If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The purpose of this study was to investigate recurrence rate and prognostic factors in a large population of patients with low back pain (LBP) up to 1 year after chiropractic care using standardized definitions.
Methods
In Switzerland, 722 patients with LBP (375 male; mean age = 44.5 ± 13.8 years) completed the Numeric Rating Scale for pain (NRS) and the Oswestry Disability Index (ODI) before treatment and 1, 3, 6, and 12 months later (ODI up to 3 months). Based on NRS values, patients were categorized as “fast recovery,” “slow recovery,” “recurrent,” “chronic,” and “others.” In multivariable logistic regression models, age, sex, work status, duration of complaint (subacute: ≥14 days to <3 months; chronic: ≥3 months), previous episodes, baseline NRS, and baseline ODI were investigated as predictors.
Results
Based on NRS values, 13.4% of the patients were categorized as recurrent. The recurrent pattern significantly differed from fast recovery in duration of complaint (subacute: odds ratio [OR] = 3.3; chronic: OR = 10.1). The recurrent and chronic pattern significantly differed in duration of complaint (chronic: OR = 0.14) and baseline NRS (OR = 0.75).
Conclusion
Recurrence rate was low in this LBP patient population. The duration of complaint before treatment was the main predictor for recurrence. The fact that even subacute duration significantly increased the odds for an unfavorable course of LBP is of clinical relevance.
Patients with low back pain had distinct clinical course patterns that were typically neither complete recovery nor constant pain. A latent class analysis of longitudinal data.
because the pattern of recurrent episodes is unpredictable and still not fully understood. Nevertheless, recurrent LBP episodes have a tremendous impact on physical and social functioning
Thus, the prevention of future relapse episodes is crucial.
Solid evidence about risk factors for recurrence of LBP is sparse because the majority of research has focused on prognostic factors for poor outcome, disability, or chronicity instead of recurrence. Those that did find prognostic factors for the recurrent course have reported conflicting information, most probably because a variety of definitions for recurrent LBP have been used. A systematic review concluded that among the studies in this field, only 38% used a specific but self-created definition for recurrence, whereas in 62% it was unclear how recurrence had been measured.
Moreover, most studies lacked a definition of recovery as part of the definition of recurrence, probably including patients with persistent pain instead of recurrent episodes.
Thus, the following consensus definition for a recurrence of an episode of LBP was reached in a modified Delphi approach: “return of LBP lasting at least 24 hours with a pain intensity of >2 on an 11-point Numeric Rating Scale (NRS) following a period of at least 30 days pain-free.”
Thus, the main aim of this study was to determine the amount of LBP recurrences using the consensus definition of recurrence, according to Stanton et al,
in patients up to 1 year after chiropractic care based on pain intensity (NRS). The second aim was to compare the recurrent patients to all other patients (grouped according to various trajectories) in terms of certain baseline factors and to investigate whether certain baseline factors increased the risk for LBP recurrence.
Methods
Study Design and Setting
This observational study is a secondary analysis of data that were collected between 2010 and 2014 in a prospective cohort study with 1-year follow-up. The Canton of Zürich ethics review board gave ethical approval for this study (EK-16/2009). All patients signed a written informed consent. All chiropractors throughout Switzerland were asked by the Swiss Chiropractic Association (ChiroSuisse) to recruit patients for this study and instructions were sent to all chiropractors by e-mail. No standardization of treatment methods or numbers was performed. Instead, chiropractors were requested to use their usual treatment method, because the aim of the study was to assess outcomes of routine chiropractic practice. Immediately before their first treatment, patients completed the NRS and the Oswestry Disability Index (ODI) in their respective chiropractic offices. The ODI was selected as the questionnaire of choice as it has been translated and validated in German and French
at the time data collection commenced. The ODI is made up of 10 sections, including 1 item on pain intensity and 9 questions on interference with daily activities such as sleeping, self-care, sex life, social life, and traveling. Each section can be answered on a scale from 0 to 5 with a total score of 50 that can be converted into a percentage. The NRS is an 11-point rating scale to assess patient’s pain intensity, ranging from 0 (“no pain”) to 10 (“the worst imaginable pain”).
