A Descriptive Study of the Centralization Phenomenon

Transcription

1 A Descriptive Study of the Centralization Phenomenon A Prospective Analysis Mark Werneke, MS, PT, Dip. MDT,* Dennis L. Hart, PhD, PT, and David Cook, BS, RN SPINE Volume 24, Number 7, pp , Lippincott Williams & Wilkins, Inc. Study Design. Occurrence and treatment responses associated with the centralization phenomenon were analyzed prospectively in 289 patients with acute neck and back pain with or without referred spinal symptoms. Objectives. To document symptom changes to mechanical assessment during initial evaluation and during consecutive visits. Using standard operational definitions, patients were categorized reliably into three inclusive and mutually exclusive pain pattern groups: centralization, noncentralization, and partial reduction. It was hypothesized that the occurrence of centralization would be less than previously reported and that the centralization group would have better treatment results. Summary of Background Data. Centralization has been reported to occur with high frequency during mechanical assessments of patients with acute spinal syndromes. When centralization is observed, a favorable treatment result is expected. Because centralization has not been defined consistently in the literature, the true prevalence and treatment responses associated with centralization have not been confirmed. Methods. Consecutive patients with neck or back pain syndromes and referred to outpatient physical therapy services were categorized into three pain pattern groups by experienced therapists trained in the McKenzie system. Changes in distal pain location were scored and documented before and after each visit. Maximal pain intensity over 24 hours, perceived functional status, and number of treatment visits were compared between groups. Results. Patients could be categorized reliably according to movement signs and symptoms. The centralization pain pattern group had significantly fewer visits than the other two groups (P 0.001). Pain intensity rating and perceived function were different between the centralization and noncentralization groups (P 0.001). There was no difference in treatment response between the centralization and partial-reduction groups (P 0.306). Prevalence of patients assigned to the three groups was 30.8% in the centralization group, 23.2% in noncentralization, and 46% in the partial-reduction group. Conclusion. Categorization by changes in pain location to mechanical assessment and treatment allowed identification of patients with improved treatment outcomes and facilitated planning of conservative treatment From *NovaCare at Southern Ocean Center For Health, Forked River, New Jersey, Focus On Therapeutic Outcomes, Great Falls, Virginia, and Automotive Motors at Thomasville, Thomasville, North Carolina. Data collected from the Office of American Rehabilitation Services, High Point, and the Office of Employee Health Clinic of Automotive Motors at Thomasville, Thomasville, North Carolina. Data Analyzed at the Office of Focus On Therapeutic Outcomes, Inc., Great Falls, Virginia. Acknowledgment date: February 3, First revision date: May 13, Acceptance date: August 18, Device status category: 1. of patients with acute spinal pain syndromes. If a proximal change in pain location is not observed by the seventh treatment visit, the results of this study support additional medical evaluation for physical or nonphysical factors that could be delaying quick resolution of the acute episode. [Key words: acute spinal pain, centralization, pain location, pain patterns] Spine 1999;24: Centralization is a clinical phenomenon observed during mechanical assessment of patients with pain in the neck or back. The centralization phenomenon (CP), as originally described by McKenzie, 17 was defined as a rapid change in the location of pain from a distal or peripheral location to a more proximal or central position. The CP could be observed in patients even if the pain was localized and spread only to the buttock. 4 In addition, the pain pattern would remain better, e.g., reduction in midline, lateral, or distal pain would not reappear if the movements or positions found to centralize symptoms were repeated or sustained while other movements were temporarily avoided. 5,16,17,28 The key and essential component to understanding centralization is this rapid response to changes in pain location as a result of clinically directed movement and positioning techniques. 7 If the appropriate mechanical procedure is not applied, then centralization will not be observed. 5,28 Medical interest in the CP has increased over the past decade because of the reported high frequency in the clinic, predictive value in identifying patients who will respond satisfactorily to conservative rehabilitation efforts, and usefulness in guiding treatment planning. 4,5,7,14,22 Although the concept of centralization appears to be straightforward, the definition of centralization as described by McKenzie 16,17 has not been consistently applied nor documented in the literature. 3,4,7,8,12,24 In addition to inconsistencies in the definition of the CP, previous research has focused primarily on this phenomenon occurring during the initial evaluation period only. Therefore, it is not clear from the literature if 1) centralization had to be noted during one or all visits, 2) referred symptoms needed to be completely or partially reduced; and 3) patients who showed no change in the location of pain during the initial evaluation period could be categorized accurately into a noncentralization group. Monitoring changes in pain location during not only the initial evaluation but also subsequent visits appears to be important in differentiating CP from variable patterns of reductions or abolition of pain from natural history alone. 676

