Manipulative and Manual Therapies in the Management of Patients with Prior Lumbar Surgery: A Systematic Review

The purpose was to identify, summarize, and rate scholarly literature that describes manipulative and manual therapy following lumbar surgery. The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and was registered with PROSPERO. PubMed, Cochrane Database of Systematic Reviews, ICL, CINAHL, and PEDro were searched through July 2019. Articles were screened independently by at least two reviewers for inclusion. Articles included described the practice, utilization, and/or clinical decision making to post surgical intervention with manipulative and/or manual therapies. Data extraction consisted of principal findings, pain and function/disability, patient satisfaction, opioid/medication consumption, and adverse events. Scottish Intercollegiate Guidelines Network critical appraisal checklists were utilized to assess study quality. recommend after most surgical interventions.

Low back pain is the leading cause of disability worldwide, impacting roughly 540 million people at any given time [1]. Lumbar surgical procedures have become increasingly more widespread over the past several decades. Surgical treatment for lumbar degenerative disc disease increased 2.4-fold from 2000 to 2009 [2], and there were 1,288,496 new posterior lumbar fusion operations reported in the United States alone between 1998 and 2008 [3].  [4,5].
The most frequent condition considered appropriate for lumbar surgery is low back pain and radiculopathy secondary to lumbar disc herniation [6], with discectomy being the most commonly performed lumbar surgical procedure [7]. Recurrence of spinal or radicular symptoms is common following surgical intervention [8][9][10]. Following lumbar discectomy for symptomatic lumbar disc herniation the 1-year and 3-year recurrence rate for leg symptoms has been estimated to be 20% and 45% and for recurrent low back pain 29% and 65% respectively [11].
Postoperative pain and potential for future operative procedure is common for those undergoing lumbar surgical procedure. Patients that undergo lumbar discectomy procedure are 2.97 times more likely to require a future lumbar fusion than individuals without prior discectomy [12]. Failed back surgery syndrome (FBSS) is a regular indicator for spinal cord stimulator implant/neuromodulation [13], though may only provide pain relief for a portion of individuals undergoing this intervention.
Turner et al. reported only 50-60% of failed back surgery patients with implanted neuromodulation reported 50% pain improvement and 40-50% continue to experience pain [13].
Many individuals with chronic pain complaints seek manual and manipulative therapy (MMT) for nonpharmacological pain management from chiropractors, osteopaths, physical therapists and massage therapists [4,6,7,11,12]. Manual therapy is the application of the practitioner's hands directly to soft tissues or joints using techniques such as mobilization, stretching, myofascial release, massage, and muscle energy techniques [14]. Manipulation is a type of manual therapy that involves the practitioner applying a high-velocity, low-amplitude manual force to a perceived hypo-mobile joint to approximate the joint near its end range of motion and to restore its physiological joint ROM [15], or alternatively through a table-assisted approach such as flexion-distraction (FD). MMT may be a potential treatment option to aid in pain reduction and functional preservation in those with a prior history of lumbar surgical intervention.
The authors are unaware of any prior systematic reviews analyzing the literature of MMT for individuals with a history of lumbar surgery. The primary aim of this study was to investigate the current relationship of MMT to the management of pain, function, patient satisfaction, and opioid/medication utilization for patients with prior lumbar operative procedures. A secondary purpose of this study was to assess the adverse events reported in the same body of literature.   Table 2 Search strategy example ((((postsurgical OR postoperative OR post-surgical OR post-operative) AND (spine OR low back OR lumbar OR lumbosacral OR "back pain" OR radiculopathy OR radicular pain OR sciatica OR disc herniation OR disk herniation OR intervertebral disc OR intervertebral disk OR spinal OR degenerative disc disease OR degenerative disk disease OR disc degeneration OR disk degeneration OR scoliosis OR spinal stenosis OR spondylolisthesis OR spondylosis OR spondylolysis OR failed back syndrome OR adjacent segment disease OR joint failure)) OR (((fusion OR decompression) OR (laser AND (surgery or surgeries))) AND (spine OR low back OR lumbar OR lumbosacral OR "back pain" OR radiculopathy OR radicular pain OR sciatica OR disc herniation OR disk herniation OR intervertebral disc OR intervertebral disk OR spinal OR degenerative disc disease OR degenerative disk disease OR disc degeneration OR disk degeneration OR scoliosis OR spinal stenosis OR spondylolisthesis OR spondylosis OR spondylolysis OR failed back syndrome OR adjacent segment disease OR joint failure))) OR (failed back surgery syndrome OR lumbar spine surgery OR microdiscectomy OR microdiskectomy OR diskectomy OR discectomy OR laminectomy OR laminotomy OR disc replacement OR disk replacement OR vertebroplasty OR kyphoplasty OR foraminotomy OR interlaminar implant OR "spinal cord stimulator" OR intrathecal drug delivery OR "extreme lateral interbody fusion")) AND (((spinal manipulation OR chiropractic OR musculoskeletal manipulations OR osteopathic manipulation OR orthopedic manipulation OR manual therapy OR manual therapies) OR (manipulative AND (therapy OR therapies OR rehabilitation)) OR ("joint manipulation" OR "joint mobilization" OR "mobilization therapy" OR "spinal mobilization" OR "soft tissue mobilization" OR flexion distraction OR myofascial OR "active release" OR Graston OR massage OR stretching techniques OR muscle stretching OR static stretching OR passive stretching OR proprioceptive neuromuscular facilitation OR "PNF stretching" OR "post isometric relaxation" OR contract-relax) OR (instrument assisted AND (soft tissue OR manipulation OR adjusting OR adjustment)) OR "manipulation under anesthesia")) NOT ("Animals"[Mesh] NOT ("Animals"[Mesh] AND "Humans"[Mesh])) AND English [la] Completed studies accepted for publication, but not yet in-print, were identified by searching clinicaltrials.gov and the World Health Organization International Clinical Trials Registry. The literature was searched with the assistance of a health sciences librarian (SW), and titles were screened independently by two different reviewers (CJD, ZAC). The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and was registered with PROSPERO (#CRD42020137314).

