Journal Logo


Award-Winning Study on Back Pain Flares Identifies Risk Factors and Potential Treatment Approaches

doi: 10.1097/01.BACK.0000767572.09950.2a
  • Free

A variety of evidence suggests that the traditional classification of low back pain (LBP) into acute, subacute, and chronic categories is inaccurate and unrealistic for many people with low back pain. These are timeline-based definitions. Yet multiple studies have identified patterns of symptoms and pain trajectories that defy these simplistic classifications.

Healthcare providers all over the world attempt to fit patients into these timeline-based categories. But in many cases, this is akin to asking someone to wear a suit of clothes that simply does not fit.

Back pain often waxes and wanes independent of timelines and can follow a variety of distinctive short-term and long-term trajectories.

Recent years have seen renewed interest in back pain flares—and whether addressing these short-term exacerbations and trajectories might reduce the overall burden of back pain for many people. Back pain flares may drive care seeking and healthcare utilization. However, flares are not currently addressed specifically in back pain guidelines and treatment pathways in most healthcare settings.

This is a vibrant research area. Many urgent questions surround back pain flares. Do they have their own defined set of risk factors? Are there treatment approaches that might reduce their burden? Is it most important to address biomechanical triggers, broader psychosocial issues, health and lifestyle issues, or all of the above? And how do the short-term trajectories of back pain flares mesh with and influence the long-term trajectories of back pain?

And as with any potential new back pain treatment target, careful research will have to consider whether focusing on flares results in any adverse effects: increased patient anxiety, misdiagnosis, overtreatment, mistreatment, and low-value treatment.

An award-winning new study suggests that back pain flares may indeed have a distinctive group of risk factors. The International Society for the Study of the Lumbar Spine (ISSLS) recently awarded its 2021 Prize in Clinical Science to an innovative study on back pain flares by Nathalia Costa and colleagues. Costa is a post-doctoral research fellow at the University of Queensland in Brisbane, Australia and a post-doctoral research associate in the Sydney School of Public Health at Sydney University.

Senior author Paul W. Hodges, DSc, MedDr, PhD, of the University of Queensland presented the new study at the virtual annual meeting of ISSLS in early June.

The study highlighted the potential importance of flares. “LBP flare has been largely ignored but is more reflective of the LBP experience than conventional definitions of acute, sub-acute and chronic LBP. This study highlights risk factors for flare, and that these differ depending on whether flare is defined by pain alone (PDF) or a broad multidimensional definition (SRF). Potential targets to reduce the intensity/frequency of LBP flares are identified, with strong indication for the potential role of sleep intervention to mitigate LBP flare risk,” according to Costa et al. “These findings highlight potentially modifiable factors to target with interventions.”

If sleep is indeed an important risk factor, this would be an encouraging development. According to recent clinical studies in other areas of musculoskeletal medicine, insomnia and other sleep problems can be treated effectively and cost-effectively among people with and without pain problems. And sleep interventions involving a cognitive behavioral approach do not require face-to-face contact with a healthcare provider—and do not involve risky drug therapies. They can be delivered at low cost via telephone, the Internet, and other remote treatment approaches.

The Context of This Research Effort

In presenting the paper at the ISSLS meeting, senior author Hodges described the context of this research effort.

“Traditionally we think about back pain as an acute, subacute, or chronic condition. based on time. And when we think about risk factors we think about factors that lead to a new episode of back pain and these are largely biomechanical. Or factors that increase the transition to chronicity, which are largely psychosocial,” according to Hodges.

“Recently people have started to reconceptualize back pain as an ongoing fluctuating condition with periods of increased symptoms which we can think of as flares.”

So the new study was an effort to identify the features that might increase the probability of experiencing a flare.

“The idea being that if we can identify those factors we might be able to reduce the occurrence of flares or reduce their severity,” according to Hodges.

“So transient exposures are what we are interested in, transient changes in psychological, social, behavioral, or biomechanical factors that might be associated with the increased probability of having a flare,” he added.

The research team was particularly interested in the role of two potential risk factors: sleep and physical activity. “Pain can interfere with sleep but it is also plausible that a night of poor sleep could cause an increase in pain,” said Hodges. “A transient change in sleep might not be enough to cause a new episode of low back pain but may be enough to cause it to flare.”

Physical activity, of course, has a complex relationship with back pain. Studying how pain fluctuates in response to transient changes in physical activity may help clarify this relationship.

Flares Have Been Poorly Defined

By way of background, researchers and healthcare providers have long acknowledged the potential importance of back pain flares. However, flares have been poorly defined, and, until recently, poorly studied.

Over the past several years Hodges et al. and other research groups have been making up for lost time.