Further baseline information such as patient’s age, sex, work status, duration of current complaint, and number of previous episodes was sent by the respective treating chiropractor to the research assistant. At intervals of 1, 3, 6, and 12 months after the initial treatment, the NRS and ODI were collected by trained research assistants from the coordinating university hospital through standardized telephone interviews (ODI up to 3 months). The research assistants were trained to do the interviews but did not know either the patient or the treating chiropractor. All patients were interviewed up to 12 months irrespective of whether or not they were still receiving chiropractic treatment.
Participants
Adult patients (≥18 years) with LBP of any duration who had not been treated with chiropractic therapy in the prior 3 months were included. Exclusion criteria were relative contraindications to chiropractic manipulative treatment, such as tumors, infections, inflammatory spondyloarthropathies, acute fractures, and severe osteoporosis. For this particular study, only patients with complete data sets for NRS at the respective points of time were included.
Variables
The outcome variable “recurrence” was defined in accordance with the consensus definition of recurrent LBP (LBP that occurred at least twice over the past year with each episode lasting at least 24 hours, with a pain intensity of >2 on a 11-point NRS and at least a 30-day pain-free interval in between episodes
Consequently, patients were defined as “recurrent” as follows: NRS ≤1 at a preceding assessment and NRS >2 at the consecutive assessment. In order to gain homogeneous groups to be compared with the recurrent patient group, the rest of the patient population was subdivided into the following subgroups: “fast recovery,” “slow recovery,” “chronic,” and “others” (for those patients who did not fit into any of these subgroups.) Patients were considered as “fast recovery” if NRS was ≤1 at 1, 3, 6, and 12 months. “Slow recovery” was defined as NRS ≤1 at 3, 6, and 12 months or NRS ≤1 at 6 and 12 months.
) at 1, 3, 6, and 12 months. Any data that did not match these definitions were defined as “others.” The latter included patients who were fluctuating or stable on a low pain intensity level (NRS <3.5) but never fell below NRS ≤1, thus not qualifying for “recurrent.”
The parameters sex, age, work status, duration of current complaint, previous LBP episodes, mean baseline NRS, and mean baseline ODI were investigated as possible predictors for recurrence. As for the categorization of duration of current complaint, there is agreement that the differentiation between acute and chronic pain should be 3 months.
Spanish Back Pain Research Network. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Spanish Back Pain Research Network. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Nevertheless, it has been reported that regardless of pain intensity, as soon as pain lasts for more than 14 days prognosis is poor and the risk for chronic disability starts to rise.
Spanish Back Pain Research Network. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Exploring the definition of acute low back pain: a prospective observational cohort study comparing outcomes of chiropractic patients with 0-2, 2-4, and 4-12 weeks of symptoms.
Consequently, the duration of complaint was categorized in the present study into acute (<14 days), subacute (≥14 days up to <3 months), or chronic (≥3 months). Work status was categorized into “working” (full time, part time, housewife), “retired,” and “not working” (unemployed, incapable to work).
