Occupational Whole Body Vibration and Lower Back

Disorders: A Review of the Published Evidence

Prepared for:

Dr. Craig Martin

Senior Medical Advisor
Workers’ Compensation Board of British Columbia
PO Box 5350 Stn Terminal
Vancouver B. C.
V6B 5L5

By:

Dr. Keith Chambers

Clinical Epidemiologist
G. K. Chambers Consulting Inc.
1281 Matthews Avenue
Vancouver, B. C.
V6H 1W5

1
Table of Contents

I. Summary

II. Background

II.A Whole Body Vibration

II.B Appraisal Criteria

II.C Purpose

III. Methods

III.A Literature Search

III.B Inclusion/Exclusion Criteria

III.C Evaluation Criteria

IV. Results

IV.A Appraisal of individual studies relevant to

review criteria:

IV.A.1 Prospective Cohort Studies

IV.A.2 Retrospective Cohort Studies

IV.A.3 Case Control Studies

IV.A.4 Cross Sectional Studies

IV.A.4.a Non Peer Reviewed Cross Sectional Studies

IV.A.4.b Peer Reviewed Cross Sectional Studies:

2
IV.A.4.b.1 Cross Sectional Studies where outcomes
are multi-factorial and whole body vibration (WBV) was not an
outcome

IV.A.4.b.2 Cross Sectional Studies with negative

results where WBV is an outcome:

IV.A.4.b.3 Cross Sectional Studies with positive

outcomes where WBV is an outcome

IV.B Appraisal of aggregated review results relevant

to study criteria:
IV.C Tables

V. Results

V.1 Critical Appraisal of studies reviewed

V.2 Assessment of the epidemiological studies

applying rules of causation

VI. References

3
I. Summary

The object this review was to review the published evidence on the
relationship of whole body vibration (WBV) to low back disorders (LBD) in
the workplace. After a literature search, 33 studies met the study criteria and
were included in the review (1-33). Most studies (73%) were cross sectional in
design and judged unable to assess causative relationships. Not a single
study, completely met basic standard critical appraisal criteria that are
normally used to assess the quality of research studies. These include
appropriate study design, adequate study period, outcomes relevant to the
hypothesis, statistically significant results, strengths of association and a lack
of obvious bias and confounding in the study results.

The results, in general, were inconsistent, often negative and often showed
stronger risks or similar risks for LBD in occupations with less WBV
exposure that were used as control groups.

Two important factors that tend to obfuscate relationships in this area of

research are a high frequency of LBD in humans and multiple factors that
potentially relate to LBD.

Review criteria utilized in this review were relatively stringent to counter the
potential for publishing bias and poor study design that other reviewers in
this area of research have also pointed out.

Future research using long-term prospective cohort studies would hopefully,

help determine causative relationships, safe levels of exposure for identified
risks and loss of work data. This information is currently missing and may
indicate a lack of ability or will to commit the necessary time or resources to
this line of enquiry on the part of researchers in this area.

The conclusion of this review is that there is insufficient evidence to

conclude that WBV is related to LBD. The more likely explanation is that
LBD are related to multiple factors, many of which are non-occupational in
nature and that the exact nature of these relationships is not understood at
this time (43).

4
II. Background

II.A Whole-Body Vibration:

Low back disorder (LBD) and low back pain are common conditions in
humans. It is estimated that the incidence of life-time LBD is in the order of
60 to 70% and because of its widespread prevalence, the socioeconomic
impact of this disorder is significant. (13,15,16). This has created a body of
literature on the nature, impact and potential causes of LBD. Some of this
research has focused on occupation as a risk factor where epidemiological
data indicates a variation in the rate of low back disorders based on type of
occupation. Over time, a number of theories have been put forward.
Alternatively, occupations requiring heavy lifting, and those, which are
sedentary in nature, have been identified as having increased risk.

Explanations for associations identified, have included systematic error due

to weak study design, job selection bias, smoking, past trauma and other non
occupational factors, as well as job related causes such as sitting in awkward
positions and heavy lifting. One such hypothesis that has been proposed is
that whole body vibration (WBV) is related to LBD.

A number of research studies and reviews have been published that have
addressed this issue. They have tried to deal with the potential for a
causative relationship. In addition, they have tried to address a number of
related issues including trying to assess a cut-off level below which whole
body vibration would not cause LBD and whether or not this relationship to
LBD would have a socioeconomic impact in terms of time off work. This is
the question central to the occupational issue, given that the incidence of
LBD is so high in the general and working population. If up to two thirds of
the workforce will have LBD at one time or another, it is more important to
know whether or not time loss from work is related to WBV, in addition to
whether or not it is a causative factor.

As an overview, the existing literature in the area of WBV in general lacks

studies of quality (4,33 ). A number of studies exist that are hypothesis
generating in nature, in that they raise the possibility that a relationship
between WBV and LBD may exist, but fall short of being able declare a

5
causative relationship. Still others have attempted to test the hypothesis of
WBV causing LBD; but lack sufficient quality of study design for the results
to be relied on. Reviews in the area of WBV and occupational risk factors
have been published (34,35), however, most of these are older reviews and
may lack critical appraisal criteria that today would normally be associated
with reviews assessing causative relationships. An unpublished review has
been produced for the Worker’s Compensation Board in 1999 (36).

A number of issues will be obvious to those reading this review, including

the wide number of definitions of outcomes appearing in the literature, that
are used to assess the relationship of WBV to back disorders. Composite
classifications such as low back disorder, general outcomes such as low back
pain and specific ones such as disc degeneration and sciatica all obfuscate
relationships. Similarly, the issue of making a non-quantitative leap from the
outcome of occupation to WBV without measurement data stands out in this
review. Other issues will be raised.

II.B Appraisal Criteria:

It is accepted in epidemiology, that the results of research studies can

represent more than one state of affairs. They can represent a true set of
results, or they can be due only to chance and not be true based on the type
of study sample and measurements obtained. A third possibility is that
results can be false, due to various sorts of systematic error. This third
possibility can be based on confounding or various biases that may arise to
systematically affect results. To deal with these issues, in any current review
of the literature, it is necessary to establish a set of criteria on which to make
judgements about the quality of published material that is to be reviewed (38).
These criteria must be able to assess the degree to which the available
literature represents a true state of affairs, a possibility of chance error or the
likelihood of being due to systematic error.

Fortunately, medical science and epidemiology in general have established

accepted criteria that can be applied to these issues. The accepted way to
deal with chance error is to reject hypotheses or study results where there is
more than a certain probability of the study results being due to chance. The
convention is to use p-values to assess point estimates and confidence limits
to give some idea of the stability of the point estimate. If the p-value is
greater than .05 or the confidence interval crosses 1 then the study results are

6
generally accepted as being not statistically different and no conclusions can
be drawn about the results.

