Post-operative, inpatient rehabilitation after lung transplant evaluation (PIRATE): A feasibility randomized controlled trial

ABSTRACT Background Postoperative rehabilitation is crucial following lung transplantation (LTx); however, it is unclear whether intensive rehabilitation is feasible to deliver in the acute setting. We aimed to establish the feasibility and safety of intensive acute physiotherapy post-LTx. Methods This feasibility trial randomized 40 adults following bilateral sequential LTx to either standard (once-daily) or intensive (twice-daily) physiotherapy. Primary outcomes were feasibility (recruitment and delivery of intensive intervention) and safety. Secondary outcomes included six-minute walk test; 60-second sit-to-stand; grip strength; physical activity; pain; EQ-5D-5L; length of stay; and readmissions. Data were collected at baseline, week 3, and week 10 post-LTx. ClinicalTrials.gov #NCT03095859. Results Of 83 LTx completed during the trial, 49% were eligible and 48% provided consent. Median age was 61 years {range 18–70}; waitlist time 85 days [IQR 35–187]. Median time to first mobilization was 2 days [2–3]. Both groups received a median of 10 [7–14] standard interventions post-randomization. A median of 9 [6–18] individual intensive interventions were attempted (86% successful), the most common barrier being medical procedures/investigations (67%). No intervention-related adverse events or between-group differences in secondary outcomes were observed. Conclusions Acute, intensive physiotherapy was feasible and safe post-LTx. This trial provides data to underpin definitive trials to establish efficacy.


Introduction
Patients post-lung transplantation (LTx) experience low levels of exercise capacity and participation in the immediate postoperative period (Wickerson, Mathur, Singer, and Brooks, 2015). Likely contributing factors include a reduction in peripheral muscle function, including size, force, and composition of limb musculature (Guerrero et al., 2005). Outpatient physical rehabilitation post-LTx carries benefit in terms of physical activity, functional recovery, and quality of life (QOL) (Fuller et al., 2017;Langer et al., 2012), and is perceived as a valued tool assisting in a return to normal life (Fuller et al., 2013). Individualized sub-acute physical rehabilitation also improves exercise tolerance in long term (>1 year) survivors post-LTx (Ihle et al., 2011). Common components of rehabilitation include endurance, strength, and flexibility focused exercise, modified according to body function, structure, activity, and participation deficits (Munro et al., 2009;Patcai, Disotto-Monastero, Gomez, and Adcock, 2013).
There are few data on the role and outcomes of rehabilitation in the acute postoperative phase following LTx (Wickerson, Mathur, and Brooks, 2010). Early rehabilitation in the acute setting has been studied in other patient groups. For patients in the intensive care unit (ICU), early mobility and rehabilitation interventions are safe, feasible, and may improve muscle strength, mobility, and hospital length of stay (LOS) (Adler and Malone, 2012). A trial of intensive inpatient physical rehabilitation in patients following traumatic injury had similar findings (Calthorpe et al., 2014). Despite this, caution must be heeded from trials in other groups. In chronic obstructive pulmonary disease (COPD), early supervised progressive physical rehabilitation during acute exacerbations resulted in higher mortality at 1 year and no change to readmissions (Greening et al., 2014). In stroke, intensive physical rehabilitation during acute recovery was not preferential to low-dose usual care, and reduced the odds of a favorable disability outcome as early as 3 months (Bernhardt, 2015). The reasons for these adverse findings are not clear, reinforcing the need to conduct trials in specific patient groups before implementing such programs in practice. The increasing age, complexity, and comorbidities of patients post-LTx further highlights the importance of studies specific to LTx (Fuller et al., 2017;Paraskeva, Levin, Westall, and Snell, 2018).
Preliminary data support the feasibility of physical rehabilitation commencing in ICU following LTx (Song et al., 2018); however, there have been no prospective, randomized trials of acute, intensive physiotherapy-led rehabilitation following LTx. We aimed to outline the feasibility and safety of standard versus intensive acute physiotherapy following bilateral sequential lung transplantation (BSLTx); and obtain preliminary efficacy data to inform future large clinical trials.

