ABSTRACT
Introduction Increasing rates of liver transplantation and improved outcomes have led to greater numbers of transplant recipients followed up in non-transplant centres. Our aim was to document long-term clinical outcomes of liver transplant recipients managed in this ‘hub-and-spoke’ healthcare model.
Methods A retrospective analysis of all adult patients who underwent liver transplantation between 1987 and 2016, with post-transplant follow-up in two non-transplant centres in the UK (Nottingham) and Canada (Ottawa), was performed.
Results The 1-, 5-, 10- and 20-year patient survival rates were 98%, 95%, 87% and 62%, and 100%, 96%, 88% and 62% in the Nottingham and Ottawa groups, respectively (p=0.87). There were no significant differences between the two centres in 1-, 5-, 10- and 20-year cumulative incidence of death-censored graft-survival (p=0.10), end-stage renal disease (p=0.29) or de novo cancer (p=0.22). Nottingham had a lower incidence of major cardiovascular events (p=0.008).
Conclusion Adopting a new model of healthcare provides a means of delivering post-transplant patient care close to home without compromising patient survival and long-term clinical outcomes.
Introduction
Liver transplantation (LT) remains the only lifesaving treatment option for patients with end-stage liver disease and acute liver failure (ALF), and for selected patients with hepatocellular carcinoma (HCC) in whom other curative treatment options have failed or are not suitable. Over the past 10 years, rates of liver transplantation have increased by more than 25% in Canada (14.4 transplants per million; Canadian Institute for Health Information (CIHI)) and 40% in the UK (15.2 transplants per million; NHS Blood and Transplant (NHSBT)). Outcomes following LT have also improved substantially. Survival rates of >90%, 85% and 70% at 1-, 5-, and 10-years post-LT, respectively, have become the accepted norm around the world.1–3
More transplants and improved long-term outcomes have led to greater numbers of patients living with a liver transplant. In 2018, there were more than 10,100 and 5,900 recipients living with liver transplant in the UK and Canada, respectively. The UK has seven liver transplant centres (0.1 per million), the same number as Canada (0.18 per million), although Canada's population is spread over nearly 10 million square kilometres. In both countries, many liver transplant recipients are followed at non-transplant centres, and these numbers are likely to continue to grow in future due to constraints of both population size and density.
A ‘hub-and-spoke’ model of healthcare, where secondary care establishments route patients to a specialised centre for specific services, and receive training, advice and support in return, have been shown to improve patient access to excellent care.4,5 The UK has a more formal hub-and-spoke model of delivering transplant care at non-transplant centres than Canada.6 This study aims to document the long-term outcomes of patients followed up in non-transplant centres in the UK and Canada in order to inform future policy guidelines.
Methods
Study population
This was a retrospective analysis of data from two tertiary care, non-transplant centres in the UK and Canada. All adult patients who underwent LT between 01 January 1987 and 31 December 2016 and followed up in either Nottingham University Hospital (UK) or The Ottawa Hospital (Canada) were eligible for inclusion. Those who underwent LT at <18 years of age and those who underwent combined organ transplantation were excluded. Data were extracted from electronic records and supplemented with manual chart review. This study was approved by the Nottingham University Hospitals NHS Trust Clinical Effectiveness Board (19-416C) and the Ottawa Health Science Network Research Ethics Board Protocol (REB #20170905-01H).
Liver transplantation selection and post-transplant follow-up
Selection of patients in Nottingham and Ottawa follow the national policy governed by the NHSBT and the Ontario provincial policy governed by the Trillium Gift of Life Network, respectively (summarised in supplementary material S1).
All Nottingham patients have been transplanted at Cambridge University Hospital since 2009 following a formal network arrangement; prior to this, patients were transplanted at various centres through informal arrangements. Following transplantation, all patients are reviewed in a joint clinic in Nottingham, attended by teams from Cambridge and Nottingham. The care is then transferred entirely to the Nottingham team during the first 6–12 months post-transplantation. However, patients who require input from the transplant team are reviewed in the joint clinic as needed. Similarly, Ottawa patients are primarily transplanted at Toronto General Hospital, with some transplants occurring in London (Ontario) and Montreal (Quebec), according to the preference of the patient / primary clinician. Following transplantation, patient care is transferred entirely to Ottawa during the first 3–12 months. Monthly meetings using video-conferencing is undertaken between the Ottawa team and the transplant centre where patients with transplant-related complications are discussed. Family physicians are responsible for regular monitoring of metabolic complications, and prescribe treatment based on individual care.
