Authors: Judith van Grafhorst, Wouter van Furth, Carmen Vleggeert-Lankamp
Categories: Article, Decompressive surgery, Functional outcome, Lumbar spinal stenosis, Mental status, Patient satisfaction, Spondylolisthesis
Source: Brain & Spine
Mental status, characterised by anxiety and depression, significantly influences physical well-being, particularly in patients with spinal stenosis symptoms.
The prevalence of depression and anxiety in our cohort. The correlation between psychological distress and physical outcome after surgery, including postoperative recovery and satisfaction.
Questionnaires evaluating anxiety and depression (HADS), functionality (ODI), quality of life (EQ-5D), and perceived recovery (Likert-scale) were sent to a randomly selected cohort of 450 lumbar spinal stenosis patients, with or without spondylolisthesis, who underwent surgery between 2007 and 2013. Results are presented, dichotomised by HADS score (score ≥8 indicating psychologically impaired) and in a Four Quadrants Model integrating functional outcomes and perceived recovery separately for psychologically impaired and non-impaired cases.
Among the 147 included patients, 32 (22%) exhibited anxiety and/or depression (impaired cases). Satisfactory outcome (perceived recovery) was reported in 29.0% of the impaired cases and 78.3% of the non-impaired cases (p < 0.001). The mean postoperative functionality score of the impaired cases was 42.46 ± 16.24, in contrast to 18.48 ± 18.25 for the non-impaired cases (p < 0.001). In the impaired group, only 12.5% achieved both a good functional outcome (ODI ≤24) and satisfactory perceived recovery, compared with 58.4% in the non-impaired group.
Patients reporting anxiety and/or depression demonstrate an inferior long-term outcome after spinal stenosis surgery compared to non-impaired patients. This clinically relevant difference underscores the importance of addressing depression and anxiety in preoperative counselling to optimize patient satisfaction and functional outcomes.
Keywords: Lumbar spinal stenosis, Spondylolisthesis, Decompressive surgery, Mental status, Patient satisfaction, Functional outcome
Lumbar spinal stenosis (LSS) is a condition in which the spinal canal is narrowed (Andaloro, 2019), usually based on degenerative spine alterations. LSS may provoke back pain and neurogenic claudication, hamper patients' walking ability and physical functioning, and lead to a decreased quality of life (Rampersaud et al., 2014; Forsth et al., 2016). Since lumbar spinal stenosis is a condition whose prevalence increases with age (Haig and Tomkins, 2010), the treatment is increasingly relevant in an ageing population (Deyo et al., 2010). Surgical treatment is generally preferred over non-surgical interventions such as physical therapy and/or pain relief using epidural steroid injections (Weinstein et al., 2008; Amundsen et al., 2000; Malmivaara et al., 2007). After surgical treatment, approximately 1/3 of patients are dissatisfied with the outcome (Strömqvist et al., 2013; Choi et al., 2014), and 7–23% of patients are even reoperated (Rampersaud et al., 2014; Forsth et al., 2016; Ghogawala et al., 2016; Ulrich et al., 2017; Lad et al., 2014; Sigmundsson et al., 2017; Javalkar et al., 2011; Deyo et al., 2011; Jansson et al., 2003). This is in line with our based on a retrospective cohort study, we concluded that two months postoperatively, 78% of patients were satisfied with their surgery and that 13% of patients were reoperated (van et al., 2023). In literature, this raises the question of what causes dissatisfaction in patients and whether satisfaction with outcomes can be predicted at baseline to optimize patient counselling.
It is essential to manage patients' expectations. A smaller gap between patients' expectations and genuine outcomes after surgery (lower expectations-actuality discrepancy) is associated with higher satisfaction (Witiw et al., 2018). Managing patients’ expectations can be effectuated by demonstrating their peers' results after lumbar spinal surgery. It is hypothesised that depression and/or anxiety are essential in both preoperative expectations and postoperative well-being and functionality.
A Four Quadrants Model integrating patient satisfaction from the Likert scale and the physical functionality from the Oswestry Disability Index (ODI) results in 4 integrated quadrants. This tool can be used in counselling patients and give depressed and/or anxious patients an insight into their global perceived outcome by reflecting on their peers’ results.
This study aims to demonstrate the prevalence of depression and anxiety in the follow-up of lumbar stenosis patients who received decompressive surgery. Additionally, the correlation between psychological distress (‘impairment') and the combination of physical outcome and satisfaction with surgery will be evaluated using a tool that can be integrated into patient counselling before surgery.