Methods
This was a randomized, assessor-blinded, parallel feasibility trial. Eligible participants were adults (≥18 years) recruited from a specialist LTx center following BSLTx for any indication. Patients were recruited once extubated, deemed medically stable by the treating physician and able to participate in physical rehabilitation as per a senior LTx physiotherapist (PT). Recruitment took place within 3 days from commencement of early mobility, which has been previously defined . Exclusion criteria consisted of: medical instability preventing mobilization (e.g. cardiovascular instability); high levels of invasive support (e.g. extra-corporeal membrane oxygenation (ECMO) and hemofiltration); low throughput LTx categories preventing representative sample recruitment such as heart-lung, and single LTx (SLTx) (Australia and New Zealand Cardiothoracic Transplant Registry, 2018;Chambers et al., 2019); and patients unable to provide informed consent (e.g. acute delirium). Participants provided written informed consent via a non-treating PT. Approval was granted by Alfred Health and La Trobe University Human Research Ethics Committees, conduct was in line with the Declarations of Helsinki and Istanbul, and this trial was registered a priori via ClinicalTrials.gov, ID: NCT03095859.
Participants were randomized to either standard care (control), or intensive PT (experimental) groups. A computer-generated block randomization sequence was used (randomization.com), with allocation concealed using opaque sealed envelopes. Randomization was stratified based on local and national recipient age (<55 years; or ≥55 years) (Australia and New Zealand Cardiothoracic Transplant Registry, 2018) as age is known to influence early morbidity and mortality post-LTx (Weill et al., 2015), and we hypothesized a priori that this may impact outcomes.
Standard care consisted of once-daily physiotherapy for approximately 30 minutes conducted by a trained LTx PT, including physical and nonphysical interventions. Physical rehabilitation included early mobility, endurance exercise, upper and lower limb strength, flexibility, and trunk mobility . Nonphysical interventions included respiratory therapies, patient/carer education, and psychosocial support. Standard care has been reported to consist predominantly of assessment, documentation, early mobility, and endurance exercise, with little time dedicated to education and focal physical exercises .
Acute physiotherapy principles included early mobilization in ICU, early oxygen weaning, optimization of ventilation and airway clearance, gait aid use to promote independent function, and the facilitation of progressive self-management (Spruit et al., 2013;Tarrant et al., 2018) (Supplementary File Method 1). Physical exercise progression was guided by Borg scale ratings of shortness of breath (0-10) and rating of perceived exertion (6-20), targeting scores of moderate (3) and somewhat hard (13), respectively (Borg, 1982).
The intensive group received standard care plus one additional 30-minute PT session per day, allowing for further progression of individualized exercise and/or completion of a more comprehensive program considering the complications often encountered during early post-LTx recovery, in line with the protocol in Supplementary File Method 1. The principles of intensity, duration, and progression were the same in both groups. This was delivered by a trained PT assistant (EQ) under supervision of the LTx research PT (BJT), with direct PT assistance as appropriate for manual handling and patient safety. Neither research PT nor assistant were involved in standard care of trial participants, while standard care PTs were not involved in delivery of experimental group interventions. From acute inpatient discharge to week 10 follow-up, both groups received routine, outpatient group-based physical rehabilitation for 1-hour, three times per week according to our previously published protocol (Supplementary File Method 2) (Fuller et al., 2017).
Additional intensive interventions were not provided on bronchoscopy or discharge days, the latter coinciding with cessation of the intervention period. In addition to routine analgesia (Supplementary File Method 3), breakthrough analgesia was delivered as required prior to intervention (typically 30 mins), guided by both numerical rating scale (0-10) and Functional Activity Score (Tong, Konstantatos, Cheng, and Chai, 2018). Intervention was provided within both ICU and ward settings, with additional intensive interventions provided either in the am or pm, no closer than 4 hours from standard care to maintain geographical and temporal separation of standard care PTs and research PT/ assistant. Medicolegal documentation was stored in a secure location so that the multidisciplinary team, including standard care PTs, remained blinded to both group allocation and delivery of experimental interventions. It was not possible to blind participants. Assessors were blinded to group allocation. This methodology has been successfully used in previous trials in our institution (Calthorpe et al., 2014). Outcomes were collected prior to randomization (baseline), then repeated at week 3 post-LTx, the usual inpatient LOS at our center, and week 10 post-LTx when patients are receiving regular sub-acute physical rehabilitation.

Feasibility
Defined by: number of patients eligible for inclusion; number of patients who consented; additional intensive physiotherapy sessions completed in the experimental group; and reasons for non-completion. The feasibility of collecting secondary outcomes at the week 3 and week 10 follow-up time points was also quantified.