This research group has conducted consensus processes involving experts and consumers to create a working definition of a back pain flare, such as the one employed in the award-winning study. “A back pain flare is a worsening of your condition that lasts from hours to weeks, is difficult to tolerate, and generally impacts your activities and emotions.” (See Costa et al., 2019)

They have documented that back pain flares should not be defined solely in terms of pain. Some self-reported back pain flares don't involve any significant increase in pain intensity. “Recent work we have done suggests that flare is more than an increase in pain. It is a multidimensional experience,” Hodges commented. (See Costa et al., 2021b.) A broad array of psychosocial and other factors come into play in eliciting flares.

Hodges and other researchers have also documented that back pain flares have a complex influence on people's lives. (See Tan et al., 2019.) (See Table I, page 93.)

Table I. - Studies on Back Pain Flares from a Research Group Centered in Australia*
Study Selected Points
Costa N et al., Low back pain flares: how to they differ from an increase in pain? (See Costa et al., 2021b.) “These findings highlight that when individuals with LBP consider they have a flare, they do not always have greater than average pain, but have worse psychosocial features. This emphasizes that flare has broader dimensions than pain alone.”
Costa N et al. What are the risk factors for low back pain flares and does this depend on how flare is defined? (See Costa et al., 2021a.) “LBP flare has been largely ignored but is more reflective of the LBP experience than conventional definitions of acute, subacute, and chronic LBP. This study highlights risk factors for flare and that these differ depending on whether flare is defined by pain alone (PDF) or a broad multidimensional definition.” Sleep stood out as a risk factor and a potential therapeutic target.
Costa N et al., What triggers a low back pain flare? A content analysis of individuals' perspectives. (See Costa et al., 2020.) Most participants identified biomedical (84.8%) triggers, endorsing physical/biological factors to explain the flare occurrence. Themes included active movements (35% of participants), static postures (28.1%), overdoing a task (5.3%), biomechanical dysfunction (4.4%), comorbidities (4%), lack of exercise (3.3%), work (1.8%), and medications (1.5%).” Only a small minority of participants identified non-biomedical factors as triggers. So, in some respects, the views of people with back pain conflict with the emerging research on flares, which suggest these are multidimensional issues.
Tan D et al., Impact of flare-ups on the lives of individuals with low back pain: A qualitative investigation.
(See Tan et al., 2019.)
“Results suggest the importance of considering that LBP flare-ups impact individuals' lives in a complex manner including psychosocial and functional effects. Clinicians should consider this complexity in their interactions with, and management of, patients with LBP flare-ups.”
Costa N et al. A definition of “flare” in low back pain: a multiphase process involving perspectives of individuals with low back pain and expert consensus. (See Costa et al., 2019.) “This multiphase study produced a definition of LBP flare that distinguishes it from other LBP fluctuations, represents consumers' views, involves expert consensus, and is understandable by consumers in clinical and research contexts:
‘A flare-up is a worsening of your condition that lasts from hours to weeks that is difficult to tolerate and generally impacts your usual activities and/or emotions.’”
Costa N et al., How is symptom flare defined in musculoskeletal conditions: A systematic review. (See Costa et al., 2018.) “Although some flare definitions began simply as ‘symptom worsening’ or ‘change in treatment,’ most evolved into multidimensional definitions that include pain, impact on function, joint symptoms, and emotional elements.
Further work is required to optimize and test the derived definitions for most musculoskeletal conditions.”
Setchell J et al., What constitutes back pain flare? A cross-sectional survey of individuals with low back pain. (See Setchell et al., 2017.) “The key finding was that many people with LBP do not consider their condition to be flared simply on the basis of a pain increase.” Other features, especially psychosocial factors, appeared to influence the definition of a flare. “These findings are important as they contrast with most commonly used definitions of a flare that focus predominantly on pain increase.”
*Editor's note: Multiple research groups have been studying back pain flares in recent years. This table of studies by a single research group provides background for the feature article of this issue. It is not meant to diminish the contributions of other research groups.

A Consensus Definition

In the award-winning study, Costa et al. looked at 126 men and women who had experienced low back pain for at least three months. They were recruited through advertisements in social media and in the local community. This was a case-crossover study, in which participants acted as their own controls.

To be included in the study all the participants had to have expectations of continuing to experience back pain in the days and months following recruitment. The study excluded subjects with spinal infection, fracture, or neoplasm; previous or forthcoming spinal surgery; rheumatoid arthritis; ankylosing spondylitis; and pregnancy over the past year. All subjects had to have a cell phone and be tech savvy enough to report symptoms, associated factors, and other data via a smartphone app.

All the subjects completed baseline assessment questionnaires regarding LBP duration, average pain intensity over the previous week, sex, and any relevant comorbidities and other characteristics.

