Investigation for Factors Affecting Body Perception Disturbance in Patients with Low Back Pain by Mechanism-Based Classification of Pain: A Cross-Sectional Study

Background Central sensitization is a pathophysiological cause of chronic low back pain and is linked with psychosocial factors. The association between central sensitization (CS) and body perception disturbance is currently unclear, and no prior studies have investigated this relationship in patients with acute or subacute low back pain. The objective of this study was to investigate potential factors that influence body perception disturbance using a mechanistic classification of low back pain. Methods This cross-sectional study was conducted at the time of initial physical therapy in patients with low back pain. During the study period, 169 patients were recruited. Pain intensity, disease duration, disability, CS, and body perception disturbance were evaluated. Patients were divided into three groups according to the pathology of low back pain, and multivariate analysis was used to examine factors affecting body perception disturbance. The dependent variable was Fremantle Back Awareness Questionnaire (FreBAQ); the independent variables were age, gender, BMI, VAS, disease duration, RDQ, and CS Inventory-9 (CSI-9). Results A total of 117 patients were included in our analysis. According to the mechanistic classification of pain, 66 (56.4%), 36 (30.8%), and 15 (12.8%) patients were categorized as having nociceptive pain (NP), peripheral neuropathic pain (PNP), and CS pain (CSP), respectively. Patients with PNP or CSP were significantly older than those with NP (p  <  0.01). FreBAQ and RDQ scores were significantly higher in patients with CSP than those with NP (p  <  0.05). The results of multiple regression analyses indicated that CSI-9 scores were significantly associated with FreBAQ (p  <  0.01). Conclusion Patients with CS syndrome and low back pain tend to have higher CSI-9 scores and be older. Body perception disturbance is influenced by CS or CS syndrome, regardless of the stage of low back pain, suggesting that patients with chronic low back pain tend to have low body image.


Introduction
Low back pain (LBP) is associated with a high physical and economic burden [1].Te lifetime prevalence of LBP is estimated to be as high as 80% [2], and recent studies have documented a consistent increase in its annual incidence; this is problematic, as there are few efective treatments for this condition [3,4].Te lack of efective treatments may be attributed to the fact that most therapeutic strategies are based on biomedical models, which are reliant on anatomical and biomechanical factors [5,6].In contrast, the bio-psycho-social model provides a more comprehensive view of LBP by accounting for interactions with physical, psychological, and social infuences.
Central sensitization (CS) has been recently reported to be involved in the pathophysiology of chronic pain [7,8].It is defned as a neurophysiological state caused by hyperexcitability of the central nervous system, and CS syndrome (CSS) has been proposed as a comprehensive disease concept in which CS is involved [9].In terms of its association with LBP, CS has been reported to be associated with both psychosocial and cognitive-behavioral factors [10].For example, CS in the acute phase of LBP may be a precursor for the transition to chronic LBP when it is combined with other psychological factors [11].Furthermore, CS has been shown to mediate the relationship between pain intensity and psychosocial factors [12].
Previous studies have described the concept of body perception disturbance (BPD) as a disease-specifc factor involved in the chronicity of LBP [13].Body perception is the ability to accurately perceive and recognize one's own body parts and movements.Wand et al. developed the Fremantle Back Awareness Questionnaire (FreBAQ) to assess BPD in patients with chronic LBP and confrmed its reliability and validity [13].While they demonstrated that BPD is associated with chronic LBP, their study only included patients with chronic LBP, as opposed to acute or subacute LBP.Tere is no previous study investigating the relationship between acute or subacute LBP and BPD, so the relationship between CS and BPD is currently unclear.Tese previous studies suggested that BPD may have an impact on the chronicity of LBP, so we hypothesized that BPD and CS are related and patients with CSS may have signifcantly higher BPD.Tis study is the frst study to examine BPD in acute or subacute LBP.
Terefore, the purpose of this study was to examine factors that potentially infuence body perception disturbance in patients with LBP, based on diferences in the mechanistic classifcation of pain.

Study Design.
Tis cross-sectional study was conducted with approval from the ethics committee of Tokyo Metropolitan University Arakawa (approval number: 20071).

