Introduction

Fusion of the spine can be performed in many different ways at various levels (cervical, dorsal and lumbosacral) to treat pathologies like trauma, infections, neoplastic, and degenerative diseases. In recent years, there has been a rapid increase in the use of spinal fusion techniques [1]. However, despite the well-defined role of fusion procedures in reducing pain and disability in selected cases [2, 3], a certain percentage of patients develop persisting spinal pain (PSP) and disability after surgery, with consequent deterioration in quality of life. In a multicenter, prospective observational study on 3120 patients enrolled in an European registry, Fletcher et al. [4] found that 56,5% of patients undergoing spinal surgery experienced mild to moderate pain 12 months after the procedure.

In the last twenty years, research has attempted to understand the causes of PSP after spinal surgery [5, 6]. These efforts have caused a transition from a dualistic model of pain, based on the concept that the mind and the body work separately, to a biopsychosocial model, where emotions and cognitions interact with neurobiological factors to generate pain experiences [7]. Based on this model, several risk factors for an unfavorable outcome after spinal surgery have been identified, among which psychological factors [8,9,10], such as preoperative anxiety [11], depression [12], and negative cognitions about pain [13]. Of these, pain catastrophizing, that is an adverse coping mechanism that produces an exaggerated response to pain, has been identified as a key factor for developing PSP after spinal surgery [14, 15].

Psychological factors are modifiable and can be a target for intervention. This intervention can be implemented in the preoperative [16] and/or postoperative [17] phase. However, despite the presence of solid evidence regarding the effect of  psychological techniques like cognitive behavioral therapy (CBT) in patients suffering from chronic lumbar pain [18], the literature is not conclusive about the effect of CBT before and/or after spinal surgery, and specifically spinal fusion surgery. In particular, there is uncertainty regarding the type of psychological intervention, the intensity needed to obtain an effect, and the type of professional that should be administering the intervention.

The main goal of this systematic review is to determine the effect of perioperative psychological interventions on pain, disability and health-related quality of life (HRQOL) in patients undergoing spinal fusion surgery at any level of the spine, at short-term (≤ 3 months), intermediate-term (> 3 months but < 12 months) and long-term (≥ 12 months) follow-up. The secondary goal was to analyze the effect on use of pain medication (MED), patient’s global impression of change (PGIC), return to work (RTW), and patient satisfaction.

Methods

This systematic review was conducted according to the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [19]. A protocol was created in advance and uploaded to the PROSPERO online platform (CRD42020187047).

Data sources and search strategy

The following databases were systematically searched by two independent reviewers (PS and WvS), without any limitations in study design, for articles in English, Dutch or Italian published prior to March 31, 2020: Pubmed, Embase, PsycINFO, Cinahl, Cochrane database of systematic reviews and Cochrane Central register of controlled trials. Initial search terms included spinal fusion, spondylosyndes*, ankylosis, pedicle screw, bone screws, cognitive behavioral therapy, behavioral therapy, acceptance and commitment therapy, mindfulness, guided imagery, and relaxation training, in various permutations.

Eligibility criteria and study selection

Articles were considered for inclusion if they met the following criteria: (1) primary research article in a peer-reviewed journal; (2) participant sample included adult patients (18-year-old or above) with a primary diagnosis of degenerative or neoplastic spinal disease, undergoing a spinal fusion surgery; (3) a psychological intervention was administered in the perioperative period (starting between 3 months before and 3 months after surgery); (4) outcomes measuring either pain, disability and/or HRQOL were reported. Qualitative and quantitative studies investigating any kind of psychological intervention explicitly targeting patients undergoing a surgical fusion on the spine were included, independently of the design of the study.

We excluded: (1) reviews, meta-analyses, letters to the editor, expert opinions, and studies that did not describe any treatment parameters; (2) articles that included patients with a primary diagnosis of idiopathic scoliosis and (3) studies that did not include analyses of clinical efficacy.