A host of issues come into play when setting criteria to assess the potential
for systematic error. The most important of which is study design. Study
designs range in terms of their ability to control for systematic error, from
controlled intervention trials to case series in which no control group is
available. In occupational medicine, where intervention trials are usually
impractical then well designed prospective cohort studies with stringent and
repeated exposure and health status measurements built into the design are
probably the most acceptable in terms of controlling for potential biases.
Studies with a decreased level of sophistication in design, such as
retrospective studies, cases control studies and cross sectional studies are
less able to establish causative relationships because of this propensity for
systematic error to affect results. Recall bias for retrospective study designs
and the inability to establish temporal relationships in cross sectional studies
are just two of many powerful biases. In the case of cross sectional studies “
cross sectional surveys, while easy and rapid to accomplish, do not establish
the temporal sequence of events necessary for drawing causal inferences”(36).
Pauline Bongers in her longitudinal study on crane operators, which is one
of the studies reviewed, mirrors this fact when she states “ Because of the
large number of dropouts due to lumbar spine disease in industry a
longitudinal study design was chosen”.

Further and more specific criteria to apply to any review of epidemiological

research would include: relevance of the exposure to the research question
being reviewed, appropriateness of choice of outcomes, adequacy of the
study period, and an assessment of the likelihood of confounding and bias
occurring. Other outcomes that could independently explain study results
should also be considered.

Finally, how positive the study results are (strength of the association),
whether or not a dose-response relationship is seen and whether or not
results are consistent across studies will impact on any assessment looking at
causative relationships. The exception is in the case of consistency if there is
a likelihood of systematic error in the study designs. In this setting,
repetitive studies may show consistent relationships that are untrue. Note
that, in general, study results tend to favor positive results as negative
studies tend not be published. This publishing bias tends to increase chance
error in results and artificially create consistency in study results.

7
II.C Purpose:

The purpose of this review is to evaluate the published literature on

occupational whole body vibration (WBV) to see if there is strong enough
evidence to conclude that a causal relationship exists between WBV and low
back disorders. Loss of work time due to this possible relationship will also
be assessed, as this is the primary outcome of interest from the point of view
of the Worker’s Compensation Board. Thresholds are the third relationship
that will be examined.

III Methods

III.A Literature Search:

A literature search of two available electronic data bases was conducted.

These were the Medline database from 1966 to 2001and EMBASE from
1998 to 2001 using the following Key Words: whole body vibration,
occupational exposure, vibration, low back pain, sciatica, and lumbar. The
search was restricted to human research and to articles in the English
language.

In addition, an unpublished report reviewing the literature and prepared for

the Worker’s Compensation Board of B.C was included (36). This report was
supplied by the WCB and was used as a guide in assembling studies in this
review.

III.B Inclusion/Exclusion Criteria:

Controlled trials and epidemiological studies, which studied occupations

involving equipment operation, or vehicles, where WBV occurs and where
low back disorders were studied as an outcome, were included. Studies
where the occupational description was too broad to identify those exposed
to WBV were excluded, as were studies where WBV was artificially
induced. These studies were excluded, as they are unlikely to duplicate the
workplace experience.

8
III.C Evaluation Criteria:

1. Study designs were assessed based on a hierarchy of study design (44,45).

In order of importance and ability to assess a causative relationship, they
were assessed as follows. The highest order looked for were controlled
trials followed by, prospective observational cohort studies, case control
studies, other retrospective observational studies, cross sectional studies
and case series in a descending order and ability to establish causative
relationships.

2. Study outcomes were assessed in their ability to address the question of

work loss and define the type of low back disorder being investigated.

3. Exposures were assessed as to the degree to which measurable WBV was

being assessed or whether the surrogate outcome of occupation was being
used. If WBV was being measured then a dose-response exposure was
assessed.

4. The study period was assessed as to whether or not it was prospective,

retrospective or at one point in time. The length of the study period was
also assessed.

5. Results were assessed as to their statistical significance, strength of

association, whether a dose-response relationship was seen, and whether
multiple factors were seen to be associated with the exposure of interest.

6. Confounding and bias was assessed for each study.

9
IV Results
A summary of the results are provided in Tables 1 and II. The results sre
presented in three parts. The first part presents summary appraisal results of
the individual studies reviewed. Only results that have a significant and
direct relationship to the causal relationship of whole body vibration and low
back disorder are reported. The second part is an aggregated summary of the
overall results of the review using appraisal criteria established in the
methods section. Finally, two tables are presented to represent an overall
view of the degree or lack of degree to which the results are positive and
consistent for the hypothesis of WBV causing LBD.

IV.A Appraisal of individual studies relevant to review criteria:

IV.A.1 Prospective Cohort Studies:

1. Pietri et al published a 1 year prospective cohort study looking at low-

back pain in commercial travelers. Risk factors identified included
smoking, psychosomatic factors and seat comfort. The weakness of this
study in terms of the research questions posed is the short nature of the
study, the lack of WBV as an outcome and the absence of data on work
loss. The multiple risk factors identified do not support a specific relation
of WBV to LBP.

2. Riihimaki et al undertook a 3 year prospective study and published in

1994, which examined sciatic pain in machine operators and carpenters v.
office workers. Previous history of low back pain (LBD) increased risk 4
fold. The rate ratio adjusted for previous history and other covariates was
1.36 (95%CL .99 -1.9) for the machine operators and 1.5 (95% CL 1.1-
2.1) for the carpenters. This argues against the WBV hypothesis as the
confidence interval crosses 1 for machine operators indicating a non
significant outcome while the larger effect is seen in the carpenters who
have the lower exposure to WBV. Outcomes of WBV and work loss
were not reported, therefore, causative relationships cannot be assessed
based on these study results.

10
IV.A.2 Retrospective Cohort Studies:

3. Thorbjornsson et al has looked at psychosocial and physical risk factors

in a cohort of subjects aged 18 to 34 years of age without low back pain
(LBP) in 1969. They were part of a larger cross sectional REBUS study
that occurred in 1969. This subset was re-examined in 1993 (n=783)
looking retrospectively at risk factors for LBP. Dissatisfaction with
leisure time and monotonous work were identified as risk factors. The
authors concluded; “in this study, factors related to work had a long term
effect only in interaction with leisure time factors”. This study did not
show increased risk due to WBV exposure in the multivariate analysis.

