 0.469) (Table S11).
Among patients taking immunotherapy, irAEs had no significant association with 1-year overall survival (p = 0.227), but were associated with a significantly higher number of days on inpatient stay (8.3 more days on average; 95% CI 4.7–11.8, p <0.001), emergency department attendance (0.5 more days on average; 95% CI 0.3–0.8, p <0.001), and blood tests/imaging (3.6 more days on average; 95% CI 1.3–6.0, p = 0.002) (Table S11).
In multivariate analysis of patients taking immunotherapy only, development of irAEs (OR 0.523, 95% CI 0.290–0.942, p = 0.031), baseline ALBI grade (OR 0.484, 95% CI 0.294–0.797, p = 0.004), AFP levels (OR 0.462, 95% CI 0.285–0.750, p = 0.002), AST (OR 0.459, 95% CI 0.276–0.762, p = 0.003), and CCI (OR 0.888, 95% CI 0.820–0.962, p = 0.044) were independently associated with lower odds of DAH180 (Table S12).

Impact of immunotherapy + TKIs and other specific treatment regimens on DAH
Patients who received both immunotherapy and TKIs averaged 269.6 ± 94.1 DAH within the first year of therapy, which was significantly higher than the DAH in patients taking immunotherapy only (mean 223.1 days) and those taking TKIs only (mean 183.3 days) respectively (both p <0.001) (Table S13). On multivariate analysis, use of both immunotherapy and TKIs was independently associated with DAH180 (OR 3.583, 95% CI 2.840–4.520, p <0.001). On further stratification by treatment sequence, initiation of either immunotherapy or TKIs first had no significant impact on DAH (272.8 ± 93.3 vs. 267.2 ± 94.7 days, respectively, p = 0.441) (Table S13).
Among patients taking immunotherapy, use of bevacizumab-containing regimens was associated with significantly higher DAH compared with the use of immunotherapy monotherapy (282.2 ± 110.0 vs. 207.1 ± 134.9 days, respectively, p <0.001). Conversely, immunotherapy combinations and monotherapy led to comparable DAH (199.6 ± 143.8 vs. 207.1 ± 134.9 days, respectively, p = 0.651). On multivariate analysis, bevacizumab-containing regimens were associated with higher probability of achieving DAH180 (OR 2.576, 95% CI 1.466–4.529, p = 0.001).
Among patients taking TKIs, lenvatinib was associated with significantly higher DAH than was sorafenib (223.2 ± 126.0 vs. 172.0 ± 128.0 days, respectively, p <0.001).