2 Centralization Phenomenon Werneke et al 677 There is a general lack of evidence to substantiate the prognostic value of centralization for patients with acute spinal syndromes. Karas et al 12 found that patients who had centralized symptoms had a significant increase in return to work at 6 months compared with patients whose symptoms did not centralize. A mix of patients with acute and chronic pain was studied, and demographics concerning acuity were not discussed. Sufka et al 24 reported that patients who had complete centralization had a significant improvement in subjective functional scores compared with those who did not have complete centralization. Generalizability of their results are difficult because of the small number of patients with acute pain (N 15) studied. In a retrospective analysis, Donelson et al 7 reported that referred symptoms to the buttock, thigh, or leg in 89% of the patients with acute pain centralized during mechanical assessment. Of this centralization group, 91% had excellent outcomes resulting in complete relief of pain and restoration to function. It is not known from this study if complete and partial centralization are equally predictive of favorable outcomes, or if patients who do not have symptoms that centralize on any visit yet who demonstrate a partial reduction in pain location over time have satisfactory outcomes. The authors of the current study are unaware of any other studies in which the clinical and predictive usefulness of the CP has been investigated specifically for patients with acute neck or back pain syndromes. Previous literature has not clearly distinguished the prognostic value of the CP from what is expected with natural history. This raises the question of CP as an important predictor of a favorable response to conservative care in the acute episode compared with the natural history of acute low back or neck pain. With no consistent definition of centralization, comparing occurrence and treatment outcomes related to the CP across different clinical trials is problematic. Therefore, the purposes of this prospective study were 1) to standardize the operational definition of CP; 2) to determine the occurrence of the CP as defined in this study in a large cohort of patients with acute neck and back pain syndromes; 3) to quantify the reliability of categorizing all patients into three pain pattern groups, i.e., centralization, noncentralization, or a partial-reduction group; and 4) to determine if there is a difference between pain pattern groups in clinical outcomes from physical therapy (visits, pain intensity, and function). The authors tested the following hypotheses: 1) the occurrence of the centralization pain pattern will be substantially less then the occurrence of CP previously reported; 2) the three pain pattern groups can be reliably distinguished; and 3) the centralization pain pattern group will have better outcomes than those of the other two pain pattern groups. Methods To evaluate CP, the authors of this study propose that Mc- Kenzie s 17 original operational definition for the CP be standardized. Patients were categorized into one of three groups based on the following operational definitions. Centralization Group: 1. A clinically induced change in location of pain/symptoms referred from the spine moves from the most distal position toward the cervical or lumbar midline. Note: For patients with only central or midline pain, the midline pain must cease during the initial visit. 2. The change in pain location or abolition of midline pain must remain better (e.g., the lateral or distal pain does not reappear) as a result of mechanical movements/positions. 3. The change in pain location initially observed on the first visit must continue its proximal movement on subsequent trials until all symptoms are abolished. Note: Midline pain must remain abolished on subsequent visits. Noncentralization Group: 1. No changes in the location of pain occur, or 2. Location of pain changes from a central to a more distal location throughout all treatment visits. Partial Reduction Group: 1. Location of pain changes from a more distal to a more central location during each visit without a progressive movement in initial pain location toward the midline at consecutive visits, or 2. No change in pain location occurs during any one visit, but the patient has a gradual decrease in pain location over subsequent visits. For patients to be categorized as centralizers, all of the criteria listed for centralizers must be met. The definitions emphasize the importance of rapid changes in pain location (not pain intensity) and progressive improvement in pain location, and the changes are induced clinically and controlled by specific movement/positioning techniques. Subjects. The patient sample consisted of 351 consecutive adult patients who met the following admission criteria: years of age, diagnosed with neck or low back pain syndromes with or without referred symptoms, referred by a physician for conservative treatment, and having symptoms of less than 6 weeks duration. The patients were referred to one of two independent medical centers for physical therapy services between January 1996 and June Before referral to physical therapy services, all patients received appropriate medical diagnostic triage, medical reassurance, medications, and temporary work restrictions if deemed necessary by the primary physician. In addition, the majority of the patients were on workers compensation and unable to resume full duty tasks. If there were no improvements, i.e., if the patient was unable to return to full duty or if significant pain persisted at the physician s follow-up evaluation and conservative management was still indicated, patients were referred to physical therapy. Patients were excluded if they 1) reported spinal pain or work loss because of spinal pain within 6 months before this episode, 2) had poor English proficiency or were unable to complete intake questionnaires independently, 3) had a history