Statistical Methods
To calculate the amount of recurrences, descriptive statistics were used. To compare the recurrent subgroup to the other subgroups, 1-way analysis of variance with post hoc Bonferroni tests were used for the continuous variables (age, previous episodes, baseline NRS, and baseline ODI). As for the categorical factors (sex, work status, and duration of complaint), χ2 tests were used and Cramer’s V was calculated to measure strength of the associations. Post hoc, the recurrent subgroup was compared with each of the other subgroups with χ2 tests. To investigate the potential risk factors for recurrence, the subgroup “recurrent” was compared with each subgroup using multivariable logistic regression analysis. Thus, 4 multivariable logistic regression analyses with “recurrent” (equal to 1 in the regression model) and “nonrecurrent” (either “fast recovery,” “slow recovery,” “chronic,” or “others”; equal to 0 in the regression model) as dependent variable were run. With a view to the independent variables, there was a large number of missing values for the parameter “previous episodes” (N = 250). Thus, for each logistic regression analysis, 2 models were run. Model 1 included 6 independent variables (age, sex, work status, duration of complaint, baseline NRS, and baseline ODI) as prognostic factors, and model 2 included 7 independent variables (“previous episodes” as an additional factor) because the large amount of missing values might lead to biased conclusions. The models were checked for multicollinearity by running linear regression analyses with the same predictors
and calculated the variance inflation factors (VIF) and the tolerance statistic. Variance inflation factor values >10 and tolerance values <0.2 were regarded as critical.
Receiver operating characteristic curves were used as measures of goodness of fit. The area under the curve (AUC) ranging from 0.5 to 1.0 was analyzed; larger values indicated a better fit. For this study, the following cutoff points for the AUC values were used: 0.5 to 0.6 indicates fail, 0.6 to 0.7 is poor, 0.7 to 0.8 is fair, 0.8 to 0.9 is good, and 0.9 to 1 is excellent.
In general, P values <.05 were considered statistically significant. When the recurrent subgroup was compared with each of the other 4 subgroups (in the post hoc tests after χ2 and in the logistic regression models), the level of significance was adapted to P < .05/4 = 0.0125 to correct for multiple comparisons. Data analysis was performed using SPSS Version 22.0 (IBM Corp., Armonk, NY).
Results
From the original database, 722 patients (375 male; mean age = 44.5 ± standard deviation 13.8 years) with complete NRS data sets could be selected for this study. Mean baseline NRS was 5.7 (±2.2) and the mean baseline score of the ODI was 12.9 (±7.8) out of 50 points. Most participants experienced a first (44.3%) or second episode (28.8%) of LBP. Duration of complaint at baseline was 36.0% acute (<14 days), 33.8% subacute (≥14 days up to <3 months), and 30.3% chronic (≥3 months). Based on the NRS values, 13.4% of the patients were classified as recurrent, 39.9% were considered recovered (fast or slow), and 7.6% were chronic. The rest did not fit in any of these categories and were classified as “others” (Table 1). Figure 1 shows the course of NRS values over the whole study period of 12 months for all subgroups.
Table 1Descriptive Statistics of the Subgroups
Fast Recovery
Slow Recovery
Recurrent
Chronic
Others
N
152 (21.1%)
136 (18.8%)
97 (13.4%)
55 (7.6%)
282 (39.1%)
Age, mean (±SD)
43.80 (±13.50)
41.09 (±13.51)
44.35 (±14.08)
50.35 (±13.53)
45.36 (±13.77)
Sex, m/f
95/57
69/67
50/47
20/35
141/141
Work status: working/off work/retired
137/4/10 missval: 1
119/6/9 missval: 2
75/7/13 missval: 2
40/3/11 missval: 1
232/18/31 missval: 1
Duration of complaint: acute/subacute/chronic
94/44/13 missval: 1
40/69/25 missval: 2
32/34/30 missval: 1
6/17/31 missval: 1
86/78/118
Previous episodes mean (±SD)
1.39 (±3.09) missval: 33
0.97 (±1.54) missval: 37
1.46 (±2.42) missval: 34
2.40 (±4.64) missval: 30
1.82 (±2.60) missval: 116
ODI at baseline mean (±SD)
12.56 (±8.25)
12.27 (±7.10)
14.50 (±8.56)
14.88 (±7.16)
12.54 (±7.52)
NRS at baseline mean (±SD)
5.46 (±2.45)
5.65 (±2.00)
5.87 (±2.28)
6.73 (±1.77)
5.67 (±2.22)
NRS at 1 months mean (±SD)
0.31 (±0.45)
3.30 (±1.87)
1.90 (±1.84)
5.54 (±1.27)
3.09 (±1.92)
NRS at 3 months mean (±SD)
0.19 (±0.38)
1.63 (±1.67)
1.75 (±1.89)
5.26 (±1.18)
2.68 (±1.74)
NRS at 6 months mean (±SD)
0.16 (±0.35)
0.30 (±0.45)
2.33 (±2.45)
5.52 (±1.25)
2.87 (±1.86)
NRS at 12 months mean (±SD)
0.14 (±0.34)
0.31 (±0.46)
3.22 (±2.24)
5.19 (±1.23)
2.57 (±1.99)
missval, missing values; N, number of patients; NRS, Numeric Rating Scale; ODI, Oswestry Disability Index; SD, standard deviation.
