The effect of early rehabilitation after lumbar spine surgery: a systematic review and meta-analysis

Abstract

Background

Evidence-based data are required to provide insightful information on the timing of rehabilitation after lumbar spine surgery (LSS).

Objectives

The aim of this study is to systematically review the outcomes of early rehabilitation interventions and conduct its meta-analysis in patients after LSS.

Patients and methods

A total of 1183 articles were retrieved through PubMed (n = 793), Web of Science (n = 721), Scopus (n = 335), and ScienceDirect (n = 83) databases. Fourteen studies were included in the systematic review. The quality analysis and risk of bias assessment of the trials included in the systematic review were performed using the Physiotherapy Evidence Database (PEDro) scoring and classification system. Narrative synthesis and standardized mean difference based pooling results were given for the systematic review and meta-analysis, respectively.

Results

The additional benefit of early rehabilitation on physical function was moderately effective (ES: − 0.62, 95% CI − 1.00; − 0.25) at the 1-month follow-up. In terms of pain, early rehabilitation provided additional improvement at 1 month (ES: 0.34, 95% CI − 0.03; 0.71), 3 months (ES: − 0.14, 95% CI − 0.37; 0.10), 6 months (ES: 0.35, 95% CI 0.04; 0.65) and 1 year (ES: 0.21, 95% CI − 0.09; 0.52) follow-up at a low level of evidence.

Conclusions

This systematic review demonstrated that early rehabilitation mainly improved disability in the early period (1-month follow-up). Regarding pain, short-term (1 month) and mid-term (6 months) follow-ups showed the most significant additional benefit. The positive effects of starting rehabilitation early after surgery on pain may have positively affected disability, specifically in the early period (1 month).

Background

The diagnosis of degenerative lumbar spinal stenosis and back pain has increased over time due to longer life expectancies, the desire for a higher quality of life, awareness of the condition, and the availability of cutting-edge imaging tools. Patients with severe lower back pain who do not respond to nonsurgical treatments for 3–6 months frequently have lumbar spine surgery (LSS) [1]. LSS is widespread in the older population and is becoming more common as the average lifespan rises [2]. Lumbar spinal fusion has emerged as the most widely utilized surgical procedure, with a rate of 13.8% for degenerative disk disease due to its superiority in terms of effectiveness [3, 4].

The typical success rate for lumbar spine procedures regarding capacity to work, neurological symptoms, and leg/back diskomfort is between 45 and 72% and reported satisfactory clinical outcomes to range from 16 to 95% [5, 6]. Questionnaires on patients’ expectations after LSS demonstrated that pain reduction and better mobility are the most expected results [7]. Numerous studies have demonstrated the efficacy of rehabilitation as the primary treatment for low back pain. However, research has demonstrated that rehabilitation after LSS is preferable to only rehabilitation including non-operative treatment but remains unclear, whereas a recent systematic review concluded that surgery might be more efficacious than unstructured care but may not be more efficacious than structured cognitive-behavioral therapy [8].

A recent meta-analysis demonstrated that standard treatment after lumbar fusion surgery does not significantly reduce disability and pain at 6 months compared to rehabilitation that combines an exercise program with cognitive behavioral therapy. Additionally, multimodal rehabilitation, which incorporates exercise therapy and cognitive behavioral training, is more effective than exercise therapy alone at reducing disability and pain-related fear [9, 10]. The most common specialized exercises are the Williams and McKenzie exercise regimens, floor exercises with the exercise ball or band, co-contraction for the transversus abdominus/multifidus muscles, and lumbopelvic stabilization. These exercise routines have been found to be both short- and long-term beneficial concerning low back issues such as persistent pain, lumbar spinal stenosis, and lumbar disk degeneration [11, 12]. According to a recent systematic review and meta-analysis, rehabilitation that includes cognitive therapy or counseling while the patient participates in an activity program has better results than exercise-only rehabilitation for lumbar fusion surgery [13].