Outcomes
The primary outcome was patient survival, defined as the time from LT to death from any cause. Death-censored graft survival, development of end-stage renal disease (ESRD), major cardiovascular events (CVE), and de novo cancer (including non-melanoma skin cancers) were secondary endpoints. ESRD was defined as glomerular filtration rate (GFR) <15 mL/min/1.73 m2 for ≥6 months or requirement of renal replacement therapy. Myocardial infarction and stroke were considered as major CVE.
Statistical analysis
Continuous variables were presented as median with interquartile range (IQR) and categorical variables were presented as number and percentage. A p value of <0.05 was considered significant.
Kruskal–Wallis, Mann–Whitney and chi-squared tests were used to compare demographics between Nottingham and Ottawa groups. Patient survival, death-censored graft survival, and development of ESRD, major CVE and de novo cancer were calculated using Kaplan–Meier analysis (log-rank test).
Difference in outcomes between Nottingham and Ottawa groups was investigated using a multivariable logistic regression model, incorporating variables with a p value of <0.10 on univariate analysis. Variables were considered to have independent association only if the p value reached Bonferroni-corrected level of significance.
All statistical analyses were performed using either Prism 8 (GraphPad, San Diego, USA) or SPSS v26 (IBM, Armonk, USA).
Results
Demographics and clinical characteristics
A total of 335 patients who underwent LT from Nottingham, UK, (132) and Ottawa, Canada, (203) were included in this study. The demographics and clinical characteristics at LT are summarised in supplementary material S2.
Nottingham patients were younger at transplantation (median 54 years (IQR 39–59) vs 56 years (IQR 49–62); p=0.007) with lower rate of history of smoking prior to transplantation (28.8% vs 45.8%; p=0.002). Although decompensated cirrhosis was the most common indication for LT in both the Nottingham (68.2%) and Ottawa (62.6%) patient groups, ALF and HCC were the second most common indications in the Nottingham and Ottawa patient groups (ALF 24.2% vs 5.9%; HCC 6.1% vs 27.6%; p<0.0001). Alcohol-related liver disease was the most common aetiology of liver disease among the Nottingham patients, while hepatitis C-related liver disease was the most common aetiology among the Ottawa patients (p<0.0001).
The number of patients undergoing LT has steadily increased in both centres during the study period. The majority of Nottingham patients underwent LT at the Cambridge transplant unit (72; 54.5%) followed by Leeds (32; 24.2%), Birmingham (21; 15.9%), King's College London (three; 2.3%), Newcastle (three; 2.3%) and Edinburgh (one; 0.8%). Similarly, the majority of Ottawa patients underwent LT at the Toronto transplant unit (140; 69.0%) followed by Montreal (37; 18.2%), London Ontario (24; 11.8%), Vancouver (one; 0.5%) and Halifax (one; 0.5%).
Patient survival
Overall, 25 (18.9%) and 24 (11.8%) patients died during a median follow-up of 9.1 years (IQR 4.3–15.6) and 6.5 years (IQR 3.5–11.2) in the Nottingham and Ottawa groups, respectively. The 1-, 5-, 10- and 20-year patient survival rates were 98%, 95%, 87% and 62%, and 100%, 96%, 88% and 62% in the Nottingham and Ottawa groups, respectively (log-rank test p=0.87; Fig 1a).
a) Estimates of patient survival following liver transplantation. b) Estimates of death-censored graft survival following liver transplantation.
Sepsis (30.8% and 25.0%) and cancer (24.0% and 12.5%) were the most common causes of death in the Nottingham and Ottawa groups, respectively.