Safety
Adverse events (AEs), including but not limited to: events related to acute physiotherapy such as wound breakdown, uncontrolled pain, musculoskeletal injury, and patient falls; and evidence of acute rejection on transbronchial biopsy (TBBx) (Crespo et al., 2018).

Statistical analysis
Analyses were performed via IBM SPSS Statistics Version 25 (©IBM, Chicago, US). Feasibility data were analyzed descriptively including frequency and percentage as appropriate. Continuous data for secondary outcomes were examined by independent samples t-test/ Mann-Whitney U test; or two-way repeated measures analysis of variance (ANOVA) with Greenhouse-Geisser correction as appropriate; and dichotomous data by generalized estimating equation combined odds ratios. For EQ-5D-5L, each dimension was scored 1-5 (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression), with higher scores indicating greater problems. Data were also dichotomized as no problems (score 1) vs problems (scores 2-5). The EQ-5D-5L health VAS was reported as a continuous scale (0-100), with higher scores indicating better health; and an overall continuous scale crosswalk index value ranging from 0 to 1.0, with a higher value indicating better health (EuroQol Research Foundation,).
Valid physical activity data included a minimum 10 hours of daytime data between 0600 and 2200 (Byrom and Rowe, 2016;Matthews, Hagstromer, Pober, and Bowles, 2012); over a minimum of 3 days (Langer et al., 2012;Rabinovich et al., 2013;Wickerson, Mathur, Singer, and Brooks, 2015); with at least one weekend day reaching inclusion threshold (Fastenau et al., 2013). Failure to reach threshold wear or device failure led to data exclusion (Walsh et al., 2016). First and last days of wear were excluded (Byrom and Rowe, 2016;Rabinovich et al., 2013;Sugino et al., 2012). In order to capture variables relevant to both the early inpatient rehabilitative phase (e.g. position transitionsstanding up), and the outpatient phase (e.g. higher-level energy expenditure and physical activity) we reported the following variables: inactive time (lying/sitting); active time (standing, stepping/walking); step count; position transitions from lying/sitting to standing, stepping or walking; and energy expenditure by metabolic equivalents (METs), by combining moderate (3 to <6) and vigorous (≥6) activity (©McRoberts BV). These variables were consistent with existing LTx literature (Langer et al., 2012(Langer et al., , 2009a.

Sample size
The primary aim was to ascertain the feasibility of implementing an intensive, acute physiotherapy program post-LTx. This trial was not powered to identify change in secondary outcomes. We recruited 40 participants, deemed a priori by the research team suitable to capture a representative range of recipient age, comorbidity burden, LTx indication, and waitlist duration to meet our aims.

RESULTS
Participants were recruited between February 26, 2019 and March 10, 2020. Participants had a median (IQR) age of 61 (49-67) years and a median time of 85 (41-201) days on the LTx waiting list. Demographic data are in Table 1. Transplant indication was predominantly interstitial lung disease and COPD at 40% and 32.5%, respectively. All but one LTx (control -transverse sternotomy) was performed via bilateral anterior thoracotomy.

Safety
There were no reported intervention-related AEs for either group. There were three AEs related to DynaPort MoveMonitor wear, resulting in pressure areas from elastic strap friction. This occurred in one participant during the day 10 assessment (participant age ≥55 years), and in two participants during the week 10 assessment (participants aged <55 years and ≥55 years).

Secondary outcomes
There were no significant between-group differences at either week 3 or week 10 in 6MWD (p = .64), grip strength (p = .44), STS (p = .28), MILOA (p = .24), or pain (p = .62) ( Table 3) Figure 3, p = .28). Perceived health by EQ-5D-5L VAS was no different between groups (p = .71), nor was health state index (p = .19). Health dimension analysis revealed no differences in any domain (Table 3; Supplementary File Table 1). There were no significant differences in DynaPort MoveMonitor outcomes at either follow-up (Table 4; Supplementary File Figures 4-8). There was no difference in initial ICU stay (mean 4 days) or total inpatient LOS (mean 21 (standard) vs 24 days). Two participants required ICU readmission with acute delirium and/or hemodynamic instability following recruitment, for 4 and 11 days in standard and intensive groups, respectively. Week 2 TBBx results were available for 17 standard, and 18 intensive group participants, with no evidence of acute cellular rejection   Table 2). Over 10 weeks, two and four participants required two readmissions in the standard and intensive groups, respectively, with one standard group participant requiring three readmissions. The most common cause was respiratory complication, including pneumothorax, anastomotic deficit, pleural effusion, and upper/lower respiratory tract infection, accounting for four (standard) and five (intensive) readmissions. Other diagnoses included cardiovascular, hematological, immunological, musculoskeletal, gastrointestinal and renal-related complications. Total, allcause acute readmission LOS was lower in the intensive group, nearing but not reaching significance (mean 8 v 4 days, p = .08) (Supplementary File Table 2).

