Introduction

Total en bloc spondylectomy (TES) is a surgery designed to achieve complete resection of an aggressive benign or malignant spinal tumor, including spinal metastasis, and to provide an adequate tumor margin [1]. The en bloc resection technique has been proven to be effective for improving the prognosis of both primary and metastatic bone tumors of the spine [2]. We previously reported excellent results of spinal metastasectomy for selected patients, and these results demonstrated that excisional surgery could potentially prolong survival [3, 4]. Furthermore, Kato et al. [5] reported that most patients were satisfied with the results of TES and maintained good ability to perform activities of daily living (ADLs) during long-term follow-up after surgery. However, no studies have evaluated the outcomes of rehabilitation early after TES. One previous study [6] demonstrated that patients with a malignant spine tumor have a short life expectancy; therefore, their length of stay for rehabilitation may be shorter to allow more time at home, thereby enhancing the quality of life. However, TES is an extensive and invasive surgical procedure that is technically demanding; therefore, perioperative systemic complications are possible [7]. The aim of TES for the treatment of spinal tumors is complete resection of the diseased vertebra. Therefore, TES is quite different from most spinal surgeries that are performed to improve symptoms and ADL performance, because the procedure can result in additional symptoms and deterioration in the ability to perform ADL due to the surgical stress associated with curative resection of the tumor. TES for spinal tumors may produce neurological dysfunction as a result of ischemic or mechanical damage to the spinal cord and nerve roots [8, 9]. Therefore, it is during the acute phase after surgery, perioperative systemic complications, especially neurological dysfunction, may disturb the recovery of physical functions and delay early rehabilitation for patients. Approximately 5–10% of cancer patients will develop spinal cord compression during the course of their disease [10]; therefore, patients are at risk for neurological deficits before surgery. Therefore, this study aimed to evaluate postoperative walking ability and ADL ability during the acute phase after surgery and to identify risk factors for poor rehabilitation outcomes early after TES.

Methods

Patients

We collected the medical data of 140 patients who underwent TES, performed by a single surgeon in our department between April 2010 and December 2017. The study population included 81 men and 59 women, with a mean age of 53.2 years (standard deviation [SD], 13.7 years) at the time of surgery. Of the 140 patients, 30 had primary tumors and 110 had metastatic tumors. The histology of primary vertebral tumors included a giant cell tumor for ten patients, an aggressive type of vertebral hemangioma for seven patients, osteosarcoma for three patients, chondrosarcoma for two patients, and other diagnoses for eight patients. The histology of metastatic tumors included kidney cancer for 36 patients, breast cancer for 14 patients, thyroid cancer for 13 patients, lung cancer for 9 patients, bladder cancer for 3 patients, colon cancer for 3 patients, liver cancer for 2 patients, and other diagnoses for 30 patients. At the time of TES, all the 110 patients had solitary and removable metastases in the spine. Seventy-two patients (65%) did not have other metastases, and 38 patients (35%) had metastases in the non-spinal skeleton or other organs. The decision to perform TES was based on patient willingness and the surgical strategy for spinal metastasis that was designed based on the treatment goals of the intervention and aggressiveness of the surgery, which was determined by a ten-point scale that considered the tumor histology data and extent of visceral and bone metastases [11]. For patients with a Tomita score of 2, 3, or 4, TES was indicated for long-term local control of the tumor. When patients had metastases from kidney or thyroid cancer, for which surgical metastatectomy is widely accepted according to current guidelines, selected patients with scores of 5 or more were considered to be potential candidates for TES [12]. We consider that the indication for patients with Tomita score of 5 or more is only for the selected patients with metastases from kidney and thyroid cancers. All patients participated in a medical rehabilitation program during the perioperative period. Figure 1 shows the medical rehabilitation program for TES at our hospital. This program aimed to recover muscle strength lost due to surgical stress and bed rest after surgery to achieve ambulation and the ability to perform ADL as soon as possible. In addition, when perioperative systemic complications occurred, exercises in the supine position was performed to maintain muscle strength and ambulation was begun as soon as any complications were resolved. Training in the supine position was provided 5 days per week; on each treatment day, patients participated in 20 min of physical therapy to maintain range of motion and muscle strength of the lower limbs.

