J Manipulative Physiol Ther. 2016 (Mar); 39 (3): 141–149 ~ FULL TEXT
Karin E. Mantel, MChiroMed, Cynthia K. Peterson, RN, DC, MMedEd, and
B. Kim Humphreys, DC, PhD
Chiropractic Medicine Doctoral Student,
Chiropractic Medicine Department,
Faculty of Medicine,
University of Zürich, Switzerland.
OBJECTIVE: The purpose of this study was to compare improvement rates in patients with low back pain (LBP) undergoing chiropractic treatment with 0-2 weeks vs 2-4 and 4-12 weeks of symptoms.
METHODS: This was a prospective cohort outcome study with 1-year follow-up including adult acute (symptoms 0-4 weeks) LBP patients. The numerical rating scale for pain (NRS) and Oswestry questionnaire were completed at baseline, 1 week, 1 month, and 3 months after starting treatment. The Patient Global Impression of Change (PGIC) scale was completed at all follow-up time points. At 6 months and 1 year, NRS and PGIC data were collected. The proportion of patients reporting relevant "improvement" (PGIC scale) was compared between patients having 0-2 and 2-4 weeks of symptoms using the ?(2) test at all data collection time points. The unpaired t test compared NRS and Oswestry change scores between these 2 groups.
RESULTS: Patients with 0-2 weeks of symptoms were significantly more likely to "improve" at 1 week, 1 month, and 6 months compared with those with 2-4 weeks of symptoms (P < .015). Patients with 0-2 weeks of symptoms reported significantly higher NRS and Oswestry change scores at all data collection time points. Outcomes for patients with 2-4 weeks of symptoms were similar to patients having 4-12 weeks of symptoms.
CONCLUSION: The time period 0-4 weeks as the definition of "acute" should be challenged. Patients with 2-4 weeks of symptoms had outcomes similar to patients with subacute (4-12 weeks) symptoms and not with patients reporting 0-2 weeks of symptoms.
KEYWORDS: Chiropractic; Low Back Pain; Treatment Outcome
From the FULL TEXT Article:
Mechanical, nonspecific low back pain (LBP) remains a very common condition in developed countries. Low back pain in general has a point prevalence between 15% and 30%.  For the individual, it means a painful dysfunction with a lifetime prevalence of about 70%.  From these patients, it has been stated that two-thirds do not completely recover after 1 year  and 7.7% develop chronic pain. [3, 4] Furthermore, not only is this a problem in terms of the individual suffering, but also their social environment and workplace have to bear their burden. For the industrialized countries, LBP is one of the most costly conditions and makes up a considerable amount of our health care costs. Bevan et al  estimated that approximately CHF 2–4 billion per year is spent for all musculoskeletal disorders in Switzerland, where LBP is a large part of this total sum. A German study calculated an average annual cost for an LBP patient of €1322.  These costs are divided into direct costs (46%)  by using health care services such as therapeutic procedures, rehabilitation, and hospital and indirect costs (54%) [1, 6] due to work absenteeism and early retirement.  With ongoing duration of symptoms, indirect costs rise, and therefore, mean costs for a chronic LBP patient are twice as high as those for an acute patient.  In addition, there are also nonfinancial costs including losing participation in social activities and family.
The management of recent-onset LBP is summarized by Koes et al  as including manual therapy, medication (paracetamol, nonsteroidal anti-inflammatory drugs, muscle relaxants, opioids), patient education and reassurance, some specific exercises, and discouragement of bed rest. Manipulation as an intervention for treating LBP has shown to be more effective than placebo  and, in particular, recommended for patients with more acute symptoms. [9–11] There are several studies on the effectiveness of spinal manipulation for LBP, [9–13] but they sometimes come to different conclusions. This is primarily due to the fact that LBP is not one homogenous condition and therefore not equally likely to respond to one therapeutic modality. 
Patients with mechanical LBP are a heterogeneous group, and it is reasonable to assume that they consist of subgroups. [12, 15–25] But these further categorizations for the so-called mechanical, activity-related, nonspecific, or idiopathic LBP are lacking. [15–18] If specific, homogeneous subgroups could be identified, this would lead to better understanding of each condition. Every single subgroup could further be matched to the best-suited therapeutic interventions [12, 14, 19, 20] and therefore should lead to better outcome.  A guideline for therapeutic decision making could be developed,  and respective patients could be identified and matched to the most appropriate treatment. [23–25] Furthermore, the patients with risk factors for inducing high health care costs could be detected.  Clinical research such as randomized clinical trials could enhance their power by using more homogenous subgroups by reducing bias or diluting the heterogeneity of the LBP condition.