Eligibility criteria
The inclusion and exclusion criteria are available in Table 3. All research designs published by peerreviewed scholarly journals in English were included in the search. Commentaries from non-peer reviewed sources (e.g., trade magazines) and other non-scholarly sources were excluded, as were writings not specific to post-surgical care provided. Case reports and case series were included to inform decision-making when no other higher level of evidence was available [16,17]. Exclusions included animal studies as done by the method from the Cochrane Handbook for Systematic Reviews of Interventions [18]. Abstracts of conference proceedings were not included due to the high rate of conference presentations that never reach full publication [19]. Articles were considered for final inclusion if they describe the practice, utilization, and/or clinical decision making related to postsurgical intervention with MMT.

Methods of review Study selection
The screening process was conducted independently by two authors, and coauthors were asked to contribute citations with which they were familiar but which might be missing from the formal search.
Citations were screened by two reviewers by reading the title and abstract for each article. Abstracts of the citations that obviously or possibly met the review criteria were saved. The full papers of each abstract were retrieved and each article was reviewed independently by at least 2 authors to verify that it met the inclusion criteria. Disagreements on eligibility were resolved by discussion and adjudicated by a third author when necessary. Articles that did not meet the criteria were discarded and a note was made as to why they were excluded. Once an article was included, the citation, study design, principal findings, surgical intervention, manual/manipulative therapy, and adverse events were extracted.