Fig. 1: Medical rehabilitation program for total en bloc spondylectomy.
figure 1

The preoperative phase (1), the program without perioperative complications (2, 3, and 5), and the program for patients with perioperative complications (4) are shown.

Surgical procedure

Our surgical technique for TES has been described in detail elsewhere [2]. Decompression of the spinal cord was accomplished by en bloc corpectomy following en bloc laminectomy. For TES of C7, T1, and L3–5 lesions, a posterior–anterior approach is required because the nerve roots at the tumor level must be preserved, and the vertebral body is resected via an anterior approach. For TES of T2-L2 vertebral lesions, a single posterior approach is employed, with the transection of the nerve roots at the tumor level during dissection and resection of the vertebral body. In some patients with large tumors that expand anteriorly to the paravertebral area, additional dissection via an anterior approach is required before the posterior TES procedure. The anterior column was reconstructed by inserting a vertebral spacer packed with a bone autograft. In all surgeries, posterior fixation extended two levels above and two below the tumor site.

Patient evaluation and data collection

The following patient data were extracted from the medical charts: age, sex, previous chemotherapy, previous radiotherapy, pathological/clinical grade of malignancy, number of resected vertebrae, nutrition (based on albumin level), comorbidities (diabetes, hypertension, and ischemic cardiac diseases), affected level of vertebrae (thoracic or lumbar), preoperative neurological function using the American Spinal Injury Association (ASIA) Impairment Scale [13], and perioperative complications occurring within 1 month of surgery (deep surgical site infection (SSI), postoperative cerebrospinal fluid (CSF) leakage, and neurological deficits). The pathological/clinical grade of malignancy was classified into three groups (slow growth, moderate growth, and rapid growth) according to the Tomita score [11]. In this study, postoperative CSF leakage was defined as drainage of colorless liquid observed in the negative pressure drainage system [14]. To evaluate ambulation status, the Spinal Cord Independent Measure indoor mobility item (SCIM item 12, ability to walk <10 m) [15] was used according to a previous study [16]. The SCIM indoor mobility item includes total assistance, wheelchair use, walking with aids, and walking without aids, and it has demonstrated excellent reliability for patients with a spinal cord injury [17]. To distinguish between individuals who could walk indoors independently and those who could not, an SCIM indoor mobility cutoff score was applied; scores of 0–3 were defined as unable to walk or dependent on assistance while walking and scores of 4–8 were defined as able to walk independently [16]. The Functional Independent Measure (FIM) is an 18-item instrument graded on a seven-point ordinal scale that is widely used to evaluate the ADL ability of the inpatient population undergoing acute rehabilitation [18]. The maximum total score is 126, and the possible item scores range from 1 (complete dependence) to 7 (independence) [18]. The FIM consists of motor scores (eating, grooming, dressing, using the toilet, sphincter control, locomotion, and transfers) and cognitive scores (communication, memory, social, and problem solving). Because no patient had cognitive dysfunctions, we used motor FIM scores (total of 91 points) to evaluate the ability to perform ADL. Because we defined the outcome of early rehabilitation as the ability to walk indoors independently and ADL score 1 month after surgery, evaluations of the SCIM indoor mobility item and motor FIM score were performed and compared before surgery and 1 month after surgery.

Statistical analyses

Continuous variables are expressed as means (SD) for parametric data, and the t-test was performed to compare abilities before surgery and 1 month after surgery or patients with and without perioperative complications. Medians (interquartile range [IQR]) were used for nonparametric data. The Mann–Whitney U-test and Wilcoxon matched paired signed-rank test were used for comparisons. Countable data were expressed as a percentage; comparisons between groups were performed using the chi-square test. Multivariate analyses using stepwise logistic regression were performed to identify factors associated with the ability to walk independently 1 month after TES. A multiple regression analysis was performed to clarify factors associated with the motor FIM score at 1 month after surgery because it was considered as a continuous variable. P < 0.05 was considered statistically significant. The JMP version 11 software program (SAS, Cary, NC) was used to perform the statistical analyses. We also used the G-power software (Franz Faul, Univesitat Kiel, Germany) to calculate the post hoc effect size (Pearson’s correlation coefficient; r) and the actual power of the sample.