2.2.
Participants.Tis study included 121 patients who received physical therapy after medical consultation at our clinic, between March and November 2021.Patients were included if they were 20-65 years old and able to provide valid responses to the administered questionnaires.LBP was defned in accordance with the criteria proposed by the Japanese medical guidelines for LBP in 2019 [10]: "pain which is located in the back side of the trunk between 12 rib and gluteal folds, lasts for at least one day, with or without unilateral or bilateral radiating pain in lower limb."Exclusion criteria consisted of the following: paralysis, tumor, infection, fresh vertebral fracture, pregnancy, those who are attending psychosomatic medicine or psychiatry, and indemnifcation problem.1).Te following parameters were extracted from the clinic medical records: age, gender, height, weight, body mass index (BMI), pain intensity, disease duration, FreBAQ, CSI-9, and Roland-Morris Disability Questionnaire (RDQ).All data were recorded during the frst physical therapy session.

Classifcation of Pain Mechanism.
Te classifcation of pain based on its underlying mechanism was determined by the physical therapist who was in charge of the frst physical therapy session for each patient.Te method of classifcation conformed with that of Nijs et al. [14] who used the following diagnostic criteria: (1) "is neuropathic pain present and able to explain the clinical picture?"; (2) "is there a disproportionate pain experience?"; (3) "is there a difuse pain distribution?";and (4) "is the CSI-9 score ≥20?".
In addition, the physical therapist accounted for physical examination results (e.g., medical interviews and neurological tests) and classifed pain into three groups (nociceptive pain (NP), peripheral neuropathic pain (PNP), and CS pain (CSP)).Mixed or uncertain pain types were excluded.A preliminary study was conducted among 10 physical therapists to evaluate the inter-rater reliability of this classifcation method; the Fleiss' kappa value was 0.638, indicating substantial agreement (Table 1).

Outcome Measurements.
Te primary outcomes were the Japanese versions of FreBAQ and CSI-9.
Te following secondary outcomes were evaluated via a questionnaire survey that was administered during the initial physical therapy session: basic patient characteristics, visual analogue scale (VAS), disease duration, FreBAQ, CSI-9, and Japanese version of RDQ.Basic patient characteristics and disease duration were obtained from the medical records.
Basic patient characteristics included age, gender, height, weight, and BMI.Disease duration was defned as the number of days elapsed from the date of pain onset to the date of the frst physical therapy evaluation.
Te VAS [15] was used to measure pain intensity, which ranged from 0 (no pain) to 100 (maximum pain imaginable).Body perception disturbance was evaluated using the Japanese version of the FreBAQ [16], which was developed by Nishigami et al. to assess body perception disturbance; its validity and reliability have been previously confrmed [13,17].Te FreBAQ is a questionnaire that can assess body perception disturbance and consists of a total of 9 questions to subfactors of neglect-like symptoms, proprioception, and body image.Te questions are answered on a 5-point scale from 0 to 4, with higher values indicating greater body perception disturbance.Although no cutof values have been reported, correlation with pain intensity, disability, depression, and pain catastrophizing has been reported.Te CSI-9 is a questionnaire used to screen for CSS and has been reported to have high validity and reliability [18].Te questionnaire consists of 9 questions with a 5-point scale from 0 to 4, with a cutof score of 20 or higher is considered suspicious for CS syndrome (10-19: mild and 20 or higher: moderate/severe).Te RDQ [19] was used to evaluate impairments in activities of daily living due to LBP.Scores range from 0 to 24 points, with higher scores refecting greater impairment.

2
Pain Research and Management

Sample Size Calculation.
Te target sample size was for 180 participants in our research period with an efect size as d � 0.21 by referencing a previous study [20] which investigated the relationship between pain intensity and its mechanism.A multiple regression analysis (using the forced entry method) was conducted to examine factors that potentially infuenced body perception disturbance.Te dependent variable was FreBAQ; the independent variables were age, gender, BMI, VAS, disease duration, RDQ, and CSI-9.

Statistical
All statistical analyses were performed using IBM SPSS Statistics version 26.0 (IBM SPSS Statistics, Version 26.0;IBM Corp., Armonk, NY, U.S.A.).Te level of statistical signifcance was set at 0.05.