One immediate priority in the search for subgroups of LBP patients and matching suitable treatments with specific subgroups is to determine with more certainty the natural history of acute, nonspecific LBP to develop a proper definition for this condition. It is difficult to observe the natural history of LBP, as it means the development of a condition in the absence of treatment.  But in many cases, treatment has been given which likely changes the natural history  to a clinical course. Generally speaking, an episode of LBP is for many a self-limiting condition which, in 90% of cases, recovers completely within 1 month.  However, the literature is confusing because there are no clear definitions of acute, low back, episodes, [28, 29] recurrence [30, 31] or recovery. Furthermore, recent studies show that acute LBP is more an episodic condition [32–36] with varying patterns but lacking an obvious end point. Within an episode, rapid improvement occurs in the weeks 1 to 4 [37, 38]; then a plateau is reached, with no further improvement seen after weeks 7 through 12. [37, 38] After 12 weeks, the level of pain and disability remains fairly constant,  but there occurs a tendency toward worsening,  probably due to a new episode. Kent et al  estimated that 30% of patients recover completely and 60% enter into a relapsing pattern. In addition, Hestbaek et al  point out that, often, outcome measures are selected for use which are easy to measure but which do not measure what they should, for example, return to work or cessation of medical consultations instead of real recovery.
New research projects are focusing on finding clearer definitions and subgroups of patients with LBP, especially in terms of duration and frequency of their symptoms. In terms of duration, the most commonly used categorization is dividing the time of pain from onset into acute (0–4 weeks), subacute (4–12 weeks), and chronic (>12 weeks).  But there are many other authors using different time cutoff points for the term acute including 72 hours,  1 week, [17, 33] 2 weeks, [10, 11, 41–43] 3 weeks, [37, 44] 6 weeks, [28, 45, 46] or even longer.
Some studies only distinguish between acute and chronic or acute and persistent. In addition, the strongest and most consistent predictor of clinical improvement over time is how quickly the LBP patient responds to treatment.  Rapid response to therapy means a more favorable short-term treatment outcome, namely, improvement at 1 week is better than at 2 weeks, and these patients do better than those who take more than 2 weeks to improve.  However, this appears to be true mainly in acute LBP patients. But does it make a difference how long the patient has suffered in this “acute” LBP group in terms of their likelihood to improve and how fast they improve? In general, do patients with symptoms longer than 2 weeks do as well as patients with a shorter onset of symptoms? This is not clearly known. It is also important to better define what time frame best captures the truly “acute” patient just as the recently published National Institutes of Health research standards for chronic LBP patients have now provided clear criteria for identifying chronic patients. 
The purpose of this study was to investigate the outcomes of patients undergoing chiropractic treatment using the heterogeneous group of “acute” low back patients to determine whether or not the time frame of 0–4 weeks as the definition of “acute” is appropriate. The specific questions to be investigated are as follows:
(1) Are there differences in clinically significant improvement for patients with either highly acute
(0–2 weeks) or midacute (2–4 weeks) duration of symptoms undergoing chiropractic treatment?
(2) Is there a similarity between the groups of midacute (2–4 weeks) and subacute (4–12 weeks)
duration of symptoms undergoing chiropractic treatment?
The purpose of this present prospective cohort study with long-term follow-up was to compare high-acute (0–2 weeks), midacute (2–4 weeks), and subacute (4–12 weeks) LBP patients regarding their improvement under chiropractic care in an attempt to clarify and provide firmer time limits regarding the appropriate definition of “acute” symptoms. Indeed, the results from this study show that the outcomes from the patients with 2–4 weeks of symptoms (midacute) are more similar to patients with subacute (4–12 weeks) duration of symptoms than they are to the high-acute (0–2 weeks) patients.
To record the primary outcome measure “improvement,” we used the PGIC scale after 1 week, 1 month, 3 months, 6 months, and 1 year. The scale contains 7 possible answers. However, only “much better” (score of 1) and “better” (score of 2) were categorized as “improved.” For this study, being “slightly better” on the PGIC scale was not considered to be improvement to avoid reporting recovery which may not be clinically relevant. The PGIC scale not only integrates the change of pain severity but also asks patients to consider their overall quality of life and limitations in daily activities in the context of their back pain experience. This is very important because it allows every patient the opportunity to report for himself/herself if he/she experienced improvement according to what is important for him/her in his/her own daily life. [47, 53–55] For the secondary outcome measures, the NRS for pain and the Oswestry pain and disability index were used. 
Comparing the baseline characteristics between the high-acute (0–2 weeks) and midacute (2–4 weeks) patients showed a significantly higher percentage of midacute patients reporting below average health (P = .049). For all of the other baseline demographic information including patient sex, age, whether or not the patient smokes, whether or not the onset of pain was caused by trauma, and presence or absence of radiculopathy, there was no statistically significant difference in percentages between these 2 groups. However, when comparing the baseline characteristics from the midacute (2–4 weeks) with the subacute (4–12 weeks) patients, no statistically significant differences at all were found, meaning that the subacute LBP group also had a higher proportion of patients reporting below average health. This was in contradistinction to patients in the high-acute (0–2 weeks of symptoms) group who reported having better average health. Therefore, below average health at baseline, although still uncommon in this patient population, may be a factor for a less favorable outcome even for patients with more acute onset of symptoms.