3 678 Spine Volume 24 Number Table 1. Patient Characteristics at Initial Evaluation Characteristic Male Female n Age (yrs) Spinal type Cervical (%) Lumbar (%) Symptom location Midline (%) Shoulder or buttock (%) Leg or arm (%) Duration of Symptoms (days) Payer Type Worker s Compensation (%) Automobile insurance (%) Other commercial (%) Medicare/Medicaid (%) 0 0 Current working status Full duty, full time (%) Full duty, part time (%) Light duty, full time (%) Light duty, part time (%) Not working (%) Physical demand level of job Sedentary (%) Light (%) Medium (%) Heavy (%) Very heavy (%) of spinal surgery, 4) were pregnant. or 5) refused to sign a consent form to participate in the study. A total of 51 patients did not meet the admission criteria or refused to participate. A total of 300 patients consented to participate. Eleven patients dropped out after the initial physical therapy evaluation and were excluded, leaving 289 patients (Table 1). Of these patients, 22.8% reported cervical symptoms and 77.2% reported low back pain. The average duration of symptoms before the initial physical therapy evaluation was 13 days. The majority of the patients (71.6%) were receiving workers compensation benefits. Eighty-seven percent were not working or were working at modified duty, and only 13% were working full duty full-time during the treatments. Of the patients remaining in the study, 80% of the men and 53% of the women had medium or heavier physical job demand duties. Physical Therapists. Five physical therapists participated in the study. Three physical therapists had years of clinical experience, and the other two therapists had 5 7 years of experience. All therapists were trained in the McKenzie evaluation and treatment methods. Three physical therapists were diplomats attaining the highest level of training by the McKenzie Institute, and two therapists who completed the basic A D courses received their credentials from the Institute. Patients were assigned to a therapist by the front-desk clerical staff member who was blinded to the methods of the study. The only criterion for assigning patients was the therapist s current patient caseload. Procedure. At the initial physical therapy consultation, each patient was asked to complete a body diagram, a pain intensity scale, and an Oswestry 9 or Neck 27 Disability Index questionnaire. Final Oswestry or Neck Disability Index questionnaires were completed by each patient immediately after the patient s discharge visit. The clerical staff member, again blinded to the administration and outcomes of the study, administered the questionnaires. The body diagram was used to record the location of the patient s pain before and after the initial evaluation and before and after each treatment visit. Each patient was instructed by the clerical staff member in a standardized manner to shade in all areas on the diagram where he or she was experiencing pain and referred symptoms when appropriate. After completing the intake questionnaires and forms, all patients received a mechanical evaluation following McKenzie s 16,17 assessment methods by one of the five physical therapists. Each patient was treated by the evaluating therapist. Subsequent treatment visits were scheduled by the clerical staff; hour intervals were recommended between visits. Exercises and manual techniques, if needed, were implemented according to the different pain responses from repeated end-range movement tests and/or positioning techniques observed during the objective examination. The movement(s) associated with pain centralization identified the direction of exercises for the patient (i.e., directional preference of exercise). Exercises performed in a specific direction were used during treatment with the goal of facilitating CP while avoiding movement patterns associated with exacerbation of symptoms. For example, if lumbar extension movements centralized pain, exercises that moved the patient toward lumbar extension and manual procedures that produced lumbar extension forces were selected as management techniques to control symptoms. Flexion movements and forces would be avoided temporarily in these patients. Different movements, positions, and/or forces were used on other patients depending on the findings of the initial evaluation. If centralization of pain was not observed and a directional preference for exercise was not found, an individualized active rehabilitation plan was developed for the patient by the treating therapist. This plan emphasized return to function, while activities that specifically peripheralized their pain were avoided. All patients received the same educational approach: empower the patient to become actively involved in his or her own recovery to reduce fear of physical activity and movement intolerance. Therapeutic modalities (e.g., ice or heat) were provided on a limited basis at the discretion of the therapist. Modalities were used to facilitate educational efforts and to enhance the patient s response to active treatment. Specific treatments varied according to the needs of each patient. The present project is a descriptive study of the CP, and, therefore, no attempt was made to standardize care beyond the above guidelines. After the patients completed each body diagram, the therapists coded the most distal pain location using a clear overlay template (Figure 1). This template and the scoring technique documenting the pain location has been described previously. 5,14 The cervical scoring grid used in this study has been recommended by Donelson (MW, personal communication, 1995), but has not yet been published. The location of the most distal pain site, shaded by the patient on the body diagram before the clinical examination and before each treatment visit, was defined as the PRE Pain Location Score (PRE PLS). After the examination and after each treatment visit, another body diagram was administered using the same standardized instructions. Using the same procedure as above, the most distal pain site was coded by the therapist (POST PSL). The coded pain location scores (0 6) were recorded on a Pain Pattern Spreadsheet (Figure 2). In this way, the operational definitions of the three patient groups could be quantified for analysis.