Data extraction
Two authors completed data extraction for each of the included studies. One author served as the primary extractor and the second served as a secondary extractor confirming the findings. Any disagreements were resolved through discussions and if necessary, a third reviewer. Data were extracted into Microsoft Word tables grouped by type of study design. Items collected on the data extraction tables included: citation with first author and publication year, surgical history, MMT intervention, principle findings comparison, adverse events, and medication discussion. For randomized controlled trial (RCT) and cohort designs, we separated principle findings into comparison, outcome measures, results, and conclusions. For studies that involved multiple surgical types within an individual patient, we classified the surgical type from least-to-most aggressive or advanced approach in the order of discectomy, laminectomy, fusion, artificial disc replacement, and spinal cord stimulator, respectively. Studies that incorporated multiple surgical types without stratifying results by type were classified as undifferentiated.

Evaluation of risk of bias
Scottish Intercollegiate Guideline Network (SIGN) critical appraisal checklists [20] were utilized to assess for risk of bias (quality). All RCTs and systematic reviews (SR) included in this study were assessed with the corresponding checklist provided by SIGN, with at least 2 authors performing each quality assessment. Disagreements were resolved with discussion and a third reviewer was incorporated as appropriate. The SIGN checklist rates each article as "high quality, low risk of bias", "acceptable quality, moderate risk of bias", and "low quality, high risk of bias".
For the SR checklist, there are 12 items to score and quality is rated as: high, low risk of bias > 9, acceptable, moderate risk of bias 6-9, and low, high risk of bias < 6 ( Table 4). For the RCT checklist, there are 10 items to score and quality is rated as: high, low risk of bias > 8, acceptable, moderate risk of bias 5-8, and low, high risk of bias < 5 ( Table 5). Table 4 Modified SIGN systematic review checklist [21] 1. 1 The research question is clearly defined and the inclusion/exclusion criteria must be listed in the paper. ( At least two people should have extracted data. 1.5 The status of publication was not used as an inclusion criterion. 1. 6 The excluded studies are listed. 1.7 The relevant characteristics of the included studies are provided. 1.8 The scientific quality of included studies was assessed and reported. 1.9 Was the scientific quality of the included studies used appropriately? 1. 10 Appropriate methods are used to combine the individual study findings.

1.11
The likelihood of publication bias was assessed appropriately.

1.12
Conflicts of interest are declared Table 5 Modified SIGN randomized trial checklist [21] 1. 1 The study addresses an appropriate and clearly focused question.

1.2
The assignment of subjects to treatment groups is randomized 1. 3 An adequate concealment method is used. 1.4 The design keeps subjects and investigators 'blind' about treatment allocation 1. 5 The treatment and control groups are similar at the start of the trial 1. 6 The only difference between groups is the treatment under investigation 1.7 All relevant outcomes are measured in a standard, valid and reliable way 1. 8 What percentage of the individuals or clusters recruited into each treatment arm of the study dropped out before the study was completed? 1.9 All the subjects are analyzed in the groups into which they were randomly allocated (often referred to as intention to treat analysis) 1. 10 Where the study is carried out at more than one site, results are comparable for all sites.

Strength of evidence
The strength of evidence for recommendations was based upon the quality and quantity of evidence available and as has been demonstrated elsewhere [21] and modified from the UK evidence report [22,23]. The criteria are outlined in Table 6 and describe high, moderate, and inconclusive strength of evidence, and favorable or unfavorable recommendation. Table 6 Rating of evidence from randomized controlled trials and systematic reviews [21][22][23] Quality and quantity of evidence Rating Consistent results found in at least 2 low risk-of-bias studies High Results of at least 1 low risk-of-bias study or at least 2 low risk-of-bias studies with some inconsistency of results

Moderate
Only acceptable-quality studies with inconsistent results, or only high-risk of bias studies Inconclusive  Table 7. The majority of included studies were case reports or case series (n = 37), followed by RCTs (n = 6), SRs (n = 3), scoping review (n = 1), narrative review (n = 1), and commentaries (n = 2). The most common reason for exclusion was due to care not involving MMT (n = 30).