4. Bongers, Boshuizen et al published a retrospective (longitudinal) cohort

study in 1988 claiming that a combination of WBV and strained or
unfavorable posture was “responsible” for the development of
intervertebral disc degeneration. 743 crane operators were compared to
662 floor workers who were employed in a Steel Company in Holland.
The review went back 10 years looking at age, nationality, shift work,
calendar year, disability pension and reasons for leaving work. The
authors claim that because more crane operators received a disability
pension for degeneration disc disease (ICD 722.6) that WBV,
unfavorable posture and climate were responsible. However, the
following issues should be noted:
• This is a study in a country where workers get a pension if they are 15%
or more disabled from their job for over 1 year and other tasks are not
available. This creates a potential for other factors such as job satisfaction
to confound results. More crane operators quit their job (27% v. 23%)
than floor workers over the course of the study. This could indicate a
relative imbalance in job related satisfaction. This is supported by the
reasons given for resigning: job dissatisfaction, financial reasons, work
reasons and shift work. These would be potentially powerful confounders
for those who might seek a pension rather than resigning their job. Crane
operators did have more shift-work supporting the potential for this bias.
It is interesting that, overall, the crane operators received more disability
pensions than floor workers (18% v.14%) indicating as above that factors
other than WBV may be driving the process.
• The two groups compared were not equivalent for factors other than
WBV. Not only did the crane operators do more shift work, they were
older and had worked longer at their job than the floor workers. While

11
the authors tried to adjust for these factors in the analysis they failed to
include and adjust for other factors such as smoking and sports activities
that would likely be different between these groups given the age, years
of employment and shift work differences.
• Less than half of both groups were still employed at the end of the study
period (1984) indicating a large loss of information about over half of the
subjects being studied. This creates an enormous potential for bias if a
larger proportion of those with degenerative disc disease occurred in one
group or the other in those who were lost to the study. This becomes a
bigger problem if the study numbers are small, as is the case in this study.
• The ability to make confident assumptions about causative relationships
depends to a large degree not only on the size of the study in terms of
sample size but also on the frequency of events. The lower the number of
events, the less one is able to make generalizable statements. In the case
of this study, the primary outcome reported by the authors was that an
“incidence density ratio (IDR) exceeded 2 for disability because of
degeneration of the intervertebral disc, after adjustment for age,
nationality, shift work and calendar year”. Looking at Table 4 one sees
that this is based on 14 cases in the crane operators and 3 in the control
group. It would be entirely inappropriate to make a case for a causative
relationship based on these small numbers even if potential biases could
be controlled for. In terms of this review, it is interesting to note that no
difference was found in disability pension rates between the two groups
for back disorders in general (90% CL .84-2.07). This comparison was
based on the larger numbers of 38 and 18 respectively.
• There are other issues that affect the ability of this study to make claims
about WBV. There is a lack of a measurable link between the occupation
and WBV. There is exposure contamination in that 30% of floor
operators were exposed to WBV 20% of the time. This relationship is
more of an assumption on the part of the authors. The retrospective
design has created an environment in which multiple biases can operate.
Multiple testing has gone on without adjustment or a logical explanation
for why the primary outcome is disc disorder and not all back disorders.
This raises the possibility of post hoc analysis. Finally, if the primary
hypothesis was originally the composite outcome of all back disorders,
which would seem more appropriate, then this is a negative study, which
argues against WBV.

5. Bongers, Boshuizen et al published a follow-up study on the above

cohort of crane operators and floor workers. This study looked at the

12
absolute number of at least one spell of sickness absence from work of 28
days or longer. The results again showed no difference between the two
groups for “ all back disorders” (90% CL .77-1.25) and this time with
larger numbers, no difference for disc disorders as well (90% CL .74-
2.07). Despite the authors’ claims, this argues against the hypothesis that
WBV causes back disorders.

6. Boshuizen, Bongers et al published a second retrospective cohort study in

1990 looking at tractor drivers from a single company and comparing
them to a reference company consisting of inspectors from another
company and technical service workers in the company where the tractor
drivers worked. The outcomes of interest were long-term sick leave and
disability pensioning due to back disorders. Issues on the potential for
biases to operate and confounding were similar to the earlier study by the
same authors. The historical nature of the study, differences between the
groups other than WBV exposure and the disability environment that
exists in Holland all are potential sources of error in interpreting results.
The authors demonstrated the reality of a selection bias with their
comparison with and without including the reference company. “ The
incidence of a first long-term sick leave due to back disorders was 3 per
100 person-years in tractor drivers and 2 per 100 person-years in the
entire reference group. However, the incidence was not substantially
increased in tractor drivers when the referents comprised only those
working at the same company, suggesting selection bias”. The incidence
of intervertebral disc disease was not significantly different in this study
as opposed to the authors earlier study (90% CL .73-5.5). Again, this
data is unstable as the numbers are so small (n= 17 v. n= 3). In terms of
disability pensioning the reported results are “ far from statistically
significant”.

7. Boshuizen, Bongers et al went on to publish a follow-up paper on this

cohort. They sent out a postal questionnaire trying to establish a
relationship to vibration dose and degrees of back pain. The response rate
was 79%. The authors found that the prevalence of self-reported back
pain is 10% higher in the tractor drivers, than in the reference population.
However, 4 of 8 crude odds ratios looking at questions relating to a dose
response relationship of vibration dose in years were not significantly
different. For example there was no difference in those treated for back
pain or those who had had a prolapsed disc. When adjusting for cofactors
such as age and duration of exposure the vast majority of odds ratios of

13
vibration exposure and type of back pain were non significantly different.
The authors point out that despite this lack of a demonstrated relationship
“the highest prevalence odds ratios are found for the more severe types of
back pain. These prevalence odds ratios do not increase with the
vibration dose”. Despite the weakness of the study design, which has the
same weaknesses, as described above plus the added weakness of a self
reported questionnaire, the study argues against a dose-response
relationship and a significant difference between groups.

IV.A.3 Case Control Studies:

8. Kelsey et al published a case control study in 1975 looking at the driving

of motor vehicles as a risk factor for lumbar disc disease (LDD). This is
an older study with a much weaker design than cohort studies. The
authors identified 223 cases of LDD from those receiving lower back x-
rays at three New Haven Hospitals during the years 1971-1973. They
then found 494 controls and were able to match 217 controls to the cases.
The cases and controls were interviewed. The object was, amongst other
things, to establish the degree to which occupations involved driving. The
response rate was 79%. A Mantel –Haenzel test of association for a
comparison between the cases and the unmatched controls was
significant and indicated a three-fold increase in risk. However, when the
control group was adjusted for group differences by matching; the
matched analysis combining surgical and probable cases produces an
odds ratio of 2.55 (p> .05) which is statistically not significant. This
study in addition to the weaknesses associated with this type of study
design as described previously, does not have WBV as an exposure but
rather whether or not the job required sitting in a motor vehicle more or
less than half time. The choice of half time appears arbitrary. In
summary, this study has no features, which demonstrate a causative
relationship between WBV, LBD and work loss.

