Introduction

Percutaneous transforaminal endoscopic discectomy (PTED) is a remarkably emerging technology for the alternative treatment of lumbar disc herniation (LDH) [1]. Meanwhile, PTED is considered as an alternative to open discectomy with or without fusion surgery, owing to minimal skin incision, less muscular injury, maximal preservation of bone structure, minimal blood loss, shorter operative time, and early return to normal daily life. Recently, several investigations have emphasized mostly on the indications and contradictions for PTED [2, 3], on new design available for clinical utilization [4, 5], on the comparison with other techniques [6], and on complications.

Anxiety is a psychological disturbance that is usually associated with chronic disorders, particularly chronic pain, which a variety of patients with LDH have suffered from. Currently, several studies have revealed that the overall prevalence of anxiety in patients with musculoskeletal diseases is high, ranging from 30 to 60% [7]. Additionally, several investigations highlighted the importance of the mental state which is a vital signal of the health condition for patients with lumbar disc degenerative disease [8]. D’Angelo et al. reported that the anxiety status was closely related to functional disability and pain intensity [9]. Moreover, other researchers also found a higher prevalence of anxiety in the low back pain patients [10,11,12]. Apart from other pre-operative factors, increasing evidence indicates that the difference documented in the results after spinal operation might be partially attributed to the pre-operative psychological status. It is also revealed in the patients who underwent cervical surgical treatments. Furthermore, PTED is an unusually challenging procedure because of the requirement of relatively long learning curve and rich clinical experience in open discectomy. To date, the assessment of pre-operative anxiety, the post-operative improvements in anxiety, and the importance of pre-operative anxiety on the prognosis involving the PTED treatment have not been thoroughly investigated. Undeniably, observational studies usually show relevance between baseline distributions across groups. Thus, propensity score matching is an effective approach to balance the differences in the confounding factors across groups, and may still not affect the results for a specific event [13].

The present study attempts to assess the changes in anxiety after PTED, to determine the related risk factors affecting the patient’s psychological state, and to compare the effect of anxiety on the outcome of PTED after propensity score matching, hence emphasizing the importance of reminding spine surgeons to take into account the patients’ psychological condition prior to the operation procedure.

Materials and methods

Patient population

After obtaining approval from local institutional review board and written informed consent from all patients, 145 consecutive patients with single-level lumbar disc herniation (LDH) undergoing PTED in our department between May 2016 and June 2017 were included in this study. All the patients were presented with characteristic symptoms and signs of LDH, aged between 18 and 60 years, and were refractory to conservative therapy for more than three  months.

The following patients were excluded: (1) those for which the reason of the symptoms was trauma, infection, or neoplasm; (2) those who suffered from lumbar canal stenosis, lumbar instability, bilateral symptoms, or could not undergo PTED; (3) those whose operative levels were not L4/5, L5/S1, or multilevel; (4) those who had psychiatric problems (such as schizophrenia and severe depression) or cognitive impairment; (5) those who had a previous surgical history; or (6) were unable to complete the follow-up or received specific treatment for anxiety on a regular basis. All patients had received the radiograms, computed tomography, and magnetic resonance images of the lumbar spine.

The included patients were stratified strictly into 2 groups according to the scores of Zung Self-Rating Anxiety Scale (SAS) [14]: a control cohort whose score was lower than 50, and the anxiety cohort including the other subjects.

Clinical assessment

The peri-operative pain intensity, disability severity, and quality of life were evaluated via the outpatient questionnaires or a phone interview scheduled by an independent investigator not related to this study. The visual analogue scale (VAS), Japanese Orthopedic Association (JOA), the 36-item Short-Form (SF-36) health survey [15], Oswestry Disability Index (ODI), and SAS were listed in the questionnaires.

The SAS scores were utilized to evaluate the degrees of anxiety, which include 20 items with favourable reliability and validity, assessing the basic features of anxiety. SAS scores ranging from 50 to 59 suggest “slight anxiety,” 60 to 69 suggest “moderate anxiety,” and more than 70 suggest “severe anxiety.”

Sample size analysis

A sample size analysis was performed for the present study. VAS was utilized as the outcome variable. In view of 5% marginal error and 80% power, a power analysis suggested that a minimum size of 18 subjects was required for each group to investigate the difference in VAS between both groups. The analysis was calculated utilizing the G*Power software (version 3.1.9.2).

