Who gets back pain? The simple answer, of course, is that most of us get back pain but there is obviously more to it than that. So, more specifically, are some people more at risk of serious back trouble or do some circumstances increase the risk?

Potential risk factors for back pain

Individual
  • Genetics
  • Gender
  • Age
  • Body build: height, weight, leg length inequality
  • Physical fitness
  • Smoking
  • Social class, education
  • Emotional distress

EnvironmentalPhysical
  • Manual handling
  • Heavy lifting
  • Bending and twisting
  • Repetitive movements
  • Static work postures and sitting
  • Driving and whole-body vibration
  • Leisure activities and sports

Psychosocial aspects of work
  • Job satisfaction
  • Work “stress”
  • High job demands and pace
  • Poor job content: low decision latitude, low job control, and monotonous work
  • Low social support
  • Job “strain”

We can broadly divide potential risk factors for back pain into individual and environmental (see above). This does not imply a dichotomy. Rather, it suggests there are both individual susceptibilities and environmental stressors that may interact.

RISK FACTORS
Before we go any further we ought to be clear what we mean by “risk.” Our first thought might be that a risk factor is something that causes back pain. So if we can remove or reduce that factor we might prevent some back pain. Sadly, risk is more complex than that and can have various implications.

Most people get back pain at some time in their lives and it is commonly recurrent. Because back pain is a recurrent problem, the most consistent and by far the strongest predictor of future back pain is the individual’s previous history. So other risk factors are usually weaker, additional influences that simply modify the natural history.

Most of the early research was on physical risk factors for back injuries, particularly at work. There have been attempts to distinguish between “hazard” and “risk.” A hazard is anything with the potential to cause harm; risk is the probability of someone actually coming to harm. Obviously, if there is no hazard, then there is no risk. However, even if there is a hazard the risk might be very low. Some hazards may have such serious consequences that we must try to eliminate them, even if the risk is very low. But at a practical level, the most cost-effective control strategies address hazards that carry a higher risk.

Many of the early studies of risk factors were cross-sectional in design. Strictly speaking, these only show statistical associations between possible risk factors and the prevalence of reported symptoms. Most were retrospective studies looking at small groups of workers and matched controls. The groups were often highly selected and not at all typical of the general population. Most studies depended on self-reports of work and of symptoms, which are often unreliable. In short, these early studies had many serious limitations.

Scientific study of risk factors and proof of cause and effect require prospective cohort studies. These measure risk factors in people who are initially free of symptoms and then study the incidence of new symptoms over time. Some factors turn out to be only risk markers that are associated with symptoms, but do not necessarily demonstrate cause and effect. For example, a cross-sectional study may show that workers in a certain job have more sickness absence due to back pain. This could be because that job causes back injuries. Or it may be the job aggravates pre-existing back symptoms. Or workers who have back pain for some other reason may have more difficulty doing that job. Or that workplace may have poor industrial relations, with high sickness absence rates, and workers who are more likely to stay off work when they have ordinary backache. True risk factors predict the development of future problems and also provide information about their etiology and causal mechanisms, e.g., certain physical demands of work. Still others may be simply a wide variety of variables, e.g., gender, which contain complex biologic, psychological, and social issues that require further analysis.

Proof of cause and effect requires strict criteria:
  • strength of association: sometimes described as the “effect size.” Weak effects may be statistically significant but are unlikely to be clinically important. For complex statistical reasons that we need not go into here, this usually requires an odds ratio (OR) or relative risk (RR) of >3–4.
  • consistency in different studies
  • biologic plausibility: does it fit our theoretic understanding? This may be difficult for a condition like non-specific back pain where we do not really understand the pathology!
  • temporal sequence of exposure and effect: which can only be shown in a longitudinal study
  • dose–response gradient: greater or cumulative exposure to the hazard increases the risk
  • specificity: it is usually only possible to demonstrate this with an uncommon exposure and an uncommon condition, e.g., asbestos and mesothelioma. It is difficult to demonstrate with a common condition like back pain
  • reversibility: stopping exposure to the hazard reduces the risk. Thus back pain that develops some time after stopping work is unlikely to be caused by that job.

Adams et all (2002) considered various risk factors for back trouble and possible relationships between them.



This started from the concept that increasing or cumulative exposure leads to increased risk of the outcome, e.g., back pain. Today, it places more emphasis on increasing exposure leading to more serious consequences. Both may be true.

Before we look at individual risk factors, it is worth repeating that we must keep them in perspective. The high prevalence of back pain means that most risk factors can only have a modest additional effect. If 50% of people get back pain at some time anyway, then most risk factors might increase that to 60–70%. What then matters is the effect size and the clinical importance of the risk.

INDIVIDUAL RISK FACTORS

Genetics
Genetic factors play a role in certain spinal disorders, such as spondylolisthesis, scoliosis, and ankylosing spondylitis. A few clinical studies suggest there may sometimes be a familial or genetic predisposition to disk prolapse. However, all of that is of little relevance to ordinary backache.

We now have various twin studies that investigate genetic factors in back pain. The evidence seems clearer for degenerative changes than for symptoms, but we must remember that the correlation between them is low.

The classic Finnish twin study ( Battie et al 1995 ) found that identical twins showed very similar magnetic resonance imaging (MRI) changes in their spines, despite different occupational histories.

Occupational exposure has little impact on degenerative changes. These magnetic resonance imagings are from 50-year-old twins. One was a farmer who had always done heavy manual work. The other was a journalist. Can you tell which is which? You have a 50% chance of guessing correctly!

This is often misquoted as showing that genetic factors determine degenerative changes in the spine. These findings are hardly surprising, as identical twins have the same body build and metabolism. But they also usually share their early lives. The main message of this study was that familial factors (which includes genetics, body build and make-up, and early environment) have more influence than occupation on the degeneration that occurs in everyone with age. The authors themselves point out that this kind of study cannot separate genetic, anthropometric, and metabolic factors from the effect of shared early environment and lifestyle. Moreover, this study still left a great deal of degenerative changes unexplained, particularly at the lower lumbar levels, which are most important clinically.

A study of British twins came to broadly similar conclusions ( Sambrook et al 1999 ). Lumbar disk height, disk bulge, and osteophytes appeared to be highly heritable, but there was a confounding influence from shared environment.

Even more fascinating are preliminary studies of the human genome. A few genes have been identified that are related to disk degeneration – those for the vitamin D receptor (Videman et al 1995 ), for collagen type IX ( Paassilta et al 2001 ), and for proteoglycans ( Kawaguchi et al 1999 ). It has been suggested that the search for genes should be extended to pathologic, physiologic, and behavioral mechanisms. It is still early days, but studies of gene–environment interactions might lead eventually to a better understanding of risk factors and causal mechanisms. Gene studies might even lead to new treatments.

Twin studies of back symptoms give inconsistent results. MacGregor et al (1999) found a large genetic contribution to the prevalence of moderately severe back pain in females. Not surprisingly, however, this was only partly explained by any genetic influence on degenerative changes on MRI. In a study of pain thresholds, MacGregor et al (1997) found that learned patterns of behavior within families were much more powerful than any genetic influence.

The Danish twin study ( Hartvigsen et al 2003 ) could not detect any significant genetic influence on short- or long-term back pain. They concluded that physical workload might be more important that any genetic effect.

A much earlier twin study by Heikkila et al (1989) found that only about 10% of “sciatica” could be explained by constitutional similarity.

In summary, it appears that certain aspects of back pain may have a genetic or at least a constitutional or familial element. We still need to disentangle these elements. Most important, there is no evidence that genetic or constitutional factors determine who is going to become a back cripple.

Gender
Most large population surveys show a slightly higher prevalence of back pain in women. However, we must interpret this against a background that women report a slightly higher level of most symptoms. This could be due as much to body awareness, pain perception, and willingness to report symptoms as to any physical difference in their backs. When it comes to actual studies of risk factors for back pain, gender consistently turns out to have a limited effect ( Burdorf & Sorock 1997 ).

The evidence on low back disability is conflicting, with no clear pattern. The evidence on work loss is also conflicting. There are biomechanical reasons to suggest that women may be more at risk of increased loading during heavy lifting than men ( Marras et al 2002 ). However, that is balanced by women generally having lighter jobs. Almost all workers’ compensation figures show more work-related back injuries and claims in men, although in some series women stay off work longer ( Waddell et al 2002 ). However, these data only cover work-related back pain in selected groups of workers. Social security data may be more representative of the general population. This shows different patterns in different countries ( Waddell et al 2002 ).

This seems to be largely a matter of more women working and becoming entitled to benefits. In Sweden, sickness absence due to back pain is higher for women than for men. However, this difference is largely explained by sick leave during pregnancy. When that is excluded, there is no difference between men and women ( Sydsjö et al 2003 ).