Psychological interventions were defined as interventions that use psychological techniques to decrease psychological distress or modify emotions and/or cognitions into a desirable direction. Interventions delivered in person, by telephone, online or by mobile applications, or a combination of these methods, were included. In addition, interventions could be delivered individually or in groups, and there was no restriction on who delivered the intervention. A brief description of these techniques is provided in Table 1. Music therapy and hypnosis were excluded. We also excluded studies where psychological intervention was combined with another active intervention (e.g., physical therapy), unless the same additional intervention was also part of the control intervention. Moreover, if the intervention was compared to other arms, as for example in randomized controlled studies (RCTs), there was no restriction on the type of comparator used.

Table 1 Definition of the most common psychological treatments considered in this review
Full size table

A meta-analysis was not performed due to the limited statistical power as a result of methodological variation and a small number of studies. Outcomes of the reviewed studies are reported as they were presented in the original manuscript.

Two authors (PS and WvS) independently screened the titles and abstracts and potential eligible studies were screened on the in- and exclusion criteria in full text. References cited in eligible studies that were considered to be relevant were also retrieved and assessed in full text. In case of disagreement between authors, a third investigator (MP) resolved the disagreement through discussion.

Data extraction

Two authors (PS and WvS) independently extracted the following data from each included study: 1) Study characteristics: lead author, publication year, country of publication, publication type and study design; (2) Patient information: population description, inclusion and exclusion criteria, sample size, mean age, percentage of women, site of pain, baseline scores of outcome variables, type of condition for which participants were operated; (3) Characteristics of the psychological and control interventions: type of intervention, group or individual intervention, targeted at at risk patients or not, intensity (number and duration of sessions, time span of treatment), mode of delivery, person delivering the intervention, efficacy of intervention, other treatments received; (4) Surgical procedure: site of surgery, indication (degenerative/neoplastic), surgical technique; (4) Outcome information: which outcome assessed, assessment instrument, time points for measurement. A descriptive statistic was obtained for each primary and secondary outcome. Differences between the two authors were resolved by a third author (MP) after discussion. The full extraction scheme can be found in Online Resources 1 and 2.

Risk of bias assessment

The risk of bias was assessed by two authors (PS and WvS) independently using a modified version of Downs and Black checklist for randomized and non-randomized studies [20]. It consists of 27 items that are distributed between five sub-scales (reporting, external validity, bias, confounding and power). Each item is scored 0 or 1, except for one item in the reporting subscale, which is scored 0 to 2. The total maximum score is 28. Justification for each score is available in Online Resource 3. Downs and Black score ranges were given quality levels as previously reported [21]: excellent (26–28); good (20–25); fair (15–19); and poor (< or = 14).

Results

Figure 1 illustrates the PRISMA flowchart for the process of study retrieval and inclusion. The search identified 1132 records. After duplicates removal, 849 were screened for potential eligibility. In total 22 papers were assessed in full text for final eligibility and 13 papers were included in the qualitative synthesis. Table 2 shows a summary of the studies included in this review. All studies were conducted between 2004 and 2017. The vast majority were RCTs, (85%). Twelve studies analyzed a psychological intervention related to lumbar fusion surgery (92%), one also to cervical fusion surgery. CBT, alone or in combination with another intervention, was used in nine studies (69%)), mindfulness-based cognitive therapy in two studies (15%), and relaxation and psychoeducation each in one study (8%).

Fig. 1
figure 1

PRISMA 2020 flow diagram for new systematic reviews which included searches of databases and registers only

Full size image
Table 2 Study characteristics
Full size table

Comparators were physiotherapy in 9 studies, while in 4 studies the characteristics of the control intervention were not specified. The rehabilitation program varied between studies: in four studies, an educational session with preoperative information about the upcoming operation and the anesthetic procedure, medication, and information about the postoperative rehabilitation and physical restrictions after surgery was performed, together with a non-structured physiotherapy program; in three studies, the patients followed a structured perioperative program, with planned sessions conducted in person and/or by telephone and a manual or diary to follow along with the study therapists, while in the remaining two studies, only non-supervised physiotherapy was performed.

The most frequently investigated outcomes were pain (92%, Visual Analog Scale-(VAS), Numeric Pain Rating Scale (NRS) and other scores), disability (77%, Oswestry Disability Index-ODI and other scores) and HRQOL (61.5%, EQ-5D, SF-36 and SF-12). MED was evaluated in six studies, PGIC in two, RTW in three, and patient satisfaction in one.

