When is the optimal time to discharge patients after liver transplantation with respect to short-term outcomes? A systematic review of the literature and expert panel recommendations
Protocol registration: PROSERO protocol ID; CRD42021245598.
ERAS4OLT.org Working Group Authorship: Alex Rothnie, London, UK, Joao Mestre-Costa, London, UK, Claus Niemann, San Francisco, CA, USA, Joerg-Matthias Pollok, London, UK, Marina Berenguer, Valencia, Spain, Pascale Tinguely, London, UK.
Authorship: All authors qualify for authorship as per the International Committee of Medical Journal Editors (ICMJE) guidelines.
Data sources: Ovid MEDLINE, Embase, Scopus, Google Scholar, and Cochrane Central.
This work was conducted in preparation for the ILTS - ERAS4OLT.org Consensus Conference on Enhanced Recovery for Liver Transplantation, January 2022, Valencia, Spain.
Abstract
Background
Several factors associated with prolonged hospital stay have been described. A recent study demonstrated that hospital length of stay (LOS) is directly associated with an increased cost for liver transplantation (LT) and may be associated with greater mortality; however, the factors associated with post-LT mortality are also related to a prolonged hospital stay, that is, those factors are confounders. Thus, the actual impact of the length of post-LT hospital stay on both short-term and long-term patient and graft survival remains uncertain.
Objectives
To identify the optimal time to discharge patients after LT with respect to short-term outcomes; readmission rate, 30–90-mortality and morbidity.
Methods
Systematic review following PRISMA guidelines and recommendations using the GRADE approach derived from an international expert panel. Initial search keywords for screening were as follows; ((discharge AND (time OR “time point” OR “time-point”)) OR “length of hospital stay” OR “length of stay”) AND ((liver OR hepatic) AND (transplant OR transplantation)). PROSPERO ID: CRD42021245598
Results
The strength of recommendation was rated as Weak, and we did not identify the direction of recommendations regarding the optimal timing after LT concerning short-term outcomes, including “Readmission rate,” six studies on 30- and/or 90-day mortality, and five studies on “30- and/or 90-day morbidity rate.”
Conclusions
Evidence is scarce to judge the optimal timing to discharge patients after LT with respect to short-term outcomes. In centers with robust outpatient follow-up, discharge can occur safely as early as post-transplant 6–8 days (Quality of Evidence [QOE]; Low | Grade of Recommendation; Weak).
1 INTRODUCTION
Survival following liver transplantation (LT) has improved steadily since the implementation of liver transplants over 50 years ago. Accordingly, the paradigm has shifted to improve other quality metrics, including cost reduction along with other health care issues.1 One of the metrics is the postoperative length of stay (LOS), used as a critical benchmark for the progress and efficiency of patient care.2
Discharge from the hospital after transplantation remains a unique challenge. In many countries, liver transplants performed for noncancer indications are often performed on hospitalized, decompensated patients with a high Model for End-Stage Liver Disease (MELD) score. In addition, the patients have multiple comorbidities and are in poor health, especially in the era when we perform LT in acute-on-chronic liver failure patients.3 Often, they have an acute or subacute illness just prior to LT.
This complexity leads to the difficulty in evaluating the optimal timing to discharge patients immediately after LT. The factors associated with impaired post-LT outcomes are also related to a prolonged hospital stay1; that is, those factors are confounders. Thus, we conducted this systematic review and evidence grading to identify the optimal time to discharge patients after LT with regards to short-term outcomes.
This work was conducted in preparation for the ILTS - ERAS4OLT.org Consensus Conference on Enhanced Recovery for LT, January 2022, Valencia, Spain.
2 METHODS
2.1 Protocol and registration
The PROSPERO ID of this project is CRD42021245598.
2.2 Eligibility criteria
Included studies reported on the timing of discharge and short-term outcomes after LT in adults with end-stage liver disease listed for and undergoing cadaveric orthotropic LT. Split livers and patients undergoing living donor LT were excluded.
2.3 Information sources
All records were screened by assessing the title and abstract by two independent investigators (AR, JMC). Controversies were resolved by consensus of the study group. All manuscripts were assessed for eligibility and data, where applicable, was extracted in an electronic form. This form enhanced the correctness of the data extraction and helped minimize missing data. The main fields extracted included bibliographic information, length of hospital stay after LT, readmission rates, morbidity and mortality at 30 and/or 90 days of follow-up.
2.4 Search
We used the bibliographic databases: Ovid MEDLINE, Embase, Scopus, Google Scholar, Clinical.Trials.gov and the Cochrane Central Register of Controlled Trials to apply the search strategy and extract all available records for our systematic review. The search date was set to April 1, 2021. Only studies reporting on duration of hospital stay or timing of discharge in association with patient outcomes, such as hospital readmission rates, morbidity, mortality, ITU and hospital stay as well as other relevant recipient outcomes were included. There was no limitation regarding year of publication. Studies reporting on pediatric populations as well as case reports or conference abstracts were excluded. Search keywords were as follows; ((discharge AND (time OR “time point” OR “time-point”)) OR “length of hospital stay” OR “length of stay”) AND ((liver OR hepatic) AND (transplant OR transplantation)).