4 Centralization Phenomenon Werneke et al 679 calculated as an effect size score. 13,23 The effect size score is calculated by subtracting the pain intensity at the initial evaluation from that at discharge, and dividing this result by the standard deviation of the pain intensity rating at initial evaluation for the group of patients. In this way, the absolute change in pain intensity score was reduced to a standardized score in standard deviation units. The suggested interpretation rating of effect size scores was followed in this study, in which 0.2 to 0.4 is small, 0.5 to 0.7 is moderate, and equal to or greater than 0.8 is large. 2 In this way, the variance of the initial evaluation scores is used to normalize the absolute change scores from initial evaluation to discharge. The effect size score was used as the data point for the analyses. As the patients improved, i.e., their pain intensity ratings decreased, the effect size scores became negative. Negative effect size scores represent reduction in pain intensity over the treatment period. Figure 1. Overlay body template. To test the inter-rater reliability for therapists to code the location of the most distal pain on the body diagrams using the Donelson et al 5 and Long 14 techniques, three therapists independently scored 30 different pain diagrams. Pain diagrams were chosen randomly by the clerical staff member. Pain diagrams were selected at random by the clerical staff member until 15 patients with neck pain syndromes and 15 patients with low back syndromes were identified and approximately 50% of each group had referred versus central pain locations. Each therapist was blinded to the scores from the other raters. One therapist (MW) reviewed the Pain Pattern Spreadsheets at discharge to categorize each patient into the centralization, noncentralization, or partial-reduction pain pattern groups based on the current authors operational definitions. The three pain pattern groups are inclusive and mutually exclusive. The therapist was blinded to the patient s name, treatment, and outcomes. The reliability of these definitions for categorizing patients was assessed by having two therapists independently review all 289 patients by using the data entered in the Pain Pattern Spreadsheets. Both therapists were blinded to the patient s name, related information, and the other rater s findings. Outcome Measures. Maximal Subjective Pain Intensity Rating. A 10-point descriptive pain scale (Scale: 0 no pain to 10 severe pain requiring management in the emergency room) was completed by the patient during the initial and each subsequent treatment visit. 11 The scale allowed patients to rate their maximal pain intensity experienced during the past 24 hours at each visit. The change in pain intensity was Back/Neck Self-rated Disability Questionnaires. The Oswestry Low Back Pain Questionnaire was administered to each patient with low back pain before treatment and upon discharge from physical therapy. 9 Patients with neck pain completed the Neck Disability Index before treatment and upon discharge from physical therapy. 27 The Oswestry and Neck Disability Index scores range from 0 to 100, and, for this study, they were considered as similar measures of perceived function and were grouped together. The pre- and post-treatment scores were reduced to effect size scores as above. As the patient s perception of his or her function ( the limitations of a patient s performance compared with that of a fit person [p. 271] 9 ) improves, the patient becomes more functional, and the score becomes lower. Negative effect size scores represent improvement in the patent s perception of his or her functional abilities. The effect size scores from the Oswestry Questionnaire and Neck Disability Index will be operationally defined as perceived function. Because both effect size scores (pain and perceived function) became negative when the patient reported improvement, all effect size scores were reported as reverse scores, i.e., positive effect size scores represented improvement, for ease of interpretation. Figure 2. Example of Pain Pattern spreadsheet.