All of these findings may reflect their social settings rather than any biologic difference between men and women. Overall, there does not appear to be any major difference in low back disability between men and women.

Women seek slightly more health care for back pain, as for all health complaints ( McCormick et al 1995 , Vingard et al 2002 : for example, see Table 19.4 ).

Sciatica does appear to be more common in men than in women ( Heliovaara et al 1987 ). Clinical reports all show more men coming to spinal surgery, although this may also be due to different referral patterns and different selection for surgery in men and women.

There is one condition that is absolutely gender-specific: pregnancy. Many women have temporary back pain during the later stages of pregnancy, possibly related to altered posture and hormonal changes in soft tissues. However, this does not appear to have any lasting effect. Several early reports suggested that women with multiple pregnancies might continue to have more back pain. More careful studies do not seem to confirm this. Ostgaard et al (1996) suggested there might have been confusion between posterior pelvic pain and lumbosacral pain. The main problem in pregnancy may be pelvic pain, which usually settles after delivery. When they distinguished this, pregnancy did not appear to influence future back pain. As epidurals became more common during labor, there were many claims that they caused chronic low back pain. However, long-term follow-up of a randomized controlled trial of epidural pain relief in labor showed no significant difference in spinal movements, back pain, or disability ( Howell et al 2002 ).

Age
Population surveys suggest that the age of first onset of back pain is spread fairly evenly from the teens to the early 40s. It is uncommon to develop ordinary backache for the first time after the mid-50s.

During the 1990s, many studies from all developed countries showed that back symptoms are also common in adolescents and teenage children aged 11–18 years ( Balague et al 1999 , Nachemson & Vingard 2000 ). These symptoms do not usually present for health care, and adult surveys suggest that most people forget about them. So we only detected them when we asked specifically about them. Burton et al (1996a) made a prospective study of 216 adolescents from 11 through 15 years of age. Only 12% of 11-year-olds said they had ever had back pain, but by age 15 it rose to adult levels of 50%. That is an annual incidence of 15%. Their back pain was often recurrent but did not deteriorate with time. Most important, however, all studies agree that it rarely causes significant disability and few seek health care ( Burton et al 1996a , Wedderkopp et al 2001 , Watson et al 2002 , 2003 ).

We should be cautious how we interpret these findings. They are based entirely on leading questions that, as every parent with kids of this age knows, do not always produce reliable results! Brattberg (1993 , 1994) carried out a longitudinal study of 471 schoolchildren aged 10, 13, and 15 years in Sweden. In each year’s survey, about 26% of children said they had back pain, but only 9% of the children reported back pain in both surveys in 1989 and 1991. King & Coles (1992) found marked variation between different European countries, ranging from 3% for girls aged 15 in Finland to 22% in Belgium. However, Hakala et al (2002) found that the prevalence in Finland increased dramatically by the late 1990s. This all suggests that there may be a major cultural element in these findings.

It might seem plausible to suggest that the search for causes of back pain should start with children and adolescents, and that preventive measures should start at that time. But we should be careful. Balague et al (1999) reviewed risk factors for back pain in children and found serious weaknesses in the scientific evidence. They found moderate evidence that competitive sports activities are associated with increased back pain in adolescents. They found some evidence that a family history of back pain, increased height, smoking, high levels of physical activity, and depression and emotional stress are also associated with reported symptoms. But many of these are behavioral phenomena that may easily be confounded with psychosocial influences on self-report. Watson et al (2002 , 2003) made one of the few studies that looked at both physical and psychosocial risk factors in children. They could not identify any mechanical risk factors, but found a stronger association with emotional difficulties and psychological problems.

There is little or no evidence that any of these adolescent risk markers have any direct biologic effect. Nevertheless, the search for a physical cause has led to a media frenzy over backpacks. There are no prospective studies looking at this issue, but there are studies that help to put it into perspective. One study in Italy found that children felt their backpacks were heavy, uncomfortable, and caused back pain. However, reports of back pain were not related to the weight of the backpack, but to the time spent carrying and subjective feelings of fatigue ( Negrini & Carabalona 2002 ). Another study in the US looked at backpack injuries coming to emergency departments ( Wiersema et al 2003 ). Only 11% of injuries in children involved the lower back, of which 59% involved carrying a backpack. However, most injuries were caused by tripping over a backpack or being hit by one!

Burton et al (1996a) suggested that we should consider most adolescent back trouble to be a normal life experience and not attach undue significance to it. Most important, there is no convincing evidence that back pain in adolescence is a risk factor for serious low back trouble in adult life. The history of adult back pain should teach us the danger of overmedicalizing back pain in children. The real risk would be if overenthusiastic intervention should turn a minor childhood symptom into a self-fulfilling medical disaster in adult life.

Schoene (2002) gave a very balanced discussion on what this means for clinical management. They provided a clinical framework for thinking about back pain in adolescence
  a. During adolescence and teenage years the prevalence of back pain increases to adult levels. Most adolescent back pain is not due to any significant medical condition and does not present for health care
  b. When adolescent back pain does present clinically, a small but important proportion will be due to serious underlying pathology (which is why it is a “red flag”). In each case, you should consider if and when further investigation is required to exclude this
  c. But, remember (a) and keep (b) in proportion.

Most population surveys show that the prevalence of back pain increases with age up to about 45–50 and then levels off or falls slightly ( Burdorf & Sorock 1997 ). However, such surveys usually focus on people of working age. Bressler et al (1999) reviewed 12 studies that gave separate data about back pain in people aged 65+ years. Methodologic weaknesses and small sample sizes gave wide variation in the estimates. Nevertheless, some studies again showed that the prevalence of symptoms fell slightly with age. Edmond & Felson (2000) studied a large cohort of 1037 Americans aged 68–100 years. The 1-year prevalence remained about 50% in those aged 68–80 and 81–100 years. It was slightly higher in women, particularly in the thoracic region. If older people did get back pain, however, it was likely to be more persistent. Twenty-two percent said they had back pain “most days.”

Self-reported disability tends to increase with age.

Self-reported restricted activities in the past 4 weeks due to back pain (as a percentage of those with back pain)

Age (years)           Men (%)           Women (%)
16–24                         11                        21
25–34                         20                        29
35–44                         25                        27
45–54                         23                        37
55–64                         32                        41

Walsh et al (1992) found that disability increased up to age 40–49 years. Social security statistics in all countries show that sickness absence and long-term disability benefits rise dramatically some time after 50–55 years of age ( Waddell et al 2002 ). This reflects all conditions. All chronic disability becomes much more common in the elderly, but we have little data for back pain. Edmond & Felson (2000) found that geriatric patients with poor general health who were confined to their homes had a particularly high prevalence of back pain and stiffness.

So, is age a risk factor in back pain? The answer is yes and no – it depends: on whether we are talking about symptoms or disability or health care. Clinically, there are similarities and differences between back pain in adolescents, adults of working age, and the elderly.

Body Build
There are many clinical myths about back pain being related to body build. Doctors and therapists can’t resist blaming obesity, or being tall, or leg length inequality.

Contrary to popular belief, most studies show that body weight, and even obesity, does not make much difference. Leboeuf-Yde (2000a) reviewed 65 studies, of which only a third showed any significant association between body weight and symptoms. Even then it was weak. She concluded that there is no clear evidence that weight actually causes back pain. There is also no clinical evidence that weight loss is an effective treatment for back pain.

Reviews by Burdorf & Sorock (1997) and Nachemson & Vingard (2000) showed no consistent relation between height and back pain.

Doctors and therapists often get excited about unequal leg length, but the literature again does not show any consistent relationship ( Nachemson & Vingard 2000 ).

In summary, contrary to individual cross-sectional reports, the evidence suggests that none of these aspects of body build is a significant risk factor for back pain or its consequences.

Physical fitness
There has been much interest in the possible role of physical fitness in back pain. Clinical evidence shows that patients with chronic back pain are less fit, but this could be effect rather than cause. The more specific idea that physically fit people get less back trouble rests mainly on a single, classic study. Cady et al (1979) found that physically fit firefighters got fewer back injuries than those who were less fit. However, that was a very select population in an unusual, high-risk situation. This study has never really been replicated. Reviews by Andersson (1997) and Nachemson & Vingard (2000) did not find convincing evidence that the level of general (cardiovascular) fitness is a risk factor for future back pain.

There are many health advantages to being physically fit. It is possible that physical fitness/strength may help to reduce the likelihood of new episodes of back pain in certain jobs. However it seems likely that fitness is more relevant if and when back pain does occur. There are theoretic reasons and some clinical evidence to suggest that fit patients make a more rapid recovery from acute back pain and are less likely to develop chronic pain and disability.