A total of 830 patients (Table 3) were included in the studies, with a mean age of 59 years for both the intervention and control group. Three papers studied the effects of a psychological intervention that was administered preoperatively, three postoperatively and six both pre- and post-operatively (not specified in one study). The intensity (duration and timespan) of the interventions varied considerably, ranging from 60 to 960 min (1 to 16 h) in a timeframe between 5 days and 6 months. There was also variability in the professional that delivered the intervention, the most frequent provider being a physiotherapist with or without a psychologist supervision (4 studies) followed by health professionals’ teams (3 studies), online support groups or internet platforms (3 studies), a psychologist (2 studies), and a surgeon (1 study). In one study [22], the intervention was targeted toward patients that presented an increased risk of developing persisting pain. Individual interventions were performed in four studies, group interventions in three studies, while in six studies this information was not available.

Table 3 Patient and intervention characteristics
Full size table

Information on relevant findings and timing of surgical outcomes in the selected studies are reported in Table 4. Follow-up duration varied considerable between the studies. Twelve studies reported a short-term follow-up (≤ 3 months), seven studies an intermediate follow-up (> 3 months but < 12 months), and six studies a long-term follow-up (≥ 12 months). Only three studies evaluated patients at short, intermediate and long-term follow-up.

Table 4 Summary of relevant findings for selected studies
Full size table

Of the thirteen studies included in the review, nine (69%) showed an effect of a psychological intervention on one or more primary outcomes. In particular, six out of 12 studies demonstrated a statistically significant reduction in postoperative pain on any timepoint; this was in 5 out of 11 at short-term follow-up [23,24,25,26,27], in 2 out of 6 at intermediate follow-up [22, 23] and in 1 out of 5 [24] at long-term follow-up.

Eight out of 11 studies demonstrated a statistically significant reduction in disability scores due to psychological therapies at short-term follow-up [22,23,24,25,26, 28,29,30]; in five of them [22,23,24, 26, 30] an effect was also demonstrated at intermediate and/or long-term follow-up. Four out of seven studies [22,23,24, 30] showed a positive effect on HRQOL; in three of these, the effect was seen at long-term follow-up [23, 24, 30], and in one [22] at short term follow-up.

Overall, pain and disability were the most frequently reported outcomes. For these outcomes, a short-term follow-up was more often at 3 months, whereas a long-term was at 12 months. At 3 months, an effect of a psychological intervention was demonstrated in 3 studies out of 6 for pain, and in 5 studies out of 7 for disability. At 12 months, in 2 studies out of 6 for pain and the same for disability.

In terms of analgesic consumption, a lower consumption was observed after preoperative CBT in one out of five studies [29], although the difference between groups was only significant on postoperative day 2. One study out of two showed an effect on RTW at long-term follow-up [23], and in another one (the only one that explored this outcome) [24], PGIC was significantly better after a postoperative psychological intervention at short-term follow-up. One study [23] explored the effect of psychological treatment on patient satisfaction, and found no difference at 2–3 years follow-up.

Despite the large heterogeneity in treatments and protocols to evaluate their efficacy in the included studies, some factors seem to influence the effect of a perioperative psychological treatment. Intensity and frequency of intervention possibly moderate postoperative pain at short term-follow-up; in the five studies where an effect on short-term postoperative pain was demonstrated [23,24,25,26,27], mean frequency and duration of the psychological treatment were 6 sessions and 546 min (range 60–960), versus 5 sessions and 476 in the other 6 where no effect was shown (range 60–1080). With regards to disability, of the four studies that did not find any effect on this outcome, two were using online platforms without verification whether the patients completed the sessions or not [27, 31]. On the other hand, more intensive interventions were related to decreased disability [22, 23, 28, 30]

Other moderators may be type and timing of intervention. Seven studies out of nine where any effect was demonstrated on pain or disability made use of CBT and included postoperative sessions [22,23,24,25, 28,29,30]. Two out of four studies that did not find any effect used interventions other than CBT (namely relaxation therapy [32] and an internet support group [31], one [33] mainly had preoperative sessions, and in another one [34] timing of the intervention was not specified.