Previously, we described the nine-years follow-up and clinical outcome of a cohort of 934 patients who underwent decompressive surgery (without fusion) for neurogenic claudication for lumbar spinal (van et al., 2023). Inclusion criteria adult patients with symptomatic lumbar spinal stenosis with or without degenerative spondylolisthesis who were operated on for this indication. Symptomatic lumbar stenosis had to be characterised by neurogenic claudication. Stenosis could be due to spondylotic degeneration, ligamentous hypertrophy, discogenic protrusion, or a combination. Exclusion criteria previous operations on the lumbar spine, spondylolisthesis due to lysis or trauma, and scoliosis defined by a Cobbs angle of 10° or more.
For this current study, questionnaires were sent to a random selection of 450 patients to get less biased and more detailed outcome information. Patient Reported Outcome Questionnaires involved the Oswestry Disability Index (ODI)(Fairbank and Pynsent, 2000; Austevoll et al., 2019; Werner et al., 2017; Tonosu et al., 2012; Parai et al., 2018), Euro-Qol-5D a health-related quality of life utility measure (EQ5-D) (Austevoll et al., 2019; Werner et al., 2017; Parai et al., 2018; EuroQol--a new facility for the, 1990; M et al., 2016), Hospital Anxiety and Depression Scale (HADS), and patient-perceived recovery. Baseline characteristics (gender, age, stenosis with or without spondylolisthesis, level of stenosis, grade of spondylolisthesis), initial satisfaction at two months (“yes” or “no”), reoperation yes or no and level of reoperation were obtained from medical files. A more extensive description of the PROMS and their used cutoff values can be found in Appendix 1.
The HADS consists of two 7-item questionnaires (Zigmond and Snaith, 1983; Bjelland et al., 2002) concerning anxiety and depression separately. Each of the 14 questions contains a four-point scale scoring from 0 to 3. The outcome scoring is individually assessed, resulting in an anxiety score and a depression score. Originally, the outcome was divided into three groups; non-impaired cases (score range 0–7), doubtful cases (score range 8–10), and impaired cases (score range 11–21)(Zigmond and Snaith, 1983; Bjelland et al., 2002). In order to dichotomise the data, clinical outcomes can be divided into ‘impaired cases and ‘non-impaired cases (Bjelland et al., 2002). An impaired case is defined as a case reporting a score of ≥8, either on the anxiety or depression scale.
To visualise the outcome of surgery, a Four Quadrant Model was presented(de et al., 2021), based on ODI (functionality) ‘success’ scores (cut-off value of ≤24) and dichotomised Likert-scale ‘perceived recovery’ scores (cut-off value of ≥6) for impaired and non-impaired cases separately (Fig. 1a and b). This resulted in four good outcome/Q1 (functional success based on the ODI as well as perceived recovery), poor outcome/Q4 (no functional success and no perceived recovery), and two intermediate outcome categories/Q 2 and 4 (Q perceived recovery but no functional success and Q3: functional success but no perceived recovery). This method is applied to and compared between the impaired and non-impaired groups.
Fig. 1a Four Quadrants Model of impaired cases. *Colour should be used in print.*Pie chart of impaired cases integrating functional outcome (ODI) with perceived recovery (Likert-scale). Good: Both functional (ODI) and perceived recovery (Likert) show a favourable outcome. Intermediate Perceived recovery, but no favourable functional outcome. Intermediate Not perceived recovery, but a favourable functional outcome. Poor: Neither measurement show favourable outcome. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Fig. 1b Four Quadrants Model of non-impaired cases*. Colour should be used in print.*Pie chart of non-impaired cases integrating functional outcome (ODI) with perceived recovery (Likert-scale). Good: Both functional (ODI) and perceived recovery (Likert) show a favourable outcome. Intermediate Perceived recovery, but no favourable functional outcome. Intermediate Not perceived recovery, but a favourable functional outcome. Poor: Neither measurement show favourable outcome. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Data were analysed using IBM SPSS 24.0. All stenosis patients, regardless of whether they had degenerative spondylolisthesis, are grouped based on their HADS scores. They are either impaired cases, doubtful impaired cases, or non-impaired cases based on the scores from the anxiety questions and/or depression questions. When dichotomised, patients are grouped into impaired and non-impaired cases. Mean scores were compared with an Analysis of variance (ANOVA), and categorical data were compared using the Chi-square test (χ^2^ test). A significance level of 0.05 was maintained when comparing the groups.
