Prognosis of Low Back Pain

This article is still missing information.

It is sometimes stated in guidelines that most patients with acute low back pain make an excellent recovery. The evidence is in fact quite conflicting, with markedly different findings across different studies. Overall the treating doctor can relay optimism, but be guarded about prognosis.

A systematic review of 11 studies performed in the US, Australia and Europe on patients with non-specific back pain found that recovery occurred in 33% of patients at 3 months, and by 1 year 65% still had pain. In studies that used total absence of pain as a criterion, 71% still had pain at 12 months. In studies that had a less stringent criteria, 57% still had pain at 12 months.

Recurrences

The Importantly, recurrences can occur after recovery. Again the literature has shown different rates for recurrence. No comprehensive New Zealand data were found. The most relevant to New Zealand are the Stanton, Machado, and da Silva studies. The former two showed recurrence rates of about 1 in 3 by one year. The latter study was much smaller and showed a much higher recurrence rate of 69%.

Study (Year)	Population & Inclusion Criteria	Definition of “Recurrence”	Recurrence Rate (% and timeframe)
Biering-Sørensen (1983)	928 general population adults (ages 30, 40, 50, 60) in Denmark; surveyed at baseline and 12 months. Included those with any prior LBP history (lifetime prevalence ~68–81%).	Any new LBP episode during 1-year follow-up (survey self-report). Frequent recurrences defined as LBP symptoms weekly or daily.	Among those with prior LBP, 23–31% had frequent LBP recurrences (daily or at least weekly) during the 1-year follow-up. (Risk of any recurrence was higher in those with recent or multiple past episodes.)
Croft et al. (1998)	463 primary care patients in UK with a new acute LBP episode; followed 1 year (South Manchester cohort). Mean age ~43; 59% single GP visit only.	“Complete recovery” = no LBP pain or disability; recurrence implied if not pain-free at follow-ups. (Outcomes measured at 3 and 12 months.)	At 12 months, only 25% were completely pain- and disability-free, meaning ≈75% had persistent or recurrent LBP symptoms in the year after the initial episode. (About 32% consulted again within 3 mo for recurrence.)
van den Hoogen (1998)	443 primary care patients in the Netherlands, with both chronic and recent onset low back pain	Monthly postal questionnaires on the course of low back pain and related disability over 12 months.	Relapse rate of 76% at 12 months, with a median number of two relapses (interquartile range 1-3), with a median time to relapse of 7 weeks (interquartile range: 5-12), and a median duration of 3 weeks for the first relapse, 2 weeks for the second and third, and 1 week for the fourth. At 12 weeks 35% still suffered with back pain, at 12 months was 10%.
Carey et al. (1999)	921 patients from North Carolina primary care practices; acute LBP at baseline. Those who were pain-free by 3 months post-episode were followed long-term (up to ~22 months).	New LBP episodes after initial recovery, assessed by interviews at 6 and 22 months. “Functionally disabling” recurrence defined by moderate functional limitation; also tracked any care-seeking for LBP.	3–6 months post-recovery: ~8–14% had a disabling LBP recurrence. By 6–22 months post-recovery (≈1–2 yrs): 20–35% had a functionally disabling recurrence (varying by provider type).
Stanton et al. (2008)	1334 primary care patients in Australia with acute low back pain, 353 recovered within 6 weeks and entered the study and observed for 1 year.	Two definitions used: (1) 12-month recall of any new LBP episode in the past year, and (2) report of LBP pain on scheduled follow-up contacts (3 or 12 mo) after recovery. (Recurrent episode defined as LBP lasting >1 day after a pain-free period.)	One-year recurrence incidence was 24% (by 1-year recall) up to 33% (by periodic follow-up measurement). Only previous episode(s) of LBP was consistently predictive of recurrence within the next 12 months (OR 1.8-2.0).
Machado et al. (2017)	832 patients seen in General Practice in Australia with acute low back pain. Of these, 469 recovered (56%) within 6 weeks and were then followed to look for recurrences.	Recurrence = a new LBP episode lasting >1 day or an episode prompting healthcare consultation, after at least 1 month pain-free. Outcomes tracked by monthly contact and interviews at 12 mo.	12-month recurrence incidence = 33% (about 1 in 3 had ≥1 recurrence). Recurrence with healthcare-seeking was 18%. Notably, patients with a history of >2 previous episodes had 3× higher odds of recurrence within a year. (No other baseline factors predicted recurrence.)
da Silva et al. (2019)	250 primary care patients in Australia who had recently recovered from an LBP episode (pain-free within past month). Mean age ~37; monthly follow-ups for 1 year.	Recurrence of an LBP episode = return of LBP (any intensity) for ≥1 day after recovery. Also tracked recurrences that limited activity and those leading to healthcare use. (Standardized 30-day pain-free period as part of definition, per consensus recommendations.)	Within 12 months after recovery, 69% of participants had at least one LBP recurrence. Moreover, 40% had a recurrence that was moderately activity-limiting, and 41% had a recurrence that prompted seeking healthcare. The median time to first recurrence was ~4.5 months. (Frequent awkward bending, >5 hours sitting per day, and >2 prior episodes were significant risk factors for recurrence.)
Medeiros et al. (2022)	238 patients (Brazil) presenting to emergency departments with acute LBP; followed for 1 year after initial episode resolution (inception cohort in São Paulo). Average age 47; stratified by acute LBP risk (STarT Back Tool).	Recurrence was assessed via two methods: (1) 12-month single recall (patient’s report at 1 year of any recurrence in past year), and (2) prospective pain assessments during follow-up (periodic pain questionnaires).	1-year recurrence rates ranged from 35% (by 12-month recall) up to 44% (by measured pain reports). When stratified by risk level: low-risk patients had ~29–37% recurrence, high-risk had ~43–56% recurrence within 12 mo. Overall, roughly “4 in 10” patients had a recurrence within a year of recovery. (Older age, higher baseline disability, and expecting persistent pain were predictors of recurrence.)