Recurrence of primary disease and graft survival
Overall, 12 (9.1%) and nine (4.4%) patients developed recurrence of non-viral primary disease following a median time of 11.8 years (IQR 4.8–15.0) and 5.1 years (IQR 4.4–14.3) in the Nottingham and Ottawa groups, respectively (log-rank test p=0.68). These include eight primary biliary cholangitis, two primary sclerosing cholangitis and two non-alcohol fatty liver disease in Nottingham, and two primary biliary cholangitis, five primary sclerosing cholangitis, one autoimmune hepatitis and one non-alcoholic fatty liver disease in Ottawa groups. In addition, 7/13 (53.8%) and 46/71 (64.8%) patients who underwent LT for hepatitis C-related liver disease were hepatitis C ribonucleic acid positive at the time of transplantation. All underwent anti-viral treatment following transplantation and five (71.4%) and 33 (71.7%) achieved sustained virologic response (SVR). Further, alcoholic recidivism was noted in 4/25 (16.0%) and 4/31 (12.9%) patients who underwent LT for alcohol-related liver disease.
Overall, 13 (9.8%) and seven (3.4%) patients underwent re-transplantation in the Nottingham and Ottawa groups, respectively. The most common indication for re-transplantation was hepatic artery thrombosis (six; 46.2%) and recurrence of primary disease (four; 57.1%), respectively. The 1-, 5-, 10- and 20-year death-censored graft survival rates were 96%, 95%, 94% and 89%, and 99%, 98%, 98% and 96% in the Nottingham and Ottawa groups, respectively (log-rank test p=0.10; Fig 1b).
End-stage renal disease
A gradual decline in renal function was evident in both the Nottingham and Ottawa groups (Fig 2a). Overall, 10 (7.6%) and 13 (6.4%) patients developed ESRD following a median time of 11.4 years (IQR 2.0–13.2) and 9.8 years (IQR 5.8–14.6) in the Nottingham and Ottawa groups, respectively. The 1-, 5-, 10- and 20-year cumulative incidence of ESRD were 0.0%, 2.3%, 2.3% and 15.4%, and 0.5%, 1.2%, 6.7% and 26.4% in the Nottingham and Ottawa groups, respectively (log-rank test p=0.29; Fig 2b).
a) Estimated glomerular filtration rate levels of patients since liver transplant. b) Estimates of the development of end-stage renal disease of patients since liver transplant.
Post-transplant metabolic syndrome and major cardiovascular events
Overall, 7/105 (6.7%) and 32/161 (19.9%) patients developed post-transplant diabetes mellitus (DM) in the Nottingham and Ottawa groups, respectively (p=0.003). Similarly, 12/106 (11.3%) and 65/159 (40.8%) developed hypertension (p<0.0001), and 3/120 (2.5%) and 38/186 (20.4%) developed dyslipidaemia (p<0.0001) following transplantation in the Nottingham and Ottawa groups, respectively. There were no significant changes in weight in both the Nottingham and Ottawa groups (Fig 3a).
a) Weight of patients since liver transplant. b) Estimates of the development of major cardiovascular events of patients since liver transplant.
Overall, four (3.0%) and 14 (6.9%) patients developed major CVE following a median time of 7.4 years (IQR 3.2–11.7) and 6.3 years (IQR 4.0–11.4) in the Nottingham and Ottawa groups, respectively. The 1-, 5-, 10- and 20-year cumulative incidence of major CVE were 0.0%, 1.0%, 1.0% and 5.4%, and 0.5%, 3.9%, 5.8% and 20.1% in the Nottingham and Ottawa groups, respectively (log-rank test p=0.008; Fig 3b).
On univariate analysis, being male (p=0.007), pre-LT DM (p=0.049), post-LT hypertension (p=0.0007), post-LT dyslipidaemia (p<0.0001) and the use of cyclosporin (p=0.002) were associated with major CVE (Table 1). In addition to the above variables, the use of tacrolimus (p=0.06) and which centre (p=0.13) were included in the multivariate analysis (seven variables); the Bonferroni-corrected level of significance in the multivariate analysis was p<0.007. On multivariate analysis, development of dyslipidaemia following transplantation (odds ratio (OR) 21.71; 95% confidence interval (CI) 5.52–85.36; p<0.0001) was the only variable associated with the development of major CVE post-LT.
