29th July 2011, Volume 124 Number 1338

David McLean, Neil Pearce, Christopher B Walls, Richard D Wigley

The role of occupational exposures in the incidence of low back pain (LBP) has always aroused much discussion. Historical texts such as The Diseases of Occupations1 and Industrial Maladies2report associations that were later discounted and then re-accepted as epidemiological techniques are explored and refined.

Most practitioners involved in the occupational field have relied on the 1997 NIOSH publication Musculoskeletal Disorders and Workplace Factors (p97–141). This was a review of epidemiological studies that reported “strong evidence” of an association between “work related lifting and forceful movements” and low back musculoskeletal disorder (LBD) and a further association with “whole body vibration” exposures. “Evidence” was also reported between “heavy physical work” and “work related postures” and low back disorder.

The advice given to employers in at risk industries, and to individual employees with significant back pain, has been complicated by recent decisions of the Accident Compensation Corporation (ACC) which have effectively denied any association between work activities and the development of LBD. Instead the hypothesis that has been offered is that this condition is genetically based, and would have occurred regardless of their occupational exposures. This hypothesis is based almost exclusively on the Twin Spine Study,3 and the ACC Review process finds the contrary scientific evidence unconvincing.

The Twin Spine Study was a series of studies of the determinants of lumbar disc degeneration in 147 monozygotic and 153 dizygotic exposure-discordant male twin pairs drawn from the population-based Finnish Twin Cohort The investigators estimated that 61% of the variance in disc degeneration in the T12-L4 region was explained by familial aggregation, while only 16% was explained by age and occupational physical loading together, and concluded that “disc degeneration appeared to be determined in great part by genetic influences”.

While these twin studies have been influential, the interpretation that most disc degeneration is primarily genetic is based on a misunderstanding

Variation is not the same as causation, and heritability is not the same as genetic determination.4 In twin studies heritability estimates are based on comparisons of the variation in disease across twin pairs, but the percentage of population variation in a disease due to a particular exposure or trait is often confused with the proportion of disease explained by this exposure or trait.

The Twin Spine Study attempts to partition the population variability observed into separate components that add up to (at most) 100%. However when we consider causation rather than variationthere is no requirement for the attributable fractions for each risk factor (genetic and environmental) to sum to 100%.

One example is phenylketonuria (PKU).5 PKU results from a single genetic variant that leads to deficient metabolism of the amino acid phenylalanine (and its heritability is essentially 100%), but the disease only occurs when both the genetic variant and the environmental exposure (i.e. dietary phenylalanine) are present. The causation, therefore, is both 100% genetic (as 100% of cases could be prevented by removing the mutation) and 100% environmental (as 100% of cases could be prevented by reducing phenylalanine in the diet).

For example, in a population where everyone smoked one pack of cigarettes a day, smoking would not account for any of the population variation in lung cancer incidence, since smoking rates were the same everywhere. The population variation would appear to be explained by genetic variation—i.e. individual susceptibility to tobacco smoke. However, 95% of lung cancer cases would be caused by smoking, and would be prevented by preventing smoking.

Moreover, most common human diseases are far more complex than PKU and result from multiple genetic and environmental risk factors. Each factor makes a separate contribution to the disease process, and there are also complex gene-environment interactions (e.g. between smoking and genes that affect susceptibility to tobacco smoke).

Subtle differences in genetic factors cause people to respond differently to the same environmental exposure, and genetic variations influence a person’s susceptibility to environmental factors. It follows that genetic risk for disease is modifiable in an environment-specific manner, and that while an individual may inherit a predisposition for a disease they will never develop the disease unless exposed to the appropriate environmental trigger(s).6

The finding of the Twin Spine Study that 61% of the variance in disc degeneration is explained by familial aggregation does not mean that 61% of chronic low back pain is caused by genetic influences.

In fact this does not tell us anything about what percentage of cases are caused by genetic factors, or what percentage are caused by environmental factors (and could be prevented by removing these environmental causes). In misinterpreting this study, ACC is confusing two completely different sets of numbers—the % of population variation that is explained by variation in the environment, and the % of cases that are caused by environmental factors.

There is ample evidence, including that from occupational epidemiological studies conducted in heterogeneous populations, of interactions between a range of factors (mechanical, traumatic, nutritional and genetic) playing a role in the disease process that results in lumbar disc degeneration and chronic low back pain.7,8

LBD, like all occupationally acquired medical conditions arises as a consequence of a multifactorial mosaic of influences. This single review by Battie et al of their own work utilising dated MRI scanning findings does not invalidate over a century of clinical observation and multiple quality epidemiological findings.

LBD is no different in this mixture of genetic and environmental influences from other occupational conditions such as allergic asthma, the sensitivity to noise induced hearing loss and angiosarcoma secondary to poly vinyl chloride exposure.

Ignoring the link between occupational tasks and subsequent clinical conditions will also break the cycle of occupational task / environment analysis and subsequent process, tool or environmental improvement leading inexorably in the medium term to increased rates of occupational conditions, worker misery and compensation demands.

David McLean
Senior Research Fellow, Centre for Public Health Research, Massey University

Neil Pearce
Professor of Epidemiology and Biostatistics, London School of Tropical Medicine and Hygiene
London, UK

Christopher B Walls
Occupational Physician

Richard D Wigley
Palmerston North

Author Information

David McLean, Senior Research Fellow, Centre for Public Health Research, Massey University, Wellington, Neil Pearce, Professor of Epidemiology and Biostatistics, London School of Tropical Medicine and Hygiene, London, UK, Christopher B Walls, Occupational Physician, Auckland, Richard D Wigley, Rheumatologist, Palmerston North

Correspondence Email



  1. Hunter D. The Diseases of Occupation, 4th Edition. Boston Little, Brown and Company, 1969.
  2. Legge T. Industrial Maladies. Oxford Medical Publications, 1934
  3. Battié MC, Videman T, Kaprio J, et al. The Twin Spine Study: Contributions to a changing view of disc degeneration. The Spine Journal 2009;9:47-59.
  4. Pearce N. Epidemiology in a changing world: variation, causation and ubiquitous risk factors. Int J Epidemiol 2011;40(2):503-512.
  5. Rothman KJ, Greenland S, Lash TL. Modern Epidemiology, 3rd Edition. Philadelphia: Lippincott-Williams, 2008.
  6. Ramos RG, Olden K. Gene-environment interactions in the development of complex disease phenotypes. Int J Environ Res Public Health 2008;5(1):4-11.
  7. Modic MT, Ross JS. Lumbar degenerative disk disease. Radiology 2007;245(1):43-61.
  8. Seidler A, Bergmann A, Jäger M, et al Cumulative occupational lumbar load and lumbar disc disease – results of a German multi-centre case-control study. BMC Musculoskeletal Disorders 2009; 10:48 doi:10.1186/1471-2474-10-48.