Older Adults

A particular note should be made about older adults, where the prognosis tends to be worse. For example, in a US cohort study of 4,665 patients who had a new primary care visit for back pain, only 16% had complete resolution of their back pain and disability by two years. Average pain intensity reduced slightly from 5.0 to 3.7 at 3 months, and then stayed relatively static to 24 months, with around half having a clinically meaningful improvement in disability and pain. Baseline characteristics are more important predictors such as female gender, higher BMI, chronic back pain, higher disability, and negative expectations for recovery.

Similar findings for older adults have been found in the Netherlands, and the BACE study which took place in the Netherlands, Brazil, and Australia.

Prognostic Risk Factors

Prognostic risk factors are broadly categorised into biological and psychosocial. An important limitation of epidemiological studies assessing this area is that only factors asked about in a questionnaire can be considered. For example, if a study only uses assessment methods looking at psychosocial factors, they won't find any biological factors. Factors that appear significant when amongst a small list of possibilities may not be significant if it is diluted in a longer list of possibilities. The most reliable studies are those that consider an initial large number of possible factors and then undertake multiple regression analyses to eliminate spurious factors.

Following the above analysis, investigators should determine how much of the variation between patients is accounted for by the significant factor. Unfortunately this is not typically done in back pain research. Very often the researchers will take a factor found to be significant, ignoring that is may only account for a small proportion of the variance amongst patients, and incorrectly conclude that it is important. In other words, many of the reported risk factors only account for a small proportion of patients.

Therefore, predictors of recurrence are often variable across studies. The psychosocial category are the strongest and potentially remediable factors. The risk factors generally include

Table 1: Prognostic Risk Factors for Chronicity

	Biological	Psychosocial
Immutable	Duration of back pain Past history of back pain Frequency of attacks Playing adult sport Female sex Previous episodes	Marital status Family status
Relatively immutable	Leg pain Severity Job demands Obesity	Job dissatisfaction Education MMPI Compensation Employment Wage Heavy lifting at work Awkward positions at work Somatisation Educational level
Potentially remediable	Work Capacity Disability Smoking BMI Inability to sit-up Physical activity	Sickness Impact Depression Poor coping skills Catastrophising Distress Rating of loads Fear-avoidance Inappropriate signs Lack of understanding
Strongest risk factors are underlined. MMPI: Minnesota Multiphasic Personality Inventory

The predictors may be similar across age groups. In older adults they include female gender, race, worse baseline clinical characteristics of back pain, leg pain, back-related disability and duration of symptoms, smoking, anxiety symptoms, depressive symptoms, a history of falls, greater number of comorbidities, knee osteoarthritis, wide-spread pain syndromes, and an index diagnosis of lumbar spinal stenosis.

It is difficult to predict who will have a recurrence after recovery. Number of previous episodes may be the only predictor.

The fear avoidance model

Note in table 1 how there is a clustering of important remediable risk factors in psychosocial category. These are indeed important predictors of chronicity. However none alone are the predominant factor, with each only accounting for a small proportion of the variance. They are not major determinants of chronicity unless several are present simultaneously. This is likely why therapies targeted at psychological factors have been disappointing.

The psychosocial risk factors don't mean that the patient has psychosomatic pain. It rather indicates that distress and illness behaviour are secondary to the physical disorder, and that psychological factors with the physical disorder combine to produce disability. Due to the profound psychosocial impacts on disability, disability is considered a psychosocial factor rather than biological factor. Pain severity only accounts for 10-14% of the variance of physical disability.

The fear avoidance model views a patients disability as a function not only of their pain but of their response to their pain. The psychological aspect of avoidance behaviour is based on the patient's experience, memory, cognitions (reasoned thoughts which may or may not be true), and beliefs (less rational thoughts related to emotions).

Some cognitions may be harmful, such as believing that bed rest is beneficial. Cognitive patterns can be reinforced if avoidance is rewarded by attention, resulting in illness behaviour. Fear is a belief. Patients may believe that pain is a sign that they will end up in a catastrophe. Beliefs can be influenced by past history of pain, previous stressful events, and personality .

Confronters view their pain as a temporary nuisance. They are motivated to return to activity, and are keen to manage their pain and get on with things. Avoiders avoid the painful activities altogether. Avoidance is counterproductive as it doesn't reduce pain, and even worse, exposure actually increases tolerance. Furthermore it only serves to reduce physical and social functioning. Fear avoidance beliefs account for around a quarter of the variance of disability. The take away from this is that with movement despite pain is helpful in recovery, generally in the form of graded exposure.