Smoking
Many studies describe smoking as a risk factor for various aspects of low back trouble. Battie et al (1991) and Goldberg et al (2000) reviewed theories about smoking. Smoking may cause chronic cough, which might influence disk prolapse and sciatica, although there is no direct evidence on this. Smoking reduces bone mineral content, so might cause osteoporosis and microfractures. It impairs fibrinolysis and promotes scar formation. It causes changes in disk nutrition. Battie et al (1991) found more degenerative changes on MRI in the disks of smokers compared with their non-smoking identical twins. Smoking could also have more indirect effects. There may be a relation between smoking, physical fitness, and body weight. Smoking is linked to how people report pain and is actually related more strongly to pain in the limbs than to pain in the neck or back. Smokers have lower physical and mental health status, and show more depressive symptoms ( Vogt et al 2002 ). Smoking varies with social class, education, and occupation. So smoking may simply be a risk marker for a complex set of demographic, psychosocial, and lifestyle factors.

Two reviews of 47 studies show that the relation between smoking and back pain is weak and inconsistent ( Leboeuf-Yde 1999 , Goldberg et al 2000 ). Leboeuf-Yde (1999) concluded that smoking is only a weak risk marker of back pain and not a cause. There is also no evidence that stopping smoking is an effective treatment. This really shows the danger of overinterpreting some of these studies!

Several early studies questioned the possible role of alcohol. Do not worry! We are delighted to say that a review by Leboeuf-Yde (2000b) showed that alcohol is not a risk factor for back pain.

Social class
There are many social influences on back pain and disability. As a very crude starting point, we might look at social class. Some surveys use a classification based on occupation:

I. professional groups such as doctors, lawyers, and scientists
II. intermediate groups such as teachers, nurses, and self-employed shopkeepers
III. skilled occupations
     IIIa: skilled non-manual, such as clerical workers
     IIIb: skilled manual, such as tradesmen
IV. partly skilled, such as process workers in industry or transport workers
V. unskilled, such as laborers and cleaners.

This classification is really twofold. It is partly occupational, with a divide between manual and non-manual. That may be why studies of social class usually show more significant findings in men than in women. It is also partly socioeconomic, and serves as a marker for all facets of social disadvantage, such as education, housing, (un)employment, and income. That applies equally to men and women.

Walsh et al (1992) provided most detail on the relation between back pain and social class.

Croft & Rigby (1994) tried to disentangle the socioeconomic influences. In men, the only correlation seemed to be with unskilled manual labor. Women showed a correlation with the lowest income category and less formal education: in them it seemed to be a question of social disadvantage.

There is a stronger association between social class and the consequences of back pain ( Waddell & Waddell 2000 ).

People in manual work are more likely to blame their back pain on work. They lose more time from work and stay off longer. People in social classes IV–V are more likely to lie down to rest and seek more health care.

Education
Dionne et al (2001) reviewed 64 studies of education as a risk factor. Most studies showed an association between lower education level and a higher prevalence of back pain, though the strength of the effect was weak. There was a stronger association with disabling back pain. There did not appear to be any association with outcome of treatment.

The problem is that it is difficult to disentangle education from other aspects of social class. In a very careful analysis, Makela (1993) concluded that education was simply an indirect measure of heavy work, work stress, and work injury. In an equally careful study, Deyo & Tsui-Wui (1987) found that education did have an independent effect. Dione et al (1995) made a longitudinal study of education and back-related disability in adults. Like most previous studies, they found that people with less than 13 years’ schooling had more disability. More interesting was what happened over the next 2 years. Disability tended to improve, particularly in those with more education, but did not improve as much in those with less education. Dione et al (1995) considered possible mechanisms, and suggested that occupational and psychological factors were more important than health care access or use. Straaton et al (1996) also found that higher education level was associated with better rehabilitation outcome.

In summary, lower social class is probably a weak risk factor for back pain. There is a stronger association with resulting disability. The relationship to social class is fairly consistent in men but less clear in women. The problem is what this means. Social class appears to be another hodge-podge variable, which is a crude measure of a host of social, educational, occupational, economic, lifestyle, and psychosocial issues, any of which could affect the consequences of back pain. It is partly a matter of heavy manual work, particularly in men. It is probably also a matter of social disadvantage in both men and women, although we do not know exactly which aspects of this disadvantage are important or how they affect back pain.

Perhaps these are more important social influences on what happens to people after they develop back pain, rather than risk factors.

Emotional distress
Patients with back pain often show emotional distress, but it is usually a secondary consequence of their pain and disability. Here, we are considering the converse: is pre-existing distress a risk factor for developing back pain?

Let us look briefly at a couple of studies that tried to disentangle cause and effect.

Mannion et al (1996) studied 403 female nurses and health workers aged 18–40 who had no previous history of “serious” low back pain, by which they meant no medical attention or work loss. Thirty-five percent did have some previous back pain that did not require medical attention or work loss. At the start of the study, they found that those with more distress were more likely to report previous back pain. They also had lower experimental pain tolerance. Over 18 months’ follow-up, 40% reported some low back pain but this was not associated with any rise in levels of distress. Twenty percent reported serious back pain and sought health care or took time off work. This latter group showed slightly increased levels of distress. Initial physical assessment did not predict those who would develop any back pain or serious back pain. Workload also had little effect, whether judged by the job description or by the workers’ own perception of their jobs. The best predictor was psychological distress, but the effect was weak and explained less than 3% of future back pain.

Burton et al (1996b) studied policemen in England and Northern Ireland, with different exposures to physical stressors (wearing heavy body armor and vehicular vibration). They also collected data on back pain history and psychosocial factors. Physical loading on the spine led to earlier first-onset back pain with a dose–response relationship. However, continued exposure to physical stress did not lead to chronic problems. Chronic pain and work loss seemed instead to depend mainly on psychosocial factors.

Linton (2000) reviewed this area. Most prospective studies show that various measures of distress are a risk factor for new onset of back pain. However, the effect is weak. Estimates vary, but it seems that psychological factors only increase the risk by about 5–10%. This does not mean that “5–10% of episodes are caused by a psychological disturbance.” It is more likely that psychological issues play a variable but generally minor role in many people. We should also remember that the outcome in most of these studies was self-reported back pain or injury. And the effect of distress on back pain is weaker than the effect on other musculoskeletal injuries, cardiovascular disability, and depression ( Manninen et al 1995 ). So it may be that distress simply makes people more likely to report symptoms. There is no good evidence linking individual psychology to the development of physical pathology in the spine.

As you would expect, Manninen et al (1995) found that mental stress only predicted non-specific low back pain, and not spinal pathology such as disk prolapse and stenosis.

In summary, emotional distress does appear to be a risk factor for the incidence of new back pain in symptom-free people. However, the effect size is weak. This does not prove that back pain is caused by psychological factors or is “psychogenic”. It seems more likely that psychological factors influence how people react to or report a bodily symptom like back pain. They also influence sickness absence and seeking health care.


ENVIRONMENTAL RISK FACTORS: PHYSICAL
We sometimes assume that physical demands on the back must be risk factors for back pain, but that is not always true. Different physical activities may either load or unload the spine, and loading may be either good or bad for the spine. Physical activities may also be good or bad for us, quite apart from any direct effect they might have on the spine.

We might argue that standing and walking are the most natural human activities, creating a standard loading on the spine. Prolonged standing and walking are not risk factors for back pain ( Hoogendoorn et al 1999 ). Indeed, natural selection would seem to make that unlikely. If standing upright had caused early hominids to develop (disabling) back pain, the experiment would have failed, and Homo sapiens would not have evolved!

Other physical activities fall into two broad categories: those that increase or decrease spinal loading compared with walking. Biomechanical measurements confirm that strenuous activities such as bending, lifting, and manual handling increase the load on the spine. Lying down has the greatest unloading effect. But some activities do not have quite the expected effect. Early biomechanical studies suggested that sitting increased disk pressure ( Nachemson & Morris 1964 ). However, this was an isolated measurement in a single disk, with possible technical limitations. It is now possible to make more sophisticated measurements of different loads on different tissues and in different positions, e.g., on the disks, facet joints, ligaments, and different muscles ( Adams et al 2002 ). These suggest that sitting, in any type of chair, may actually unload the spine relative to standing ( Althoff et al 1992 ). Even some apparently strenuous tasks, such as working with the arms overhead, can lead to relative unloading ( Burton et al 1994 ). Thus, normal physical activities at work and leisure expose our spines to both loading and relative unloading.

It is also wrong to suppose that all spinal loading is harmful. Quite the contrary, some loading is essential for spinal health.