Discussion
We analyzed the safety and feasibility of intensive, twicedaily PT compared to standard care in 40 adults following BSLTx in the acute postoperative period, observing no discernable intervention-related AEs. Recruitment was feasible, with 49% of patients post-LTx eligible for inclusion, and 98% of those eligible providing consent. Additional intensive interventions were successfully delivered on 86% of occasions. Collection of secondary outcomes at week 3 presented challenges, primarily data collection occurring earlier than protocol due to assessor availability, space, and resource requirements. Physical activity data were negatively impacted by device-related failure, application safety, and adherence.
Feasibility trials are important precursors to larger efficacy trials to ascertain intervention uptake and appropriateness for definitive assessment. However, there is no consensus on ideal methodology (Arain, Campbell, Cooper, and Lancaster, 2010). A set of general areas of focus have previously been proposed to assess in any feasibility trial (Bowen et al., 2009). Consistent with these areas, we had already established local acceptability and perceived demand, and created a multidisciplinary reviewed protocol prior to trial commencement . This allowed for the evaluation of the implementation and practicality of acute, intensive PT post-LTx. Successful delivery of intensive PT in this trial was higher than that reported in acute care patients recovering from critical illness (65%) (Berney, Haines, Skinner, and Denehy, 2012), and previously defined feasibility in pulmonary rehabilitation (75%) (Chaplin et al., 2017). This trial also builds on retrospective LTx literature suggesting feasibility of early physical rehabilitation beginning in the ICU (Song et al., 2018).
The absence of intervention-related AEs in a complex cohort recovering from major thoracic surgery is reassuring, consistent with literature regarding the safety of early mobilization and rehabilitation in both ICU (Tipping et al., 2017) and LTx observational studies (Song et al., 2018;Tarrant et al., 2018). Previously published literature has proposed the scope of acute LTx PT practice (Wickerson et al., 2016), including early mobility, progressive functional rehabilitation, exercise training, and oxygen titration. However, the absence of international guidelines outlining evidence-based practice during postoperative care implies that clinicians should continue to adopt a safe, individualized, teamorientated approach, following the general principles of safe mobility and physical activity translated from general ICU and acute care populations (Berney, Haines, Skinner, and Denehy, 2012;Hodgson et al., 2014). Secondary outcomes were feasible to collect, and although there were no between-group differences, the data reported here will provide valuable information for powering future efficacy trials. Encouragingly, overall STS repetitions at week 10 (n = 19) showed continual improvement from week 3 (12), exceeding the proposed minimum detectable change (5.22) (Tarrant et al., 2020). Daily step count at week 10 also exceeded reported values at 3 months post-LTx (range 3451-5194), as did METs >3 (range 38-69) (Langer et al., 2012), however, unsupervised physical activity was not objectively recorded outside of MoveMonitor wear times. Water damage despite thorough education, safety concerns regarding pressure areas, and adherence issues leading to data not reaching inclusion thresholds all led to physical activity data loss, warranting further consideration of device choice in acute care.
Based on mean postoperative LOS, the day 10 MoveMonitor wear period occurred within the inpatient admission; however, a transition between the inpatient setting and home/local accommodation occurred in 35% of participants, while the MoveMonitor was still active. During the week 10 wear period, both sub-acute rehabilitation attendance and timing of inpatient  readmissions may also have varied between participants. These factors should be considered when appraising physical activity results. Pre-LTx 6MWD data was not able to be standardized regarding time performed pre-LTx. The IQR of the week 3 follow-up was within one week of the protocol-defined time-point, however this occurred earlier than week 3 in 73% of patients. Although COVID-19 did not negatively impact primary outcome data collection, intervention delivery, or week 3 follow-up, physical capacity outcomes were unable to be collected at week 10 in a small number of participants due to infection control limitations. Affected participants continued with sub-acute, homebased physical rehabilitation with weekly remote exercise diary monitoring, following the same principles of incenter rehabilitation as equipment and environment allowed (Janaudis-Ferreira et al., 2019). Future researchers should consider the implementation of both remote interventions and assessments.
