Data were retrospectively extracted from a prospectively-collected center registry of LTx patients and electronic medical records from the ICU. Consecutive patients who underwent LTx between March 2015 and February 2016 were assessed for inclusion in this study. Among these patients, those who received early PR initiated in the ICU according to our institutional eligibility criteria (Figure 1) were selected for this study. Patients were excluded if PR was initiated 2 weeks after admission to the ICU post-transplant. The usual procedures for LTx have been described previously [16]. All patients were admitted to the ICU from the operation room. This study was approved by the Institutional Review Board of Severance Hospital of Yonsei University College of Medicine, with a waiver of individual consent due to the retrospective study design (IRB No. 4-2013-0770).

In 2015, a structured rehabilitation program was developed at our center, with a focus on improving rehabilitation and facilitating mobility. The program was designed to be initiated within the ICU for patients with the following conditions or treatments: (1) LTx recipients; (2) prolonged mechanical ventilator use after surgery; (3) prolonged mechanical ventilator use in chronic obstructive pulmonary disease; and (4) prolonged mechanical ventilator-dependent chronic respiratory failure. The PR was developed by a core rehabilitation team, consisting of a pulmonary/critical care physician, specialist nurse, and physical therapist. On admission to the ICU after LTx, all recipients received physiotherapy provided by the ICU nursing staff at least once daily from day 1 post-transplant. Physiotherapy was individually tailored but primarily included frequent chest percussion and range of motion (ROM) of the upper and lower extremities lasting 5 to 10 minutes, and a position change every 4 hours. In addition, as soon as LTx recipients were considered able to receive PR, initial assessment of the functional grade for mobilization capacity and subsequent establishment of a rehabilitation plan were conducted by a specialized physical therapist. A qualitative assessment of the use of steroids or neuromuscular agents was made for each patient.

Our PR protocol involved four levels of mobilizing therapy, consisting of 60 minutes of exercise each day, 5 days a week. The functional grade of mobilizing capacity was recorded weekly until PR was over, as illustrated in Figure 2, and the corresponding level of PR was delivered by an attending specialized therapist. In brief, subjects of grade 1 were confined to bed and received bedside physical therapy, including passive ROM, active-assistive ROM, active ROM, resistance exercise, sitting exercise leaning against the backrest, and bed cycling. Subjects who were able to sit on the edge of the bed (grade 2) received dynamic sitting balance training. The ability to stand with one’s knees supported by others (Bobath standing) or to stand with a walker at the bedside was grade 3. Advancement to the final grade was based on the ability to walk (grade 4, gait group).

Continuous variables were analyzed with Mann-Whitney Utests and summarized as median and interquartile range. Categorical data were analyzed with Fisher exact tests or Pearson chi-square tests, depending on the distribution, and were summarized as absolute frequencies and percentages. All analyses were performed with the SPSS version 16.0 software (SPSS Inc., Chicago, IL, USA).

During the study period, 26 patients had LTx and PR after ICU admission. Of these 26 patients, 22 were eligible for the study: they had PR sessions of at least 7 days that began in the ICU within 4 weeks after transplant. Patient demographics and preoperative status are summarized in Table 1. The median age of the 22 patients was 53 years (range, 25 to 73 years), and 14 (64%) patients were males. Indications for LTx were idiopathic pulmonary fibrosis (n=12), collagen vascular disease interstitial lung disease (n=1), chronic pulmonary obstructive disease (n=1), bronchiectasis (n=2), bronchiolitis obliterans syndrome after allogeneic stem cell transplantation (n=3), acute interstitial pneumonitis (n=2), and primary pulmonary hypertension (n=1). Emergency status as assessed by the Korean Lung Allocation Score [17] is summarized in Table 1. Preoperative extracorporeal membrane oxygenation (ECMO) support was used in seven patients (32%). The median ischemic time was 282 minutes (range, 185 to 325 minutes), and all patients received bilateral LTx using ECMO for intraoperative cardiopulmonary support. All patients were intubated and transferred to the ICU after LTx; 12 (55%) were weaned off of ECMO in the operating room.

Eight patients (36%) had received rehabilitation prior to LTx with a median of 26 days (range, 2 to 100 days). Of the eight patients who had preoperative PR, three had received hematopoietic stem cell transplantation. All 22 patients started PR at a median of 7.5 days (range, 1 to 29 days) after LTx (Table 2). We did not observe any rehabilitation-related complications during follow-up. Median time from weaning of intraoperative ECMO support and PR initiation was 6 days (range, 1 to 21 days). Thirteen patients (59%) were on mechanical ventilator support when PR was initiated with a median of 11.5 days (range, 2 to 43 days), while eight patients were on tracheostomy status. Neuromuscular disorder was observed in six patients. At the time of PR, the initial functional grade of mobilizing capacity according to our institutional grading system (Figure 2) was grade 1 in nine patients, grade 2 in seven patients, grade 3 in four patients, and grade 4 in two patients (Table 2). Patients received a median of 6.5 weeks (range, 1 to 36 weeks) of rehabilitation before hospital discharge or cessation for safety reasons. At the end of PR, the functional grade was grade 1 in six patients, grade 3 in two patients, and grade 4 in 14 patients. Fourteen of the 22 patients were able to walk with or without assistance, and 13 of them maintained grade 4 until discharge; the eight remaining patients never achieved grade 4. In comparison to those who failed to achieve and maintain grade 4, patients with grade 4 had significantly lower scores on the modified Medical Research Council dyspnea scale (P=0.074) and the New York Heart Association functional class (P=0.022). These patients also had a significantly shorter median duration of propofol (8 days vs. 14 days, P=0.02) and remifentanyl (16 days vs. 23 days, P=0.004) infusions during immediate postoperative management. The use of other drugs, such as high dose steroids, used within 1 month prior to PR, muscle relaxants, or midazolam, did not differ between grade 4 patients and patients of lower grades. Other pretransplant clinical characteristics also did not differ between these groups.

Median duration of postoperative ECMO support was 2 days (range, 1 to 9 days); median duration of postoperative mechanical ventilation was 13 days (range, 3 to 149 days). Six patients received second ECMO support due to bleeding (n=3), pneumonia (n=2), or cardiac arrest (n=1). The median ICU stay was 13 days (range, 3 to 149 days), and the median hospital stay was 43 days (13 to 221 days). The ICU stay was shorter in grade 4 patients (6.5 days) than in patients of lower functional grades (32.5 days, P<0.001). Likewise, the hospital stay was shorter in grade 4 patients (50.0 days) than in patients of lower functional grades (102.5 days, P=0.035). Within 6 months from LTx, there were 10 cases of mortality, and after a median follow-up of 13.4 months (range, 7.0 to 16.0 months) after LTx, 13 patients (59%) had died (Table 2). Infection (77%) was the most common cause of death in our cohort.