9. Heliovaara et al published a case control study in 1987 looking at lumbar

disc disease or sciatica. 592 men and women were compared to 2140
matched controls. As expected the risk of being hospitalized for herniated
disc was lowest in professional and related occupations relative to all
other occupations. Relative risks were elevated for occupations exposed
to WBV and those not exposed to WBV. The maximal risk was “ highest
among blue collar workers in industry and among motor vehicle drivers”.

14
This is a hypothesis generating study and no claims are made about WBV
nor causal relationships. The authors point out that the occupational
variation may “not be due to disease incidence but rather to greater
likelihood of hospitalization”.

IV.A.4 Cross Sectional Studies:

The largest number of studies, in this review, are of a cross sectional design
which, in general, as stated above, cannot be used to establish causation.

IV.A.4.a Non Peer Reviewed Cross Sectional Studies:

Four identified studies were excluded, as they did not appear to be published
in a peer reviewed journal (39-42).

IV.A.4.b Peer Reviewed Cross Sectional Studies:

IV.A.4.b.1 Cross Sectional Studies where outcomes are multi-factorial

and whole body vibration (WBV) was not an outcome:

10. Frmoyer et al reported in 1979 that pregnancy, smoking, anxiety,

depression, emotionally stressful occupations, lifting, carrying, pulling,
pushing and twisting in addition to truck driving are all related to low
back pain. There was no mention of WBV as an outcome and this study
is a hypothesis generating study with no attempt to look at causation. It
supports the notion of causes of LBD being multi-factorial.

11. Heliovaara et al published a large cross sectional study in 1991 looking at

several factors for their “ association with the prevalence of chronic low-
back syndromes, sciatica, and low back pain (LBP)”. Approximately,
5,000 subjects from the large Mini-Finland Health Survey project were
examined and 5.1% were found to have sciatica while 11.6% were found
to have LBP. Previous traumatic back injury, smoking, body height, and
physical and mental stress at work all increased risk. The authors
concluded “ many factors, independent of each other, determine the
occurrence of chronic-low back syndromes”. This study does not support
the notion that WBV is a major cause of chronic-low back syndromes.

15
12. Saraste et al published in 1987 showing that low back pain was related to
lower education, physically heavy, monotonous or repetitive work and
smoking. This supports a multifactorial cause of LBD. WBV was not an
outcome looked at.

13. Nayha et al published in 1991 and reported that neck and shoulder pain
was related to physical work in slaughtering reindeer but that “the
association of back pain with the amount of work was weak’. WBV was
not an outcome.

14. Burdorf et al published in 1993 and looked at sedentary workers using

crane operators and straddle-carrier drivers as the occupations of interest.
He compared them to office workers. Burdorf concluded that “ in both
occupations the daily exposure to whole-body vibration was low, and
therefore not considered an important risk factor for low back pain”. The
study “suggested” that non-neutral trunk posture may be a risk factor but,
appropriately, stopped short of describing this as a causative relationship.
This study does not support the use of these types of occupations in
trying to identify a relationship between WBV and LBD.

15. Guo et al has published a study in 1995 showing that back pain is not
only common, accounting for 25% of WCB cases but that it is common
in a wide variety of jobs both with WBV exposure and in occupations
where WBV is not present. The highest risk is for construction workers
for males and for nurses’ aides in females. Industrial truck and tractor
operators had similar but slightly lower relative risks of back pain than
carpenters or construction workers with reputed lower WBV exposure.
WBV was not an outcome of interest in this study.

IV.A.4.b.2 Cross Sectional Studies with negative results where WBV is

an outcome:

16. Bongers, Boshuizen et al published a study in 1991 looking as a primary

hypothesis at the dose response relationship of WBV to self-reported
back pain. While younger less exposed drivers complained of more
regular back pain than the reference group; “ with increasing age, the
difference in the prevalence of self-reported regular back pain between
drivers and the reference group disappeared”. This argues against a dose
response relationship, although, the authors feel that a “health-based

16
selection” bias may account for the unusual findings. Adjusted for age,
there were no differences in 14-day sick leave, treatment for back pain or
disc herniation between groups, which argues against a work loss
hypothesis due to WBV.

17. Miyashita et al published in 1992 showing that fork lift operators (n=44)
had more back pain, than office workers (p< .001). However, this
difference was not seen in the two other groups looked at which had
larger sample sizes (power shovel operators n=184, bulldozer operators
n=127) and which also were exposed to WBV. Further, as 10 primary
outcomes were assessed in 4 groups (30 tests), without a description of
the statistics that were employed, then this result in forklift drivers must
be discounted because of multiple testing without adjustment. The
authors looked at Raynaud’s phenomenon, as a measure of local
vibration, and found that this was more prevalent in the control group of
office workers.

18. Masset et al in 1994, published a large cross sectional study looking at

low back pain (LBP) in blue-collar workers in Belgium. The study
showed the high prevalence of LBP in workers in that “66% of the
workers, younger than 40 years, already had experienced lumbar
problems during their lives”. The authors undertook multivariate logistic
regression to show “ exposure to whole body vibrations were not
associated with an increased prevalence of LBP”. The largest association
was for heavy efforts of the shoulders (OR=1.62, p =.01) with a very
small increase in risk for vehicle driving (OR=1.15, p<.01).

19. Xu et al in 1997 published another large study that was a re-analysis of a

Danish survey. The full multivariate model adjusting for age and
educational level showed a odds ratio of 1.23 (p=.122) for WBV as a risk
factor for LBP. This is statistically a non significant finding. The
unadjusted and reduced model gave a similar result in that the 95%
confidence limit included 1.

20. Johanning et al looked at subway train operators and found higher rates
of hearing loss, gastrointestinal problems and sciatica. “ A cumulative
dose-response relationship could not be statistically demonstrated”. This
study was particularly weak in the low response rate and the dissimilarity
of the control group (switchboard operators).

17
21. Magnusson et al published a large two country study (USA and Sweden)
looking at truck drivers, bus drivers and sedentary workers. The
outcomes of interest were based on measuring WBV exposure, stress at
work, work loss and a variety of complaints including low back, neck and
shoulder pain. As expected, truck drivers had the highest exposure to
WBV both daily and long term. However, work loss due to low back pain
was lowest in this group in both countries. Not only was it significantly
less than bus drivers but it was lower than sedentary workers (3.8 days v
18 days and 4.8 days). Approximately, half of the entire cohort
complained of LBP. There was a non-significant increase in the percent
of truck drivers that had LBP compared to sedentary workers (56% v.
42%). The authors concluded “inability to work seems affected by stress
at work”.