Results
Figure 1 shows the study fow and the number of subjects excluded due to missing data at each stage of the analysis and the reasons for this exclusion.During the study period, 169 patients with LBP were recruited.Fifty-two patients were excluded due to the following reasons: did not provide informed consent (n � 3); unable to complete the assessment due to technical problems (n � 43); and mixed pain type (n � 4).No patients were classifed with an uncertain pain type.Terefore, a total of 117 patients were included in our analysis.
Patient characteristics and outcomes according to pain type are summarized in Table 2. Patients with PNP or CSP were signifcantly older than those with NP (p ≤ 0.001).Gender, height, weight, and BMI were not signifcantly diferent among pain groups (p � 0.549, 0.918, 0.496, 0.348) (Table 2).To clarify the infuence of age, a parallelism test was conducted with age as a covariate.While all outcomes were parallel, this did not reach statistical signifcance.Tus, one-way ANOVA and Kruskal-Wallis tests were conducted to examine diferences among pain groups.
Comparisons of outcome measures among pain groups are shown in Tables 2-5  Pain Research and Management indicated that CSI-9 scores were signifcantly higher in patients with CSP than those with NP or PNP (p ≤ 0.001) (Tables 4 and 5).Also, FreBAQ and RDQ scores were signifcantly higher in patients with CSP than those with NP (p � 0.043, 0.034) (Tables 2-5).Tere were no signifcant intergroup diferences in disease duration and VAS (p � 0.546, 0.214) (Table 2).Additionally, one-way ANOVA was also conducted to examine diferences among groups divided by disease stages.However, there is no signifcant diference in FreBAQ (p � 0.496) (Table 6).
Te results of the multiple regression analysis using the forced entry method are summarized in Table 7. CSI-9 was a signifcant independent predictor of FreBAQ (p ≤ 0.001).Te standardized partial regression coefcient, which indicates the degree of infuence from independent variables, was 0.409 for CSI-9.Te rate of contribution from the regression formula (R 2 ) was 0.280.Te variance infation factor was <2 for all independent variables.

Mechanistic Classifcation of Pain.
According to a classifcation of pain based on its underlying mechanism, 66 (56.4%), 36 (30.8%), and 15 (12.8%) patients had NP, PNP, and CSP, respectively.Tese results were generally similar to those reported by Smart et al. [20] (NP, 55%; PNP, 22%; CSP, 22%); however, the proportion of patients with CSP in the present study was slightly lower.Tis distribution in pain type may be attributed to the pathology of the pain and disease duration.Chronic pain is generally defned as pain persisting for >3 months [21].CS is considered to be a form of chronic pain that becomes more prominent with time.However, in practice, the diagnosis of chronic pain and CS is not only based on the disease stage but also on the integration of various fndings.In a study conducted among patients with musculoskeletal disorders, Tanaka et al. [22] reported that 15.17% of patients had mild CS (CSI score of 30-39) and 11.00% had moderate-to-severe CS (CSI score of >40).Te number of patients with CS was slightly lower than that in our study.However, their study was only assessed CSI-9 scores, and the actual number of patients with CS and CSS was expected to be diferent.Terefore, based on these previous studies, we believe that the results of the present study are externally valid for patients with LBP.Patients with PNP or CSP were signifcantly older than those with NP.Te reason for this diference may be related to the pathophysiology of chronic pain.Apkarian et al. [23] reported that the dorsolateral prefrontal cortex is dysfunctional in patients with chronic LBP.Furthermore, aging causes a decline in function in this region.[24] Antonella et al. reported that elderly individuals have lower pain thresholds [25] and are more prone to CS. Tus, patients with CSP are more likely to be older and have a dysfunctional dorsolateral prefrontal cortex.

Comparison of Outcomes among Pain Groups.
Te signifcant diferences in CSI-9 scores among the pain groups can be attributed to the fact that CSI-9 is used to screen for CS and CSS.Te higher its scores, the more its pathology can be closer to CS.So, the results of CSI-9 are directly related to the classifcation results.Tis may be due to the inclusion and assessment of patients with LBP at their frst physical therapy session.Chronic LBP is defned by the persistence of pain for at least 3 months, and CS becomes increasingly prominent over time.However, in the present study, patients were divided into groups based not only on disease duration alone but also by accounting for a range of physical fndings; this may have explained the lack of diference in disease duration among groups.On the other hand, there was no signifcant diference in VAS scores among the groups.Tis fnding difers from that of other studies [20].Shigetoh et al. reported that CS mediates pain intensity and psychological factors [12].Tus, the absence of a signifcant diference in pain intensity may be due to the low prevalence of psychological factors in some patients with CSP.In addition, the present study included patients with acute and subacute LBP; pain intensity and CS may not be related during these earlier stages of LBP.As current evidence indicates that pain intensity is associated with CS in patients with chronic LBP, it can be inferred that the strength of this association increases over the disease course.