The participants then used a smartphone app to report data on flares, potential flares, and potential risk factors three times per day for 28 days.

“In the morning they were asked about sleep quality, sleep duration, and pain intensity,” according to Hodges. “In the afternoon they were asked about intensity of pain. In the evening they were asked about pain intensity and we asked them if they had experienced a flare according to our definition.”

The new study looked at back pain flares defined in two different ways, one based on self-report and one based on a standardized measure of pain severity.

“First, self-reported flares (SRF) were identified by positive response to the question asked each evening regarding the participants' own interpretation whether they had experienced a flare,” according to Costa et al.

“Second, pain-defined flares (PDF) were identified as a pain increase of 2 or more points on the 11-point numerical rating scale above the pain averaged across all days without a self-reported flare,” they added.

For their scientific analysis the authors only selected flares (SRFs/PDFs) preceded by at least 3 days without a flare to compare exposure to the potential risk factors across pre-flare/pre-no flare periods of that duration. The case period was defined as the three days prior to an SRF/PDF, according to the study. Similarly, the control period was defined as the three days that preceded a day with no flare.

The researchers gathered data on a variety of potential risk factors: sleep quality and sleep duration, pain and timing of pain, rumination about pain, pain self-efficacy, physical activity, fear of physical activity, fatigue, disability, involvement in paid work, medication, and other treatments.

Previous work by this group has suggested psychosocial factors play an important role in flares. However, Hodges noted that gathering sufficient data on these factors is challenging in a study conducted via cell phones over the Internet. There is a limit to the number of questions study participants can field three times a day.

The research team asked study participants single questions from a series of more comprehensive questionnaires related to psychological variables, fear-avoidance beliefs, pain self-efficacy, and other factors. While the larger questionnaires have been validated, the use of single questions from them has not. Hodges noted that the research group is planning a follow-up study that will employ a more refined series of questionnaires “so we can study these factors more fully, now that we have definitions of flares and the ability to track them more fully.”

Study Assessed a Variety of Potential Risk Factors

According to the study results, risk factors differed if the flare was defined by pain alone (pain-dependent flare) or self-reported flare.

The risk factors for a pain-dependent flare (i.e. at least a 2-point increase on an 11-point pain scale) included pain in the morning, afternoon, and/or evening. An increase in pain by this definition increased the risk of a flare for up to three days.

“One interpretation is that pain-defined flare simply represents the peak of a progressive increase in daily pain over 1-3 days and may simply reflect that pain fluctuates and sometimes exceeds a threshold used to define a flare,” Costa et al. explained.

When flare was defined using a broader definition (i.e. patient self-reporting a flare), only pain in the preceding morning increased flare risk. “There were no increased odds of a flare 2-3 days later and high afternoon or evening pain did not increase the odds of self-reported flare. This suggests a different mechanism for self-reported flares.”

Poor Sleep a Risk Factor for Both Definitions of Back Pain Flare

Poor sleep quality was a risk factor under both definitions. But it had a broader impact on flares defined by self-report.

“Only the category of ‘very poor’ sleep preceded a pain-defined flare, whereas subtle sleep deviations increased the odds of self-reported flare up to 2 days later. This distinction between PDF and SRF may relate to the negative impact of poor sleep on features considered in broader dimensions of self-reported flares—sleep quality impacts mood, affective anticipatory brain mechanisms (i.e. responses to threat/danger) and emotional brain regulation,” according to Costa et al. (Table I).

“What we saw is that the risk of a pain-defined flare is reduced by having a good night's sleep but increased by having a night of very bad sleep,” Hodges observed.

“The risk of a self-reported flare was again reduced by having a night of good sleep—and is increased even by having a “fairly good” night's sleep as opposed to a “very good” night's sleep. So a subtle change [in sleep quality] can increase the risk of a self-reported flare.”

Interestingly, the number of hours of sleep did not appear to be a risk factor for flares. Sleep quality trumped sleep duration.

What about physical activity? The researchers found that physical activity during leisure time increased the risk of a pain-defined flare but did not increase the risk of a self-reported flare.

“How do we interpret this finding?” Hodges asked. “It may relate to a tradeoff. If a person is performing activity for leisure they may accept that that will increase their pain and therefore not consider it a flare.”

The overall conclusions? “LBP fluctuates over time and it is important to differentiate between types of fluctuation across LBP trajectories,” according to Costa et al. “Risk factors for LBP flare depend on how it is defined. As risk factors for pain-defined flares and self-reported flares differ, it is plausible that outcomes of trials of treatment efficacy and prognosis might be influenced by how flare is defined. We argue that self-reported flare is likely to provide a measure that is more meaningful for a patient.”