22. Riihimaki et al published a study in 1989, looking at low back trouble

(LBT), in machine operators (exposed to WBV), carpenters (not exposed
) and office workers (not exposed). The lifetime incidence of LBT was
the same for machine operators and carpenters (90%). It was lower for
office workers (75%). Occasional LBT occurred equally in all three
groups. Frequently occurring LBT was reasonably similar in machine
operators and carpenters (30% v. 23%). The 12 month prevalence of LBT
was “ 82% for the machine operators, 79% for the carpenters, and 61%
for the sedentary workers”. These results argue against WBV as a
causative factor in LBT. This study has undertaken multiple comparisons
using univariate Chi square statistics without adjusting for these
comparisons. Therefore, any interpretation of p-values given in this study
are invalid other than to hypothesis generate.

23. Froom et al published a study in 1984 looking at helicopter pilots in

Israel. While the authors found that helicopter pilots had a fourfold
increase in spondylisthesis (SLL) over cadets or transport pilots, they
concluded that “ although SLL is associated with LBP, the pain has little
clinical significance”.

24. Walch et al published in 1989, a retrospective postal survey on 545

randomly selected adults. The outcomes looked at were factors related to
low back pain. 65% of men had had LBP. A Cox proportional Hazards
model was used to examine lifetime LBP, low back pain on the birthday
prior to onset of symptoms and unremitting low-back pain on the

18
birthday prior to symptoms. Lifting or moving weights of 25kg. or more,
were significant in all three categories. Driving a truck, tractor or digger
did not increase risk for any of these outcomes adjusted for other
activities. Similarly, using vibrating machinery had 95% confidence
limits that cross 1 for all three analyses. The authors conclude that LBP is
related to “heavy lifting and prolonged car driving”. This argues against
WBV in heavy equipment operators.

25. Burdorf et al published a small sample size study in 1990 on 49 crane

operators comparing them to 281 workers in a control group in the same
factory. Only 67% of crane operators agreed to be in the study (n=33)
and so no conclusions can be drawn between the index and the control
groups. What is relevant is the authors’ multivariate logistic regression
showing the “ strong influence of back straining factors in the past
(previous employment) on the present probability of LBP. The authors
conclude the study provides evidence “ to strongly recommend persons
with a history of back complaints not to seek employment as crane
operators”.

IV.A.4.b.3 Cross Sectional Studies with positive outcomes where WBV

is an outcome:

26. Brendstrup et al published a study in 1987 on low back trouble in fork-

lift drivers. 240 fork lift drivers were compared to two comparison
groups ( unskilled workers and the working Glostrop men). The
outcomes of interest were the frequency of LBT in the preceding year
and absence from work. The frequency of LBT was not significantly
different when compared to unskilled workers but was higher than the
Glostrop working men. Work absence was higher for fork-lift drivers
than for the other two groups. The weakness in this study over and above
the weak cross sectional design is the lack of evidence of the
appropriateness of using the Glostrop working men as the control group.
In fact the study “ set up a reference group which resembled the fork-lift
drivers as much as possible” in the second unskilled workers control
group. In this comparison, no difference was detected. This weakens the
claims of the author that “fork-lift driving may be a contributory cause
for low back trouble”. WBV was not an outcome studied.

19
27. Dupuis and Gerlett published in 1987, comparing 352 operators of earth-
moving machines to 315 workers who were not exposed to WBV. This
study demonstrates the problem with cross sectional design in that 71%
of operators had already had x-rays of the spine v. only 48% in the
control group by the time of the study. Clearly, the potential for an
uncover bias with respect to the amount of investigations or an
incomparability of groups are at issue here. If more operators have had x-
rays then more disease of the spine exists or is likely to be found in this
group. All subsequent analyses follow from this, especially the
comparisons for those x-rayed in both groups. The problem is we have no
idea if this difference is real. It may be due health selection factors, in
that those seated may work may be able to work with a back disorder,
while more physical jobs such as a labor job may require one to change
jobs, if a back disorder develops. This study appears, on appraisal, not to
be set up to make claims about a causative relation of WBV to LBD.

28. Bongers, Boshuizen et al in 1990 published the results of a survey

questionnaire on 163 helicopter pilots compared to 297 non-flying air
force officers. This study showed that total flight time was related in a
dose response relationship to risk for chronic back pain. This relationship
became significant over 2000 hours of flight time. The authors were
unsure if this effect was due to constrained posture or a combination of
this and WBV.

29. Liira et al published a study in 1996 on the 1990 Ontario Health Survey,
looking at long-term back problems and physical work exposures. This
study involved a household-based population survey and only “reports
the prevalences of reported long-term back problems”. This study is,
therefore, hypothesis generating by its nature and not attempting show
causative relationships. A logistic regression model was used for the
analysis showing significant odds ratios for a variety of factors including
age, smoking, blue collar occupations, working in awkward positions,
bending and lifting, lifting light objects, lifting heavy objects and
operating vibrating equipment. The authors emphasize that “ Low back
pain is multi-factorial in origin”.

30. Burdorf et al published a further study in 1991, looking at 114 concrete

manufacturing workers exposed to heavy physical work, bending and

20
twisting and vibration due to operating vibrotables. These concrete
workers were compared to 52 maintenance engineers. The prevalence of
back pain in the 12 months preceding the investigation was 59% in the
index group and 31% in the control group. 44% of concrete workers had
the onset of back pain after starting work at the present factory v. 31% of
the control group. This indicates support for the notion that a health
selection bias such as the one discussed in the study Burdorf in 1990 was
driving the results. Burdorf concluded in this study that working in a
bent/twisted position may contribute to the prevalence of back pain and
that WBV, “due to operating vibrotables, is a second risk factor for back
pain”. The sample size in this study is too small to generalize any
conclusions drawn to settings outside of this study.

31. Bovenzi et al published a study in 1994, on 1155 tractor drivers looking

at LBP. They were compared to 220 revenue officers. Dose response
relationships for postural loads and vibration were looked at for a variety
of factors. The interviews were conducted at the residence of the subjects
by teams. Postural load questions about awkward positions at work were
highly related to LBP but the results cannot be interpreted. This is
because of the potential for recall bias. Those interviewed with back
problems would be much more likely to remember sitting in awkward
positions at work than those without back pain in a dose response type of
fashion. Secondly, the control group of revenue officers would likely be
very different in terms of socioeconomic factors, lifestyle factors and
previous employment factors. This tends to invalidate conclusions about
a causative relationship between WBV and LBP that might otherwise be
derived from this study. The authors found a host of risk factors for
lifetime LBP, transient LBP and chronic LBP including age, occupation
as a tractor driver as compared to a revenue officer, previous trauma and
educational level. This was in addition to a dose-response relationship for
total vibration dose. Trends were seen for the prevalence of recent and
chronic LBP as well as for sciatica. Sick leave was increased but a non-
significant trend related to vibration dose was seen. The same relations
were seen for postural load. It is interesting that in this study known
factors for risk such as smoking and previous jobs at risk were not
significant.