The timing of the rehabilitation therapy is a crucial consideration. A study showed that ambulation within 8 h after elective cervical and LSS improved outcomes such as less complication rate, shorter hospital stays, lower 90-day readmission, and lower urinary retention rate compared to the patients who ambulated between 8 and 24 h [14]. Systematic reviews have demonstrated the significance of the timing of rehabilitation following procedures other than LSS. For instance, early rehabilitation following spinal cord injury was related to better functional outcomes and shorter hospital stays, according to a recent review [15]. Additionally, Greenwood et al.'s comprehensive review and meta-analysis showed that rehabilitation reduces short- and long-term impairment and fear avoidance behavior after lumbar fusion surgery. However, the effect of early rehabilitation after LSS has not been thoroughly evaluated [16].

More evidence-based data for better patient outcomes in rehabilitation practice would emerge from a systematic review and meta-analysis that provides insightful information on the timing of rehabilitation after lumbar spine surgery (LSS). To date, no systematic review has focused on the effectiveness of early rehabilitation after LSS. Additionally, a more thorough evaluation is required to highlight existing exercise alternatives and rehabilitation strategies that do not involve exercises that can be performed throughout the postoperative period of lumbar surgeries.

Aim of the work

The aim of this study is to systematically review the outcomes of early rehabilitation interventions and conduct its meta-analysis in patients after LSS.

Patients and methods

Search strategy

"Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA)" and "Cochrane Handbook for Systematic Reviews of Interventions" guidelines were considered for the methodological design of the review [17, 18]. Between November 2022 and January 2023, the literature search was performed through PubMed, Web of Science, Scopus and ScienceDirect databases with the specific keywords presented in “Appendix”. The "Medical Subject Headings (MeSH)" database was used to identify keywords. The terms "Lumbar surgery", "Early rehabilitation", "Enhanced rehabilitation", "Accelerated rehabilitation", and "Fast-track rehabilitation" were combined with Boolean operators to focus on studies concentrating on early rehabilitation after LSS. The search was performed independently by two separate researchers of the study.

Eligibility criteria

Before the screening procedures, the study's investigators determined the inclusion and exclusion criteria to ensure that the studies included in the systematic review had a more homogeneous sample and methodology. Inclusion criteria for the review were: (1) studies focusing on the effectiveness of early rehabilitation after LSS, (2) studies with a randomized controlled design. Exclusion criteria for the review: (1) studies focusing on the efficacy of rehabilitation before LSS, (2) studies with other non-randomized controlled research designs and designs, (3) articles published in a language other than English, (4) duplicate publications, (5) publications for which the full text was not available, (6) studies focusing on the efficacy of medical interventions other than rehabilitation after surgery.

Study selection and data extraction

The datasets containing the independent searches of two researchers were imported into Rayyan (QCRI, Qatar) software. Rayyan is a practical and automated article management tool for systematic reviews. Owing to this software, duplicate records can be detected automatically [19]. On the other hand, it is possible to manually mark the inclusion of trials in the review with "yes", "no", and "maybe" commands on the title/summary.

The two investigators who performed the screening evaluated the trials' eligibility by considering the study's inclusion/exclusion criteria through the Rayyan software. When two investigators disagreed on trial selection, a consensus was reached by considering the opinion of an experienced investigator who is an expert in the field of neurosurgical rehabilitation and knowledgeable about the systematic review methodology. The CONSORT flowchart of the systematic review is presented in Fig. 1. "Author, purpose, gender, sample, sample size, intervention, assessment and outcomes sections of the included studies were recorded (Table 1).

Fig. 1

PRISMA flow diagram of the study

Table 1 Characteristics of the included studies

Quality and risk of bias assessment

The quality analysis and risk of bias assessment of the trials included in the systematic review were performed using the Physiotherapy Evidence Database (PEDro) scoring and classification system. The primary purpose of selecting the PEDro tool was to include specific items to audit the design of trials, including rehabilitation interventions. PEDro scoring was performed independently by the two investigators of the study. In case of disagreement, a consensus was achieved by obtaining the opinion of a third expert academic. PEDro addresses the level of evidence of the trials with 11 items, including eligibility criteria, random allocation, concealed allocation, baseline comparability, blind subjects, blind therapists, blind assessors, adequate follow-up, intention-to-treat analysis, between-group comparisons, point estimates and variability. Both items are scored with "Yes" (1-point) or "No" (0-point). The first question (eligibility criteria) is not included in the scoring. PEDro scores are classified as "excellent (9–10 points)", "good (6–8 points)", "moderate (4–5 points)", and "poor (0–3 points)". The validity and reliability of PEDro have been demonstrated [20].