Systematic reviews
Three SRs met the inclusion criteria. One of the 3 was high quality [25], 1 was acceptable quality [26], and 1 was low quality [27] (Table 8). Two of the 3 included SRs described physical therapy and rehabilitation intervention in patients with undifferentiated lumbar surgery for degenerative conditions [25,26], and 1 described care following lumbar fusion surgery [27]. Two of the 3 describe physical therapy (PT) and rehabilitation including, but not specific to, MMT [26,27], and 1 specifically described neural mobilization techniques [25].  The high-quality and acceptable-quality reviews addressed rehabilitation after a variety of lumbar surgical types (e.g. discectomy, laminectomy, fusion). The high-quality SR investigated neural mobilization and included 69 studies, of which only 1 study that was postoperative low back pain [28], and concluded that inpatient neural mobilization in the 3 days following lumbar operation did not add benefit to usual care [25]. The acceptable-quality review analyzed inpatient PT including 4 studies, of which 1 was relevant to MMT [28] and it was the same study identified by the neural mobilization SR [26].
Following lumbar fusion, a low-quality review found insufficient evidence to make an argument for or against the inclusion of joint mobilization, nerve mobilization, or soft-tissue mobilization for lumbar fusion postoperative rehabilitation [27]. Despite insufficient evidence, among other treatments, the study authors recommended joint mobilization of the thoracic spine and hips to maintain posture and increase functional mobility, early neural mobilization to improve ROM by decreasing nerve tension, and soft-tissue mobilization to decrease post-surgical pain and swelling around the incision site.
Randomized controlled trials Table 9 provides the RCTs risk of bias as high, acceptable, and low-quality studies and Table 10 presents the evidence. Of the 6 RCTs, 3 were pilots and were underpowered to make any conclusions regarding efficacy and were not rated for quality. Of the 3 remaining studies, 2 were rated highquality [28,29] and 1 was rated acceptable-quality [30].   Following lumbar surgery (undifferentiated), one RCT [29] compared a control group of selfmanagement to 2 PT groups, a "spinal stabilization exercise group" and a "mixed-physical therapy group" including Maitland, manual therapy, spinal mobilization, and soft-tissue mobilization among other PT techniques. They found no between-group differences as measured by the numerical pain rating scale or Roland-Morris Disability Questionnaire. The study did not control for specific interventions utilized by physical therapists in treatment. This RCT was rated high quality (low-risk-of bias) by the SIGN checklist.
A second high-quality RCT was described above and investigated inpatient use of neural mobilization following undifferentiated lumbar surgery [28]. Their study found no between-group differences for global perceived effect, the visual analog scale for pain, McGill Pain Questionnaire, Quebec Disability Scale, or return-to-work.
The last RCT studied outcomes after L5 laminectomy in a 5-arm trial comparing control (no treatment) to postoperative physical agents, joint mobilization, low-tech exercise, and high-tech exercise [30].
This study was graded acceptable-quality as it did not adequately address group assignment randomization, blinding of the investigators or patients, and handling of missing data (intention-totreat). In their study, active approaches were the most effective for the improvement of functional measures of chronic low back pain, with low-tech exercise having the longest interval of chronic low back pain relief. Joint mobilization increased lumbar extension ROM but did not impact objective outcomes for spinal function.

Literature reviews, case reports, and commentaries
There was a large body of lower-level studies that were not assessed for quality. This included 1 scoping review [31], 1 narrative review [32], 14 case series [33][34][35][36][37][38][39][40][41][42][43][44][45][46], 23 case reports , and 2 commentaries [70,71]. Ten of the case reports described 53 cases following discectomy, 16 reports described MMT in 143 cases post-laminectomy, 16 reports described MMT care for 67 cases after fusion, 1 report discussed post-surgical treatment in 8 cases after artificial disc replacement, and 1 report discussed care in 3 cases following implantation of spinal cord stimulators. There were multiple instances where a single report described FBSS cases from following more than 1 surgical intervention. The narrative review discusses lumbar fusion with relevance to chiropractors and the scoping review analyzed rehabilitation protocols directed at the lumbar spine in the perioperative periods. The findings from these studies and case reports are presented in Tables 11-13.