Results

Patient characteristics

The mean operative time was 475.8 min (SD, 172.1 min). The mean intraoperative blood loss was 600.6 mL (SD, 686.1 mL). Fifty patients (36%) underwent previous chemotherapy, and 39 patients (28%) underwent previous radiotherapy. The tumor was located in the thoracic spine in 102 patients (73%); it was located in the lumbar spine in 38 patients (27%). Preoperative nutrition based on the albumin level was 3.7 g/dL (SD, 0.1 g/dL). Diabetes was found in 13 patients (9%), hypertension was found in 24 patients (17%), and ischemic heart diseases were found in 4 patients (3%). The malignancy grades of the tumors based on the Tomita system were slow growth for 45 patients (32%), moderate growth for 50 patients (36%), and rapid growth for 45 patients (32%). Before surgery, 89 patients (64%) were considered to have a neurologically normal ASIA grade E. Preoperative neurological deficits were found in 51 patients (36%). Twenty-three patients (16%) were considered to have ASIA grade D, 20 patients (14%) were considered to have ASIA grade C, 5 patients (4%) were considered to have ASIA grade B, and 3 patients (2%) were considered to have ASIA grade A. There were no surgery-related or all-cause deaths during the study period.

Perioperative complications

Within 1 month after surgery, postoperative CSF leakage was observed in 29 of the 140 patients (21%), and SSIs were observed in 11 patients (8%). Postoperative neurological deficits with depreciation of at least one ASIA score occurred in 41 patients (29%). Twenty patients (14%) underwent additional surgery because of SSI, CSF leakage, and wound dehiscence.

Rehabilitation outcomes early after TES

Ambulation status, represented as the SCIM indoor mobility score (total of 8 points), was significantly decreased at 1 month after surgery compared with before surgery (score, 8 [IQR, 5–8] versus 6 [IQR, 2.3–8]; P < 0.01; r = 0.37; power = 0.99). Before TES, 109 of 140 patients (78%) could walk independently (score, 8; IQR, 8–8). At 1 month after TES, 92 of 140 patients (66%) could walk independently (score, 8; IQR, 6–8) (P < 0.01). At 1 month after surgery, the motor FIM score (total of 91 points), which reflected the ability to perform ADL, was also significantly lower than that before surgery (score, 76.6 [SD, 3.7] versus 68.2 [SD, 27.2]; P < 0.01; r = 0.38; power = 0.99) (Table 1). Table 2 shows the influence of preoperative neurological deficits and perioperative complications on rehabilitation outcomes early after TES. The SCIM indoor mobility score and motor FIM score were significantly decreased for patients with postoperative CSF leak, SSI, preoperative neurological deficits, and postoperative neurological deficits. Because of their emergency surgery, we could not directly evaluate the walking ability and ADL scores for two patients. Therefore, we evaluated these scores using medical records.

Table 1 Comparison of ambulation status and ADL score before surgery and 1 month after surgery.
Table 2 Influence of preoperative neurological deficit and perioperative complications for rehabilitation outcomes 1 month after TES.