To examine the contribution of mental status to the success of the surgery, the ODI and the Likert scale will be plotted against each. First, the cohort is split into impaired and non-impaired cases by either a HADS-anxiety score of 8 or higher and/or a HADS-depression score of 8 or higher, followed by dividing the patients by the ODI cut-off value of 24 and the Likert scale cut-off value of 6. By doing so, four groups are those with a “good outcome” with a favourable outcome on both the ODI and Likert scale, those with “intermediate outcomes” with either a favourable outcome on the ODI or Likert scale, and those with a “poor outcome” without a favourable outcome on either scale.
From the 450 questionnaires sent out, 33 were returned as undeliverable. A total number of 147 patients returned the questionnaires, resulting in a response score of 35%. Demographics are reported in Table 1; only the mean age of 74 years in the responders’ group significantly differed from the mean age of 77 years in the whole group.
Dividing the cohort into the three groups for anxiety and three groups for depression results in 13 anxiety cases, 11 borderline-anxiety cases, 123 non-anxiety cases, 11 depression cases, 13 borderline-depression cases, and 123 non-depression cases, with some overlap in (borderline) anxiety and depression cases (Table 2). If the cohort is divided into two groups, 32 impaired cases (HADS subscale score ≥8) (21.8%) and 115 non-impaired cases are present.
Considering the overlap between the six groups (three based on anxiety and three based on depression), we will continue with the two groups; impaired and non-impaired. The mean age in the impaired group was 74.5 ± 9.5 and in the non-impaired group 75.5 ± 9.4 (p = 0.987; Table 3).
The mean ODI score of the impaired and non-impaired cases is 42.46 ± 16.24 and 18.48 ± 18.25, respectively (p < 0.001; Table 4). The impaired cases demonstrate ‘ODI-success’ in 12.5% of patients, and the non-impaired cases show ‘ODI-success’ in 66.4% (p < 0.001).
Comparing the ODI scores between the three HADS-anxiety and HADS-depression groups leads to the following outcomes; the anxiety cases had a higher ODI score (40.59 ± 14.26) as well as the borderline (45.90 ± 15.90) compared with the non-cases (19.95 ± 19.15; p < 0.001; Table 7a). Two anxiety cases and one borderline-anxiety case scored ‘ODI-success’, in contrast to 45 non-anxiety cases (p < 0.001). A comparable trend is demonstrated in the HADS-depression groups. The depression cases had a higher ODI score (51.09 ± 12.54) as well as the borderline cases (37.57 ± 16.68) compared with the non-cases (19.81 ± 18.76; p < 0.001; Table 7b). When the ODI is dichotomised by the cut-off value of 24 for success, none of the depression cases scores successfully compared with 2 of the borderline-depression cases and 77 of the non-cases (p < 0.001).
The mean EQ-5D score for the impaired cases is 0.546 ±, 0.288 whilst the non-impaired cases have a mean EQ-5D score of 0.791 ± 0.222 (p < 0.001; Table 4). By dichotomising the EQ-5D by a score of 0.68, 40.6% of impaired cases scored successfully versus 85.1% of non-impaired cases (p < 0.001).
When the HADS-anxiety and -depression are divided into three groups, similar trends are seen. The HADS-anxiety cases score worse on the EQ-5D (mean 0.439 ± 0.331) than the borderline anxiety cases (0.607 ± 0.255) and the non-anxiety cases (0.78 ± 0.224; p < 0.001; Table 7a). Again, the same accounts for the HADS-depression groups; The depression cases score worse (0.446 ± 0.277) than the borderline cases (0.597 ± 0.323) and the non-cases (0.779 ± 0.227, p < 0.001; Table 7b).
Perceived recovery is reported in 29.0% of the impaired cases and 78.3% of the non-impaired cases (p < 0.001). Perceived recovery in the HADS-anxiety group is reported in 15.4% of the cases, 36.4% of the borderline cases, and 76.2% of the non-cases (p < 0.001; Table 7a). Perceived recovery in the HADS-depression groups is reported in none of the cases, 30.8% of the borderline cases, and 77.2% of the non-cases (p < 0.001; Table 7b).