Evidence synthesis and meta-analysis

The review results were presented, considering the principles of narrative synthesis when pooling was not possible. The procedures of "developing a preliminary synthesis, exploring relationships within and between studies, and determining the synthesis's robustness" were regarded during the synthesis. Then, the results were shown, considering the qualitative and quantitative characteristics of the trials. In the meta-analysis section, numerical data on pooling were presented. Meta-Mar software (Philipps-Universität Marburg, Germany) calculated effect size and associated statistics [28]. The "Standardized Mean Difference (SMD)" was calculated regarding the "mean, standard, and sample size" of the relevant pooled parameter. Unknown standard deviation and confidence interval values were calculated according to the "Cochrane Handbook" guidelines [17]. "SMD, CI, weighted average effect size and p-value" values were given for each parameter pooled for meta-analysis. The heterogeneity of the measurements was analyzed with "I2, Tau2, and Chi2". Meta-analysis results were schematized with Forest plots.

Results

A total of 1183 articles were retrieved through PubMed (n = 793), Web of Science (n = 721), Scopus (n = 335), and ScienceDirect (n = 83) databases. Fourteen studies were included in the systematic review. After excluding duplicate and irrelevant studies for systematic review, 37 articles were analyzed according to the eligibility criteria. We excluded 24 studies that did not meet the eligibility criteria. Finally, 13 studies were included in the systematic review (Fig. 1).

Quality analysis and risk of bias results

The median score calculated for the PEDro total score of the 13 studies included in the systematic review was 5 (range = 3–8) [22,23,24,25, 29,30,31,32,33,34,35,36,37]. According to the PEDro classification, there were 5 "good" [24, 25, 29, 31, 34], 6 "moderate" [22, 23, 30, 33, 35, 37], and 2 "poor" [32, 36] evidence-level studies. All studies provided details on eligibility criteria and random allocation [22,23,24,25, 29,30,31,32,33,34,35,36,37]. Seven studies stated that allocations were concealed [23,24,25, 29, 31, 33, 34]. Most studies (nine) provided information on the homogeneity of the groups in terms of baseline comparability for assessment parameters [23,24,25, 29,30,31, 34,35,36]. None of the studies mentioned the identity of "therapists and subjects". Only three studies reported that the assessors were blind [24, 31, 35]. Nine studies reported appropriate monitoring procedures [22, 25, 29,30,31,32, 34, 36, 37]. Two studies calculated the intention-to-treat analysis [25, 29]. Only one study did not provide data on between-group comparison [36], point estimates and variability (e.g., intergroup comparison, SD, CI) [32]. Regarding items, the median value for total scores calculated from the scores of 14 studies was 9. Accordingly, items 3, 5, 6, 7, and 9 were below the median value (Table 2).

Table 2 PEDro scores of the trials

Study characteristics

A total of 1658 patients were available in 13 studies included in the systematic review [22,23,24,25, 29,30,31,32,33,34,35,36,37]. Four studies included Lumbar Microdiskectomy [22, 24, 35, 36], 3 Lumbar Fusion Surgeries [23, 29, 34], 2 Lumbar disk Herniation Surgeries [25, 32], one Microsurgical Lumbar disk Herniation [31], one study included "Lumbar Microdiskectomy or Percutaneous Endoscopic diskectomy", one "Percutaneous Transforaminal Endoscopic diskectomy [30] and one Robot-Assisted Minimally Invasive" and "Minimally Invasive Internal Lumbar Spine Fixation" [33]. All the studies focused on the effectiveness of early rehabilitation. Ten ODI, 9 VAS, 3 SF-36, 2 BDI, EQ-5D, 2 FABQ, two muscle strength and one each 50 Foot Walking Test, 6MWT, BBQ, BI, Complication Rates, Reliability/Expectation Questionnaire, Multifidus and Longissimus Muscle Cross Sectional Area (mm²), CSQ, CST, Early Retirement, EQ-5D-3L, Global Perceived Impact Scale, Health Care and Productivity Costs (euros), Intraabdominal Pre-Activation Pattern (seconds) Long Term Curative Effects (Excellent/Good/Bad), Lordosis index (MRI), Lumbar curvature (MRI), Lumbar Function Scale, Orebro Musculoskeletal Pain Screening Questionnaire, Pain Coping Inventory, Patients' 1st and 4th and 4. Days (Complete/Partial/Non-Compliant), Postoperative Conditions (Drainage Time, Time from Placement to Removal of Surgical Plasma Drainage Tube, Time to Lying on the Floor for the First Time After Surgery, Time from Completion of Surgery to Return to the Ward, Time to Get Out of Bed and Standing on Lumbar Support; Time of Postoperative Hospitalization, Time from Completion of Surgery to diskharge.), PSFS, QALY, Questionnaire (Remaining Sciatica, Sick Leave Days, Questionnaire (Working Status, Sick Leave, External Healthcare Use, Analgesic Use, Treatment Satisfaction, Frequency of Education and Reoperation Rates), Percentage Return to Work Questionnaire (%), Return to Work Rate, Return to Work (weeks), Roland's Disability Questionnaire, Sacral Tilt Angle (MRI), Satisfaction with Procedure), SES, Short Form McGill Pain Questionnaire, Less than 12, Short-Term Curative Effects (Excellent/Good/Bad), Spinal Stability (Lateral X-Ray), SRH (Table 1) [22,23,24,25, 29,30,31,32,33,34,35,36,37].