2.5 Study selection
Both comparative and single cohort studies, retrospective or prospective, describing patient length of hospital stay after LT in relation to short-term outcomes, including readmission rates and morbidity, were included.
2.6 Quality of studies and recommendations grading
The “Grading of Recommendations Assessment, Development and Evaluation” (GRADE) approach was used for grading quality of evidence (QOE) and strength of recommendations.4 The GRADE system was designed to provide a comprehensive and structured approach to rating the QOE for systematic reviews, and to grade the strength of recommendations for development of guidelines in health care. We applied the modified GRADE approach for QOE assessment derived from systematic reviews using estimates summarized narratively.5 The QOE was rated separately for each outcome. The direction and strength of recommendation was assessed individually by all authors and disagreements resolved by consensus.6, 7
3 RESULTS
3.1 Study selection
We screened 3028 studies, out of which 27 studies were assessed for eligibility. After excluding 13 studies (one due to written in a language other than English and 12 due to the measures of the outcomes not relevant to the research question), we finally included 14 studies in the review. The flow diagram is shown as Figure 1.
3.2 Study characteristics
The characteristics including study type, size, and outcomes are presented in Table 1.
| Study | Study type | No. of patients | Main short-term outcomes assessed |
|---|---|---|---|
| Chung-Ping, 2019 | Retrospective cohort (Propensity Score Matched) | 421 | LOS with Fast Track Protocol |
| Estimated cost savings from use of Fast Track Protocol | |||
| Brustia, 2020 | Systematic review Delphi Study single-center pilot study | 45 | Create a checklist for safe discharge after LT |
| % completion of checklist at the time of discharge | |||
| Smith, 2008 | Retrospective | 521 | Rate of Prolonged LOS (>30 days) |
| Rudich et al., 1999, CA US | Single-center, retrospective | 40 | LOS |
| Length of ICU stay | |||
| Rao JH et al., 2019, China | Single-blind RCT | 128 | Length of ICU stay |
| The median ICU stay was 2 days (range 1–7 days) in the | |||
| Postoperative complications | |||
| Amiri M et al., 2020, Iran | Single-center, retrospective | 161 | LOS |
| Washburn WK et al., 2009, CA US | Two-center, retrospective | 1455 | LOS |
| Toledo, 2013 | Unicenter interventional prospective trail | 98 | LOS |
| Readmission rate | |||
| Krishnan, 2013 | Retrospective analysis of multi-institutional UNOS STAR data | 38,896 | LOS |
| Tovikkai, 2016 | Retrospective analysis of United Kingdom Transplant Audit database | 3,772 | LOS |
| Loh et al., 2019 | Retrospective cohort, quasi-experimental with control groups formed using propensity-score matching | 17,660 | LOS |
| Estimated cost savings for reduced LOS | |||
| Rodriguez-Laiz et al., 2021 | Prospective cohort study. Follow-up period not stated but appears to be 7 years for earliest enrolled patients | 236 | LOS |
| Readmission rate | |||
| Rana et al., 2017 | Retrospective cohort study utilizing UNOS data for all liver transplant recipients >18 years | 73,635 | LOS |
| Prolonged LOS rates (>30 days) | |||
| Pinero et al., 2015 | Retrospective cohort study of consecutive adult liver transplant recipients (12 years) | 289 | Early discharge from hospital (before 8th day post LT) |
| Acute cellular rejection (ACR) rates | |||
| Rate of short-term rehospitalization (during first 30 days after hospital discharge) | |||
| Overall liver transplantation costs |
3.3 Results of individual studies
The result of each study is summarized in Table 2. Six studies included data on “Readmission rate,” six studies on “30- and/or 90 day mortality,” and five studies on “30- and/or 90-day morbidity rate.”