32. Bovenzi et al published a smaller study in 1992, comparing 234 bus

drivers to 125 maintenance workers and found bus drivers were at

21
increase risk for LBP. However, “low back symptoms occurred at
exposures that were lower than the health based exposure limits proposed
by the International Standard ISO 263/1”. Prolonged sitting in a
constrained position was also increased risk. In the control group, only
8% had previous jobs with heavy physical demands compared to 21.3%
in the bus drivers. This once again demonstrates the difficulty in
interpreting cross sectional studies in terms of finding control groups that
are similar to index groups except for the exposure of interest. If the
groups are not similar as in this case, validity of the results cannot be
ensured.

33. Netterstrom et al conducted a study that was published in 1989 on bus

drivers in Denmark. The prevalence of frequent low back pain was
somewhat higher in bus drivers (57%) than in the control group (40%).
Increased risk was also shown for lumbar disc herniation. The authors
conclude that “sedentary position and whole-body vibration may be
contributory causes” but point out that “this might lead to a notion that
people with LBT applied for jobs as bus drivers to avoid heavy manual
labour. That this should be the case cannot be ruled out, especially since
the prevalence of frequent lumbar pains was relatively high in the
younger groups as well”. WBV was not an outcome nor was loss of time
from work. Once again, the cross sectional nature of the design and the
non comparability of the control group (all Danish men) somewhat limit
the interpretation of these findings.

IV.B Appraisal of aggregated review results relevant to study criteria:

There were 33 epidemiological studies identified that appeared to be

published in peer reviewed journals, dealing with the potential causal
relationship of whole body vibration, occupation and various defined low
back disorders. The majority of studies identified, 73% (24/33), were of a
cross sectional design. There were 2 case control studies and 5 retrospective
cohort studies. Only 6% or 2/33 of the identified research studies were
prospective cohort studies. Of the prospective cohort studies, (the design
most capable of assessing causative relationships of studies reviewed), 1
study was of too short a duration to generalize results as the study period
was only 1 year’s duration. Therefore, only 1 study or 3% of those accepted
to be reviewed were prospective and of sufficient duration to be assessed as
not being classed as weak in terms of trying to answer the research question
of a causative relationship of WBV to LBD.

22
With regards to outcomes studied, most studies (28/33 or 85%) did not study
WBV directly, but used the surrogate outcome of occupation. Of those that
did look at WBV directly (5/33 or 15%), all were of the cross sectional
design. Over the entire 33 studies reviewed, negative outcomes in terms of
not reaching statistical significance were reported in a surprising number of
studies for a variety of comparisons. Of those studies reviewed, 17/33 (52%)
or half of studies reported negative findings. For example in the only
prospective cohort study with an adequate study period, Riihimaki showed
that carpenters, who had presumably less WBV exposure than machine
operators, were more at risk than machine operators for sciatica. This
represents a negative result for the hypothesis of WBV. For the machine
operators, the relative risk for sciatica was 1.36 with 95% confidence limits
of .99-1.87. This is a negative study result from a statistical significance
point of view. Bongers, Boshuizen et al in their series of retrospective cohort
studies show statistically negative results for all back disorders and disc
disorders. Cross sectional studies reviewed also reported negative results.
Boshuizen’s study in 1991 on tractor drivers, showed a lower incidence of
leave due to back trouble over 1-year in drivers v. the reference group (45 v.
56). In this study, the higher prevalence of self-reported regular back pain in
younger drivers was not seen in older drivers. Xu’s random sample of 5185
Danish employees gives an odds ratio of 1.23 for “vibration affecting the
whole body” with a highly non-significant p-value of .122.

Most studies (31/33 or 94%) were unable to demonstrate a dose-response

relationship or the design did not permit a dose-response relationship to be
investigated. Of the 2 studies that did show a dose-response relationship,
both were cross sectional in design. Further, the potential for confounding,
study bias and multi-factorial relationships as alternate explanations for the
relationship of low back disorder to occupation were evident in a majority of
studies (82%). For example, selection bias was indicated as a reasonable
explanation of results in several studies including those published by
Boshuizen, Netterstrom and Burdorf. Low back disorders appeared often to
occur at increased frequency in situations where WBV was occurring at low
exposures (eg. Bovenzi) or in occupations with less WBV exposure than the
occupation being investigated (eg. Riihimaki, Walch). The literature
reviewed, as a body, indicated a number of other non work related factors
that were associated with LBD in workers: leisure time factors, smoking,
pregnancy, anxiety, depression, lower education, age, and pre-existing low
back disorder. Similarly, a number of work related exposures were identified
as being potential alternate explanations for associations between

23
occupations and LBD. These included sedentary work, heavy physical work,
monotonous work, repetitive work, awkward sitting, bending, lifting, stress
at work and blue-collar occupations in general.

Loss of work information was assessed in 8/33 or 24% of studies. Of these

several studies reported negative results. Bongers et al failed to show
significantly more sickness absence from work of 28 days or longer for all
back disorders nor for disc disorders. Froom also found that LBP due to
spondylolithesis had little clinical significance in terms of loss of work,
while Magnusson found that truck drivers had less loss of work due to LBP
than sedentary workers. Of the other 2 authors who published studies
dealing with work loss, Guo showed a similar risk of work loss due to back
pain for equipment operators in comparison to 15 other high risk
occupations most of which did not involve WBV exposure. The final author,
Brendstrup, found a higher absence from work for fork-lift drivers.

Strengths of association for WBV exposure were not tested for in most
studies. In the 6 studies where this association was looked, at there was a
high degree of variability with a range in relative risk of no association to
39.5.
Strengths of association for occupation were looked at in more than half of
the studies. Again, there was a wide range in results ( RR’s 1.15 to 4.8).