As this trial was single-center, aspects of the protocol may not be generalizable to other LTx centers due to inherent variations in activity, transplant procedure, and models of care. Dependent of resources and service delivery models, the rehabilitation protocol utilized in this trial (Supplementary File Method 1) may not be feasible in all centers. We also excluded patients following SLTx based on the low proportion of these procedures nationally, hypothesizing that trial scope would not allow for representative recruitment of this relatively selective sub-group exhibiting differences in age, morbidity, and survival characteristics (Australia and New Zealand Cardiothoracic Transplant Registry, 2018; Chambers et al., 2019). However, SLTx activity was higher than anticipated at 18% over the trial period, more consistent with international activity in 2017 of 19% (Chambers et al., 2019). It is possible that this sub-group may respond differently to post-LTx physiotherapy. A generalizable, multi-center protocol should be considered in future definitive trials.
Investigation of intensive physical rehabilitation in the ICU, notably in patients requiring ECMO, is also needed based on a 14% incidence of post-LTx ECMO during recruitment. This group is known to have physical function deficits, leg complications, and prolonged LOS (Hayes et al., 2018), compounding the already delayed recovery of muscle strength post-LTx that may impact on both physical capacity and health-related QOL (Walsh et al., 2013). Cumulative, unplanned acute readmissions in the first year post-LTx are also correlated with an increased risk of mortality (Lushaj et al., 2016;Mollberg et al., 2017). These factors strengthen the case for potential intervention periods to extend into the sub-acute setting, with follow-up periods beyond 10-weeks.
The planning of complex PT intervention during acute care often requires an element of individualization and ongoing adaptation, however there are common traits post-LTx. Barriers to physical activity in patients following LTx can include perceived physical limitation, energy levels, fear, and comorbidities (van Adrichem et al., 2016). Facilitators include motivation, coping strategies, education regarding the consequences of inactivity, routine/habit formation, and goal setting. Structured incorporation of these aspects into future trials should be considered.
Increases in service delivery ultimately require an increase in staffing to deliver additional interventions, whether physiotherapist or physiotherapy assistant. This may be cost-effective if the intensive intervention is able to deliver enhanced patient and health system outcomes. For future trials to investigate efficacy, care must be applied when choosing an appropriate outcome from which to power a definitive trial. Although there lacks data for acute post-LTx PT, similar trials in trauma have demonstrated efficacy with lower total PT contact time (124 min), over a shorter inpatient LOS (<8 days) (Calthorpe et al., 2014). This is especially important in the acute setting with less time for potential PT intervention, highlighted by the median of nine additional interventions completed by the experimental group in this feasibility trial.
Lung transplantation aims to improve QOL as well as life expectancy (Weill et al., 2015); however, it is currently unknown whether the intensity of acute physiotherapy has any impact on QOL (Vermuelen, van der Bij, Erasmus, and TenVergert, 2007). While the EQ-5D-5L, and EQ-5D variant have been used and shown to be responsive both pre-to post LTx , and post-LTx (Goetzmann et al., 2018;Wietlisbach et al., 2020), a LTx-specific QOL outcome (Singer et al., 2014), or the more widely utilized Short-Form 36 (Janaudis-Ferreira et al., 2019) may also be considered appropriate. Pain in the postoperative period may also have a significant impact on physical activity levels (Richard et al., 2004), which in turn may be better predictors of health outcomes than physical capacity testing in patients following LTx (Walsh et al., 2016). These factors should be considered by LTx researchers.

Conclusion
This study has shown that it is feasible to recruit to a trial of intensive physiotherapy in the acute setting post-LTx, with high rates of eligibility and consent. Twice daily, individualized, supervised physiotherapy showed no increased risk of AE, with acceptable delivery success in a high-volume specialist LTx center. These results will underpin definitive clinical trials adequately powered to assess the impact of intensive acute physiotherapy on physical function and capacity, exercise participation, and/or QOL in patients following LTx.