For all other variables (intervention provider, modality of the intervention and tailoring of the intervention toward patients at risk) there was large heterogeneity and not enough data to draw any conclusion.

Risk of bias assessment

The vast majority of the included studies had a low risk of bias. Eleven out of the 13 studies had a score > 20 on the modified Downs and Black checklist, meaning a good or excellent quality level (Table 2).

Discussion

This systematic review is the first that summarizes the effects of psychological interventions on pain, disability, and quality of life in adult patients undergoing surgical fusion of the spine. Compared to recent reviews [35, 36], we focused on interventions that had a clearly defined psychological component, and excluded studies that analyzed patients undergoing less painful surgical procedures, like discectomy or lumbar decompression. Our work suggests that a psychological intervention in the perioperative period leads to a postoperative reduction of disability in the short-term follow-up compared to usual care, including rehabilitation with or without an educational component, and may lead to a postoperative reduction of pain. There is less evidence that these interventions have an additional effect at intermediate and long-term follow-up. Health-related quality of life was analyzed in seven papers; the majority of these showed some effect of a psychological intervention, and interestingly in three out of four papers this effect was seen at long-term follow-up.

Only few studies examined the effect of psychological intervention on MED and RTW. One study observed a lower consumption of analgesics after pre-operative CBT at postoperative day 2 [29]. However, studies assessing longer term effects found no effect on analgesic use [23, 26, 27]. Regarding RTW, one study [23] reported that significantly more patients were employed 2 to 3 years after surgery and fewer were on long-term sickness leave after postoperative CBT with relaxation, but another study found no effect [28].

PGIC and patient satisfaction were only studied in one study each [23, 24], and therefore no conclusions can be drawn for these aspects.

PSP after spinal fusion procedures pose significant problems in terms of quality of life, and in recent years many efforts have been made to identify patients at risk. The exact mechanism that leads to the development of chronic pain remains uncertain: however, psychosocial factors like catastrophizing, anxiety and depression have been shown to modulate postoperative pain in these patients [37], and therefore a perioperative psychological intervention targeting these factors could be effective in reducing the incidence of PSP.

Although various researches have shown a positive effect of psychological treatments like CBT on low back pain in general [38], few have specifically investigated the effect of these interventions applied around the time of spinal surgery. In a recent meta-analysis, Janssen et al. [35] explored the effectiveness of prehabilitation programs in adult patients suffering from degenerative diseases, and scheduled for spine surgery. After meta-analysis, they found no additional effect of prehabilitation on any outcome when compared to usual care. Analyzing the type of interventions, authors stated that most of them included CBT. However, seven studies out of the 15 included (46%) did not include a psychological component targeting cognitive aspects, and where mainly based on education and counseling. Moreover, there was considerable heterogeneity regarding the surgery type, as 6 studies (40%) included patients submitted to simple decompression procedures and microdiscectomy, other than lumbar fusion. Previous research [39] has suggested that psychosocial factors, and pain catastrophizing in particular, are dynamic constructs, subject to change, in association with the improvement of pain and disability. As such, it is possible that patients treated for lumbar spinal stenosis or disk herniations encounter a different evolution of postoperative pain, and therefore of pain catastrophizing. This different postoperative pain pattern makes comparison of the results between different (preoperative) surgical conditions more difficult.

In another recent meta-analysis, Parrish et al. [36] summarized the evidence using CBT pre- and/or post-operatively in patients undergoing different types of lumbar spinal surgery, including disk surgery, decompression and fusion. The authors collected data on different patient-reported outcomes (PROs) 2–3 months postoperatively and at final assessment of each study. Out of 11 studies included in the qualitative analysis, six favored CBT. Comparators were qualified as “usual care” only in 9 studies, “exercise group” in one study, and “education” in another one. Notably, when comparing CBT to these interventions, the largest effect size was shown shortly after surgery for disability and back pain, while at long-term follow-up differences for psychological health outcomes and HRQL were more significant.