The groups significantly differed in sex (χ2 [4] = 12.65, P = .013, Cramer’s V = 0.13), work status (χ2 [8] = 15.86, P = .044, Cramer’s V = 0.11), duration of complaint (χ2 [8] = 113.93, P < .001, Cramer’s V = 0.28), age (F [4,717] = 5.00, P = .001), NRS at baseline (F [4,717] = 3.57, P = .007), and ODI at baseline (F [4,717] = 2.39, P = .050). The recurrent subgroup consisted of approximately equal numbers of acute, subacute, and chronic patients, which significantly differed from those in the fast recovery (majority of patients acute: χ2 [2] = 27.63, P < .001, Cramer’s V = 0.33) and the chronic (majority of patients chronic: χ2 [2] = 12.71, P = .002, Cramer’s V = 0.29) subgroups. In all other investigated parameters, the recurrent subgroup did not differ from the other subgroups (Table 1).
Previous episodes were not a predictor for recurrence, and model 1 is presented in Table 2. No multicollinearity was detected in the regression models (VIF values: 1.02-1.72; tolerance statistic: 0.58-0.98). The recurrent subgroup could be fairly discriminated from the fast recovered (AUC = 0.75) and the chronic (AUC = 0.76) subgroup but only poorly from the slow recovered subgroup (AUC = 0.65) and the subgroup classified as “others” (AUC = 0.61). Therefore, duration of complaint was the most important factor. Patients with a subacute problem were at higher odds than patients with an acute problem to experience recurrence instead of fast recovery (OR = 3.31). Patients presenting with a chronic problem were at higher odds than patients with an acute problem to follow a recurrent instead of a fast recovery pattern (OR = 10.10) and were in turn at lower odds to experience a chronic instead of a recurrent course (OR = 0.14). In addition, high pain intensity at baseline slightly increased the odds for a chronic pattern instead of a recurrent course (Table 2).
In this population of patients with LBP undergoing chiropractic care, recurrences during the first year were sparse. The regression models could distinguish the recurrent patients from those who recovered quickly and those who became chronic, mainly based on the duration of the current complaint. The chronic patients and those with a subacute problem (≥14 days) at baseline were at considerably higher odds to experience recurrence rather than a fast recovery. Also, a chronic, but not a subacute problem at baseline increased the odds for a chronic course. High pain intensity at baseline more likely resulted in a chronic rather than in a recurrent course.