Risk factors for the ability to walk independently 1 month after TES

Results of the univariate analysis of risk factors for ambulation ability 1 month after surgery are shown in Table 3. Factors including number of resected vertebrae, radiotherapy history, nutrition, preoperative ASIA score, the incidence of SSI, postoperative CSF leak, and postoperative neurological deficits demonstrated a significant effect on the ability to walk independently 1 month after surgery in the univariate analysis. In this study, both univariate and multivariate analyses were used to evaluate risk factors. Five significant factors (preoperative ASIA score, preoperative ADL score, incidence of SSI, postoperative CSF leak, and postoperative neurological deficits) were extracted using a stepwise regression analysis. Multivariate analysis results demonstrated that the preoperative ASIA score (odds ratio [OR], 5.23; 95% confidence interval [CI], 2.07–15.99), incidence of SSI (OR, 15.27; 95% CI, 2.26–127.7), postoperative CSF leak (OR, 13.42; 95% CI, 2.93–78.82), and postoperative neurological deficits (OR, 4.86; 95% CI, 1.33–19.99) were significantly associated with the ability to walk independently at 1 month after TES (Table 4). At 1 year after surgery, we could perform follow-up assessment of 110 patients (79%) of the 140 patients who underwent TES, including the 24 who had postoperative neurological deficits. The assessment showed that 107 of 110 patients (97%) could walk independently and 3 who were AIS A before surgery maintained their neurological deficits.

Table 3 Univariate analysis of the relative risk of the ability to walk independently 1 month after TES.
Table 4 Multivariable stepwise logistic regression of significant risk factors for nonambulatory status 1 month after total en bloc spondylectomy.

Risk factors for independence of ADL 1 month after TES

Table 5 shows the results of a stepwise multiple regression analysis of the motor FIM score 1 month after TES. There was no multicollinearity between the variables. Preoperative ASIA scores, SSI incidence, postoperative CSF leak, and postoperative neurological deficits were independently associated with the motor FIM score (P < 0.01).

Table 5 Stepwise multiple liner regression analysis for motor FIM score 1 month after total en bloc spondylectomy.

Discussion

Several researchers have reported the rehabilitation outcomes of patients with spine tumors. McKinley et al. [6] and Scivoletto et al. [19] demonstrated that the functional improvement of patients with spinal tumors who underwent chemotherapy or radiotherapy was less than that of patients who sustained a traumatic spinal cord injury because they may have complications caused by the systemic manifestations of cancer. Furthermore, Quan et al. [20] demonstrated that palliative surgery for spinal metastases was effective for achieving rapid overall improvements in axial and radicular pain, neurological deficits, and ambulatory status. However, no study has evaluated the outcomes of rehabilitation after TES.

In the present study, walking ability and ADL scores were significantly lower at 1 month after TES than they were before TES. Furthermore, the multivariate analyses indicated that preoperative neurological deficits and major complications including CSF leakages, SSI, and neurological deterioration were risk factors for poor rehabilitation outcomes during the acute phase after TES. Previous studies [21, 22] reported that the perioperative complication rate is associated with TES and its independent risk factors. However, this is the first study to report the influence of perioperative complications on the recovery of ambulation and ability to perform ADL during the acute phase after TES.

In contrast, walking ability and ADL scores after TES were significantly lower than the preoperative scores. However, 66% of patients could walk independently at 1 month after TES. The mean clinical change in the motor FIM score, indicating the ability to perform ADL, was 8.4 (SD, 20.6) (before TES; 76.6 [SD, 3.7]; 1 month after TES, 68.2 [SD, 27.2]); this change was not large. A previous study [23] demonstrated that the minimal clinically important difference in the motor FIM score is 17 points. In this study, 44 (31%) patients experienced a change in the preoperative motor FIM score of more than 17 points, and 96 patients (69%) experienced a change in the motor FIM score of less than 17 points. Therefore, at only 1 month after surgery, 66% of patients could walk independently indoors and 69% of patients maintained their ADL score.