5 680 Spine Volume 24 Number Treatment Visits. The number of visits for physical therapy was documented. The number of visits was taken at face value and not reduced to an effect size score. Data Analyses. Reliability Assessment. Agreement among the three judges assessing the coding of the location of the distal pain using the Donelson et al 5 and Long 14 techniques and agreement between two judges classifying the patients into pain patterns (centralization, noncentralization, or partial reduction) were quantified with a Cohen s kappa weighted for chance. 10,25 Patient Characteristics at Initial Evaluation. Descriptive statistics were used to describe the demographic and clinical values at initial evaluation (independent variables). Differences between each independent variable were analyzed statistically by a Chi-squared statistic (ordinal or categorical data) or by one-way analysis of variances (ANOVAs; interval data, P 0.05) to determine the homogeneity of the sample. Outcome Variables. The effect of each independent variable or pairs of independent variables on the dependent variables of visits and change scores for pain and perceived function were assessed via one- or two-way ANOVAs with Sheffe post hoc analyses (P 0.05). Results Reliability The ability of the three judges to code reliably the most distal location of pain shaded on body diagrams following Donelson et al 5 and Long s 14 operational definitions was excellent (k 0.917, Judge 1 to Judge 3; k 0.917, Judge 2 to Judge 3; and k 1.0, Judge 1 and Judge 2). The ability of two judges to categorize patients reliably into the three pain pattern groups according to changes in pain location over time was also excellent (k 0.96). Patient Characteristics at Initial Evaluation Age, gender, duration of symptoms, payer type, physical demand level of the job, and current work status were not different across pain pattern groups. Symptom location was different across pain pattern groups (Chisquared 10.0, P 0.04). The number of patients reporting midline pain was smallest; more patients reported leg or arm pain, and the majority of patients reported shoulder or buttock pain. This distribution or prevalence of patients by symptom location across pain pattern appears to be clinically logical. However, using standardized deviates (observed expected)/(square root of expected value), there were more patients classified in the centralization group with shoulder/buttock pain (1.2 standardized deviate) compared with those with arm/leg pain ( 1.9 standardized deviate). In addition, there were more patients classified in the noncentralization group with arm/leg pain (1.5 standardized deviate) compared with those with shoulder/buttock pain ( 1.3 standardized deviate). Occurrence of Pain Pattern The frequency for each pain pattern was similar for patients with neck pain and those with back pain. Each pain pattern showed the following frequencies: 30.8% centralization (31.2% back, 24.6% neck), 23.2% noncentralization (21.7% back, 24.6% neck), and 46% partial reduction (43.7% back, 46.4% neck). Relation Between Patients With Neck or Low Back Pain The effect size scores for change in perceived function and pain intensity were first calculated for patients with neck pain and those with low back pain by pain pattern group to determine if spinal type and pain pattern affected the outcome variables. Number of visits and the change in pain intensity scores and perceived function had only one significant factor (pain pattern; P for all three dependent variables). Spinal type (neck or low back) and the interaction of spinal type by pain pattern were not significant (P , two-way ANOVAs). Therefore, it was decided to group all patients together regardless of spinal type (neck or low back) and to recalculate the effect sizes for change in pain intensity and perceived function. Changes in Perceived Function, Pain Intensity, and Visits by Pain Pattern Number of visits (F 36.8, P 0.001) and the effect size scores for perceived function (F 69.6, P 0.001) and pain intensity (F 37.0, P 0.001) were all different across pain pattern groups. Effect size scores were moderate to large. Patients categorized as centralizers averaged fewer visits ( , mean standard error) than noncentralizers ( visits) and those in the partialreduction group ( visits). There was no difference in number of visits between the partial-reduction and noncentralizer groups (P 0.88). Patients categorized into centralization or partial reduction had greater improvements in pain intensity and perceived function than those in the noncentralization group (pain, F 37.0, P 0.001; perceived function, F 69.6, P 0.001). For both dependent variables, there was no difference between the centralization and partial-reduction groups (P 0.14 for pain, P for perceived function). Partial-Reduction Group Analysis Patients who did not meet the current authors criteria for the centralization or noncentralization pain patterns were assigned to the partial-reduction group. This group represents those patients who demonstrated only partial and proximal changes in referred symptoms throughout the treatment period. Two distinct groups of pain location changes over time were observed in this category. First, 28.6% of the partial-reduction category demonstrated a proximal change in pain location during the first visit (a partial centralization response). This proximal reduction in pain score was maintained during treatment. Second, 95 patients (71.4%) in this category demonstrated no proximal changes in pain location (a noncentralization response) during the initial assessment or during any visit (i.e., PRE PLS was equal to POST