Using a standardized definition of recurrence, a recurrence rate of 13.4% was low. Recurrence was defined as NRS ≤1 at a preceding assessment and NRS >2 at the consecutive assessment, which only includes the aspect of pain intensity. However, because of lack of information, the aspect of pain duration as required in the definition by Stanton et al
(ie, “return of LBP lasting at least 24 hours with a pain intensity of >2 on an 11-point NRS following a period of at least 30 days pain-free”) was not included. In a similar study, Stanton et al
found a recurrence rate of 33%. They asked their patients at 3 and 12 months after recovering from an acute LBP episode whether they had a recurrence of LBP lasting for more than 24 hours. This definition included the duration aspect of the recurrence definition but disregarded pain intensity. Thus, patients with NRS ≤2 in the new LBP pain episode were classified as recurrent, which might explain the lower recurrence rate in the present study. In fact, the present study might underestimate the true recurrence rate because there were only 5 assessments during the follow-up and some recurrences might have been missed. Nevertheless, the comparable study by Stanton only assessed LBP at 2 points in time (at 3 and 12 months after recovery) and found a higher recurrence rate. Similarly, a recent study by Hancock et al
used the definition of recurrence as “return of LBP lasting at least 24 hours with a pain intensity of 3 or more on a 0-10 numerical pain rating scale.” This definition corresponded to the first part of the consensus definition by Stanton et al
but did not, to our understanding, consider the second part (“following a period of at least 30 days pain-free”). Thus, in contrast to the present study, that definition of recurrence did not include recovery. This implies that patients with fluctuating and persistent pain might be included, which was claimed to be a lack of many studies on LBP recurrence.
reported a recurrence rate of 54%, which corresponds well to the “recurrent” subgroup plus the “others” (fluctuating and persistent mild pain) subgroup in our study. Obviously, using a stringent definition considerably reduces the amount of LBP recurrences. However, this should not hide the fact that the majority of LBP patients in the present study did not become pain free within a year. About 60% (classified by NRS) of the patients either belonged to the recurrent, chronic or “others” (fluctuating or persisting mild pain) subgroup, indicating that they were not constantly pain free. On the other hand, less than 10% (8%) developed severe chronic pain. These results are in line with a recent study by Kongsted et al,
Patients with low back pain had distinct clinical course patterns that were typically neither complete recovery nor constant pain. A latent class analysis of longitudinal data.
which found, using latent class analysis, that most patients do not become pain free within a year and a few report constant severe pain. The finding of the present study that the recurrent pattern could be distinguished from the fast-recovering and the chronic courses but not from those patients who slowly recovered or were classified as “others” (fluctuating or persistent low pain) could either reflect that these groups differed in factors that were not investigated in the present study or that there were actually 3 main subgroups, namely a fast recovery subgroup (about 21%), a chronic persistent moderate or severe pain subgroup (about 8%), and a subgroup of patients who experience LBP as a chronic problem but not constantly on a high level (about 71%). Consequently, it might be hypothesized that at least some of the patients who experienced slow recovery in the present study might follow a recurrent pattern later, beyond the follow-up period for the present study. Failure to improve in the initial period after care seeking has previously been reported to negatively affect the prognosis of LBP.
Duration of current complaint emerged from this study as the most powerful factor to distinguish between recurrence and fast recovery and chronic course, respectively. Interestingly and in accordance with other studies,
Spanish Back Pain Research Network. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Exploring the definition of acute low back pain: a prospective observational cohort study comparing outcomes of chiropractic patients with 0-2, 2-4, and 4-12 weeks of symptoms.
the prognosis deteriorated as early as 2 weeks. Patients presenting with a subacute problem, lasting for more than 14 days at baseline, were at higher odds for a recurrent course, whereas the odds for a chronic course were higher only for patients presenting with a chronic problem (≥3 months) at baseline. Downie et al
reported that pain duration of more than 5 days was a factor that negatively affects prognosis. Similarly, duration of the current episode emerged as the most consistent factor for prognosis after 1 year in a study by Bekkering et al
Consequently, every attempt should be made in this initial phase to help the patients experience improvement. This includes not only adequate physical treatment but also addressing coping behaviors,
Information and advice to patients with back pain can have a positive effect. A randomized controlled trial of a novel educational booklet in primary care.
Interestingly, the number of previous episodes did not emerge as predictor for recurrence from the present study, which is in contrast to the studies by Stanton et al
was 1.04, which indicates a limited influence of the previous episodes on recurrence. Still, this discrepancy might be explained by the differences in the definition of “recurrence” that are reflected in the recurrence rates. Furthermore, the patient sample in the study by Hancock et al
consisted mainly of acute patients, whereas the present study included acute, subacute, and chronic patients about equally. Lastly, higher pain intensity at baseline was associated with a chronic rather than a recurrent course, when assessed by NRS. Accordingly, pain intensity at baseline identified persistent high pain but not fluctuating pain from rapid recovery in the study by Downie et al.
