In this article, Sue Hewitt discusses the exposure-response relationship for hand-arm vibration (HAV) (i.e. the relationship between vibration exposure and the health effects that such exposure can cause) and discusses some of the reasons why it will probably always be ill-defined.
The Exposure-response Relationship for Hand-arm Vibration
The original exposure-response relationship for hand-arm vibration, given in ISO 5349:1986[1] was based on the work of Tony Brammer[2] [3] and was the outcome of a meta-analysis of many studies, most of which were of Japanese foresters in the 1960s. These workers were using high vibration chain saws in cold conditions for prolonged periods on a daily basis. The exposure-response relationship developed by Brammer as given in ISO 5349 relates only to vascular symptoms (finger blanching); the risk of health effects of exposure being assessed purely on the basis of the symptoms of damage to the circulation of the fingers, hands and arms. The British Standard BS 6842[4] published in 1987 contained a more cautious interpretation of this relationship, which did not include as much information as the 1986 version of ISO 5349. By 2001, when the current version of ISO 5349-1[5] was published, there was significantly more information on the health effects, whilst the exposure-response relationship itself was addressed with less precision and more qualifying statements and notes. This change reflected the uncertainty in the exposure-response relationship at the time of publication, which still persists today.
The current exposure-response relationship still relates only to the time taken before the vascular symptoms start to occur. This is known as the latent period. The relationship is not applicable to the rate at which symptoms of finger blanching will progress and develop once they have started. It also does not relate to the neurological symptoms (numbness and tingling) or the musculoskeletal effects. Due to the techniques that need to be applied to epidemiological studies when developing an exposure-response relationship, the original work only looked at the vascular symptoms, and yet it is now widely recognised that it is the neurological symptoms that are the most disabling and have the greatest effect on quality of life.
Another important point with regard to the exposure-response relationship for HAVS is that it is based on studies of populations of workers and therefore should only be applied in terms of populations of workers. Again, this is due to the techniques applied and the assumptions made in the development of the model. In BS 6842:1987 a daily exposure of 2.8 m/s2 A(8) (based on the dominant or highest axis of vibration) is the exposure at which the most susceptible 10% of workers in an exposed population may be expected to start exhibiting symptoms after 8 years. This level was adopted as the Action Level in the first HSE guidance on hand-arm vibration[6] published in 1994. However, the relationship should not be interpreted as predictive of the likelihood that any one individual will begin to exhibit symptoms, because there are a multitude of additional factors that will influence the development of health outcomes in each individual.
Our review of the literature in 2013 concluded that the quantitative relationship suggested in ISO 5349-1:2001 is not universally applicable. Some studies reported findings that were in broad agreement with the original work, whilst others showed, in roughly equal proportions, that the ISO 5349-1:2001 relationship either over-or under-predicted the risk of vascular HAVS. This suggests that although the relationship is broadly indicative, there are many unstudied factors and sources of bias and/or uncertainty and these have affected the outcome of subsequent research.
The most powerful studies for predicting long-term health effects are longitudinal in nature, i.e. involving long-term observations of the exposure patterns of workers and the development of health outcomes in the exposed workers. These studies may, however, suffer from significant attrition rates, for example when workers change their occupation (as is likely if they begin to suffer from significant effects) and no longer form part of the exposed population. This may introduce bias. Long-term studies are also relatively uncommon because of cost and their time-consuming nature.
More often, studies are cross-sectional in design, involving the investigation of a vibration-exposed population at one point in time. Studies of this type involve the need to make exposure estimates based on the reconstruction of vibration exposure patterns, depending often on workers recalling their work patterns and tool usage from many years previously. They also require an estimation of the likely vibration magnitudes from the tools used in the past, since these can no longer be measured. The nature and degree of uncertainty between the elements for estimating retrospective cumulative exposures may obscure their respective relevance in defining the exposure-response relationship. A possible source of bias in cross-sectional workplace investigations is again that a survivor population is studied, with the most severe cases no longer in the workplace, thus introducing a bias in the results towards detecting a weaker exposure-response relationship than exists in reality.
A further, more general source of potential bias in any literature study, not just of the exposure-response relationship for HAV, is the favouring and reporting of studies with positive findings, compared to studies with negative outcomes.