The same mechanical loading that can deform and damage spinal tissues also stimulates growth and repair. Adaptive remodeling strengthens bone, collagen, and muscle. Risk may then be more a matter of certain patterns or levels of loading that exceed the capacity for repair ( Adams et al 2002 ). The different rates at which spinal tissues are able to adapt to increased mechanical demands could mean that poorly vascularized tissues such as disks and ligaments might be more vulnerable. This may be important when levels of physical activity are suddenly increased, e.g., starting a new job or in sport.

It has been suggested that there might be a U-shaped risk between spinal loading and back pain. This was originally an attempt to explain contradictory findings that both heavy physical work and light activities such as sitting were associated with back pain. There is some biomechanical face-validity to this as a model for back injury. There is a theoretic argument that it reflects the capacity of spinal tissues to adapt to loading and the balance between repair and damage. The concept is even philosophically attractive, of “moderation in all things” with virtuous roots as far back as Aristotle! However, some of these early findings no longer appear valid. Overall, the current epidemiologic evidence on risk factors for back pain is not entirely supportive of U-shaped risk. Nevertheless, the value of this model is to reinforce the idea that loading is not always harmful and some loading is essential for health.

Physical demands of work
The most important question is whether occupational loading leads to mechanical overload damage to the spine. In vitro experiments certainly show that certain levels of loading (both sudden loads and cyclical loading) can produce mechanical disruption of vertebrae, end plates, and disks. But does that level of loading occur in life and is there any evidence of damage in vivo?

Brinckmann et al (1998) made one of the most careful studies. They made precise measurements of spinal X-rays from 355 workers who had been exposed to extreme physical demands and 737 unexposed controls. They showed that very heavy lifting and handling, particularly in miners working in confined underground conditions, led to reduced disk height. Substantial exposure to whole- body vibration on unsprung seats had the same effect. However, these X-rays were from historic archives. The jobs and the physical demands far exceeded what would be permitted in any North American or European country today. Ergonomic improvements and current regulations mean that today’s jobs rarely involve the kind of physical demands likely to cause any such lasting damage. This was most starkly demonstrated in Brinckmann et al’s study by the contrast between the effect of vibration on operators with unsprung seats and the lack of effect on those with damped seats.

The study was not designed to assess the relation between overload damage and symptoms, but some data happened to be available. One cohort with overload damage did not have any higher prevalence of back pain. Another cohort with a high prevalence of back pain did not show overload damage. So even mechanical overload damage did not necessarily result in symptoms.

Even if modern work does not cause any structural damage, it is still important to ask whether it is a risk factor for back pain. Since the classic study by Magora (1970) , there have been hundreds of studies on the relation between physical demands of work and back pain. There are also many good reviews ( Burdorf & Sorock 1997 , Bigos et al 1996 , Hoogendoorn et al 1999 , Videman & Battie 1999 , Waddell & Burton 2000 , Adams et al 2002 ).

Videman & Battie (1999) reviewed the influence of occupation on lumbar degeneration. This is perhaps the most authoritative statement from leading world experts. They concluded that there is evidence that occupational exposure can influence disk degeneration. However, this is a weak effect that explains a very small portion of the degeneration found in the adult population. Further, the lack of a clear dose–response relationship casts doubt on any strong causal link. Contrary to popular belief, occupational loading does not appear to play a dominant role in disk degeneration.

The UK Occupational Health Guidelines ( Carter & Birrell 2000 , Waddell & Burton 2000 ) tried to summarize the evidence on the complex relationships between physical demands of work and back pain.
1. Most adults (60–80%) experience LBP at some time, and it is often persistent or recurrent. It is one of the most common reasons for seeking health care, and it is now one of the commonest health reasons given for work loss.
2. There is strong epidemiological evidence that physical demands of work (manual materials handling, lifting, bending, twisting, and whole body vibration) can be associated with increased reports of back symptoms, aggravation of symptoms and “injuries”.
3. There is limited and contradictory evidence that the length of exposure to physical stressors at work (cumulative risk) increases reports of back symptoms or of persistent symptoms.
4. There is strong evidence that physical demands of work (manual materials handling, lifting, bending, twisting, and whole body vibration) are a risk factor for the incidence (onset) of LBP, but overall it appears that the size of the effect is less than that of other individual, non-occupational and unidentified factors.
[Note: evidence statements 2 and 4 are not incompatible. Whilst the epidemiological evidence shows that low back symptoms are commonly linked to physical demands of work, that does not necessarily mean that LBP is caused by work. Although there is strong scientific evidence that physical demands of work can cause individual attacks of LBP, overall that only accounts for a modest proportion of all LBP occurring in workers.]
5. There is moderate scientific evidence that physical demands of work play only a minor role in the development of disc degeneration.
6. There is strong epidemiological and clinical evidence that care seeking and disability due to LBP depend more on complex individual and work-related psychosocial factors than on clinical features or physical demands of work.

Manual handling
Manual materials handling involves various combinations of lifting, moving, carrying, and handling physical loads. It is difficult to separate manual handling per se from generally heavy manual jobs. An important subgroup involves patient handling by nurses and other health workers. Patients are hardly “materials”, but the principles are the same!

There is strong and consistent evidence that workers in jobs involving manual handling report more back pain ( Burdorf & Sorock 1997 , Hoogendoorn et al 1999 ). The effect size is weak–moderate (RR or OR ranging from about 1.5 to 3). The UK Labour Force Surveys of the 1990s consistently showed that manual workers report all musculoskeletal complaints more than non-manual workers. They also had more persistent symptoms 3 years after stopping work and attributed them to work ( Jones et al 1998 ).

There is limited and inconsistent evidence on manual handling as a risk factor for disk prolapse or sciatica.

Most workers’ compensation data suggest that men with heavy manual jobs report more back injuries at work. The data are less clear for women. Nurses and certain other groups of health workers report more back pain and injuries, but they are a special group.

Almost all data sets show that workers with heavy manual jobs lose more time from work with back pain. Some, but not all, studies show that workers with heavy manual jobs have more spells off work with back pain. Others show that, when they are off, they return to work more slowly. There is wide variation in long-term disability and early retirement rates in different jobs but, surprisingly, this does not correlate well with the physical demands of work. However, these data do not tell us whether or not heavy work is the cause of more disabling back pain. It could equally be effect. It may simply be more difficult to do a heavy job when you have back pain, whatever its cause.

In summary, workers in heavy manual jobs do get more back trouble, but we must be careful how we interpret this.

Lifting
Ideally, we would like to identify which physical activities in heavy manual work might cause back trouble. From biomechanical studies, lifting, bending, and twisting are most likely to damage the spine. These are also the activities that have been studied most in the workplace.

Industrial accident and workers’ compensation statistics certainly show that back injuries are reported more commonly in jobs that involve:
  • heavy lifting
  • lifting objects which are bulky or must be held away from the body
  • lifting from the floor
  • frequent lifting.

The more general role of lifting as a risk factor for back pain is less clear. It is difficult to separate any specific effect of lifting from manual handling and heavy physical work in general. The reviews by Burdorf & Sorock (1997) and Hoogendoorn et al (1999) were unable to reach any definite answer. Perhaps we should simply accept that, in principle, frequent heavy lifting carries about the same risk as manual handling.

It might seem possible that handling unexpectedly heavy or asymmetric loads would carry a higher risk of injury. In fact, a recent biomechanical study found no evidence to support this ( van der Burg et al 2000 ). It seems the neuromuscular apparatus is robust enough to cope.

The lack of clear epidemiologic evidence means that lifting and handling guidelines and regulations are based on theory and consensus. In theory, ability to lift and the risk of injury will vary with individual strength. Heavy lifting that exceeds the person’s ability may carry greater risk. However, it is difficult to set safe limits. In addition to the weight, we must also consider the frequency and rate of lifting, the level of the lift, and the position of the body. The distance between the load and the body greatly increases the forces on the back. Lifting standards set by the US National Institute for Occupational Safety and Health ( Waters et al 1993 ) or the UK Health and Safety Executive’s guidance ( HSE 1992 ) are simply based on experience and consensus.

Despite popular clinical belief, there is limited epidemiologic evidence on lifting as a risk factor for disk prolapse or sciatica.

There is little separate information on pushing and pulling, although heavy manual jobs often involve these activities as well. In one careful prospective study, Hoozemans et al (2002) found a limited relation between pulling and pushing and low back complaints.

Bending and twisting
There is strong biomechanical evidence that lifting combined with bending and twisting has the potential to injure spinal structures. The risk of disk prolapse is especially high with combined loading and twisting. Twisting alone, without lifting, probably does not carry much risk. This is probably because of the anatomic limitations to vertebral rotation.