Discussion
This study utilized a large real-world cohort to study the disease time burden in patients with advanced HCC. We established the benefits of immunotherapy over TKIs, because the former was associated with significantly higher DAH within the first year of treatment. The differences in DAH were primarily driven by improved survival with immunotherapy. This concurs with data from the immunotherapy registration trials, which showed superior overall survival with immunotherapy over TKIs.22,23 Our data also showed the potential DAH benefits of bevacizumab-containing immunotherapy regimens, supporting the synergistic effects of bevacizumab with immunotherapy. Although clinical trials generally included selected patient cohorts,7 our study provided real-world data, and demonstrated robust findings in sensitivity and subgroup analyses. Furthermore, we demonstrated that the survival benefits from immunotherapy translated to higher DAH, which could improve patients’ HRQoL.
When assessing disease time burden in detail, distinct patterns were observed between the immunotherapy and TKI groups. Patients taking immunotherapy spent significantly fewer days on inpatient stay or in attending the emergency department, which might be attributable to lower rates of disease progression and clinical deterioration.22,23 Conversely, immunotherapy was associated with more days on day procedures and investigations. This is an expected finding, given that immunotherapy requires intravenous administration (day procedures) and frequent reassessment.2,7 By contrast, TKIs are administered orally and require less scheduled healthcare utilization. Among patients with survival <6 months, TKIs were associated with significantly higher DAH than was immunotherapy, highlighting the disparate patterns in healthcare utility after controlling for patient’s overall survival. Although clinicians choose therapies based on patient characteristics and treatment efficacy, patients might choose therapies based on disease time burden and tolerability. Our data are beneficial for patient counselling, because they enable patients with HCC to understand their projected healthcare utilization before receiving systemic therapies. In turn, this enables shared decision-making between clinicians and patients.
Our study further assessed predictors of DAH in patients with HCC taking systemic therapies. Outside of immunotherapy use, older age was associated with higher DAH. A potential explanation is that young-onset HCC has more aggressive tumor behaviour, which results in poorer prognosis and higher healthcare utilization.24 Previous HCC therapy was another factor associated with higher DAH. We hypothesize that patients with previous HCC therapies, compared with patients who first present with advanced HCC, might have a lower tumor burden because of intensive surveillance and better residual liver function. By contrast, higher ALBI grade, MELD score, and AFP levels, which are all factors representing more severe disease, were independently associated with lower DAH.
The complicated profiles of systemic therapies can influence healthcare utilization and patient acceptability. In particular, irAEs in immunotherapy can have substantial yet diverse impacts on HRQoL. irAEs can manifest with severe multiorgan involvement, which necessitates prolonged therapies. Although uncommon, refractory irAEs causing mortality have also been reported.25 Conversely, irAEs can represent target engagement by immunotherapy, which is associated with improved prognosis and survival in specific cohorts.26 Overall, the impact of irAEs on HRQoL remains poorly understood, and this topic has not yet been studied in HCC populations.[27], [28], [29] In our real-world cohort, irAEs had no significant impact on survival, but led to significantly higher time burden on inpatient stay, emergency department attendance, and investigations. Development of irAEs led to significantly lower DAH among patients taking immunotherapy. These findings provide additional considerations for therapy selection, and patients must be clearly counselled on the potential complication profiles of immunotherapy.
As HCC systemic therapies further develop, treatment regimens and algorithms will likely become more complex. Combination therapies with immunotherapy + TKIs have shown positive results in clinical trials, and could be increasingly utilized by clinicians.30,31 Switching between immunotherapy and TKIs upon disease progression is also recommended in international guidelines.[9], [10], [11] Patients who have experienced both immunotherapy and TKI s had significantly higher DAH compared with those taking only immunotherapy or only TKI in our study. This might be attributable to survival bias, given that patients with longer survival and better general condition are more likely to experience treatment regimen changes. By contrast, patients with rapid disease progression and deterioration might not receive second-line systemic therapies, and might instead be put on best-supportive care. In future studies, DAH should be specifically studied in patients with combination treatment or with drug-class switching. These data would be valuable in understanding the disease time burden in these unique patient groups.
In 2022, the AASLD Practice Metrics Committee defined important HCC outcomes as outcomes that are impactful for patients or clinicians, meaningful across populations, and can facilitate change.32 As an explainable metric that can guide treatment decisions, DAH clearly fits the criteria as an important HCC outcome. Thus, DAH should be reported in upcoming HCC systemic therapy trials, because it would provide valuable data to improve the HRQoL of patients with advanced HCC.14 Although DAH is a useful HRQoL metric, the inclusion of other HRQoL metrics, such as the Chronic Liver Disease Questionnaire (CLDQ)33 and the Short form Health Survey (SF-36),34 would provide a more comprehensive overview of HRQoL in patients with HCC in future studies.
A limitation of this study is that granular details on HCC characteristics (e.g. tumor size and number) were not available from our territory-wide cohort. Nonetheless, systemic therapies (immunotherapy or TKIs) are generally used in advanced HCC in our locality. Neoadjuvant or adjuvant treatment are only utilized in trial settings, and should not be captured in our territory-wide database.2 Thus, our cohort represents a relatively homogenous population of patients with advanced HCC. By utilizing our territory-wide database, we were able to maximize patient recruitment (n = 4,677) to generate comprehensive data on DAH.
Another limitation is the potential underdiagnosis of irAEs through diagnostic coding, which is widely acknowledged in immunotherapy research.35,36 Despite potential underdiagnosis, our study demonstrated the negative impact of irAEs on DAH. We hypothesize that the negative impact of irAEs might in fact be more prominent than reported by us. Experts have called for the improvement of coding systems for irAEs,35 and enhanced reporting will be paramount for future studies in immunotherapy.
To conclude, immunotherapy was associated with significantly reduced disease time burden in patients with advanced HCC when compared with TKI treatment. However, this HRQoL benefit was dampened if irAEs occurred with immunotherapy. Our data have important HRQoL implications, and could influence patients’ decisions over which therapy to receive.