When integrating the functional questionnaire (ODI) and the perceived recovery (Likert scale), a good outcome was achieved in 12.5% of the impaired cases compared with 58.4% of the non-impaired cases. The intermediate outcome was observed in 15.6% of the impaired cases and 27.5% of the non-impaired cases. A significantly larger proportion of poor outcomes was seen in the impaired cases (68.8%) compared with the non-impaired cases (13.9%; Table 5a, Table 5b, Fig. 1a and b).
When integrating the quality-of-life questionnaire (EQ-5D) with perceived recovery (Likert scale), significant differences in the quadrants are found in the non-impaired group (Table 6b). However, there are no significant differences between the quadrants in the impaired group; a good outcome is found in 15.6% of patients, an intermediate outcome is found in 32.5% of patients, and a poor outcome in 43.8% of patients (Table 6a). The difference between the two Cross Tables is significant, with a p-value of < 0.001 (Fig. 2a and b).
Fig. 2a Four Quadrants Model of impaired cases. *Colour should be used in print.*Pie chart of impaired cases integrating quality of life (EQ-5D) with perceived recovery (Likert-scale). Good: Both quality of life (EQ-5D) and perceived recovery (Likert) show a favourable outcome. Intermediate Perceived recovery, but no quality of life. Intermediate Not perceived recovery, but a favourable quality of life. Poor: Neither measurement show favourable outcome. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Fig. 2b Four Quadrants Model of non-impaired cases*. Colour should be used in print.*Pie chart of non-impaired cases integrating quality of life (EQ-5D) with perceived recovery (Likert-scale). Good: Both quality of life (EQ-5D) and perceived recovery (Likert) show a favourable outcome. Intermediate Perceived recovery, but no quality of life. Intermediate Not perceived recovery, but a favourable quality of life. Poor: Neither measurement show favourable outcome. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
21 of the 147 (14%) patients had a second surgery. No significant differences in the distribution of the HADS groups were found between reoperated and non-reoperated patients (Table 8).
The data presented in this study convincingly confirm the hypothesis that patients suffering from spinal stenosis symptoms and additionally being burdened by anxiety and/or depression are likely to demonstrate less favourable postoperative recovery. The outcome squares representation of results facilitates a clear insight into the differences between groups. This insight can be used as a basis for developing a prediction model in the future which can be used as a valuable tool in patients’ counselling and surveillance.
This cohort study describes long-term outcomes after lumbar spinal stenosis surgery. The mean follow-up is nine years. In this Dutch population, 21.8% of patients reported feelings of depression and/or anxiety, which is higher than the 12.1% prevalence of anxiety and depression in the general Dutch population of 65 years or older (Breslau et al., 2021; van der et al., 2022). Do note that those percentages are obtained by different questionnaires than those used in our study, the Mental Health Index or Inventory, whilst we used the HADS. According to van der Velden et al. (van der et al., 2022), physical disease is a risk factor for anxiety and depression symptoms (aOR 1.51 (95% CI 1.17–1.94; p < 0.01). This could explain the higher prevalence in our population and makes it even more important to acknowledge this silent problem amongst lumbar spinal stenosis and spondylolisthesis patients.
Kashlan et al. (2020) analysed the Quality Outcome Database for patients undergoing Meyerding 1-grade degenerative lumbar spondylolisthesis surgery in the United States. They have found a comparable prevalence of depressive and/or anxiety disorders (25.6%) in their population. Patients were divided into depressed and/or anxious based on the patient's medical record at the time of enrolment. Kashaln et al. described that preoperative depressed patients were significantly less likely to achieve an improvement of at least 20% in ODI at three months. This was not observed in the case of an anxiety disorder. However, at two years follow-up, there was no significant difference in the odds of achieving an improvement of at least 20% in ODI in the case of a depressive or anxiety disorder. Therefore, Kashlan et al. (2020) concluded that depressed patients take longer to achieve the goal of an improvement of at least 20% in ODI but do eventually get there. Even though depressed and/or anxious patients take longer to achieve an ODI of 20% less, according to Kashlan et al., a difference in functional outcomes remained. Our study demonstrates that this conclusion cannot be extended towards ‘no difference’ between depressed and/or anxious patients after nine years of follow-up.