6 Centralization Phenomenon Werneke et al 681 PLS). However, a partial reduction of pain still was observed over time as a gradual decrease in the pain location score during consecutive visits (e.g., PRE PLS at the fifth visit PRE PLS at the first visit). A total of 51.6% demonstrated this improvement by the third visit, 73.7% by the fifth visit, and 92.6% by the seventh visit. If a patient did not demonstrate a decrease in pain location score by the seventh visit, no significant outcome changes measured by pain intensity or perceived function were noted. In addition, there were no differences between patients demonstrating a change in either type of pain location in the partial-reduction category. Discussion This study supports McKenzie s 16,17 original definition of centralization as a clinical phenomenon occurring in patients with acute spinal syndromes. This phenomenon is observed as an immediate and proximal change in pain location in response to clinically directed movement and position techniques. This proximal movement not only was observed during the initial assessment but progressed rapidly throughout subsequent treatment visits, until all pain and referred symptoms were abolished. The centralization pain pattern commonly was observed for patients with either neck or low back pains with and without referred symptoms. Centralization could be identified reliably from the other two pain patterns, because rapid proximal responses were 1) clearly differentiated from either no change/increases in pain location (noncentralization) or 2) slow and at times variable reductions in pain location over time (partial reduction). For the majority of patients in the partial-reduction group, the pain location could not be influenced by clinically directed procedures during any single visit, yet reduced scores were noted between visits. The authors of the current study believe such responses could have occurred as part of the natural recovery expected in an acute episode with just the simple passage of time. If confirmed, the addition of the partial-reduction group could strengthen research differentiation between true clinical effects of movement evaluation or treatment versus natural history. There has been little consensus in the literature on the definition and documentation of the CP. Karas et al 12 found that 73% of their cohort showed centralized symptoms during the first or second visit. Pain intensity and pain location changes were operational in their centralization definition. They did not discuss the manner in which changes in pain location were recorded and whether these changes were observed in all patients regardless of acuity. Delitto et al 3 and Erhard et al 8 defined centralization as a proximal change in pain location. They limited this definition to patients with referred symptoms below the buttock. Of their patients with acute back pain, 79% and 92%, respectively, had central and/or buttock pain only and were classified by change in pain intensity during initial assessment. Donelson et al 4 subsequently reported that the CP occurred in the research projects by Delitto et al 3 and Erhard et al 8 at a high frequency (55 61%). In a retrospective investigation, Donelson et al 7 reported that the symptoms in 89% of patients with acute back pain with referred symptoms into the buttock, thigh, or leg centralized. They based the CP on a change in the perceived location of pain. Exactly how changes in pain location were documented was not discussed. In a follow-up prospective study, Donelson et al 5 developed an overlay template to record carefully changes in pain location. They found that nearly half of their subjects, representing a mix of patients with acute and chronic pain, had symptoms that centralized during the initial evaluation in response to sagittal end-range movements, and that 58% had symptoms that centralized when nonsagittal movements were applied. 6,7 Williams et al 28 also used an overlay template to document changes in pain location and found a 48% CP prevalence rate in a group of patients with acute and chronic low back pain. Lastly, Sufka et al, 24 analyzing a small cohort of patients, reported a complete centralization occurrence rate of 83% in patients with acute pain and a rate of 73% in patients with subacute pain. Patients were categorized as centralizers only if all symptoms moved to a central location; however, midline pain did not have to be abolished. All other changes in pain behavior including partial centralization responses were categorized into a noncentralization group. In addition, they did not report the manner in which pain location changes were measured, nor did they elaborate on the exact pattern in which pain location changes occurred during the 14-day assessment period. To standardize the definition of centralization in the current study, simplistic operational definitions that are inclusive and mutually exclusive were proposed and evaluated. The pain pattern categorizations used in this study were based solely on numeric changes in pain location scores occurring before and after each visit and throughout the course of treatment. Occurrence The centralization pain pattern occurred in 30% of the patients in this study, which is considerably lower than that previously reported, supporting the authors first hypothesis. There are several reasons that could explain these conflicting results. First, monitoring changes in pain intensity as reported in several of the above studies was subjective and could account for elevated centralization occurrence rates. Interobserver documentation of the site of the patient s pain and changes in the site from repeated movements has been shown to be reliable. 15,21 Pain intensity was not a consideration in the definition of centralization in the current study. Second, in previous studies in which the overlay template was not used, pain intensity determined centralization for patients with midline and shoulder/buttock referred symptoms. When pain intensity was used to define CP, occurrence rates were high compared with the findings of the current study. By using the overlay template,

