Study design
This study was designed to retrospectively review the impact of the characteristics and variable factors and overall survival of patients with deterioration and good recovery of ADL from admission to discharge in newly diagnosed patients with GBM who were enrolled in the database of the Medical Record of Kagoshima University Hospital. The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethical committee of Kagoshima University Graduate School of Medical and Dental Sciences.
Patients population
A total 110 patients with newly diagnosed GBM who were admitted to the neurosurgery department between January 2011 and October 2016 were analyzed in this study. All the patients underwent surgical tumor resection. Subsequently, they received rehabilitation services, including physical, occupational, and speech therapies during hospitalization for neurologic, physical, and psychological impairments. However, five patients did not receive inpatient rehabilitation after surgery. Therefore, a total of 105 patients with GBM were included in the study. First, we investigated the BI score of all patients at admission, at 7 days after surgery, and at discharge. The median BI score of all patients was 65 at admission (interquartile range [IQR]: 35 – 90), 30 at 7 days after surgery (IQR: 0 – 60), and 65 at discharge (IQR: 20 – 90). Furthermore, we calculated the change of BI score from admission (pre-operation) to discharge, and the median change of BI score was -5 (IQR: -20 – 5) (Supplementary Figure 1). Therefore, the present study was categorized into three groups by a quartile of change of BI score: the 1st quartile (≤ -25, n = 25, deterioration group), the interquartile range (n = 55, no remarkable change group). the 3rd quartile (≥ 10, n = 25, good recovery group).
Outcome measures
The patient characteristics, such as the patient’s age, sex, KPS at admission, extent of resection, type of main treatment (TMZ, radiation, bevacizumab, TMZ concomitant radiation and bevacizumab), tumor location/hemisphere/size, length of hospital stay, duration from initial symptoms to surgery, initial symptoms (motor paralysis and weakness, cognitive dysfunction, headaches, visual field defect, dysphagia, and fatigue), surgery (surgery time, bleeding and transfusion volume, infusion volume, and fluid balance), adverse events during chemoradiotherapy, and duration from surgery to chemoradiotherapy were determined from the hospital’s medical records. If we could not identify the data of initial symptoms in the medical records, we determined by hearing it from the patients or their families. However, the initial symptom data were missing in 5 and 3 patients in the deterioration and good recovery groups, respectively. The extent of resection was defined as gross and near total (the entire or > 95% of the enhancing tumor was resected), partial (< 95% tumor resection) and biopsy only. We excluded patients with biopsy only (deterioration group, n = 5; good recovery group, n =2) in the analysis of intraoperative factors.
With respect to the postoperative events, we defined the presence or absence of fever (≥ 38°C), infection, pneumonia, intracerebral hemorrhage, and motor paralysis of the lower limb that occurred within 7 days after surgery. Furthermore, the rate of interruption and discontinuation of chemoradiotherapy was examined as a postoperative event. The Brunstrom recovery stage (BRS) of the lower limbs is used to assess motor paralysis [23]. The deterioration of motor paralysis after the surgery was defined as a decrease of ≥ 2 stages. The presence of postoperative events and motor paralysis was assessed by rehabilitation medical doctors and physical or occupational therapists.
In addition, we examined rehabilitation-rerated factors (length up to rehabilitation onset after surgery, length up to sitting and walking training onset, change of motor paralysis, severe cognitive disorder or depression, and patients who discontinued rehabilitation).
The severity of the adverse effects was evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE) version 3.0 (grade 0: no adverse event, grade 1: mild, grade 2: moderate, grade 3: sever or medical intervention is needed, grade 4: life-threatening, grade 5: death). Hematology and other adverse events, including the cognitive function, constipation, and fatigue were assessed more than once per week during chemoradiotherapy. Severe cognitive impairment and depressive symptoms were defined as depression grade 1 or higher and severe cognitive impairment as grade 3 or higher.
Furthermore, we followed the overall survival of patients with GBM. Overall survival was defined as the period from the time of surgery to the date of death or the date of the last follow-up for patients who were still alive. The overall survival time was calculated in months for all the patients.
Treatment and inpatient rehabilitation
Standard radiation therapy was started within 2 weeks after tumor resection, and concomitant chemotherapy (TMZ) was started on the first day of radiotherapy. Subsequently, adjuvant TMZ was started 4 weeks after the end of radiotherapy and was delivered for a period of 5 days every 28 days. The rehabilitation intervention was performed for 40 minutes to 1 h a day for 5 days a week to improve the functional impairment. Rehabilitation programs were provided by means of sitting, standing and walking training as early as possible after resection based on the current trend of postoperative early mobilization in the ICU [24, 25].
Statistical analysis
Statistical analyses were performed using both parametric and non-parametric tests after the Shapiro-Wilk test. Subsequently, the three groups were compared using either a one-way analysis of variance (ANOVA) or Kruskal-Wallis test, followed by Bonferroni’s post hoc tests for multiple comparisons. A one-way ANOVA was used to analyze the age, tumor size, and length of hospital stay. The Kruskal-Wallis test was used to examine the KPS at admission and BI score. Comparisons between patients with deterioration and good recovery groups were performed using Student's t-test or Mann-Whitney U test. The student's t-test was used to analyze the length from the initial symptoms to surgery, surgery time, bleeding volume, transfusion volume, infusion volume, fluid balance, length from surgery to chemoradiotherapy, length up to rehabilitation onset, length up to sitting, and walking training onset. The Friedman’s test was used to examine the BI score and motor paralysis at admission, after surgery, and discharge followed by Bonferroni’s post hoc tests for multiple comparisons. The chi-square test or Fisher's exact test was used for categorical variables. Cohen’s effect size was used to evaluate intergroup differences [26]. Stepwise multiple regression analysis was employed to determine the predictor variable associated with the improvement of ADL from admission to discharge. The independent variables were adjusted for predictor variable (age, extent of resection, length from initial symptoms to surgery, KPS at admission), intraoperative factors (fluid balance), postoperative factors (change in motor paralysis after surgery, fever, and fatigue), and the rehabilitation factors (length up to walking training initiation). The selection of these independent predictor variable was based on previous reports or the factors that were observed to be statistically different between the deterioration and good recovery groups. As the data of 16 patients were missing, the data of a total of 89 patients were used for a stepwise multiple regression analysis. The Kaplan-Meier overall survival time distributions were compared between patients in the deterioration and good recovery groups using the log-rank test. The follow-up of the overall survival of 6 patients (deterioration group; n = 3, no remarkable change group; n =3) was not possible. Therefore, the overall survival values of the 98 patients were analyzed. The 95% confidence intervals (CIs) and corresponding p values were provided. The statistical significance was set at p < 0.05, and data are expressed as mean ± standard deviation (SD). Statistical analyses were performed using SPSS version 26 (IBM, Armonk, NY, USA).