Statistical analysis

The analysis making use of propensity score matching is generally performed in retrospective cohort studies to adjust for known confounding biases. The procedure for this analysis was calculated by SPSS version 22.0 (IBM Statistics, New York, USA). Anxiety and control cases were matched based on propensity score with the caliper lower than 0.02, and anxiety and control patients were then matched using a 1:1 nearest neighbor matching method. Multivariate analysis was applied to verify covariates that may influence the outcome. Standardized differences were utilized to compare baseline characteristics between the two groups. Significant differences were found between those two groups relating to age (P = 0.003), gender (P = 0.039), and segment distribution (P = 0.028). The post-matching balance of related baseline covariates was re-assessed (Table 1). Before and after matching, continuous and categorical values between both groups were compared using t test or Chi-square test, respectively. Correlations among the SAS score with the related factors were analyzed through a correlation analysis. P value < .05 was considered statistically significant.

Table 1 Baseline characteristics of patients before and after propensity score matching
Full size table

Results

Patient characteristics

For the initial data, 145 eligible patients were recruited. Of the 145 patients, 106 complied with the inclusion criteria: 34 with the anxiety symptom and 72 without it. At baseline, there were statistical differences between both groups with regard to demographic characteristics. After propensity-score matching, the anxiety cohort and the control cohort each included 26 patients (Fig. 1). No significant differences between the 2 groups in clinical and demographic characteristics were detected, showing a successful matching process (Table 1).

Fig. 1
figure 1

Eligible patients and those included in propensity-score matching

Full size image

Results in the propensity-matched cohorts

Pre-operatively, there were ten (38.5%) patients suffering from mild anxiety and 16 (61.5%) patients suffering from moderate anxiety in the anxiety cohort. There were statistical improvements in the pain and disability for all patients at the final follow-up after PTED operation in each cohort (P < 0.05). The clinical results are shown in Fig. 2. For the patients with anxiety, significant improvements were noted between the pre-operative and the two year post-operative VAS (leg pain and low back pain) scores (P < 0.05), the SF-36 scores, ODI scores, JOA scores, and SAS scores (P < 0.05). Similarly, a significant increase in the aforementioned scores two  years after the surgery was also observed in the control group (P < 0.05). Although significant improvements in pain, functional activity, neurological condition, and psychological state have been achieved in the two groups at each post-operative follow-up, there were still statistical differences in the terms of VAS (leg pain and low back pain) scores, SF-36 mental scores, ODI scores, and SAS scores between the 2 groups during follow-up(P < 0.05).

Fig. 2
figure 2

Line graphs showing a comparison of VAS, SF-36, ODI, and SAS scores between anxiety cases and control cases. Change in trend of VAS-low back pain (a); change in trend of VAS-leg pain (b); change in trend of SF-36-physical (c); change in trend of SF-36-mental (d); change in trend of SAS (e); change in trend of ODI (f); change in trend of JOA (g); VAS, SF-36, ODI, JOA, and SAS scores had significantly improved for all patients. *P < 0.05 between 2 groups. VAS indicates visual analogue scale; SF-36, 36-item Short-Form Health Survey; Oswestry disability index, ODI; Japanese Orthopedic Association, JOA; SAS, Zung Self-Rating Anxiety Scale

Full size image

The peri-operative data are summarized in Table 2. There was no statistical difference between the two groups in terms of operation time and intra-operative blood loss (P > 0.05). Pre-operative JOA (leg pain and low back pain) score, pre-operative SF-36 (physical and mental) score, pre-operative ODI score, and pre-operative SAS score in both groups demonstrated significant differences (P < 0.05). No complication was documented in either groups (P > 0.05).

Table 2 Pre-operative parameters after propensity score matching
Full size table

Correlation between pre-operative anxiety and related parameters

Among the matched patients, the disease duration, the degree of pain (the VAS score), the level of neurological deficit (the JOA score), and the quality of life (the SF-36 score and ODI score) were related to the pre-operative level of anxiety (the SAS score) in the correlation analysis. However, age, gender, BMI, and segment distribution had no significant impact on the degree of anxiety. The results indicated that disease duration, VAS score, JOA score, SF-36 score, and ODI score were closely linked to the SAS score (Table 3).