7 682 Spine Volume 24 Number small changes in pain location could be documented accurately, regardless of the original pain location. The overlay template would allow clinicians to document the CP by noting changes in pain location alone without relying on variations in pain intensity. The authors of this study recommend that an overlay template be used in future research projects investigating the CP. Third, CP, as defined in this study, required that a proximal change in pain location clinically induced during the first assessment was maintained after the visit (i.e., POST PLS PRE PLS) and that a similar progressive proximal movement was observed on subsequent visits. Such precise documentation of changes in pain location and observations of different pain patterns over time have not been reported previously. Future scientific investigations are recommended to confirm the current findings on the occurrence of CP for patients with acute spinal pain syndromes. A comparison of the prevalence rates in the current study with those of previous studies was not possible because operational definitions were different. By using the same operational definitions in the future, this limitation will be minimized. Outcomes The observation of the CP in patients with acute spinal syndromes has been purported to be of prognostic value in determining outcomes with conservative care. Two studies investigating the CP and outcomes involved either a mix of patients with acute and chronic pain or just patients with chronic pain. 12,14 Donelson et al 7 investigated CP and its relation to treatment outcomes for patients with acute low back pain. In the Donelson et al study, 91% of the patients with acute pain in whom symptoms centralized had excellent outcomes with relief of pain and full functional recovery. Sufka et al 24 also reported greater improvements for patients with acute pain whose symptoms centralized. The results of the current study support their findings. The authors of the current study found that patients in the centralization category had fewer visits with greater improvement in pain intensity and perceived function compared with patients categorized as noncentralizers. These findings complement McKenzie s 17 original description on centralization, i.e., if a rapid change in pain location is observed in certain patients, then a complete recovery can be expected in several days. Patients in the centralization group also had fewer visits than those in the partial-reduction group. However, these two groups did not significantly differ in decreases in maximal pain intensity or improvement in perceived function. This was an unexpected finding. Patients whose peripheral symptoms were not completely abolished and patients whose pain location scores slowly reduced only between visits (not during visits) had outcomes that were as satisfactory as those of patients in the centralization category. The partial-reduction group achieved these same outcomes over a slightly longer treatment period. This finding raises the question of whether the CP is an important predictor of a favorable response to conservative care in the acute episode when compared with the natural history of acute low back pain. Nachemson 18 has stated that the majority of patients will improve with no treatment within the first few weeks after onset. The real value of the CP appears to be in documenting its absence, i.e., peripheralization or no change in pain location during treatment (noncentralization pain pattern). In the current study, if the patient s most distal pain location score increased or did not change by the seventh treatment visit, further improvement in pain intensity and perceived function was not found. This is important when considering the management of acute spinal syndromes. Pain intensity and functional limitations are documented outcome measures monitoring the impact of spinal syndromes on each patient. 26 Patients who fail to show reduced pain and improved function during the first 4 weeks of conservative treatment are at higher risk for the development of chronic disability. 19,20 Recent guidelines have suggested that these patients receive additional medical evaluation for either physical or nonphysical factors, which could be delaying quick resolution of the present episode. 1 Limitations The research project reported here was a descriptive study investigating the CP in patients with acute spinal pain who were referred to physical therapy. Prospective analysis tracked the occurrence and relation of different pain pattern categories with treatment responses. There was no attempt to randomize different treatment trials, a control group was not present, and treatment approaches were not specifically controlled (only guided ) and varied depending on pain location response to mechanical evaluation. In addition, symptom location was different across pain pattern groups. Specifically, more patients with leg/arm pain were classified in the partial reduction category. Therefore, variations in treatment and initial symptom location can be considered confounding factors when investigating associations between pain category and outcomes at discharge in this study. The predictive value of centralization could not be assessed using the design, methods, and results in this study. There was no longitudinal patient follow-up evaluation to document subsequent outcomes. It is not known if the treatment responses observed in this project will be maintained over time. Analyzing centralization as a risk factor in patients with acute spinal pain for the development of chronic disability remains an important issue and merits future investigation. Conclusions Centralization is a clinically induced phenomenon observed during McKenzie mechanical assessments and treatments. Centralization results in rapid and proximal changes in pain location throughout treatment. Al-