This finding is similar to what was found in this review, in which the effect of a perioperative psychological treatment is more evident on pain and disability in the short-term, and on HRQL in the long-term. Many variables may influence quality of life after spinal surgery [40]. Among them, causal linkages have been identified between different types of patient outcome measurements [41]. Psychological symptoms as well as functional status seem to have a strong influence on HRQOL at long-term. As psychological interventions in combination with perioperative rehabilitation have been demonstrated to significantly increase level of physical activity one year after spinal fusion [33, 42], it is likely that these interventions may have a positive influence on quality of life.

Moreover, meta-analysis of all included RCT outcomes in the review from Parrish et al. suggests that the effect of CBT may be sustained over time, showing effects for the majority of the outcomes at long-term follow-up. This finding could be only partially confirmed in the present review, given that for disability three [23, 24, 26] out of five studies found that the effect persisted at 12 months. Another difference with the present study concerns the design, as Parrish et al. focused only on RCTs with one type of psychological intervention (CBT) before lumbar spinal surgery, while our search included various kinds of psychological interventions in patients submitted to more invasive surgical procedures, represented by fusions. These procedures are associated with a higher rate of complications compared to non-instrumented surgery [43], that may increase the postoperative healthcare cost [44] even if the incidence of PSP was found to be similar in a recent study [45].

Previous systematic reviews and metanalyses aimed at assessing the effects of psychological interventions in the context of cardiac, abdominal or orthopedic surgery. Most of these reviews found psychological interventions to be effective [46,47,48,49], while others showed mixed results with no effect on postoperative pain [50,51,52]. A recent meta-analysis [53] explored the utility of a perioperative psychological intervention in different types of surgery including orthopedic surgery, general surgery, cardiac surgery, urologic surgery and spinal surgery. Authors analyzed data from 1880 patients included in 21 RCTs, the majority of them (57%) using CBT, and found significant effects of a psychological treatment on both acute and persistent pain and disability, (> 3 months). Effect sizes were small to moderate and there was substantial heterogeneity in the effects.

This important heterogeneity has also been found in our review, despite the fact that we restricted the analysis to patients submitted to spinal fusion surgery. In particular, significant variations were detected in timing and intensity of intervention, as well as in the professional who delivered the intervention (physiotherapist, psychologist or other team members). In 3 out of 13 studies, the intervention was delivered online through a platform, or an internet support group. No firm conclusions on the influence of these factors can be drawn, but our qualitative analysis did show a tendency that studies in which postoperative efficacy was detected had used psychological interventions with a higher intensity and frequency of sessions [22,23,24, 54].

Moreover, the present review suggests that psychological interventions obtain a long-term effect on pain and disability when they are implemented in the postoperative period, and when they are combined with an exercise therapy protocol. Abbott et al. [23] in a randomized trial on 107 patients scheduled for lumbar fusion were able to show a significant improvement in various PROs including disability at 12 months using a psychomotor therapy protocol in the postoperative period, with 3 sessions over 12 weeks. This protocol included both a physical component as well as CBT. Similarly, Monticone et al. [24] showed a greater improvement in disability using a postoperative protocol that included CBT and exercise therapy over 4 weeks, in comparison to standard care.

Limitations

In this study, we were limited by many factors that precluded a meta-analysis, mainly related to heterogeneity in the psychological protocols.

Moreover, as already underlined by Parrish et al. [36], postoperative PROs like pain, disability and quality of life are influenced by many perioperative factors. Important information regarding surgical technique, perioperative complications and comorbidities were lacking in most of the studies that were included and represent a further confounding effect that could not be assessed.

Conclusion

Patients undergoing spinal fusion surgery are at risk of developing persisting spinal pain. Since psychological interventions have been shown to reduce post-surgical pain after various types of surgery, these patients could benefit from such an approach. Despite numerous limitations related to the heterogeneity of protocols, the present review suggests that among psychological treatments, CBT and mindfulness-based cognitive therapy are likely to have a positive effect on pain and disability in the short term after surgery, while few data were available to demonstrate an effect in the long-term. Furthermore, to achieve this effect, the data suggest that the protocol should be performed with high intensity and frequency, preferably in the postoperative phase. However, more research is necessary to furtherly elucidate what specific aspects of CBT make it most effective in this patient population.