Burdorf & Sorock (1997) found 10 studies giving strong and consistent evidence that frequent bending and twisting is a risk factor for back pain. The effect size is weak–moderate (RR or OR ranging from about 1.3 to 2.8).

Repetitive strain
Repetitive strain injury is currently fashionable, particularly in a medicolegal context. It usually affects the upper limb, although the pathology and the diagnosis itself are hotly disputed. Several legal claims about occupational back pain have explored the same concept.

It is certainly possible to produce fatigue failure due to repetitive strain in the laboratory. There are, however, major differences between such experiments and clinical back pain. Most of the biomechanical studies are on bones and disks, which are probably not the source of most work-related back pain. There are some hypotheses about how this might apply to soft tissues but little experimental data. The many thousands of rapid cycles required to produce failure are quite different from the pattern of repeated everyday movements in work. In vitro experiments also fail to allow for biologic adaptation and healing in response to repeated strain.

There is little clinical evidence to support the idea of repetitive strain injury to the back. Most claimants have already done the same tasks over long periods without symptoms. There is nothing new or changed in their job when back pain develops. The symptoms are subjective and are the same as common, ordinary backache. No one has defined any specific clinical syndrome or objective pathology with repetitive strain injury. When back pain is present, such repetitive activities may aggravate symptoms, but yet again this is not proof of cause and effect. Burdorf & Sorock (1997) could only find three studies of repetitive work, and two out of three found no association.

In summary, repetitive strain injury seems to be more of a medicolegal concept than a clinical or pathologic reality.

Static work postures and sitting
Several early cross-sectional studies suggested that there was an association between sitting and back pain. This was linked to biomechanical theories about raised disk pressure, but we have already seen these findings were suspect. Moreover, this is static loading and any pressure is very low compared with that required to cause experimental damage. We have already discussed the more sophisticated U-shaped model of risk. Despite these theories, there is no actual biomechanical evidence that sitting damages the spine.

Hartvigsen et al (2000) reviewed 35 epidemiologic studies on sitting. Only eight had a satisfactory experimental design. Only one showed any significant relation between prolonged sitting at work and low back pain. Seven out of eight showed no effect. They concluded that the extensive evidence now available does not support the popular belief that sitting is a risk factor for back pain.

Seating has fluctuated greatly over the centuries in different cultures, from upright to slouched positions ( Pynt et al 2002 ). There is no evidence that any type of seat or sitting position makes any difference to the risk of back pain. The current epidemiologic evidence demolishes debate about the best form of seating and the “ideal” seated posture. Biomechanical arguments for or against different positions now seem pretty irrelevant! So choice of seat is entirely a matter of subjective comfort.

In summary, sitting is not a risk factor for back pain. Prolonged sitting in one position may aggravate back pain that is already present, whatever its cause. Experience suggests that it is reasonable advice to change position and get up and move about regularly. But all of that is more a matter of coping with back pain rather than anything to do with risk or cause.

Driving and exposure to whole-body vibration
Driving is different from ordinary sitting, because it involves exposure to whole-body vibration. The dominant frequency of vibration in many vehicles is 4–6Hz, which is also the resonant frequency of the spine ( Pope et al 1991 ). Most of the biomechanical evidence is about whole-body vibration, but most of the epidemiologic evidence is about driving. However, driving exposes us to more than just vibration ( Heliovaara 1999 ). It involves static and sometimes awkward postures with variable lumbar support. It requires use of the legs with imposed loads on the spine. Perhaps most important, there is exposure to transmitted shocks from the road, jolting, and various accelerations. Unfortunately, the epidemiologic evidence cannot distinguish the possible risks of driving and of whole-body vibration.

Kjellberg et al (1994) and Wickstrom et al (1994) made an extensive review of the health effects of whole-body vibration. They concluded that there was extensive evidence of an association with low back pain. However, at that time there was insufficient evidence to establish the exposure–response relationship.

The Finnish twin study ( Battié et al 2002 ) found no association between lifetime driving exposure and disk degeneration. Videman et al (2000) took a more extreme example. They looked at top rally drivers who were regularly subjected to severe whole-body vibration and compared them with normal controls. This was a small study, but they did not find any MRI evidence of increased degenerative changes.

Lings & Leboeuf-Yde (2000) reviewed the more recent epidemiologic evidence. They concluded that there is strong evidence that driving is a risk factor for back pain and limited evidence for disk prolapse. However, there is only weak evidence on a dose–response relationship. The effect size is moderate: Burdorf & Sorock (1997) found RR or OR generally ranging from about 1.5 to 3.9. Lings & Leboeuf-Yde (2000) concluded that there is good reason to reduce exposure to whole-body vibration to the lowest practical level. Modern, damped, vehicle seats probably achieve this. There is little evidence of harm occurring on such modern seats. Perhaps being deliberately provocative, they suggested this is no longer an important problem.

Leisure activities and sports
Hoogendoorn et al (1999) reviewed 17 studies of sports and physical activity during leisure time. The results were inconsistent. There is no clear evidence that most sports activity or total physical activity during leisure time are risk factors for back pain. Most important, leisure activities such as swimming, walking, running, cycling, golf, or physical exercise do not appear to carry any risk. Overall, the prevalence of back pain is no higher in those who are physically active or take part in general athletic activities. On the contrary, as we saw earlier, improved physical and mental health and physical fitness are likely to be beneficial.

There is limited evidence that certain strenuous sports such as weightlifting and gymnastics may carry an increased risk of disk degeneration and vertebral damage ( Sward et al 1990 , 1991; Videman et al 1995 ). Some high-level and competitive sports may also be associated with an increased prevalence of back pain. That link could, however, involve both physical and psychosocial factors.

In summary, apart from certain high-level and competitive sports, most leisure and normal sporting activities seem likely to do more good than harm. Indeed, the clinical evidence shows that exercise and sports activities are the best treatment for back pain!

ENVIRONMENTAL RISK FACTORS: PSYCHOSOCIAL ASPECTS OF WORK
Work obviously imposes physical demands on workers, but it also imposes psychosocial demands.

During the 1990s, there was a lot of interest in job satisfaction and its possible influence on back pain. The classic Boeing Study ( Bigos et al 1991 ) found that “hardly ever” enjoying the job was one of the few predictors of reporting a back injury over the next 4 years. However, it was actually a very weak predictor. The only reason it got so much attention was because so many of the findings of the Boeing Study were negative. There are now a large number of studies of job satisfaction and reported back pain, injury claims, seeking health care, and sick listing with back pain ( Burdorf & Sorock 1997 , Hoogendoorn et al 2000 ).

Over the past 20 years, there has been a great deal of research into more detailed psychosocial aspects of work:
• work “stress”
• high job demands and pace
• poor job content: low decision latitude, low job control, and monotonous work
• low social support from fellow workers or supervisors
• job “strain.”

High job demands and conflict at work produce stress. Poor control over work and poor social support make it harder to cope with stress. This led to the “demand–control” theory that the level of job “strain” depends on the balance between high demands vs low control and support ( Karasek 1979 , Karasek & Theorell 1990 ). These concepts were originally developed for cardiovascular disease, but have since been applied to musculoskeletal disorders.

There are now at least 70 studies on psychosocial aspects of work and spinal pain, many of them from Scandinavia. Fortunately, there are also good reviews ( Burdorf & Sorock 1997, Davis & Heaney 2000, Hoogendoorn et al 2000 , Linton 2001 ).

Most of the studies in this area are cross-sectional, with all their limitations. Hoogendoorn et al (2000) could only find 11 cohort and two case–control studies. Davis & Heaney (2000) reviewed the methodologic problems in this field. “Psychosocial aspects of work” are by definition subjective. So individual perceptions or psychosocial reactions to the job are what matter, rather than any more “objective” measures of social or organizational characteristics of the job. There is a particular problem with potential confounding. We rely on self-report of both the risk factor and the outcome. Few studies allow for the physical demands of work. As we have already seen, we must be cautious how we interpret such complex associations.

Summary of evidence
The following is a brief summary of the evidence on each of these psychosocial aspects of work as risk factors for the onset of back pain.

Job satisfaction
There is strong and consistent evidence that job satisfaction is a risk factor for reported back pain ( Burdorf & Sorock 1997, Davis & Heaney 2000 , Hoogendoorn et al 2000 ). The effect size is weak (RR or OR generally ranging from about 1.4 to 2.4). Part of the problem may be attempting to measure such a complex psychosocial issue by simple questions.

Job “stress”
Stress is now fashionable and the subject of intense professional, occupational, and legal debate. This is not the place to enter the fray that surrounds this area and we will limit ourselves to the evidence on stress as a risk factor for back pain.