Predictors of major cardiovascular events
Cancer
Overall, 22 (16.7%) and 33 (16.3%) patients developed de novo cancer following a median time of 8.2 years (IQR 4.9–13.7) and 5.9 years (IQR 3.2–9.8) in the Nottingham and Ottawa groups, respectively. The 1-, 5-, 10- and 20-year cumulative incidence of de novo cancer were 1.5%, 5.2%, 17.4% and 34.1%, and 1.5%, 7.5%, 19.5% and 49.9% in the Nottingham and Ottawa groups, respectively (log-rank test p=0.22; Fig 4). Skin cancer was the most common cancer in both Nottingham (22.7%) and Ottawa (42.4%) groups. This was followed by colon (13.6%) and haematological (13.6%) cancers in the Nottingham group, and lung (9.1%) and pancreas (9.1%) cancers in the Ottawa group.
Estimates of the development of de novo cancer in liver transplant patients.
Two of the nine (22.2%) and 4/56 (7.1%) patients who underwent LT for HCC developed recurrence of HCC following transplantation in the Nottingham and Ottawa groups, respectively (log-rank test p=0.15). The median duration to HCC recurrence was 6.2 and 6.6 years, respectively.
Discussion
This study represents the largest reported experience on the outcomes of LT recipients managed in non-transplant centres. It includes two centres; one in the UK and the other in Canada. This study reports on broader relevant post-transplant clinical outcomes (ESRD, major CVE and de novo cancer) in addition to conventional patient and graft survival outcomes. Despite the heterogeneities in the patient groups and policies, clinical outcomes were similar in Nottingham and Ottawa.
The 1- and 5-year patient survival, according to the 2017/2018 annual report published by the NHSBT, which includes 8,428 liver transplants recipients, were 90–94% and 80–81%, respectively.7 Similarly, the 1-, 3- and 5-year survival rates, according to the 2012 annual report published by the Canadian Organ Replacement Register, which includes 5,654 liver transplants recipients, were 92%, 82–88% and 78–83%, respectively.8 Interestingly, the survival rates of patients followed up in Nottingham and Ottawa were numerically better than their respective national rates. Although exploring the reasons behind such difference is beyond the scope of this study, it would not be unreasonable to speculate that it may likely be due to the relatively smaller study cohort and the potential for a risk-averse approach to patient selection in non-transplant centres. In addition, ‘sicker’ patients are likely to be kept within the liver transplant centre programme for follow-up; however, this reflects accepted clinical practice and is not necessarily a weakness of our study. Rather, this lends additional evidence that the majority of patients who undergo uneventful liver transplantation can be safely followed up outside liver transplant centres.
The recurrence of autoimmune liver diseases following LT ranges from 10% to 50%.9–11 In line with the previous literature, the recurrence rates of autoimmune liver diseases in our cohorts were 10/40 (25%) and 8/39 (21%) in Nottingham and Ottawa, respectively. However, it is important to note that in conforming to the national practices neither Nottingham nor Ottawa undertake routine screening for recurrence with protocol liver biopsy or surveillance imaging. In our study, the overall post-LT hepatitis C SVR rates were just over 70% in both centres. Although the SVR rates are markedly lower than the direct-acting antiviral (DAA) SVR rates, it is important to highlight that a substantial period of our study predates DAA treatment. However, the overall re-transplantation rates of our study were comparable to that of previous literature from both Europe and North America.12–14
ESRD, major CVE and de novo cancer are being increasingly used in the literature, along with the patient and graft survival rates, in reporting outcomes following organ transplantation both in adult and paediatric populations.15–17 Our findings are on par with a recent study that reported rates of ESRD and major CVE of 9% and 5%, respectively.14 However, the development of de novo malignancy rates was higher in both our centres compared with previously reported rates of 5.6%–8.7%.14,18,19 We speculate that our relatively smaller study cohort and the potential for missing data in large database studies may have contributed to those differences.