Another significant obstacle to the successful development of the exposure-response relationship for HAVS is that diagnosis and staging of HAVS are based on the symptoms as reported by the sufferer, rather than on any objective measures of damage or physiological deficit. This means that diagnosis is open to misclassification through confounding conditions, which are relatively common in the general working population. For example, roughly 5 – 10% of the adult working population suffer from symptoms of finger blanching and as many as 15% suffer from symptoms of tingling and numbness, despite never having been exposed to hand-transmitted vibration of any type. For this reason, it is essential that an occupational health physician is employed for formal diagnosis, in order to exclude any other potential causes for the reported symptoms. It is also important to establish that there is sufficient history of exposure to hand-arm vibration for this to be causative of the symptoms. Ultimately, if all other possible causes for the symptoms can be ruled out, then they may be due to the vibration exposure.
The outcome of the review of the literature was that we did not find any strong evidence of a precise quantitative relationship between exposure to vibration and health outcomes, either for vascular or neurosensory HAVS. There was some evidence suggesting possible limited reversibility of vascular HAVS after cessation of exposure, but the limited evidence concerning neurosensory HAVS did not indicate any reversibility of the condition.
Our review found some studies that had considered whether a daily vibration exposure level at which the latent interval corresponds to a working lifetime could be established. If such a level could be established this would represent a level to which an operator could theoretically be exposed throughout his or her working lifetime without developing any symptoms. However, although ISO 5349 -1:2001 records that reports of ill health are rare below 2 m/s2 A(8) and not known below 1 m/s2 A(8), our review found no new evidence to support the existence of a ‘no effect level’ for vibration exposure, other than the somewhat obvious zero exposure level.
A further point of note relates to the setting of the Exposure Limit Value (ELV) in the Physical Agents (Vibration) Directive, 2002[7] which was implemented in the UK by the Control of Vibration at Work Regulations 2005[8]. Loosely based on information in a 1984 paper by Brammer[9], the ELV of 5 m/s2 A(8) (vector sum, or vibration total value) was alleged in part[10] to have been selected on the assumption that a typical working lifetime would be approximately 25 years and with the intention of preventing progression of the disease to a later stage by the end of the typical working lifetime. (Other non-scientific factors would no doubt also have influenced the final selection of the ELV). With the increasing longevity of the population, the age at which manual workers are able to retire is becoming older and the working lifetime is consequently being extended. In view of this, a working life of 40+ years might be more typical. Such an increase in the duration of lifetime exposure may increase the likelihood of reaching later stage disease by retirement age (albeit only if, for example, health surveillance is absent or inadequate).
The exposure-response relationship is likely to remain ill-defined due mainly to the great differences between individuals in susceptibility to occupational disease, but also because thankfully, these days large populations of workers exposed to high vibration magnitudes for most of the day are no longer available for study. In view of this, perhaps what we should be trying to achieve now is an improved understanding of what characteristics of vibration are most damaging, so that designs of tools, work processes, and maybe even PPE, can be better informed.
The vagaries of the exposure-response relationship for hand-arm vibration are yet another reason why the priority should always be to reduce exposures to as low as reasonably practicable in order to prevent harm.
For any more information on the exposure-response relationship for hand-arm vibration please contact [email protected].
[1] ISO 5349:1986 Mechanical vibration. Guidelines for the measurement and assessment of human exposure to hand-transmitted vibration.
[2] Brammer, A., Relations between vibration exposure and the development of the vibration syndrome. Vibration Effects on the Hand and Arm in Industry, Brammer, A, 1981: p. 283-290.
[3] Brammer, A., Dose-response relationships for hand-transmitted vibration. Scandinavian Journal of Work, Environment and Health, 1986. 12(4): p. 284-288.
[4] BS 6842:1987. Guide to measurement and evaluation of human exposure to vibration transmitted to the hand.
[5] ISO 5349-1:2001 Mechanical vibration. Measurement and evaluation of human exposure to hand-transmitted vibration. General requirements.
[6] HSE Hand Arm Vibration. HS(G)88. HSE Books. 1994.
[7] Minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (vibration) (sixteenth individual Directive within the meaning of Article 16(1) of Directive 89/391/EEC) Directive 2002/44/EC of the European Parliament of the Council of 25 June 2002 http://eur-lex.europa.eu/
[8] The Control of Vibration at Work Regulations, 2005. Statutory Instrument no. 1093.
[9] Brammer, A., Model of the onset of white fingers during the exposure of the hands to vibration from chain saws. Archives des maladies professionnelles, 1984(5): p. 315-322.
[10] Brammer, A., Personal Communication (2011).