And this investigation, like several other recent studies, suggests a strong relationship between back pain and sleep. It would be positive news for people with back pain—and their healthcare providers—if sleep turns out to have a strong protective effect against flares. If it did, it could alter both the short- and long-term trajectories of low back pain and related disability.

Post-Presentation Discussion

A short but lively discussion section followed the ISSLS presentation by Hodges. A BackLetter editor asked whether there is evidence that back pain flares influence care-seeking and healthcare utilization.

“There is,” Hodges responded. “There is a body of work looking at the impact of flare in working environments and the impact of flare on people's lives. There are a series of qualitative projects that show very strongly that flare is one of the most troublesome factors that people experience. So there is good data to show that it is a feature that causes people to seek care. One of the problems is that most [research] has not investigated back pain in this way.”

Most studies on the development and progression of low back pain have not considered flares at all, according to Hodges. So there is a strong need for further research on pain trajectories, to better understand “these phasic periods of waxing and waning of symptoms.” He is confident that there will be a productive research effort in this area “now that we have definitions of flares and the ability to track them more closely.”

At least two questioners asked how the researchers were able to distinguish the relative contributions of various risk factors, since several of them appear to be interrelated.

“Congratulations,” said researcher Anne Mannion. “Just wondered—are all the odds ratios univariate? Many variables are likely intercorrelated. Sleep and exercise and fatigue... We are told that exercise promotes sleep, yet exercise and sleep had opposing effects.

“All of our analyses were univariate but we do completely take on board the idea that physical activity and sleep are interrelated,” Hodges answered. “And we have just completed a series of studies with mice where we manipulated sleep and physical activity independently to address exactly that question.”

Valerio Tonelli Enrico asked whether it might be useful to study inflammatory changes in the days leading up to a back pain flare.

By way of background, many researchers have hypothesized that systemic inflammation may play a role in the development and persistence of low back pain. And there is some evidence of distinctive patterns of inflammatory biomarkers in nonspecific low back pain. (See Morris et al., 2020.) So exploring these patterns would be a natural adjunct to research on back pain flares.

“Absolutely, great question,” said Hodges. He said there are already ongoing studies of inflammatory mediators during flares. “We have done a longitudinal study with very coarse measurements over multiple months and we are about to commence a study looking at measurements of daily inflammatory mediators from blood.”

What About Generalizability?

One important question that was not addressed during the brief discussion session at ISSLS—or in the published study—concerns the generalizability of these study results.

This study recruited a highly selected group of participants: those who were expecting their symptoms to wax and wane, those with cell phones, and those with the tech savviness and organizational skills to report data three times a day to researchers.

People meeting these criteria are likely to represent a limited proportion of people with back pain.

In the United States, for example, these selection criteria would likely exclude many older people with back pain, people of limited socioeconomic and educational status, rural residents, and those with limited health literacy and organizational skills. So the results should be interpreted cautiously until studies look at a variety of back pain populations. This is not a criticism of this innovative study. It is simply a challenge for researchers going forward.

Disclosures: None declared.


Costa N, et al., A definition of “flare” in low back pain: A multiphase process involving perspectives of individuals with low back pain and expert consensus, Journal of Pain, 2019; 20(11):1267–75.
Costa N, et al., How is symptom flare defined in musculoskeletal conditions: A systematic review, Seminars in Arthritis and Rheumatism, 2018; 48(2):302–17.
Costa N, et al., ISSLS Prize in Clinical Science 2021: What are the risk factors for low back pain flares and does this depend on how flare is defined, European Spine Journal, 2021a; 30:1089–97.
    Costa N, et al., Low back pain flares: How do they differ from an increase in pain?, Clinical Journal of Pain, 2021b; 37(5):313–20.
    Costa N, et al., What triggers a low back pain flare? A content analysis of individuals' perspectives, Pain Medicine, 2020; 1;21(1):13–20.
      McCurry SM, et al., Effect of telephone cognitive behavioral therapy for insomnia in older adults with osteoarthritis pain: A randomized clinical trial [published online ahead of print February 22, 2021], JAMA Internal Medicine. doi:10.1001/jamainternmed.2020.9049.
        Morris P, et al., A systematic review of the role of inflammatory biomarkers in acute, subacute and chronic non-specific low back pain, BMC Musculoskeletal Disorders, 2020, 3;21(1):142.
        Setchell J, et al., What constitutes back pain flare? A cross-sectional survey of individuals with low back pain, Scandinavian Journal of Pain, 2017; 17:294–301.
          Tan D., Impact of flare-ups on the lives of individuals with low back pain: A qualitative investigation, Musculoskeletal Science & Practice, 2019; 43:52–7.
          © 2021 Wolters Kluwer Health, Inc. All rights reserved.