Table 3 Correlations between SAS score and pre-operative-related factors after propensity score matching
Full size table

Discussion

PTED has become a well-accepted minimally invasive treatment for LDH as a result of the advancement in endoscopic equipments [16]. However, it was found that some patients did not achieve the excellent improvements in the symptoms and the prognosis as expected through PTED surgery in our clinical practice. These inconsistencies have led us to assess the psychopathological profile of patients independently influencing the outcomes. A matter of debate in relation to anxiety has less frequently been investigated despite growing evidence of its basic implication on the clinical outcomes. Anxiety is a psychological disturbance relevantly influencing PTED because the existence and the degree of pre-operative psychological disturbance have independently been associated with a significantly poor response to operation. Previous investigations have indicated that there is a connection between psychological disorders and lumbar surgery, cervical surgery [17], or chronic spinal pain [18, 19]. Nonetheless, there is a great disparity in our cognition of the relationship between anxiety and PTED. In the current study, 38.5% of patients presented mild anxiety while 61.5% of patients had moderate anxiety. Even though PTED is a relatively novel surgery with undeniable advantage, pre-operative anxiety may still be present and affect the success of the operation. Given retrospective analysis in nature with lack of a comparable matched control group, the propensity score matching method was applied to compare the clinical outcomes between anxiety and control patients by PTED procedure. The propensity score is a balancing score that enables us to make the distribution of the measured baseline variables similar between the groups.

The effect of anxiety on the prognosis

This study revealed that the patients with anxiety achieved significant improvement in the pain and quality of life after the PTED treatment, while psychopathological scores decreased rapidly. These changes may strongly be linked to the extent of relief in pain and disability. However, the clinical outcomes in the anxiety cohort were still inferior to those in the control cohort. One possible interpretation for the phenomenon might be pre-operative anxiety and pain intensity. Previously, a clinically related connection between anxiety sensitivity and pain magnitude was found [20, 21], which suggests a strong interplay between pre-operative mental status and patient-reported outcome after PTED. Moreover, Dunn et al. also reported that pre-operative anxiety may play an essential role in regulating post-operative pain, and earlier, psychotherapy can decrease pain magnitude significantly [22]. Another possible reason may be the influence of disease duration and pain chronicity. Anecdotally, a prolonged duration of symptoms is often related to a poor prognosis for recovery. It is also proposed that persistently activated peripheral nociceptors might finally result in the neural plasticity, over a specified duration, in either the central nervous system (“central sensitization”) or the dorsal root ganglions (“peripheral sensitization”) [23, 24]. Apart from a continuous severity of pain, it may be elucidated that a significantly prolonged disease duration would be required for plastic changes. Moreover, D’Angelo suggested that the pre-operative presence of anxiety is a prognostic indicator for the post-operative persistence of pain [9]. Along with the complexity of the nervous system and the regulating mechanisms of pain, additional attention to psychological status should be paid.

Risk factors for pre-operative anxiety

Our results after adjustment by propensity score showed a strong relationship between the status of anxiety and the disease duration, the level of pain (the VAS scores), neurological deficits (the JOA scores), and the quality of life (the SF-36 scores and ODI scores). These findings, which demonstrate that pain, low quality of life, and disease duration were associated with anxiety, are similar to those of previous studies. The relationship between the mood condition and pain has not been comprehensively elucidated. Heikkinen et al. revealed a major correlation between the LDH and anxiety [25]. Furthermore, Jess et al. indicated that patients with shorter durations of pain might have relatively superior results in the short term, emphasizing the increased advantage in enrolling patients in the treatment in the early stages [26]. Our results also suggested that long disease duration in the anxiety cohort had poor clinical prognosis than that in the control group. Frequently, LDH with long disease duration may lead to chronic pain and influence the quality of life [27]. Patients experiencing chronic pain and/or neurological deficits often endured several problems that affect daily life, including mental stress, sleep disturbance, and social disaffiliation, indicating that the relationship between anxiety and LDH should be deeper and complete. Notwithstanding, some researchers insisted that early intervention is critical for the efficacy of LDH irrespective of prolonged disease duration.

Limitations

There were naturally some limitations in this study. Despite the statistical methodology employed in this study, the data utilized in the propensity score matching were collected retrospectively and limited to the integrity of our data. Additionally, there is the likelihood of unmeasured confounding socioeconomic factors, such as educational degree, occupational characteristics, insurance, and employment status. Finally, even though it exceeds the size that was estimated based on the preliminary study for detecting a statistical difference of clinical outcome, the sample size for subjects included in the present study might be relatively small. Therefore, further clinical studies with a larger sample size will be necessary.

Conclusion

Our study demonstrated that patients with anxiety achieved significant improvement, while anxiety significantly improved after the PTED operation. Anxiety had a strong correlation with pain, neurological deficit, the quality of life, and disease duration, and had negative effects on the prognosis. Our findings indicated that surgeons should pay extra attention to the anxiety in patients with LDH to not only promote the relief of pre-operative symptoms but also to ameliorate the psychological condition. Early intervention is critical for LDH with anxiety irrespective of disease duration.