In a review focusing on early prognosis of LBP, all reviewed studies identified patient rating of pain as an important factor for predicting a negative outcome.
The present study used, as far as possible, stringent definitions for recurrence and recovery, which is a clear strength. However, information about pain duration of a recurrent episode was not available, and thus, the definition for recurrence could only be fulfilled in terms of pain intensity, not duration. Furthermore, psychosocial factors have not been investigated and some recurrences might have been missed because the number of assessments was limited, which are clear limitations to the study. Although the results might be refined by more numerous assessments, some of the main results were comparable to studies that used weekly assessments.
Patients with low back pain had distinct clinical course patterns that were typically neither complete recovery nor constant pain. A latent class analysis of longitudinal data.
Lastly, the presented data do not reflect the natural course of LBP as the patients underwent chiropractic treatment, which was not standardized for methods or number of consultations. However, this was purposely chosen because it better reflects clinical practice. Nevertheless, future studies on LBP recurrence using a standardized definition should include psychosocial assessments and should further investigate how the recurrent pattern differs from the fluctuating course and from slow recovery. For example, patients who experienced fast and slow recovery from an LBP episode should be followed for longer than 1 year in order to compare the recurrence rates. The findings of this study were only for people from Switzerland and therefore may be limited to this population. Similar studies should be completed in other countries to determine if these data can be extrapolated to other populations.
Conclusion
The recurrence rate of LBP using a stringent definition of recurrence was found to be low in this chiropractic LBP patient population. Nevertheless, the vast majority of patients were not pain free after 1 year. The recurrent course could be distinguished from the fast recovering and chronic patterns, but the differences with respect to the others subgroups were minor. The duration of complaint before treatment was the main predictor for recurrence. Of importance, a subacute duration, defined in the present study as longer than 14 days, significantly increased the odds for an unfavorable course of LBP, which is of clinical relevance.
Funding Sources and Conflicts of Interest
The following were funding sources for this study: Balgrist Hospital Foundation, Chirosuisse Foundation, Uniscientia Foundation, and European Academy for Chiropractic. No conflicts of interest were reported for this study.
Contributorship Information
Concept development (provided idea for the research): B.K.H.
Design (planned the methods to generate the results): B.W., B.K.H.
Supervision (provided oversight, responsible for organization and implementation, writing of the manuscript): B.W.
Data collection/processing (responsible for experiments, patient management, organization, or reporting data): B.K.H.
Analysis/interpretation (responsible for statistical analysis, evaluation, and presentation of the results): B.W., C.K.
Literature search (performed the literature search): C.K., B.W.
Writing (responsible for writing a substantive part of the manuscript): C.K., B.W.
Critical review (revised manuscript for intellectual content, this does not relate to spelling and grammar checking): C.K., B.W., B.K.H.
Practical Applications
•
Duration of complaint before treatment is an important predictor for recurrence of LBP.
•
A duration of complaint of >14 days significantly increases the odds for an unfavorable course of LBP.
References
Hoy D
March L
Brooks P
et al.
The global burden of low back pain: estimates from the Global Burden of Disease 2010 study.
Patients with low back pain had distinct clinical course patterns that were typically neither complete recovery nor constant pain. A latent class analysis of longitudinal data.
Spanish Back Pain Research Network. The transition from acute to subacute and chronic low back pain: a study based on determinants of quality of life and prediction of chronic disability.
Exploring the definition of acute low back pain: a prospective observational cohort study comparing outcomes of chiropractic patients with 0-2, 2-4, and 4-12 weeks of symptoms.
Information and advice to patients with back pain can have a positive effect. A randomized controlled trial of a novel educational booklet in primary care.
30121-X&id=gr1.jpg){kind=link}