Multiple studies and systematic reviews have been performed to define preoperative risk factors for unsatisfactory lumbar spinal stenosis surgery outcomes. McKillop et al.'s systematic review concluded that preoperative depression is likely a prognostic factor for dissatisfying postoperative lumbar spinal stenosis symptom severity and disability (McKillop et al., 2014). These findings align with the systematic review by Aalto et al. concluding that preoperative depression is associated with worse treatment satisfaction and more severe symptoms (Aalto et al., 2006). Yamamoto et al. demonstrated that not preoperative depression but anxiety was a significant prognostic factor for patient satisfaction in their multivariate analysis (Yamamoto et al., 2022). However, none of the aforementioned studies has addressed the long-term postoperative presence of anxiety or depressive status. In our study, we demonstrated that patients who report anxiety and/or depression in the follow-up after LSS surgery are significantly more dissatisfied with the surgical outcome in comparison to the non-anxious and/or depressed patients. This is in line with one-year postoperative results reported by Falavigna et al. (2015); postoperative patients who reported being depressed exhibited more dissatisfaction. Remarkably, Falavigna et al. also demonstrated that patients who did show depression before surgery but not after surgery had similar physical outcomes to patients without preoperative depression. These data arouse the assumption that lumbar stenosis patients’ physical burden might be correlated to their psychological distress. In line, Wagner et al. (2020) reported that preoperative psychological distress improves after surgery, concluding that patients can significantly benefit physically and mentally from surgery. No analysis was performed on the correlation between the outcome of LSS surgery and psychological outcome—unfortunately, no factors to predict which depressed patients will recover and which will not were yielded.
Previous research has shown that the relationship between depression and functional outcome is most likely bidirectional; Falavigna et al. and Lebow et al. (2012) both concluded that some preoperative depressed patients returned to good mental health postoperatively, with a decline in the prevalence of depression (28.6%–17.6% (Falavigna et al., 2015)). However, with Falavigna's relatively short follow-up period of 1 year (compared with ours of 9 years), the question of “what came first, the anxiety and/or depression or the physical problems” remains. There is a possibility that patients report having fewer mental problems in the first year following their surgery because of the sudden relief of their pain and gain in functionality. Still, after that first year, their physical complaints return, as well as their psychological problems. Therefore, in developing a prediction model integrating anxiety and/or depression, long-term follow-up data seem to give more realistic outcome data.
The two-dimensional representation of the data gives a clear and straightforward insight into the outcome of surgery for depression/anxiety patients and those who are non-depressed/anxious. This is helpful in preoperative counselling of patients who want to be well-informed before undergoing LSS surgery. Eventually, we want to create a prediction model; a patient fills in a HADS questionnaire before surgery, and based on the score, it is obvious which of the four quadrants-figure applies to the patient's surgical outcome.
Our study had some limitations. Firstly, there were no preoperative PROMs. Hence, we could not track PROM changes before and after surgery. However, the postoperative PROMs that we found correspond with the PROMs found in the literature(Forsth et al., 2016; Wagner et al., 2020). Secondly, this study has a response rate of 35%. This response rate is acceptable because a somewhat older study population is considered with a relatively long follow-up. It accurately represents the population of interest (Forsth et al., 2016).
Patients who underwent spinal stenosis surgery and report being depressed and/or anxious have significantly worse clinical outcomes nine years postoperatively. Therefore, the HADS is an important prognostic factor that can be used in patient counselling.
This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
The data of this research is not available online.
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests.Dr. Vleggeert-Lankamp reports a relationship with Cervical Spine Research Society that board membership and funding grants. Dr. Vleggeert-Lankamp reports a relationship with Netherlands Neurosurgical Society (NVvN) that board membership. Dr. Vleggeert-Lankamp reports a relationship with EUROSPINE The Spine Society of Europe that board membership. Dr. Vleggeert-Lankamp reports a relationshiigsp with Advisory Board Rijndam Rehabilitation that consulting or advisory. Dr. Vleggeert-Lankamp reports a relationship with Covidien that funding grants. Dr. Vleggeert-Lankamp reports a relationship with YM Fund that funding grants. Dr. Vleggeert-Lankamp reports a relationship with Achmea Health Insurance that funding grants. Dr. Vleggeert-Lankamp reports a relationship with Nuvasive that speaking and lecture fees. Dr. Vleggeert-Lankamp reports a relationship with Stryker Spine that speaking and lecture fees. Dr. van Furth reports a relationship with Sella Therapies that Co-Founder/CEO.
Handling F Kandziora
The following is the Supplementary data to this
The data of this research is not available online.