| Study | Readmission rate | 30- and 90-day mortality | 30- and 90-day morbidity rate |
|---|---|---|---|
| Brustia, 2020 | 6.7% | NA | NA |
| Smith, 2008 | NA | 30 day mortality was exclusion criteria for study, 90 day not reported | Morbidity calculation for study (PLOS vs. nonPLOS). Biliary 17.6 versus 5.3% |
| Rudich et al., 1999, CA US | Late discharged recipients (vs. early) had higher rehospitalization rate (20 of 25 versus 6 of 15, respectively, p = .0102). | NA | NA |
| Rao JH et al., 2019, China | Re-admission rate was 0 for both groups. | Mortality rate was 0 for both groups. | 90-day morbidity rate was 19% for fast-track group and 27% for non-fast track group, respectively, without significant difference. |
| Amiri M et al., 2020, Iran | NA | NA | 30/161 (18.6%) among all patients. |
| Washburn WK et al., 2009, CA US | NA | NA | NA |
| Toledo, 2013 | 30-day readmission rate before intervention 53%, after intervention 48%; 90-day readmission rate before intervention 65%, and after intervention 56% | No impact on mortality rate at 30 days and 1 year (data not presented) | NA |
| Krishnan, 2013 | NA | Recipients predicted as high risk for prolonged LOS by all stages of the HALOS-ND also have a higher risk of mortality | NA |
| Loh et al., 2019 | NA | NA | Fast-tracking associated with 2.5-3.2 fewer days of post-LT LOS; Reduction in post-LT LOS yielded an estimated cost-savings of $2,062,930 - $2,639,270 |
| Rodriguez-Laiz et al., 2021 | 30d readmission rate: Fast-track – 29%; Regular discharge – 45% (p = .015) | NA | NA |
| Pinero et al., 2015 | Early re-hospitalization: | NA | Acute cellular rejection (ACR) rate within 3 months after LT |
| Early discharge – 21% (7/34) | Early discharge – 18% | ||
| Late discharge – 24% (53/223); (p = .43) | Non-early discharge – 42%; (p = .006) | ||
| ACR rate after hospital discharge | |||
| Early discharge – 15% | |||
| Non-early discharge – 11%; (p = .32) | |||
| Median overall LT costs | |||
| Early discharge - $23,078 | |||
| Non-early discharge - $28,986 (p < .0001) |
3.4 Quality of evidence
The summary of findings for the main outcomes, including the QOE assessment according to the GRADE approach are summarized in Table 3.
| Summary of findings | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Number of studies | |||||||||
| RCT | Observational comparative | Observational non-comparative | Effect from comparative studies | Limitations | Inconsistency | Indirectness | Imprecision | Publication Bias | Quality of evidence (GRADE) |
| Outcome 1: Re-admission rate | |||||||||
| 1 | 4 | 1 | Earlier discharge was related to lower readmission rate in intervention group in some studies | Very serious | Serious | Not serious | Serious | Not likely | Low ●●○○ |
| Outcome 2: 30- and 90-day mortality rate | |||||||||
| 1 | 4 | 0 | Most of the studies which addressed this issue (5/14) did not suggest significant difference | Very serious | Serious | Not serious | Serious | Not likely | Low ●●○○ |
| Outcome 3: 30- and 90-day morbidity rate | |||||||||
| 1 | 4 | 0 | Most of the studies which addressed this issue (5/14) suggested lower morbidity rate, in only some of the parameters | Very serious | Serious | Not serious | Serious | Not likely | Low ●●○○ |
3.5 Recommendations
The recommendation regarding “When is the optimal time to discharge patients after liver transplantation?” is summarized in Table 4 according to the GRADE approach: Quality of the evidence, balance of desirable/undesirable outcomes, and values and preferences. The panel of experts considers a benefit for early discharge of straightforward low risk cases but according to lacking evidence a certain day was not identified. The strength of recommendation was rated as weak (Table 4).
| Question: When is the optimal time to discharge patients after liver transplantation | |||
|---|---|---|---|
| Decision domain | Judgement | Reason for Judgement | |
| Yes | No | ||
| Balance between desirable and undesirable outcomes (estimated effects), with consideration of values and preferences (estimated typical). Given the best estimate of typical values and preferences, are you confident that the benefits outweigh the harms and burden or vice versa? | ✓ | Evidence is scarce to judge the balance of desirable and undesirable outcomes but suggests patients with uncomplicated postoperative courses can be discharged in a timely manner (e.g., 8 days per Pinero et al, in centers with robust outpatient follow-up). | |
| Confidence in the magnitude of estimates of effect of the interventions on important outcomes (overall quality of evidence for outcomes). Is there high, moderate or low-quality evidence? | ✓ | Evidence is scarce for defining what is the optimal time for patient discharge. Reduce re-admission rate: Low ●●○○. Reduce 30- and 90-day mortality rate: Low ●●○○. Reduce 30- and 90-day morbidity rate: Low ●●○○ | |
| Confidence in Values and Preferences, and their Variability. Are you confident about the typical values and preferences and are they similar across the target population? | ✓ | We cannot be confident: Scenario for patient discharge differs amongst programs from different regions therefore criteria for a “safe discharge” will also vary between different target populations. These estimations of the impact of the timing of discharge on clinical outcomes are scarce, mainly driven by clinical experience (that is, retrospective studies), derived from varying clinical, ethnical, age and gender backgrounds. | |
| Resource implications. Are the resources worth the expected net benefit from following the recommendation? | ✓ | Early safe discharge is a cost savings for health care systems per only one paper (Chung-Ping et al, 2019), but the implementation of such interventions potentially vary across programs from different regions and the evidence regarding its impact is yet scarce, so we cannot draw robust conclusion | |
| Overall Quality of Evidence: Very Low | |||
| Recommendation: Patients with uncomplicated postoperative course can be discharged safely in a timely manner (e.g., in centers with robust outpatient follow-up, discharge can occur safely as early as POD 6–8 days) Weak: | |||
4 DISCUSSION
In relation to donor and recipient risk factors, intra- and postoperative complications the length of Intensive Care Unit (ICU) and hospital stay may differ. For the decision of the optimal time to discharge patients after LT, a definition of functional recovery is required.