24
25
26
V Discussion

A number of theories into causes of low back disorder have been proposed
over the last few decades. Some have been tested in animal models and
under laboratory conditions. Epidemiological research looking at these
associations in populations has had the advantage of looking at the issue
under more real working and living conditions, as opposed to these
artificially created conditions such as laboratory conditions. These
epidemiological studies tend to support the notion that the causes of low
back disorder are multi-factorial and their relationships to outcomes poorly
understood (43). This review supports that point of view. This area of research
is complicated not only by the existence of a multitude of potential causes
and their inter-relationships; but also by the extremely high frequency of
LBD in humans. Since this condition appears to be ubiquitous in human
beings, it is extremely difficult to isolate putative exposures from other
potential causes and confounders. In relation to occupation, there is a strong
tendency for workers to self select employment on the basis of psychosocial
and physical factors that, in themselves, predict on the incidence of low back
disorder. Any attempt to assess causes in the workplace, must adequately
control for these factors, which include, pre-employment low back
conditions, leisure activities including sports that are differentially played by
various occupations, smoking, family history, level of education, depression
and a host of other factors not related to employment.
This review has focused on critically appraising the existing epidemiological
literature on low back disorder and whole body vibration as a potential
cause. Given this environment of high frequency of the disorder, the
apparent multiplicity of potential causes and the importance of the question
in terms of cost to society, it was seen as important to assess the quality of
the studies reviewed and their ability to draw conclusions about causation. In
addition, it was also seen as important to apply the normally accepted
criteria for assessing causation. It would have been relatively easy to be less
stringent and be open to making the mistake of attributing causation to
relationships, which in fact are mere associations.

V.1 Critical Appraisal of studies reviewed:

Most studies were cross sectional in design and, therefore, by themselves

unable to make assessments about causative relationships. As described in
the results, many confounders and bias’s were uncovered supporting the
notion that these were weak designs for assessing causation. Many times,

27
results were negative and conflicted with the hypothesis that WBV caused
LBD. Many times associations other than WBV were uncovered.
Case control and retrospective cohort studies had similar weaknesses in
design with potential bias’s explaining results seen, negative results, and a
lack of a link of the outcomes to WBV as an exposure.
Finally, the studies, with a design most likely to be able to draw causative
conclusions about the relationship of WBV to low back disorders, also had
serious methodological flaws. The first study by Pietri was of too short a
study period (1 year). In addition, it had irrelevant outcomes in that there
was no information on WBV or work loss. The study also identified other
risk factors such as smoking, psychosomatic factors and seat comfort. The
other prospective cohort study by Riihimaki also failed to study WBV or
work loss as outcomes of interest and reported statistically negative results.
This study showed that carpenters with less WBV tended to have increased
risk than machine operators.
In conclusion, not one study met the review criteria of; adequate design and
study period, appropriate outcomes assessed, inferentially positive results,
and a lack of obvious bias and confounding.

V.2 Assessment of the epidemiological studies applying rules of

causation:

1. Strength of Association

Most strengths of association seen were relatively weak as demonstrated by

the high frequency of relative risks (RR) whose confidence limits crossed 1
and so were statistically non significant findings. Even despite these
negative statistical results, most RR’s reported were in the range of 1.2 to
2.0. This is relatively weak when compared to RR’s seen in the
epidemiological literature on such health hazards as smoking.

2. Consistency of results

There was a good deal of inconsistency in results to support the theory that
WBV causes LBD. The reporting of negative results was common. In
addition, unexpected results, where occupations with lower WBV were at
higher risk for LBD, were reported several times. Explanations for a number
of outcome results including self selecting for occupations involving driving
for those with pre-existing LBD argued against a consistent relationship.

28
3. Temporal Relationship

As most studies were cross sectional in design, the ability to assess temporal
relationships of exposure to WBV to the development of LBD was lacking.
This problem of the inability to establish a temporal relationship was echoed
in many of these studies, most notably those studies where job selection bias
was demonstrated. The 2 prospective studies, able to examine temporal
relationships, without issues related to recall bias or cross sectional design
weakness, also failed to adequately demonstrate temporal relationships for
other reasons, including the length of the study period and the demonstration
of the confounding effect of a previous history of LBD.

4. Dose Response relationship

Only 2 of 33 studies demonstrated a dose-response relationship. As

discussed above these relationships were demonstrated in studies with cross
sectional designs that contained a number of design flaws that made the
results inconclusive in terms of interpretation.

5. Specificity

There was a clear lack of specificity of a one-to-one relationship between

WBV and LBD. It was not demonstrated that WBV exposure was both
“necessary and sufficient”(37) to cause LBD in any of the studies reviewed.

In summary, it appears that current epidemiological literature does not

support the notion that whole body vibration causes low back disorder. This
inability to demonstrate a causative relationship despite decades of research
by multiple researchers is remarkable. Criteria to assess future research in
this area must be stringent, as with this much research activity, there is likely
a strong bias in terms of publishing positive results. This publishing bias
could strongly skew study results in favor of the relationship of WBV to
LBD. Future research should be limited to prospective long-term cohort
studies, which use sophisticated measurements of both exposure to WBV
and the degree of LBD over the course of the study. A crucial measurement
would be identifying subjects as being condition free at the start of the study.
This would require a commitment in terms of time and resources, on the part
of investigators, that has, to this point in time, been missing in this area of
research.

29
VI References

1 Pietri F, Leclerc A, Boitel L, Chastang JF: Low-Back pain in

commercial travelers. Scand J Work Environ Health 1992; 18:52-58.

2 Riihimaki H, Viikari-Juntura E, Moneta G, Kuha J: Incidence of

sciatic pain among men in machine operating, dynamic physical work,
and sedentary work. Spine 1994; 19: 138-142.

3 Thorbjornsson BCO, Alfredsson L, Fredriksson K, Koster M:

Psychological and physical risk factors associated with low back pain:
a 24 year follow up among women and men in broad range of
occupation. Occup Environ Med 1998; 55: 84-90.

4 Bongers PM, Boshuizen HC, Hulshof CT, Koemeester AP: Back

disorders in crane operators exposed to whole-body vibration. Int
Arch Occup Environ Health 1998; 60: 129-137.

5 Bongers PM, Boshuizen HC, Hulshof CT, Koemeester AP: Long-term

sickness absence due to back disorders in crane operators exposed to
whole-body vibration. Int Arch Occup Environ Health 1988; 61:59-
64.

6 Boshuizen HC, Hulshof CT, Bongers PM: Long-term sick leave and
disability pensioning due to back disorders of tractor drivers exposed
to whole-body vibration. Int Arch Occup Environ Health 1990;
62:117-122.

7 Boshuizen HC, Bongers PM, Hulshof CT: Self-reported back pain in

tractor drivers exposed to whole-body vibration. Int Arch Occup
Environ Health 1990; 62:109-115.

8 Kelsey JL, Hardy RJ: Driving of motor vehicles as a risk factor for
acute herniated lumbar intervertebral disc. American Journal of
Epidemiology 1975; 102:63-73.