There is actually limited epidemiologic evidence that job “stress” is a risk factor for reported back pain ( Burdorf & Sorock 1997 , Davis & Heaney 2000 ). The effect size is weak (RR or OR of the order of 1.3–2.1). Part of the problem may be attempting to measure such a complex psychosocial issue by simple questions.

High mental demands
There is inconsistent evidence on high mental demands and work pace. Some studies report positive findings, but as many fail to show a significant association. Burdorf & Sorock (1997) , Davis & Heaney (2000) and Hoogendoorn et al (2000) all concluded that it is not possible to demonstrate that job demands are a risk factor for back pain.

As already noted, however, there is a practical difficulty defining and measuring “job demands” and “work stress.”

Poor job content: low decision latitude, low job control, and monotonous work
There is inconsistent evidence on decision latitude, job control, and monotonous work. Different reviews reach different conclusions. Even if there is any effect, it appears to be weak. Hoogendoorn et al (2000) and Davis & Heaney (2000) concluded that there is insufficient evidence on poor job content as a risk factor for back pain.

Low social support
There is strong and generally consistent evidence that low social support from fellow workers and supervisors is a risk factor for reported back pain. The effect size is weak (RR or OR generally ranging from about 1.3 to 1.9).

Job “strain”
There is insufficient evidence to support the demand–control theory in back pain or musculoskeletal disorders.

Altogether, it is surprising that psychosocial aspects of work seem to have such a weak effect. Perhaps it is because we have quite crude methods of measuring what are really complex psychosocial issues. Perhaps it is because most studies and reviews look at each aspect individually.

Bartys et al (2001) showed that there might be a cumulative effect. Individual and work-related psychosocial issues appear to interact, which is exactly what you would expect. Preliminary results from their prospective study are also encouraging. Such interactions appear to be a promising area for further research

Even more fascinating are possible interactions between physical demands and psychosocial aspects of work. A few years ago, there was an argument about which were more important risk factors in back pain, but that was naive. Both may play a role. So the real question is whether and how they might have an additive or interactive effect.

Davis & Heaney (2000) provided one of the most thoughtful reviews of these complex relationships. They suggested three potential links. First, physical demands and psychosocial factors could each contribute independently to the onset or consequences of back pain. These might also have an additive effect. Second, psychosocial factors may modulate the relation between physical demands and back pain. For example, poor psychosocial conditions might reduce ability to cope with physical demands that would otherwise be tolerated. Third, physical demands and psychosocial aspects may co-vary. Many jobs involve both greater physical demands and poorer psychosocial conditions. Until recently, few studies investigated both physical demands and psychosocial aspects, which raises the possibility of confounding. Most studies of psychosocial aspects did not adjust for physical demands.

Biomechanical risk factors might cause back pain through excessive loading or repetitive loading. Psychosocial aspects of work were originally thought of as “stressors.” There are various theories about the possible biologic effects of stress, but there is no convincing evidence that stress is a direct cause of physical pathology in the back. However, psychosocial factors might have more indirect effects on the biomechanics of the back. First, psychosocial factors could influence spinal loading by changes in muscles, forces exerted, and trunk movement. Suffice to say at this point that they could involve changes in muscle tension, muscle activity, and patterns of movement. Second, these muscle changes or the neurohormonal changes that occur with stress could influence metabolic activity in various tissues of the back. Third, psychosocial factors could influence the neurophysiology of pain in various ways. Finally, psychosocial factors might influence the reporting of low back pain.

There are several tantalizing studies ( Devereux et al 1999, Hoogendoorn et al 2002 ), but at present there is limited evidence for any of these mechanisms. This is clearly an area where much more research is needed. Whether or not these mechanisms turn out to be important for the initial cause of back pain, they have major implications for how we manage back pain at work.

CONCLUSIONS
What does this information about risk factors mean in practice?

Approaches to prevention and control
Occupational back pain is an enormous problem, and the ideal answer would be to prevent it. Biomechanic and ergonomic approaches aim to reduce back injuries by controlling physical hazards and potential risk factors. This primary prevention may be an unrealistic goal ( Burton 1997 ).

Reviews by van Poppel et al (1997) and Linton & van Tulder (2001) could not find good evidence on the effectiveness of primary prevention. Historically, this approach seems to have helped control more extreme physical demands and risks of work in previous generations. But there is little evidence that modern work is damaging to the back. So it is not surprising that there is also little evidence that this approach is effective in reducing the incidence of back trouble. And as back pain is so common, perhaps the goal of primary prevention is unrealistic.

Recent clinical developments also raise questions about that approach to risk. On the one hand, we try to prevent back trouble by reducing physical risk factors. On the other hand, modern treatment and rehabilitation aim to increase physical activity levels and challenge the musculoskeletal system.

This leads to a different approach to risk. We must continue to reduce more extreme hazards that might lead to damage. But controlling the physical demands of modern work is probably more a matter of comfort and enabling workers with back pain to cope (whatever the cause of their pain; Hadler 1997 ). This is secondary prevention – reducing the consequences of back pain, even if we cannot prevent it in the first place. As back pain is almost universal, and its natural history is to recur, this may be more realistic.

Physical demands of work remain important: manual handling, lifting, bending and twisting, and exposure to vibration. Ergonomics still has a role here. Ergonomics aims to improve the “fit” between people, the things they use, and the way they use them. Information about human abilities, attributes, and limitations is also used to improve the design of equipment and tasks. The goal is to maximize comfort and safety for workers, by preventing excessive fatigue, discomfort, or stress. Occupational health often uses the same approach to enable workers to remain at work, return to work when they have back pain, and reduce recurrences.

We must also recognize the importance of psychosocial as well as physical demands of work. Physical risk factors may be most important for the initial onset of back pain. But psychosocial issues are probably even more important for its impact and consequences, for management, and for chronic pain and disability. Addressing psychosocial aspects of work and providing support may be just as important as modifying the physical demands.

What should we tell patients?
The review of individual risk factors suggests that most of us are going to get back pain at some time in our lives. It does not make much difference whether we are male or female, young or old, tall and thin, or small and fat. There is not a lot we can do about these personal characteristics in any event, but we do not need to worry about them. We may all be fated to have some back pain, but there is nothing in our genes that dictates it will inevitably lead to chronic pain and disability.

This has implications for what you really need to know about back problems. Too often, you will be told that you have back pain because you are too tall, too fat, the wrong build, or their legs are of unequal lengths. This is nonsense, and it is a dangerous message because it implies your back pain is inevitable and there is nothing that they or we can do about it.

Advice about work is a critical part of managing back pain, and that advice depends on whether work is a risk. Sadly, too much advice is based on old myths that current evidence shows are wrong. Many patients and health professionals are firmly convinced that heavy manual work must somehow cause back injury or degenerative changes. So further exposure might cause further damage and hinder recovery or lead to chronic pain and disability. Strong scientific evidence now explodes these myths.

Back pain is certainly work-related to the extent that people of working age commonly get back pain and it impacts on their work. Physical demands at work are clearly associated with occupational back pain . Extreme loading may cause lasting damage, but that is rare in modern work. Occupational exposure can affect disk degeneration, but the effect is weak. Physical demands of work may provoke episodes of back pain, but that only accounts for a small portion of such a common bodily symptom. Work may aggravate back pain, whatever its cause. And back pain may make it more difficult to meet certain physical demands. The influence of cumulative exposure remains uncertain, but it seems not to be related to persistent back trouble. Altogether, there is little convincing evidence that work is physically harmful to the back. On the contrary, work is generally good for people with back pain.

All too often, doctors and therapists tell patients that their back pain is due to their job. So they advise them to take time off work, change to lighter work, give up their job, and even to retire early. This review shows that there is very little evidence to support such advice. It is usually not possible to say with any certainty that a patient’s back pain is due to his or her job, or that the job is bad for his or her back. Too often, we give such advice glibly without adequate thought for the impact on our patients and their families. Try to imagine if someone casually told you to give up your job, for no very good reason – except they “thought” it might be good for you. How would that affect you? Education, knowledge, and insight would probably allow you to discount such advice.

It is too important a matter to make these decisions lightly on such flimsy evidence. It is rarely justified to advise patients to stay off work, change their job, or give up work completely because of ordinary backache. Advice such as that can easily become self-fulfilling.