Multiple Regression Analysis for Body Perception
Disturbance.Te multiple regression analysis indicated that CSI-9 was an independent and signifcant factor associated with BPD in patients with LBP.Tese results are consistent with those of previous studies on patients with chronic LBP.Tus, CS may be associated with BPD during the early acute and subacute phases of LBP, and this association becomes increasingly evident following the transition to the chronic stage.Nevertheless, the rate of contribution of CSI-9 was low (28.0%),suggesting the potential infuence of other factors.
LBP is a complex condition with a multifactorial etiology.A bio-psycho-social model has been previously proposed to account for the efects of psychological, social, and biophysical factors [4].However, we did not evaluate outcomes related to psychological factors that are believed to be involved in pain chronicity [10,11].Terefore, although it was clear that CSS was associated with FreBAQ, the contribution of this factor was limited.
In the one-way ANOVA, FreBAQ scores were higher in patients with CSP than those with NP.Bogduk et al. [26] investigated the cause of LBP using nerve blocks and found that NP was the major contributor.Terefore, while NP may be predominant in the early stages of LBP, the mechanism of pain generation may change with chronicity.Psychosocial factors such as kinesiophobia and pain catastrophizing may be facilitators of CS [27] and CSP; they have also been shown to afect FreBAQ.In this study, there is no signifcant difference in FreBAQ among groups divided by disease stages.Furthermore, the contribution of CSI-9 was higher than that of disease duration in multiple regression analysis for Fre-BAQ.Tese results may indicate that the factors leading to chronicity are more infuenced by body perception disturbance than by the disease duration itself.Tus, these factors may be used to predict chronicity [28].
Tere were some limitations in this study.First, we could not correct sufcient sample size.Te target sample size was set for 180 participants in our research period with an efect size as d � 0.21 by referencing a previous study that investigated the relationship between pain intensity and its pain mechanism.Second, this study was a single-institutional study, so external validity was unclear.Tird, this study was a cross-sectional study, which does not provide a causal association.We need to conduct further investigation about the chronicity of LBP or CS with a longitudinal study.

. Conclusion
Te results of this study suggest that patients with CS or CSS and low back pain tend to have higher CSI-9 scores and be older.Body perception disturbance is infuenced by CS or CSS, regardless of the stage of low back pain, thus refecting that patients with chronic low back pain tend to have low body image.

Figure 1 :
Figure1shows the study fow and the number of subjects excluded due to missing data at each stage of the analysis and the reasons for this exclusion.During the study period, 169 patients with LBP were recruited.Fifty-two patients were excluded due to the following reasons: did not provide informed consent (n � 3); unable to complete the assessment due to technical problems (n � 43); and mixed pain type (n � 4).No patients were classifed with an uncertain pain type.Terefore, a total of 117 patients were included in our analysis.According to the mechanistic classifcation of pain determined during the frst physical therapy session, 66 (56.4%), 36 (30.8%), and 15 (12.8%) patients were categorized as having NP, PNP, and CSP, respectively.Patient characteristics and outcomes according to pain type are summarized in Table2.Patients with PNP or CSP were signifcantly older than those with NP (p ≤ 0.001).Gender, height, weight, and BMI were not signifcantly diferent among pain groups (p � 0.549, 0.918, 0.496, 0.348) (Table2).To clarify the infuence of age, a parallelism test was conducted with age as a covariate.While all outcomes were parallel, this did not reach statistical signifcance.Tus, one-way ANOVA and Kruskal-Wallis tests were conducted to examine diferences among pain groups.Comparisons of outcome measures among pain groups are shown in Tables2-5.Te results of the one-way ANOVA

Table 1 :
Cross tabulation of the results for mechanism-based classifcation of pain.

Table 2 :
Demographic data in each group.

Table 3 :
Outcome measurements in each group.

Table 4 :
Diference among groups by one-way ANOVA for mechanism-based classifcation.

Table 6 :
Diference in FreBAQ among groups by one-way ANOVA for disease stages.

Table 7 :
Te result of multiple regression analysis for FreBAQ.