Quantitative synthesis results

Regarding pain parameters evaluated by VAS, the advantage of early rehabilitation (min 3 months, max 1 year) was emphasized in 5 of 9 studies [24, 29,30,31, 34]. Four studies emphasized that early rehabilitation did not contribute more to pain (min 1 month, max 7 years) [22, 33, 35, 36]. Four of the nine studies that evaluated ODI-based physical function reported that early rehabilitation (min 3 months, maximum 1 year) provided significantly more improvement [23, 24, 29, 34]. Five studies reported no additional benefit from early initiation of rehabilitation (min 1 week, max 3 years) [22, 25, 30, 33, 36]. Most of the studies (four) reported that early rehabilitation had no additional positive effect on quality of life. Two studies showed that early rehabilitation did not positively affect depression and fear avoidance beliefs [29, 35, 36]. Detailed results of the studies are presented in Table 1.

Meta-analysis results

Of the seven homogeneous studies, five evaluated pain evaluated by VAS [25, 29, 31, 33, 36], six assessed function by ODI [23,24,25, 29, 33, 36], and 3 included quality of life measurement by EQ-5D and SF-36 [23, 24, 29]. The additional benefit of early rehabilitation on physical function was moderately effective (ES: − 0.62, 95% CI − 1.00; − 0.25) at the 1-month follow-up. However, at 3 months (ES: 0.06, 95% CI − 0.17; 0.29), 6 months (ES: 0.09, 95% CI − 0.15; 0.33) and 1 year (ES: 0.08, 95% CI − 0.21; 0.37) follow-up, the contribution of early rehabilitation to physical function was at a low level of evidence. In terms of pain, early rehabilitation provided additional improvement at 1 month (ES: 0.34, 95% CI − 0.03; 0.71), 3 months (ES: − 0.14, 95% CI − 0.37; 0.10), 6 months (ES: 0.35, 95% CI 0.04; 0.65) and 1 year (ES: 0.21, 95% CI − 0.09; 0.52) follow-up at a low level of evidence. Finally, early rehabilitation was found to have a small effect size at 3 months (mental component) (ES: 0.13, 95% CI − 0.20; 0.47) and 1 year (general quality of life) (ES: − 0.04, 95% CI − 0.33; 0.25) follow-up (Figs. 2, 3, 4).

Fig. 2

Forest-plot of the VAS score at 1 month, 6 months and 1 year

Fig. 3

Forest-plot of the function (ODI) at 1, 3, 6 months and 1 year

Fig. 4

Forest-plot of the quality-of-life score at 3 months and 1 year

Discussion

This systematic review demonstrated that early rehabilitation mainly improved disability in the early period (1-month follow-up). Regarding pain, short-term (1 month) and mid-term (6 months) follow-ups showed the most significant additional benefit. There is insufficient evidence for the effectiveness of early rehabilitation in terms of quality of life and psychosocial status. The positive effects of starting rehabilitation early after surgery on pain may have positively affected disability, specifically in the early period (1 month). Future trials should elaborate on which types of exercises may be more effective in early rehabilitation.

