Article Text
Abstract
Objectives: To investigate prospectively the relation between vibration-induced white finger (VWF) and measures of cumulative (lifetime) exposure to hand-transmitted vibration (HTV).
Methods: Two hundred and forty-nine HTV workers and 138 control men of the same companies participated in a 3-year follow-up study. The diagnosis of VWF (Raynaud’s phenomenon in the controls) was based on the medical history, the administration of colour charts and the results of a cold test. Tool vibration magnitudes were expressed as root-mean-square (r.m.s.) acceleration, frequency-weighted according to international standard ISO 5349-1 and also unweighted over the frequency range 6.3–1250 Hz. From the vibration magnitudes and exposure durations, alternative measures of cumulative vibration dose were calculated for each HTV worker, according to the expression: dose = Σaimti, where ai is the acceleration magnitude on tool i, ti is the lifetime exposure duration (hours) for tool i, and m = 0, 1, 2 or 4.
Results: The incidence of VWF varied from 5 to 6% in the HTV workers versus 0 to 1.5% for Raynaud’s phenomenon in the controls. After adjusting for potential confounders, measures of cumulative vibration dose derived from total operating hours and high powers of unweighted acceleration (ie, , with m>1) gave better predictions of the occurrence of VWF than dose measures calculated from frequency-weighted acceleration (ie, ). These findings were observed in the entire sample of HTV workers, in those with no VWF at the initial investigation, and in those with normal cold test results at baseline.
Conclusions: This prospective cohort study suggests that measures of cumulative vibration doses constructed from unweighted r.m.s. acceleration perform better for the prediction of VWF than dose measures calculated according to the recommendations of current standards. These findings should contribute to the improvement of the ISO frequency weighting for evaluating the severity of hand-transmitted vibration.
Statistics from Altmetric.com
Vibration-induced white finger (VWF), a secondary form of Raynaud’s phenomenon, is one of the main components of the hand–arm vibration syndrome.1 This vasospastic disorder occurs in workers with prolonged exposure to vibration from handheld tools and is considered a prescribed occupational disease in most of the industrialised countries. In addition to some individual and environmental factors (eg, smoking, cold exposure), it is believed that the onset and the severity of VWF depend on the physical characteristics of hand-transmitted vibration (HTV). There is epidemiological evidence that the frequency, magnitude and duration of vibration are associated with the occurrence of VWF in professional users of a great variety of vibratory tools.1 2
What this paper adds
To assess the severity of exposure to hand-transmitted vibration (HTV), the international standard ISO 5349-1 (2001) recommends to weight vibration magnitude (r.m.s. acceleration) according to a frequency-weighting curve which assumes that the effect of vibration acceleration decreases in inverse proportion to frequency from 16 to 1250 Hz.
Since the exposure-response relationship for vibration-induced white finger (VWF) suggested by ISO 5349-1 is derived from epidemiological studies of cross-sectional type, the aim of this prospective cohort study of the occurrence of VWF in forestry and stone workers was to compare exposure-response models in which the measures of vibration exposure were constructed with and without the current ISO frequency weighting and with different time dependencies.
This study showed that the strength of the association between VWF and HTV exposure varied according to the job title of the exposed workers, the diagnostic criteria for VWF, and the metrics of HTV exposure.
Measures of vibration dose constructed from unweighted r.m.s. acceleration over the frequency range 6.3–1250 Hz perform better for the prediction of VWF than dose measures derived from r.m.s. acceleration frequency-weighted according to the current standards.
There is experimental and epidemiological evidence for an improvement of the vibration frequency weighting and the time dependency for VWF currently recommended by the international standard ISO 5349-1.
The international standard ISO 5349-1 recommends to weight vibration magnitude (acceleration) according to a frequency-weighting curve (defined Wh) which assumes that the effect of vibration acceleration decreases in inverse proportion to frequency from 16 to 1250 Hz.3 In an informative annex, ISO 5349-1 suggests an almost linear relationship between daily vibration exposure (expressed in terms of 8 h energy-equivalent frequency-weighted acceleration magnitude according to a second power time dependency) and the number of years of exposure for a probability of developing VWF in 10% of an exposed worker population.3 It should be noted that the tentative ISO dose-response relationship is basically derived from epidemiological studies of cross-sectional type which are known to suffer from several drawbacks and are unsuitable to explore causal associations between risk factors and health disorders.
The aim of this prospective cohort study was to investigate the relations between alternative measures of cumulative vibration exposure and the occurrence of VWF outcomes over time in Italian professional users of vibratory tools who were recruited in a 4-year research project entitled “Risks of Occupational Vibration Injuries (VIBRISKS)” and funded by the European Commission.4
In this study, the effects of vibration frequency, magnitude and duration were investigated by comparing exposure-response models in which the measures of vibration exposure were constructed with and without the current ISO frequency weighting and with different time dependencies.