Boriani et al. [24] suggested that en bloc resection can improve the prognosis of aggressive benign and low-grade malignant tumors in the spine; however, complications are not rare and are possibly fatal. Because of its demanding procedures, and because it is used for patients who often have a complicated medical history, such as cancer, the perioperative complications rates with TES are high compared with those associated with other spinal surgeries [25]. In the present study, postoperative CSF leakage was observed in 21%, SSIs were found in 8%, and postoperative neurological deficits occurred in 29% of 140 patients. A previous study [24] reported a complication rate of 34% for patients who underwent en bloc spinal resection, which was similar to the complication rate of our study. Among the risk factors for poor recovery of ambulation, number of resected vertebrae, radiotherapy history, nutrition, preoperative ASIA score, SSI incidence, postoperative CSF leakage, and postoperative neurological deficits had significant effects on the ability to walk independently 1 month after surgery in the univariate analysis performed during this study. According to a previous study, a higher complication rate was observed for cases that required multisegmented resections [26]. Moreover, we previously found [21, 22] that a history of irradiation was a significant independent risk factor for SSI and postoperative CSF leakage. Preoperative malnutrition has been reported to be associated with increased postoperative complication rates and hospital stays [27]. These factors might have influenced the high complication rate and the functional outcomes although they were not significant in the multivariate analysis of this study. The multivariate analysis indicated that postoperative CSF leakage, SSI, and postoperative neurological deficits were associated with ambulation and the ability to perform ADL at 1 month after TES. In addition, the preoperative ASIA score was an independent risk factor for walking ability and ability to perform ADL after surgery. Postoperative CSF leakage causes further complications and prolonged hospitalization [28]. Moreover, SSI is a serious complication for patients who undergo major surgery, such as TES [29]. Because these complications require prolonged bed rest and immobilization, disabilities such as weakness, disuse, and inability to dependently perform ADL occur. Recently, it has been shown that only 4 days of limb immobilization can lead to an ~10% decrease in the mean muscle fiber cross-sectional area in both young and elderly men [30]. Disabilities due to deconditioning and neurological deficits are considered major causes for inpatient rehabilitation [31]. Several studies [32, 33] reported that the level of preoperative paralysis was associated with postoperative ambulation, and ambulatory function before surgery was the main determinant of postoperative gait function. Our study presented similar findings. Hence, as part of the informed consent process, surgeons should discuss the likelihood of postoperative deterioration in the ability to perform ADL and the delay of early rehabilitation with their patients. This information could be effective for patients and their families to predict the delay in ADL recovery and discharge when these factors occur in the perioperative period. Therefore, this study provides quantitative measures of these factors that can be incorporated in the decision-making process. In this study, although high complication rates and deterioration after surgery were observed, the rate of postoperative complications was similar to that reported in a previous study [24]. Approximately 70% of patients could maintain their walking ability and ability to perform ADL at only 1 month after surgery. Moreover, the walking abilities of 97% of patients who could be assessed at a year after surgery were considered independently except patients who were AIS A before surgery. These results indicated that the patients who underwent TES recovered their ability to perform ADL including walking within 1 year after surgery, and that surgery is appropriate and indicated even for metastatic tumor patients with a favorable prognosis [5].

The present study had several limitations, including its retrospective design, diversity of the patients’ backgrounds, lack of pain data, which could affect the ADL outcome, the possible influence of confounding. Furthermore, this study was performed at a single center, as is often the case for such extraordinary surgery. Inaccuracies owing to incorrect documentation may have existed and problems were caused by missing data. In addition, we could not clarify the long-term effects of TES on rehabilitation outcomes. Further studies are needed to identify the effects of the early recovery of ambulation and ability to perform ADL on the long-term physical function of patients after TES. Despite these limitations, the present study is the first to report the influence of preoperative neurological deficits and perioperative complications on the recovery of ambulation and ability to perform ADL during the acute phase after TES. The results indicated that perioperative neurological dysfunction and perioperative complication are independent risk factors for poor outcomes of early rehabilitation.

In conclusions, walking ability and ADL scores were significantly lower 1 month after TES than before TES; however, these differences were not clinically meaningful. Multivariate analyses revealed that the factors influencing delayed recovery of ambulation and ability to perform ADL during the acute phase after TES were perioperative complications and preoperative neurological deficits. Risk factors identified in this study should be considered when performing medical rehabilitation program after TES. In the informed consent process, surgeons should discuss the likelihood of postoperative deterioration in the ability to perform ADL and the delay of early rehabilitation.