Although the indications for LT, namely decompensated cirrhosis, HCC and ALF, are broadly similar, Nottingham and Ottawa follow different listing polices. For example, while modified Milan criteria is used in the UK, total tumour volume / alpha-fetoprotein criteria are used in Ontario for the selection of patients with HCC. Further, the allocation of donor organ policies also differ between Nottingham and Ottawa. In addition, there were intrinsic differences between the Nottingham and Ottawa cohorts (supplementary material S2). The Nottingham cohort was relatively younger with significantly higher proportion of patients with ALF, likely due to paracetamol-related ALF being more common in the UK.20–23 Further, due to the geography, the distance and travel time to transplant centres are substantially higher in Canada than in the UK. However, those differences, did not impact the overall long-term outcomes of LT recipients in Nottingham and Ottawa, which is encouraging.
Our data combined with excellent patient and graft survival outcome results from other non-transplant centres within UK, promises a way forward for delivering pre- and post-transplant patient care close to home without compromising patient wellbeing.24,25 Video-conferencing technology has already been successfully used to increase access to care in underserved areas in Canada, most notably in the treatment of hepatitis C infection and is increasingly used in the management of transplant recipients in non-transplant centres. Further, outreach clinics are becoming part of day-to-day clinical service provision. Trainees are also being exposed to transplant hepatology as part of their regular training and majority of them continue their career in non-transplant centres, a resource that could be optimally exploited in providing patient care close to home.26,27 These would likely mitigate the barriers to transplant service, while helping to reduce the clinical burden on transplant centres.28 It would also improve access to healthcare and equalise service provision nationally. All of this would help to remove some of the patient care burden from transplant centres and allow transplant services to grow beyond their geographical limits.29 Further, this may improve organ donation rates from regions away from transplant centres as people see more visible ‘benefit’ to the local population, and thereby overcome the societal behaviours and beliefs that currently create barriers to donation.30 Cost analysis, based on NHS reimbursement tariff per transplant, shows financial sustainability of providing post-transplant service in non-transplant centres such as Nottingham (personal communication to Martin W James).
Our study has certain strengths and limitations. It is the largest study of its kind and it reports on both conventional and broader relevant clinical outcomes of LT recipients. Our study compares two centres in two different jurisdictions with different transplant polices, protocols and patient populations. This study reports on outcomes over a long period, during which surgical techniques, pre- and post-transplant care, and use of immunosuppression have undergone changes, as in any other field of medicine, and is unavoidable. Although the differences in patient populations and significant changes in patient care may have impacted the study results, demonstration of comparable outcomes despite these unavoidable differences is a major strength of this study. On the other hand, the retrospective nature of our study is a natural limitation, due to the risk of bias inherent to study design and the unavailability of data for some variables, such as donor characteristics and intra-operative factors that may have impacted the outcomes. Further, the differences in post-transplant outcomes between centres could not be investigated adequately due to lack of data on all potential risk factors that may have contributed to these differences (eg maintenance immunosuppression dose, protocols for the use of statins and anti-hypertensives etc). Implementation of similar transplant service provision models will depend on individual healthcare systems and the availability of clinicians trained in transplant medicine, and it would be wrong to assume that similar service provision could be adopted in other countries.
Conclusion
This is the largest non-registry study to investigate outcomes of LT recipients managed at non-transplant centres. It clearly demonstrates that satisfactory long-term outcomes comparable to that of national and international outcomes are achievable in non-transplant centres, despite the differences in transplant policies, protocols and healthcare systems. This study provides promising preliminary data to build towards future changes to public healthcare policy; namely, additional resources to increase local care for post-transplant patients, while increasing access to subspecialists via virtual care. This is particularly important in the era of COVID-19, when travel and in-person healthcare visits are discouraged. Future studies should aim to provide prospective comparisons of different models of post-LT care and include additional ‘non-traditional’ outcomes such as cost-effectiveness and patient satisfaction, which would greatly assist in effecting these policy changes.
Supplementary material
Additional supplementary material may be found in the online version of this article at www.rcpjournals.org/clinmedicine:
S1 – Summary of selection criteria for first adult liver transplantation in the UK and Ontario, Canada.
S2 – Demographics and clinical characteristics of Nottingham and Ottawa patients.
- © Royal College of Physicians 2021. All rights reserved.
References
Article Tools
Jump to section
Related Articles
Cited By...
- No citing articles found.