References

Adams M A, Bogduk N, Burton K, Dolan P 2002 The biomechanics of back pain. Churchill Livingstone, Edinburgh

Althoff I, Brinckmann P, Frobin W, Sandover J, Burton K 1992 An improved method of stature measurement for quantitative determination of spinal loading: Application to sitting postures and whole body vibration. Spine 17: 682–693

Andersson G B J 1997 The epidemiology of spinal disorders. In: Frymoyer J W (ed.) The adult spine: principles and practice, 2nd edn. Lippincott-Raven, Philadelphia, pp 93–141

Balague F, Troussier B, Salminen J J 1999 Non-specific low back pain in children and adolescents: risk factors. European Spine Journal 8: 429–438

Bartys S, Tillotson M, Burton K et al 2001 Are occupational psychosocial factors related to back pain and sickness absence? In: Hanson M (ed.) Contemporary ergonomics 2001. Taylor & Francis, London, pp 23–28
Battie M C, Videman T 2003 Genetic transmission of common spinal disorders. In: Herkowitz H (ed.) The lumbar spine, 3rd edn. Lippincott, Williams & Wilkins, Philadelphia (in press)

Battie M C, Videman T, Gill K et al 1991 Smoking and lumbar intervertebral disc degeneration: an MRI study of identical twins. Spine 16: 1015–1021

Battie M C, Videman T, Gibbons L E, Fisher L D, Manninen H, Gill K 1995 Determinants of lumbar disc degeneration. A study relating lifetime exposures and MRI findings in identical twins. Spine 20: 2601–2612

Battie M C, Videman T, Gibbons L E et al 2002 Occupational driving and lumbar disc degeneration: a case–control study. Lancet 360: 1369–1374

Bigos S J, Battie M C, Spengler D M et al 1991 A prospective study of work perceptions and psychological factors affecting the report of back injury. Spine 16: 1–6

Bigos S J, Holland J, Webster M et al 1996 Prevention and risks of reporting occupational back problems: a methodological literature analysis. American Academy of Orthopedic Surgeons Report. AAOS, Rosemount, Illinois

Bombardier C, Kerr M S, Shannon H S, Frank J W 1994 A guide to interpreting epidemiologic studies on the etiology of back pain. Spine 19 (18S): 2047S–2056S

Brattberg G 1993 Back pain and headache in Swedish school children: a longitudinal study. Quality of Life Research 3: 157–162

Brattberg G 1994 The incidence of back pain and headache among Swedish school children. Quality of Life Research 3: S27–S31

Bressler H B, Keyes W J, Rochon P A, Bradley E 1999 The prevalence of low back pain in the elderly: a systematic review of the literature. Spine 24: 1813–1819

Brinckmann P, Frobin W, Biggemann M, Tillotson M, Burton K 1998 Quantification of overload injuries to thoracolumbar vertebrae and discs in persons exposed to heavy physical exertions or vibration at the work-place. Part II. Occurrence and magnitude of overload injury in exposed cohorts. Clinical Biomechanics 13 (suppl. 2): S(2)1–S(2)36.

Burdorf A, Sorock G 1997 Positive and negative evidence of risk factors for back disorders. Scandinavian Journal of Work and Environmental Health 23: 243–256

Burton A K 1997 Back injury and work loss: biomechanical and psychosocial influences. Spine 22: 2575–2580

Burton A K, Tillotson K M, Boocock M G 1994 Estimation of spinal loads in overhead work. Ergonomics 37: 1311–1322

Burton A K, Clarke R D, McClune T D, Tillotson K M 1996a The natural history of low back pain in adolescents. Spine 21: 2323–2328

Burton A K, Tillotson K M, Symonds T L, Burke C, Mathewson T 1996b Occupational risk factors for first-onset and subsequent course of low back trouble. A study of serving police officers. Spine 21: 2612–2620

Cady L, Bischoff D, O’Connel E 1979 Strength and fitness and subsequent back injuries in firefighters. Journal of Occupational Medicine 21: 269–272

Carter J T, Birrell L N (eds) 2000 Occupational health guidelines for the management of low back pain at work – principal recommendations. Faculty of Occupational Medicine, London. Available online at: www.facoccmed.ac.uk

Croft P R, Rigby A S 1994 Socioeconomic influences on back problems in the community in Britain. Journal of Epidemiology and Community Health 48: 166–170

Davis K G, Heaney C A 2000 The relationship between psychosocial work characteristics and low back pain: underlying methodological issues. Clinical Biomechanics 15: 389–406

Devereux J J, Buckle P W, Vlachonikolis I G 1999 Interactions between physical and psychosocial risk factors at work increase the risk of back disorders; an epidemiological approach. Occupational and Environmental Medicine 56: 343–353

Deyo R A, Tsui-Wu Y-J 1987 Functional disability due to back pain. Arthritis and Rheumatism 30: 1247–1253

Dione C, Koepsell T D, Von Korff M, Deyo R A, Barlow W E, Checkoway H 1995 Formal education and back-related disability: in search of an explanation. Spine 20: 2721–2730

Dionne C E, Von Korff M, Koepsell T D, Deyo R A, Barlow W E, Checkoway H 2001 Formal education and back pain: a review. Journal of Epidemiology and Community Health 55: 455–468

Edmond S L, Felson D T 2000 Prevalence of back symptoms in elders. Journal of Rheumatology 27: 220–225

Goldberg M S, Scott S C, Mayo N 2000 A review of the association between cigarette smoking and the development of nonspecific back pain and related outcomes. Spine 25: 995–1014

Hadler N M 1997 Back pain in the workplace. What you lift or how you lift matters far less than whether you lift or when. Spine 22: 935–940

Hakala P, Rimpela A, Salminen J J, Virtanen S, Rimpela M 2002 Back, neck and shoulder pain in Finnish adolescents: national cross-sectional surveys. British Medical Journal 325: 743–745

Hartvigsen J, Leboeuf-Yde C, Lings S, Corder E H 2000 Is sitting while at work associated with low back pain? A systematic, critical literature review. Scandinavian Journal of Public Health 28: 230–239

Hartvigsen J, Kyvik K O, Leboeuf-Yde C, Lings S, Bakketeig L 2003 Ambiguous relation between physical workload and low back pain: a twin control study. Occupational and Environmental Medicine 60: 109–114

Heikkila J K, Koskenvuo M, Heiovaara M et al 1989 Genetic and environmental factors in sciatica. Evidence from a nationwide panel of 9365 adult twin pairs. Annals of Medicine 21: 393–398

Heliovaara M 1999 Editorial: Work load and back pain. Scandinavian Journal of Work and Environmental Health 25: 385–386

Heliovaara M, Impivaara O, Sievers K et al 1987 Lumbar disc syndrome in Finland. Journal of Epidemiology and Community Health 41: 251–258

Hoogendoorn W E, van Poppel M N M, Bongers P M, Koes B W, Bouter L M 1999 Physical load during work and leisure time as risk factors for back pain. Scandinavian Journal of Work and Environmental Health 25: 387–403

Hoogendoorn W E, van Poppel M N M, Bongers P M, Koes B W, Bouter L M 2000 Systematic review of psychosocial factors at work and private life as risk factors for back pain. Spine 25: 2114–2125

Hoogendoorn W E, Bongers P M, de Vet H C W, Ariëns G A M, van Mechelen W, Bouter L M 2002 High physical work load and low job satisfaction increase the risk of sickness absence due to low back pain: results of a prospective cohort study. Occupational and Environmental Medicine 59: 323–328

Hoozemans M J M, van der Beek A J, Frings-Dresen M H W, van der Woude L H V, van Dijk F J H 2002 Low-back and shoulder complaints among workers with pushing and pulling tasks. Scandinavian Journal of Work and Environmental Health 28: 293–303.