8 Centralization Phenomenon Werneke et al 683 though common, centralization does not occur as frequently as previously reported for patients with acute spinal syndromes. Previous studies did not adequately differentiate between the CP and the natural reduction or abolition of pain that can occur with the passage of time. By precisely documenting changes in pain location before and after the initial evaluation and each subsequent visit, three distinct pain pattern categories were identified. Centralization and the partial-reduction pain patterns are equally common and associated with good treatment results. If a reduction in pain location score was not observed by the seventh treatment visit, no further improvements in pain and perceived function were noted. Patients categorized into the noncentralization group appeared to be at higher risk for chronic disability and should receive multidisciplinary evaluation in conformity with recent guidelines for the management of acute spinal syndromes. Acknowledgment The authors thank Marybeth Behringer PT, Dip MDT, clinical manager and supervisor, for supporting all phases of this research project until its completion, and Charles Moulton PT, Dip MDT, Patrick Duffy, PT, Cred MDT, and Paul Weiss, PT, Cred MDT, for their assistance with data collection and support for the overall research efforts. References 1. Agency for Health Care Policy and Research. Acute low back problems in adults, clinical practice guide No 14. Rockville, MD: U.S. Department of Health and Human Services, 1994: Cohen J. Statistical Power Analysis for the Behavior Sciences. New York, NY: Academic Press Inc, Delitto A, Cibulka M, Erhard R, Bowling RW, Tenhula JA. Evidence for use of an extension-mobilization category in acute low back syndrome: A prescriptive validation pilot study. Phys Ther 1993;73: Donelson R, Aprill C, Medcalf R, Grant W. A prospective study of centralization of lumbar and referred pain: A predictor of symptomatic discs and annular competence. Spine 1997;22: Donelson R, Grant W, Kamps C, Medcalf R. Pain response to sagittal endrange spinal motion: A prospective, randomized, multi-centered trial. Spine 1991;16(Suppl):S Donelson R, Grant W, Kamps C. Pain response to sagittal end-range spinal motion: A multi-centered, prospective, randomized trial. Presented at the International Society for the Study of the Lumbar Spine, Heidelberg, Germany, Donelson R, Silva G, Murphy K. Centralization phenomenon: Its usefulness in evaluating and treating referred pain. Spine 1990;15: Erhard R, Delitto A, Cibulka M. Relative effectiveness of an extension program and a combined program of manipulation and flexion and extension exercises in patients with acute low back syndrome. Phys Ther 1994;74: Fairbank JCT, Couper J, Davies JB, O Brian JP. The Oswestry low back pain disability questionnaire. Physiotherapy 1980;66: Fleiss JL. Measuring nominal scale agreement among many raters. Psychol Bull 1971:76; Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity: A comparison of six methods. Pain 1986;27: Karas R, McIntosh G, Hall H, Wilson L, Melles T. The relationship between nonorganic signs and centralization of symptoms in the prediction of return to work for patients with low back pain. Phys Ther 1997;77: Kazis LE, Anderson JJ, Meenan RF. Effect sizes for interpreting changes in health status. Med Care 1989;27(Suppl):S Long A. The centralization phenomenon: Its usefulness as a predictor of outcome in conservative treatment of low back pain: A pilot study. Spine 1995; 20: McCombe P, Fairbank J, Cockersole B, Pynsent P. Reproducibility of physical signs in low-back pain. Spine 1989;14: McKenzie R. The Cervical and Thoracic Spine: Mechanical Diagnosis and Therapy. Waikanae, New Zealand: Spinal Publications Ltd, 1990: McKenzie R. The Lumbar Spine: Mechanical Diagnosis and Therapy. Waikanae, New Zealand: Spinal Publication Ltd, 1981: Nachemson AL. The natural course of low back pain. In: White AA, Gordon SL, eds. Symposium on Idiopathic Low Back Pain. St. Louis, MO: CV Mosby Company, 1982: Nordin M, Skovron ML, Hiebert R, et al. Early predictors of delayed return to work in patients with low back pain. J Musculoskeletal Pain 1997;5: Roland M, Morris R. A study of the natural history of low-back pain: Part 1. Development of a reliable and sensitive measure of disability in low back pain. Spine 1983;8: Spratt K, Lehmann T, Weinstein J, Sayre H. A new approach to the low back physical examination: Behavioral assessment of mechanical signs. Spine 1990; 15: Stankovic R, Johnell O. Conservative treatment of acute LBP: A prospective randomized trial: McKenzie method of treatment vs. patient education in mini back school. Spine 1990;15: Stratford PW, Binkley JM, Riddle DL. Health status measures: Strategies and analytic methods for assessing change scores. Phys Ther 1996;76: Sufka A, Hauger B, Trenary M, et al. Centralization of low back pain and perceived functional outcome. J Orthop Sports Phys Ther 1998;27: Uebersax JS. A generalized kappa coefficient. Educational and Psychological Measurement 1982:42; Von Korff M, Saunders K. The course of back pain in primary care. Spine 1996;21: Vernon H, Mior S. The Neck Disability Index: A study of reliability and validity. J Manip Phys Ther 1991;14: Williams M, Hawley J, McKenzie R, Wijmen P. A comparison of the effects of two sitting postures on back and referred pain. Spine 1991;16: Address reprint requests to Mark Werneke, MS, PT, Dip. MDT NovaCare at Southern Ocean Center For Health 730 Lacey Road Forked River, NJ 08731