Strength Of Evidence
The strength of evidence is rated and grouped by prior surgical type and criteria are described in Table 6.

Discectomy
Evidence was inconclusive because of a scarcity of studies and is insufficient to recommend or discourage application of MMT in treatment plans following lumbar discectomy.

Laminectomy
Evidence was inconclusive regarding spinal mobilization (Grade III or IV Maitland) following L5 laminectomy but is favorable for improving lumbar extension ROM without improving pain and function outcome measures. Evidence is insufficient to recommend or discourage application of MMT in treatment plans following lumbar laminectomy.

Fusion
Evidence was inconclusive because of a scarcity of studies and is insufficient to recommend or discourage application of MMT in treatment plans following lumbar fusion.

Disc replacement
Evidence was inconclusive because of a scarcity of studies and is insufficient to recommend or discourage application of MMT in treatment plans following lumbar total disc replacement.

Spinal cord stimulator
Evidence was inconclusive because of a scarcity of studies and is insufficient to recommend or discourage application of MMT in treatment plans following spinal cord stimulator implantation.

Discectomy
We found no trials that specifically investigated MMT following discectomy. There were 3 pilot studies published by Kim et al. investigated OMT versus active control following microdiscectomy [72][73][74].
Two of these studies were of the same patient data (short-and long-term follow-up), and all 3 described greater improvements in pain and disability following OMT. There were 10 case reports and series describing the care of 54 patients following discectomy. All but one intraoperative report [34] involved care provided by chiropractors. Favorable responses were reported with spinal manipulation [33,37,51,59], FD manipulation [38,41,51,52], manual therapy [52,56], and manipulation under anesthesia of the sciatic nerve [34] or spinal joints [36]. Following discectomy, a scoping review suggested early passive and active hip and knee flexion exercises to reduce time to independent mobility and return-to-work [31].

Fusion
We found no trials that specifically investigated MMT following discectomy. We identified 16 case reports or series describing MMT for 67 patients with history of lumbar fusion. Three of these reports were from the medical profession [34,42,67], 1 from massage [68] and the rest were chiropractic specific. Favorable response to care was noted following spinal manipulation [42,44,47,62,69], FD manipulation [38,41,[53][54][55]58], massage [67,68], neural mobilization both post- [61] and intraoperative [34], and spinal manipulation under anesthesia [43]. A literature review outlined types of lumbar fusion operation, common adverse events, and described chiropractic fusion related literature while calling for clinical trials to assess the safety and efficacy of care [32].

Disc Replacement
No trials and only 1 case described MMT following lumbar total disc replacement [39]. O'Shaughnessy et al. described management of 8 cases with spinal manipulation. As a safety measure, the authors incorporated flexion-extension radiographs to ensure intersegmental stability and patients were positioned in a preloaded manipulative setup to determine tolerance. Disability and fear-avoidance was improved in 75% (6/8) and 63% (5/8) of cases respectively.

Spinal cord stimulator
No trials and only 1 case report described MMT following spinal cord stimulator [40]. This report outlined chiropractic management of 3 cases through a combination of spinal manipulation, FD, and myofascial release. One of the patients could not tolerate positioning for spinal manipulation and as a result, was not performed. Two of the 3 cases reported favorable outcomes and one had no benefit from care.
Postsurgical undifferentiated (lumbar discectomy, laminectomy, or fusion) Two of the 3 randomized controlled trials enrolled patients following a variety of different lumbar surgical procedures (discectomy, laminectomy, and fusion) and did not breakdown their results by surgical type [28,29]. The studies were both early postoperative, and neither study found significant improvement by incorporating MMT. The study by Mannion et al. did not specifically require MMT as part of the intervention group [29]. In a scoping review of lumbar surgery perioperative rehabilitation, Marchand et al. found that passive and active hip and knee flexion exercises reduced time to independent mobility and return-to-work. Commentaries by Walker [70] and Shapiro [71] discussed complications related to, and the role of manipulation for, individuals with FBSS.