In the early period, muscle strength, activities of daily living training, core stabilization, balance and gait training can provide more gains in the physical functions of individuals after lumbar surgery [13, 26]. In addition, earlier progress in joint range of motion may lead to less disability. Rehabilitation practices aimed at reducing pain after lumbar spine surgery may contribute more to improving function [16]. However, excessive training on the range of motion in the early period may cause an increase in the pain level of individuals. On the other hand, it is also comprehended that individuals have few gains in disability levels due to avoidance of functionality, fear of movement, and increased fear-avoidance beliefs to avoid pain [38, 39]. In this respect, our meta-analysis is unique to emphasize the gains in pain and function more clearly. In particular, we interpreted that improvement in early disability may be due to improvement in early and mid-term pain because the effect size in individuals' mid- and long-term functional improvements was low. However, since psychological and social multidimensional parameters (kinesiophobia, fear-avoidance, compliance, satisfaction) may affect physical function, more comprehensive psychosocial evaluations should be evaluated in future trials.

Analyzing the quality and bias risk of the studies

The median quality score of the studies included in the systematic review was moderate. Failure to mention the allocation procedure in some of the studies may have increased the risk of bias. However, the lack of blinding primarily decreased the methodologic quality. The fact that therapists and patients were not blinded in any study may suggest intervention bias. The use of assessor blinding in only three studies may suggest a suspicion of outcome bias. However, it should be emphasized that the effect of assessor bias is weakened when considering that most measurements were patient-reported outcome measures. Especially in studies with sensitive measurements such as muscle strength and physical performance tests, it should be emphasized that assessor blinding will reduce the risk of bias in order to ensure protocol integrity. Future studies should consider CONSORT or STROBE procedures regarding bias and randomization procedures [21, 40, 41].

Analyzing study characteristics

The types of surgery in the enrolled studies varied. In this respect, it should be considered that the difference in surgical procedures may have partially influenced the results regarding the rehabilitation procedure. The most commonly used surgical technique appears to be lumbar microdiskectomy [22, 24, 35, 36]. This finding suggests that especially minimally invasive methods are in the majority, and this advantageous situation for rehabilitation may produce more efficient results in terms of early rehabilitation.

The most preferred assessments in the studies are function and pain with ODI and VAS, respectively. The ODI is the gold standard scale used for many years in evaluating the lumbar region. The validity and reliability of VAS in postoperative patient follow-up have been emphasized in detail [42, 43]. Another point to be mentioned about the characteristics of the studies is that evaluations with non-standardized questionnaires were performed in some studies. Since the validity and reliability of non-standardized instruments have not been established, the consistency and responsiveness of the results are questionable [27]. On the other hand, the absence of studies addressing quality of life, psychosocial status, objective clinical measurements, and heterogeneous methodologies is indicative.

Effectiveness of the early exercise interventions

Pain

In more than half of the studies that addressed pain with VAS, the additional contribution of early rehabilitation was emphasized at follow-up from 3 months to 1 year [24, 29,30,31, 34]. In the meta-analysis, the pain was effective at a low-moderate level of evidence at 1- and 6-months follow-up [25, 29, 31, 33, 36]. In one study, early rehabilitation did not provide more effective results than the control group at a 6-month follow-up regarding pain assessed with the Örebro Musculoskeletal Pain Screening Questionnaire [25]. Early gains in terms of pain may also positively affect the disability levels of individuals. In this respect, the additional improvement gained at the 1-month follow-up suggests the effect of early rehabilitation on individuals regaining their physical functions in the early period. Considering that individuals complain of more pain in the acute period after lumbar spine surgery, short-term pain gain makes early rehabilitation advantageous. On the other hand, the maintenance of similar improvements in pain in the 6 months confirms the advantage of early rehabilitation in terms of pain in the medium term.