9 Heliovaara M: Occupation and risk of herniated lumbar intervertebral

disc or sciatica leading to hospitalization. Journal Chron Dis 1987; 40:
259-264.

30
10 Frymoyer JW, Pope MH, Costanza MC, Rosen JC: Epidemiologic
studies of low back pain. Spine 1980; 3:419-423.

11 Heliovaara M, Makela M, Knekt P, Impivaara O: Determinants of

sciatica and low-back pain. Spine 1991;16:608-614.

12 Saraste H, Hultman G: Life conditions of persons with and without

low back pain. Scan J Rehab Med 1987; 19: 109-113.

13 Nayha S, Videman T, Laakso M, Hassi J: Prevalence of low back pain

and other musculoskeletal symptoms and their association with work
in Finnish reindeer herders. Scand J Rheumatol 1991; 20: 406-413.

14 Burdorf A, Naaktgeboren B, de Groot HC: Occupational risk factors

for low back pain among sedentary workers. J Occup Med 1993;
35:1213-1220.

15 Guo H, Tanaka S, Cameron LL, Seligman PJ: Back pain among

workers in the United States: national estimates and workers at high
risk. Am J Ind Med 1995; 28: 591-602.

16 Froom P, Froom J, Van Dyk D, Caine Y: Lytic spondylolisthesis in

helicopter pilots. Aviation Space Environ Med 1984; 55: 556-557.

17 Riihimaki H, Tola S, Videman T, Hanninen K: Low-back pain and

occupation: a cross-sectional questionaire study of men in machine
operating, dynamic physical work, and sedentary work. Spine 1989;
14: 204-209.

18 Walsh K, Varnes N, Osmond C, Styles R: Occupational causes of

low-back pain. Scand J Work Environ Health 1989; 15: 54-59.

19 Burdorf A, Zondervan H: An epidemiological study of low back pain

in crane operators. Ergonomics 1990; 33: 981-987.

20 Bongers PM, Hulshof CT, Dijkstra L, Boshuizen HC: Back pain and
exposure to whole body vibration in helicopter pilots. Ergonomics
1990; 33: 1007-1026.

31
21 Johanning E: Back disorders and health problems among subway train
operators exposed to whole body vibration. Scand J Work Environ
Health 1991; 17: 414-419.

22 Miyashita K, Morioka I, Tanabe T, Iwata H: Symptoms of

construction workers exposed to whole body vibration and local
vibration. Int Arch Occup Environ Health 1992; 64: 347-351.

23 Masset D, Malchaire J: Low back pain: epidemiologic aspects and

work related factors in the steel industry. Spine 1994; 19: 143-146.

24 Magnusson M, Pope M, Wilder DG, Areskoug B: Are occupational

drivers at an increased risk for developing musculoskeletal disorders?
Spine 1996; 21: 710-717.

25 Xu Y, Back E, Orhede E: Work environment and low back pain: the

influence of occupational activities. Occupational and Environmental
Medicine 1997; 54: 741-745.

26 Brendstrup T, Biering-Sorensen F: Effect of fork-lift truck driving on

low back trouble. Scand J Work Environ Health 1987; 13: 445-452.

27 Dupuis H, Zerlett G: Whole-body vibration and disorders of the spine.

Int Arch Occup Environ Health 1987; 59: 323-336.

28 Netterstrom B, Juel K: Low back trouble among urban bus drivers in

Denmark. Scand J Soc Med 1989; 17: 203-206.

29 Boshuizen HC, Bongers PM, Hulshof CT: Self-reported back pain in

fork-lift truck and freight-container tractor drivers exposed to whole
body vibration. Spine 1992; 17:59-65.

30 Liira JP, Shannon HS, Chambers LW, Haines TA: Long-term back
problems and physical work exposures in the 1990 Ontario health
survey. Am J Public Health 1996; 86:382-387.

31 Burdorf A, Govaert G, Elders L: Postural load and back pain of

workers in the manufacturing of prefabricated concrete elements.
Ergonomics 1991; 34: 909-918.

32
32 Bovenzi M, Zadini A: Self-reported low back symptoms in urban bus
drivers exposed to whole body vibration. Spine 1992; 17: 1048-1059.

33 Bovenzi M, Betta A: Low-back disorders in agricultural tractor

drivers exposed to whole body vibration and postural stress. Applied
Ergonomics 1994; 35: 231-241.

34 Bovenzi M, Hulshof CT: An updated review of epidemiological

studies on the relationship between exposure to whole-body vibration
and low back pain. Int Arch Occup Environ Health 1999; 72: 351-
365.

35 Lings S, Leboeuf-Yde C: Whole-body vibration and low back pain: a

systemic, critical review of the epidemiological literature 1992-1999.
Int Arch Occup Environ Health 2000; 73: 290-297.

36 Teschke K, Nicol AM, Davies H, Ju S: Whole body vibration and

back disorders among motor vehicle drivers and heavy equipment
operators: a review of the scientific evidence. Unpublished 1999.

37 Mausner JS, Kramer S: Mausner and Bahn epidemiology- an

introductory text. Philadelphia, Saunders, 1985, pp 177; 180-191.

38 Industrial Diseases Standing Committee: Protocol for the assessment

of medical/scientific information. Workers’ Compensation Board of
British Columbia 1993.

39 Chernyuk V: Effects of whole body vibration on disease of the lumbar

section of the spine in agricultural machinery operators. Gigienda
Truda 1992; 28: 75-77.

40 Barbieri G, Mattioli S, Grillo S, Geminiani AM: Spinal diseases in an

Italian tractor drivers group. Agricultural Health and Safety:
Workplace, Environment, Sustainability. CRC Press 1995; pp 319-
332.

41 Ruppe K, Mucke R: Functional disorders at the spine after longlasting

whole body vibration. Advances in Industrial Ergonomics and Safety.
London, Taylor and Francis, 1993, pp 483-486.

33
42 Kompier M, de Vries M: Physical work environment and
musculoskeletal disorders in the busdriver’s profession.
Musculoskeletal Disorders at Work. London, Taylor, 1987, pp 17-22.

43 Crites Battie M, Videman T, Gibbons LE, Fisher LD: 1995 Volvo

Award in clinical sciences: Determinants of lumbar disc degeneration:
A study relating lifetime exposures and magnetic resonance imaging
findings in identical twins. Spine 1995; 20: 2601-2612.

44 Sackett D, Haynes B, Tugwell P: Clinical Epidemiology. A Basic

Science for Clinical Medicine. Little Brown and Company, 1985.
45 Checkoway W, Pearce N, Crawford-Brown D: Research Methods in
Occupational Epidemiology. Oxford University Press, 1989.

34