Regarding clinically relevant “improvement” comparing high-acute (0–2 weeks) and midacute (2–4 weeks) LBP patients undergoing chiropractic treatment, a significantly higher percentage of high-acute patients reported they were much better or better on the PGIC scale at 1 week, 1 month, and 6 months (P ≤ .05). The only significant difference in the proportion of patients reporting clinically relevant “improvement” between the midacute (2–4 weeks) and subacute (4–12 weeks) patients was at 1 week, with a higher percentage of midacute patients reporting improvement. At the other data collection time points, there was no significant difference, again strongly suggesting that LBP patients with 2–4 weeks of symptoms are similar to patients usually categorized as “subacute” (ie, 4–12 weeks of symptoms).
The high-acute LBP patients also reported a significantly higher mean NRS change scores at all data collection time points compared with the midacute group, indicating that they experienced more pain reduction. Similarly, the Oswestry change scores for the high-acute LBP group also had significantly higher change scores at 1 week, 1 month, and 3 months (ie, at every time point for this outcome measure), indicating a better outcome in terms of their disability levels. This is not unexpected because it is well known that acute spine pain patients usually report higher baseline pain and disability levels compared with patients with a longer period of suffering. [47, 54, 55] However, when comparing the NRS change scores between midacute (2–4 weeks) and the subacute (4–12 weeks) LBP patients, the midacute group only had statistically significantly higher pain reduction at 1 week. At all the other time points, the NRS change scores were quite similar between these 2 groups. Although the Oswestry change scores showed that the midacute LBP patients did significantly better at 1 week, 1 month, and 3 months compared with the subacute LBP patients, this significant difference, particularly at 1 and 3 months, may have been due to the rather large sample sizes. By 1 and 3 months after the start of treatment, the mean Oswestry scores were already very low, and the actual difference in the Oswestry change scores between the 2 groups was only 1.3 points at 1 month. This likely represents a floor effect for the Oswestry questionnaire in this patient population. At 3 months, the difference in the change scores was only 2.7 points. Consequently, one could argue that although these differences were statistically significant, were they truly clinically relevant?
The results of this study demonstrate that patients with midacute (2–4weeks) duration of symptoms are more similar to subacute (4–12 weeks) LBP patients than to high-acute (0–2 weeks) LBP patients. Currently, the most commonly used definition of acute for research purposes is symptoms between 0 and 4 weeks,  which may not reflect an appropriate categorization for this what appears to be a heterogeneous LBP subgroup. The results from this current study support a few previous studies which also used 0–2 weeks as “acute” pain. [10, 11, 41–43] The reasons why the midacute LBP patients respond to treatment more similarly to subacute patients should be explored. Are they already developing neuroplastic changes in their brains relating to future chronicity? Should specific treatment interventions addressing specific psychosocial issues or catastrophizing be introduced earlier? Unfortunately, because the Oswestry questionnaire was used in this study because of its availability in German, issues of anxiety and depression could not be evaluated. The Bournemouth questionnaire, which has subsequently been translated and validated into German, would have been a better outcome measure for this group of patients.
There were varying numbers of patients included for the various data collection time points. This is due to the fact that strict date limits were set as to when each telephone call should be done. The largest amount missing is particularly at the first telephone interview after 1 week of receiving chiropractic care because of the narrow time window allowed to call all the patients. In addition, because this was an ongoing study, the collection time points at 6 months and 1 year after first chiropractic treatment were not reached for all patients at the time of data analysis.
It is relevant to mention that the Oswestry questionnaire is probably not the best outcome measure to record mechanical LBP. Originally, this questionnaire was designed for more severe and surgical LBP patients. From the maximum of 50 points, we had mean baseline scores of 15.6 for the high-acute, 12.4 for the midacute, and 10.4 for the subacute LBP patients. This suggests that the Oswestry questionnaire for this group of patients is less appropriate. Nevertheless, it was used in this study because there were translated and validated versions in German and French available. Since September 2013, there is now a German version of the Bournemouth questionnaire for LBP available,  which is more appropriate for this patient group, as it was developed from the biopsychosocial model and includes 7 relevant domains such as pain, physical disability, and some of the most important psychological aspects of being an LBP patient such as anxiety, depression, and social and work interference questions.
As this is a prospective outcomes study and not a randomized clinical trial, the outcomes of the patients cannot be definitively attributed to the chiropractic treatment.
Studies [58, 59] have already reported the disadvantage of using telephone interviews compared with written questionnaires or text messaging because they may have a false-positive effect on the outcomes. To diminish this effect, the data were collected by anonymous research assistants at the university to avoid direct contact between patients and their treating chiropractic office.
Furthermore, the 44 chiropractors contributing patients for this study are probably a population of chiropractors who are more interested in research than the average DC in Switzerland. Whether or not this might influence patient outcomes is unknown.
The time period 0–4 weeks as the definition of acute should be challenged. Patients with 2–4 weeks of symptoms had outcomes similar to patients with subacute (4–12 weeks) symptoms but not with patients reporting 0–2 weeks of symptoms.
The time period of 0–4 weeks of symptoms is too long
for an accurate definition of acute LBP.
Patients with 0–2 weeks of LBP were significantly more
likely to report clinically relevant improvement and
reduced pain and disability scores compared with
patients with 2–4 weeks of symptoms.
Patients with 2–4 weeks of symptoms report treatment
outcomes similar to patients with 4–12 weeks of symptoms
Return to the LOW BACK PAIN Section