Adverse events
None of the clinical trials reported patient dropout in any treatment groups including MMT. Each of the pilot trials reported patients lost to outcome, but no side effects or complications were reported [72][73][74]. None of the case reports or series reported any serious adverse events such as loss of bowel or bladder function, stroke, fracture or hospitalization [75]. The case series describing intraoperative neural mobilization reported 61% (19/31) patients noted increased pain post MMT and 29% (9/31) required additional exploratory surgery [34]. Mild lumbar soreness was reported by several case reports for various MMTs and surgical types [39,42,51]; however, mild soreness is commonly reported following manual therapy in patients without history of surgery [76][77][78]. One study reported increased lower extremity pain in 2 of 8 patients being treated with spinal manipulation following lumbar total disc replacement [39].

Medications
None of the adequately powered trials used pharmacologic prescription or utilization as an outcome, thus no conclusions or recommendations can be determined regarding the ability of MMT to reduce or impact patient usage of medication. One pilot trial assessed anti-inflammatory, analgesic, and muscle relaxant medication usage as a secondary outcome and found that patients assigned to osteopathic manipulative rehabilitation after microdiscectomy used less medication than the control group [72]. A few case reports similarly described patient medication reduction or elimination through the utilization of MMT [47,48,56,67], but most did not comment on any change in medication. A recent systematic review and meta-analysis revealed an inverse association between chiropractic care and opioid receipt in veterans with spinal pain [79] and multiple cohort studies of health insurance claims data displayed a significantly lower likelihood of filling opioid prescriptions for recipients of chiropractic care than nonrecipients [80,81]. Although promising, it is not clear if this relationship persists in the post-surgical population.

Limitations
This review is limited by the evidence that is available and underscores the knowledge gap and the need for high-quality trials to allow for recommendations for or against MMT following a variety of lumbar surgeries. Although numerous case reports describe favorable outcomes with MMT, the limited number of RCTs and the absence of cohort studies with comparison make it impractical to make recommendations for most MMT. There is insufficient evidence to make any recommendations following most surgical procedures. Two of the 3 sufficiently powered RCTs included multimodal care, were heterogeneous of design, and all 3 were perioperative, making the findings impossible to pool and challenging to generalize to outpatient settings with patients presenting months to years postprocedure. Further, none of these trials specifically investigated or included spinal manipulation as an intervention. The current literature to guide clinicians relies heavily on case reports, with which there is a strong prospect of positive publication bias and likely under-reporting of adverse events [82].
The increased utilization of surgical intervention to address lumbar degenerative conditions and high rate of spine pain recurrence necessitate the need for studying MMT as a non-pharmacological treatment option post-operatively. Further study is needed which emphasizes pragmatic application of MMT within study designs. RCTs and longitudinal cohorts with comparison or control groups could shed light on the relative safety or dosage of MMT that is appropriate to reach maximum therapeutic benefit. There is a need to assess the impact of MMT on prescription medication utilization. Lastly, there is a need for studies that stratify the response to MMT by surgical type. There may be betweengroup differences for treatments depending on surgical-type history. Although low-level studies suggest favorable outcomes associated with MMT in the postsurgical patient, no conclusions can be drawn from the evidence related to timing, dosage, tolerance, or safety of MMT after lumbar surgery.

Conclusions
The findings of this review will help to inform practitioners of MMT (chiropractors, physical therapists,

Declarations
Ethics approval and consent to participate Since our study is a systematic review, an ethical review is not required.

Consent for publish
Not applicable Availability of data and materials Not applicable. The data used for analysis was retrieved from published studies listed in our manuscript.