Howell C J, Dean T, Lucking L, Dziedzic K, Jones P W, Johanson R B 2002 Randomised study of long term outcome after epidural versus non-epidural analgesia during labour. British Medical Journal 325: 357–360

HSE 1992 Manual handling: guidance on regulations. Manual handling operations regulations 1992. HMSO, London

HSE 2000 Management of health and safety at work: management of health and safety at work regulations 1999 – approved code of practice and guidance (L21). HSE Books, Norwich

Jones J R, Hodgson J T, Clegg T A, Elliott R C 1998 Self-reported work-related illness in 1995: results from a household survey. HSE Books, Her Majesty’s Stationery Office Norwich

Karasek R A 1979 Job demands, job decision latitude and mental strain: implications for job redesign. Administrative Science Quarterly 24: 285–308

Karasek R A, Theorell T 1990 Healthy work. Basic Books, New York

Kawaguchi Y, Osada R, Kanamori M et al 1999 Association between an aggrecan gene polymorphism and lumbar disc degeneration. Spine 24: 2456–2460

King A, Coles B 1992 The health of Canada’s youth: views and behaviours of 11-, 13- and 15-year olds from 11 countries. Health and Welfare Canada, Ottowa: (data quoted in Waddell et al 2002, p. 3)

Kjellberg A, Wickstrom B O, Landstrom U 1994 Injuries and other adverse effects of occupational exposure to whole-body vibration. Arb Halsa 41

Leboeuf-Yde C 1999 Smoking and low back pain: a systematic literature review of 41 journal articles reporting 47 epidemiologic studies. Spine 24: 1463–1470

Leboeuf-Yde C 2000a Body weight and low back pain: a systematic literature review of 56 journal articles reporting on 65 epidemiologic studies. Spine 25: 226–237

Leboeuf-Yde C 2000b Alcohol and low back pain: a systematic literature review. Journal of Manipulative and Physiological Therapeutics 23: 343–346

Lings S, Leboeuf-Yde C 2000 Whole body vibration and low back pain: a systematic, critical review of the epidemiological literature 1992–1999. Archives of Occupational and Environmental Health 73: 290–297

Linton S J 2000 A review of psychological risk factors in back and neck pain. Spine 25: 1148–1156

Linton S J 2001 Occupational psychological factors increase the risk for back pain: a systematic review. Journal of Occupational Rehabilitation 11: 53–66

Linton S J, van Tulder M W 2001 Preventive interventions for back and neck pain problems: what is the evidence? Spine 26: 778–787

MacGregor A J, Griffiths G O, Baker J, Spector T D 1997 Determinants of pressure pain threshold in adult twins: evidence that shared environmental influences predominate. Pain 73: 253–257

MacGregor A J, Andrew T, Snieder H, Sambrook P, Spector T D 1999 A genetic model for lower back pain: a population-based MRI study of twins. Arthritis and Rheumatism 49: 5146 (abstract)

Magora A 1970 Investigation of the relation between low back pain and occupation. Industrial Medicine 39: 28–37, 504–510; 41: 5–9

Makela M 1993 Common musculoskeletal syndromes. Prevalence, risk indicators and disability in Finland. ML 23. Publications of the Social Insurance Institution, Finland

Manninen P, Riihimaki H, Heliovaara M, Makela P 1995 Mental distress and disability due to low back and other musculoskeletal disorders – a ten year follow up. Presented to the 22nd annual meeting of the International Society for the Study of the Lumbar Spine, Helsinki

Mannion A F, Dolan P, Adams M A 1996 Psychological questionnaires: do ‘abnormal’ scores precede or follow first-time low back pain? Spine 21: 2603–2611

Marras W S, Davis K G, Jorgensen M 2002 Spine loading as a function of gender. Spine 27: 2514–2520

McCormick A, Fleming D, Charlton J 1995 Morbidity statistics from general practice. Fourth national study 1991–1992. Office of Population Censuses and Surveys Series MB5 no. 3. HMSO, London, pp 1–366

Nachemson A, Morris J M 1964 In vivo measurement of intradiscal pressure. Journal of Bone and Joint Surgery 46A: 1077–1092

Nachemson A, Vingard E 2000 Influences of individual factors and smoking on neck and low back pain. In: Nachemson A, Jonsson E (eds) Neck and back pain: the scientific evidence of causes, diagnosis and treatment. Lippincott Williams & Wilkins, Philadelphia, pp 97–126

National Research Council 1999 Work-related musculoskeletal disorders: report, workshop summary and workshop papers. National Academy Press, Washington, DC. Available online at www.nap.edu

National Research Council & Institute of Medicine 2001 Musculoskeletal disorders and the workplace. National Academy Press, Washington, DC

Negrini S, Carabalona R 2002 Backpacks on! Schoolchildren’s perceptions of load, associations with back pain and factors determining the load. Spine 27: 187–195.

NIOSH 1997 Musculoskeletal disorders and workplace factors. A critical review of epidemiologic evidence for work-related musculoskeletal disorders of the neck, upper-extremity, and low back. NIOSH, Cincinnati

Ostgaard H C, Roos-Hansson E, Zetherstrom G 1996 Regression of back and posterior pelvic pain after pregnancy. Spine 21: 2777–2780

Paassilta P, Lohiniva J, Göring H H H et al 2001 Identification of a novel common genetic risk factor for lumbar disc disease. Journal of the American Medical Association 285: 1843–1849.

Pope M H, Wilder D G, Krag M H 1991 Biomechanics of the lumbar spine: A. Basic principles. In: Frymoyer J W (ed.) The adult spine: principles and practice. Raven Press, New York, pp 1487–1501

Pynt J, Higgs J, Mackey M 2002 Milestones in the evolution of lumbar spinal postural health in seating. Spine 27: 2180–2189

Rothman K J, Greenland S 1998 Causation and causal inference. In: Rothman KJ, Greenland S (eds) Modern epidemiology. Lippincott-Raven, Philadelphia, pp 7–28

Sambrook P N, MacGregor A J, Spector T D 1999 Genetic influences on cervical and lumbar disc degeneration: a magnetic resonance imaging study in twins. Arthritis and Rheumatism 42: 366–372.

Schoene M 2002 Back pain in children and adolescents: is medicine clinging to an outmoded view? The Back Letter 17(3): 25, 32–34.

Straaton K V, Maisiak R, Wrigley J M, White M B, Johnson P 1996 Barriers to return to work among persons unemployed due to arthritis and musculoskeletal disorders. Arthritis and Rheumatism 39: 101–109

Sward L, Hellstrom M, Jacobsen B et al 1990 Back pain and radiologic changes in the thoraco-lumbar spine of athletes. Spine 15: 124–129

Sward L, Hellstrom M, Jacobsson B, Nyman R, Peterson L 1991 Disc degeneration and associated abnormalities of the spine in elite gymnasts. Spine 16: 437–443

Sydsjö A, Alexanderson K, Dastserri M, Sydsjö G 2003 Gender differences in sick leave related to back pain diagnoses: influence of pregnancy. Spine 28: 385–389.

van der Burg J C E, van Dieen J H, Toussaint H M 2000 Lifting an unexpectedly heavy object: the effects on low-back loading and balance loss. Clinical Biomechanics 15: 469–477

van Poppel M N M, Koes B W, Smid T et al 1997 A systematic review of controlled clinical trials on the prevention of back pain in industry. Occupational and Environmental Medicine 54: 841–847

Videman T, Battie M C 1999 Spine update: the influence of occupation on lumbar degeneration. Spine 24: 1164–1168

Videman T, Sarna S, Battié M C et al 1995 The long-term effects of physical loading and exercise lifestyles on back-related symptoms, disability, and spinal pathology among men. Spine 20: 699–709

Videman T, Simonen R, Usenius J-P, Osterman K, Battie MC. 2000 The long-term effects of rally driving on spinal pathology. Clinical Biomechanics 15: 83–86

Vingard E, Mortimer M, Wiktorin C et al 2002 Seeking care for low back pain in the general population. Spine 27: 2159–2165

Vogt M T, Hanscom B, Lauerman W C, Kang J D 2002 Influence of smoking on the health status of spinal patients: the National Spine Network Database. Spine 27: 313–319

Waddell G, Burton A K 2000 Occupational health guidelines for the management of low back pain at work – evidence review. Faculty of Occupational Medicine, London. Available online at: www.facoccmed.ac.uk

Waddell G, Waddell H 2000 Social influences on neck and back pain and disability. In: Nachemson A, Jonsson E (eds) Neck and back pain: the scientific evidence of causes, diagnosis and treatment. Lippincott, Williams & Wilkins, Philadelphia, pp 13–55

Waddell G, Aylward M, Sawney P 2002 Back pain, incapacity for work and social security benefits: an international literature review and analysis. Royal Society of Medicine Press, London

Walsh K, Cruddas M, Coggon D 1992 Low back pain in eight areas of Britain. Journal of Epidemiology and Community Health 46: 227–230

Waters T R, Putz-Anderson V, Garg A, Fine L J 1993 Revised NIOSH equation for the design and evaluation of manual lifting tasks. Ergonomics 36: 749–776

Watson K D, Papageorgiou A C, Jones G T et al 2002 Low back pain in schoolchildren: occurrence and characteristics. Pain 97: 87–92

Watson K D, Papageorgiou A C, Jones G T et al 2003 Low back pain in school children: the role of mechanical and psychosocial factors. Pain (in press)

Wedderkopp N, Leboeuf-Yde C, Andersen L B, Froberg K, Hansen H S 2001 Back pain reporting pattern in a Danish population-based sample of children and adolescents. Spine 26: 1879–1883

Wickstrom B O, Kjellberg A, Landstrom U 1994 Health effects of long-term occupational exposure to whole-body vibration: a review. International Journal of Industrial Ergonomics 14: 273–292

Wiersema B M, Wall E J, Foad S L 2003 Acute backpack injuries in children. Pediatrics 111: 163–166