Abbreviations
AFP, alpha-fetoprotein; ALBI, albumin–bilirubin score; ASCO, American Society of Clinical Oncology; AST, aspartate aminotransferase; CCI, Charlson Comorbidity Index; CDARS, Clinical Data Analysis and Reporting System; CLDQ, Chronic Liver Disease Questionnaire; DAH, days at home; HR, hazard ratio; HCC, hepatocellular carcinoma; HRQoL, health-related quality-of-life; ICD-9-CM, International Classification of Disease 9th Revision Clinical Modification; irAE, immune-related adverse event; MELD, Model for End-Stage Liver Disease; OR, odds ratio; SBRT, stereotactic body radiation therapy; SF-36, Short form Health Survey; TACE, transarterial chemoembolization; TKI, tyrosine kinase inhibitor.

Financial support
This study was supported by the Simon KY 10.13039/100007419Lee Foundation.

Authors’ contributions
Study concept and design: RWHH, MSHC, WKS, LYM. Data acquisition: RWHH, MSHC, LL, XM, CLC, CKHW, ICKW, KSC, MFY. Data interpretation: RWHH, MSHC, LL, XM, CLC, CKHW, ICKW, KSC, MFY, LYM. Data analysis: RWHH, MSHC, LYM. Drafting of manuscript: RWHH, MSHC. Critical revision of manuscript: WKS, LYM. Overall study supervision: LYM. Preparation of the manuscript, and have seen and approved the final version: all authors.

Data availability
Data from this study are available from the corresponding authors on reasonable request.

Conflicts of interest
CLC reports receiving grants from 10.13039/100009945Merck KGaA and Taiho and personal fees from 10.13039/100030732MSD, 10.13039/501100003769Eisai, Taiho, Varian, and 10.13039/501100003769Eisai outside the submitted work. ICKW reports funding from 10.13039/100002429Amgen, Bristol Myers Squibb, 10.13039/100004319Pfizer, Janssen, Bayer, GSK, 10.13039/100004336Novartis, the Hong Kong Research Grants Council, the Hong Kong 10.13039/100018696Health and Medical Research Fund, the 10.13039/501100000272National Institute for Health Research in England, the 10.13039/501100000780European Commission, and the 10.13039/501100000925National Health and Medical Research Council in Australia, outside the submitted work. He is also a non-executive director of Jacobson Medical in Hong Kong and a consultant to IQVIA and World Health Organization. MFY is an advisory board member and/or received research funding from 10.13039/100006483AbbVie, Arbutus Biopharma, Assembly Biosciences, Bristol Myer Squibb, Dicerna Pharmaceuticals, GlaxoSmithKline, Gilead Sciences, Janssen, 10.13039/100004334Merck Sharp and Dohme, Clear B Therapeutics, Springbank Pharmaceuticals; and received research funding from 10.13039/100014931Arrowhead Pharmaceuticals, Fujirebio Incorporation and 10.13039/100017981Sysmex Corporation. WKS received speaker’s fees from 10.13039/100024223Echosens, is an advisory board member for and received speaker’s fees from Abbott, received research funding from 10.13039/100004325AstraZeneca, 10.13039/100006396Alexion Pharmaceuticals, Boehringer Ingelheim, 10.13039/100004319Pfizer, and Ribo Life Science, and is an advisory board member for and received speaker’s fees and researching funding from Gilead Sciences. LYM received research funding from Gilead Sciences and 10.13039/100016545Roche Diagnostics. The remaining authors have no conflict of interests to declare.
Please refer to the accompanying ICMJE disclosure forms for further details.