The complexity of the procedure “liver transplantation” including recipients suffering from significant comorbidities, varying organ quality and post-transplant complications, maybe the reason that so far the question for the optimal time to discharge patients after LT could not be directly answered. Compared to LT, other major surgery both abdominal and thoracic surgery, might be less multifactorial and more easily predictable and therefore, LT is the last procedure for the establishment of an ERAS-protocol.
Estimations of the impact of the timing of discharge on clinical outcomes are lacking evidence and are mainly driven by clinical experience. Only one randomized controlled trial (RCT),8 11 observational comparative studies and one observational noncomparative study were identified for analysis of this special question. Hence, by reason of lacking literature an expert-based recommendation is provided.
Risk factors for longer hospital stays were dependent on recipients' comorbidities and initial graft function.9 In predictive models, post-transplant length of hospital stay was estimated considering different recipient factors, donor factors and post-transplant variables.10-15 The differences in predictors detected were likely due to center-specific practices such as candidate and donor organ selection and matching, and therefore predictors at one center may not have the same impact at another.12
The safety and efficacy of an ERAS-protocol might be measured by parameters such as early readmission rate, morbidity, mortality and cost-effectiveness. In the reviewed literature, early readmission rate was comparable between the population with early versus regular/late discharge from hospital with a trend toward lower readmission in the ERAS group.8, 14, 16-19 In daily clinical practice, a definition of functional recovery after LT would be a requirement for a standardized timely discharge protocol. As a first step, a checklist was established on confounders indicating the recipients´ postoperative course for a safe discharge policy.19 The identification and treatment of susceptible risk factors that result in a longer hospital stay could further improve outcome. Functional recovery after LT could be defined by standardized and consensual checklist criteria to assess readiness for hospital discharge. Unfortunately, a one-size-fits-all checklist may not be appropriate for each hospital depending on the availability of affiliated outpatient facilities and such checklists may require to be modified according to the local needs and infrastructure.
Besides comparable readmission rates in groups following an ERAS procedure also 30- and 90 day morbidity and mortality were not increased.8, 16, 17 Therefore, ERAS in LT can be judged as a safe procedure.
Standardization and optimizing the timing for discharge from hospital might also reduce costs directly attributable to LT during hospitalization. A cost-saving effect was shown in one single center study17 but was less than expected in an analysis from the SRTR.20 The reason might be the low granularity of data in the registry, as one could assume that shortened LOS would translate into significant cost savings.
4.1 Limitations
Evidence is scarce for defining what is the optimal time for patient discharge following LT. Moreover, the scenario for patient discharge differs among programs from different regions and is dependent on the environment outside the hospital like availability of facilities for rehabilitation closely connected to the transplant center. Therefore, criteria for a safe and early discharge will also vary between different target populations.
5 CONCLUSION
Evidence is scarce to judge the balance of desirable and undesirable outcomes but studies suggest that patients with uncomplicated postoperative courses can be discharged safely in a timely manner (QOE: Low/Grade of Recommendation: Weak). Despite the heterogeneity of study populations overall a beneficial effect of early/timely discharge from hospital was demonstrated. In centers with robust outpatient follow-up, discharge can occur safely as early as post-transplant 6–8 days. There is obvious need for future studies defining functional recovery after LT taking into account the availability of different outpatient facilities.
ACKNOWLEDGMENT
This manuscript was prepared for the ERAS4OLT.org Consensus Conference 2022, which is partially funded by the International Liver Transplant Society (ILTS).
CONFLICTS OF INTEREST
The authors declare no conflicts of interest.
AUTHOR CONTRIBUTIONS
Concept/design: TT, GB, DAR, MS, Data analysis/interpretation TT, GB, TR, AO, AA, AM, DAR, MS, GB, Drafting article TT, GB, MS, GB, Critical revision of article TT, GB, TR, AO, AA, AM, DAR, MS, GB, Approval of article TT, GB, TR, AO, AA, AM, DAR, MS, GB
Open Research
DATA AVAILABILITY STATEMENTS
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