Disability

The most apparent gain in physical function was observed in the early period (1-month follow-up) with moderate evidence [33, 36]. Early rehabilitation efficacy was not noticed at an adequate level of evidence in later periods. Four of the studies evaluating physical function with ODI emphasized the superiority of early rehabilitation for a disability [23, 24, 29, 34], while the other five studies reported no additional contribution [22, 25, 30, 33, 36]. A study reporting the evaluation results with the Lumbar Function Scale reported the advantage of early rehabilitation [37]. The fact that improvements in physical function were reported only in the early period may be due to decreased pain in the early period. Although there is no additional advantage for individuals to start rehabilitation early in the middle and late periods, it may be valuable in clinical practice for secondary parameters such as independence in daily life and the shortening of hospitalization in the early period.

Quality of life and psychosocial status

According to the systematic review results, most studies noticed no additional contribution of rehabilitation to quality of life [29, 35, 36]. Meta-analysis results also supported these findings with a low effect size [23, 24, 29]. Since it is understood that gains in quality of life may occur in the long term, early rehabilitation was interpreted as usual when pain and disability outcomes were considered. It should be noted that long-term gains in disability and pain are similar in quality of life.

Psychological status is related to pain and the general condition of individuals. Studies showed that early rehabilitation did not positively affect depression and fear avoidance beliefs. Given the complex relationship of psychological state with other parameters such as pain, disability, satisfaction, complexity, kinesiophobia and heterogeneous study designs, it is difficult to make precise predictions. Future studies should focus more on secondary psychological parameters such as patient-oriented satisfaction and kinesiophobia [38, 39].

Study limitations

Some databases (e.g., EMBASE) could not be searched because the authors did not have access. Non-standardized assessment tools in the studies may have provided some results of questionable validity and reliability. The effects of surgical techniques on the study could not be addressed. Since 14 studies were included, an exclusion criterion related to the surgical procedure would have reduced the number of studies included in the meta-analysis, reducing the efficiency of effect size analyzes. However, the possible effect of surgical procedures may be a limitation affecting the study results. Finally, different rehabilitation protocols applied within the scope of early rehabilitation may suggest heterogeneity in the studies considered in pooling analyzes.

Conclusions

This systematic review demonstrated that early rehabilitation mainly improved disability in the early period (1-month follow-up). Regarding pain, short-term (1 month) and mid-term (6 months) follow-ups showed the most significant additional benefit. There is insufficient evidence for the effectiveness of early rehabilitation in terms of quality of life and psychosocial status. The positive effects of starting rehabilitation early after surgery on pain may have positively affected disability, specifically in the early period (1 month). Future trials should elaborate on which types of exercises may be more effective in early rehabilitation.

Appendix: Keywords

Search strategy PubMed

Search ID#	Search terms	Search options
S1	Lumbar surgery AND Early rehabilitation	Boolean/Phrase
S2	Lumbar surgery AND Enhanced rehabilitation	Boolean/Phrase
S3	Lumbar surgery AND Accelerated rehabilitation	Boolean/Phrase
S4	Lumbar surgery AND Fast-track rehabilitation	Boolean/Phrase

1. 793 references were included from PubMed

Search strategy web of science

Search ID#	Search terms	Search options
S1	Lumbar surgery AND Early rehabilitation	Boolean/Phrase
S2	Lumbar surgery AND Enhanced rehabilitation	Boolean/Phrase
S3	Lumbar surgery AND Accelerated rehabilitation	Boolean/Phrase
S4	Lumbar surgery AND Fast-track rehabilitation	Boolean/Phrase

1. 721 references were included from Web of Science

Search strategy Scopus

Search ID#	Search Terms	Search Options
S1	Lumbar surgery AND Early rehabilitation	Boolean/Phrase
S2	Lumbar surgery AND Enhanced rehabilitation	Boolean/Phrase
S3	Lumbar surgery AND Accelerated rehabilitation	Boolean/Phrase
S4	Lumbar surgery AND Fast-track rehabilitation	Boolean/Phrase

1. 335 references were included from Scopus

Search strategy ScienceDirect

Search ID#	Search terms	Search options
S1	Lumbar surgery AND Early rehabilitation	Boolean/Phrase
S2	Lumbar surgery AND Enhanced rehabilitation	Boolean/Phrase
S3	Lumbar surgery AND Accelerated rehabilitation	Boolean/Phrase
S4	Lumbar surgery AND Fast-track rehabilitation	Boolean/Phrase

1. 83 references were included from ScienceDirect

Total result literature searches: 1183 references