Materials and methods
Subjects and medical investigation
In autumn 2003 to winter 2004, the occurrence of VWF and the cold response of digital arteries were investigated in seven public companies of forestry workers and one private company of stone workers. The overall study population included 274 forestry workers and 36 stone workers. Of the former, 232 workers operated brush and chain saws in the forest, while the remaining 42 subjects used both saws and pneumatic hammers and breakers for maintenance work in the mountain (tracks, footpaths, roads). The saws used by the forestry workers were equipped with anti-vibration devices. The stone workers operated both rotary and percussive tools for marble processing.
From autumn 2004 to winter 2007, three follow-up investigations were carried out in the same calendar period as for the baseline survey. Figure 1 shows the flow chart of the cohort study. One hundred and seventy-seven workers (57.1%) participated in three follow-up surveys, and 72 (23.2%) in one or two follow-up investigations. All subjects continued to work with vibratory tools during the follow-up. Of the 61 subjects lost to follow-up and the 72 workers who did not participate in all follow-up investigations, 15 were excluded from the study because of cardiovascular or metabolic disorders, 28 had changed their place of residence, 64 were retired, 10 refused to participate in the follow-up, 1 was dead, and 15 could not be identified. At the cross-sectional survey, these subjects did not differ significantly from those with complete follow-up with respect to age, smoking and drinking habits, measures of vibration exposure, and prevalence of vibration-induced vascular disorders.
One hundred and forty-three workers employed at the same companies (maintenance operators, inspectors, supervisors) and unexposed to HTV, served as controls at the cross-sectional survey. Ninety-nine control subjects (69.2%) participated in three follow-up surveys, and 39 (27.3%) in one or two follow-up investigations (fig 1). These latter and the five subjects lost to follow-up dropped out of the study for reasons similar to those reported for the HTV workers.
All subjects gave signed informed consent to the study, which was approved by the local health authorities.
The HTV workers and the controls were interviewed by trained occupational health physicians on their personal, work and health histories using a structured questionnaire developed within the project VIBRISKS.4 Both the HTV workers and the controls were questioned about smoking, alcohol consumption, metabolic, cardiovascular, and neurological disorders, previous musculoskeletal injuries, and use of medicines. Ex-smokers were classified as no smokers if they had stopped smoking for at least 2 years. The same time period was applied for ex-drinkers to be classified as no drinkers. Each subject underwent a physical examination focused on the vascular, neurological and musculoskeletal systems of the upper limbs.
Vascular outcomes
For the purpose of this study, VWF was defined according to two different diagnostic criteria:
(1) A clinical criterion based on a reliable history of white finger symptoms assisted by colour charts (VWF(H)) (see below).
(2) A more restrictive criterion based on a positive history of white finger assisted by colour charts and supported by abnormal findings at a cold test with measurement of finger systolic blood pressure (VWF(HCT)).
The anamnestic diagnosis of VWF at the medical interview was made according to the minimal requisites established at the Stockholm Workshop ’94: (a) positive history of cold provoked episodes of well-demarcated blanching in one or more fingers after excluding primary Raynaud’s phenomenon or other probable causes of secondary Raynaud’s phenomenon; (b) first appearance of finger blanching after the start of occupational exposure to HTV and experience of finger blanching attacks during the last 2 years.5
To confirm a positive history of white finger symptoms, colour charts were administered to the subjects. The colour charts consisted of a series of photographs illustrating various degrees of blanching, cyanosis, or redness of the fingers and hands, according to the scheme proposed by Maricq and Weinrich,6 partially modified.7 The procedure for the administration of colour charts has been described in detail elsewhere.7 A clinical diagnosis of white finger was considered positive when the subject, in addition to subjective symptoms, indicated the photographs displaying well-demarcated blanching of the fingers. White patching of hand palm, cyanosis of fingers, or acrocyanosis alone were not considered to be sufficient for a diagnosis of Raynaud’s phenomenon.5 Individuals reporting white finger symptoms at the medical interview but not supported by colour charts were not classified as cases of Raynaud’s phenomenon.
Cold provocation test
The cold test was performed with the subject in a supine position after a rest period of 20–30 minutes in a laboratory room with an ambient temperature of 20–22°C. The cold test consisted of strain-gauge plethysmographic measurement of finger systolic blood pressure (FSBP) in a test finger cooled to 30°C and 10°C and in a reference non-cooled finger according to the recommendations of the international standard ISO 14835-2.8 After five minutes of ischaemic cooling, FSBP was measured by a strain gauge in the distal phalanx of the test and reference finger and the following FSBP index was calculated:
FSBP%10° = (FSBPt,10°×100)/[FSBPt,30° − (FSBPref,30° − FSBPref,10°)] (%)
where FSBP%10° is the change of systolic blood pressure in the test finger at 10°C (FSBPt,10°) as a percentage of the pressure at 30°C (FSBPt,30°), corrected for the change of pressure in the reference finger during the examination (FSBPref,30° − FSBPref,10°).
In the controls of this study, the lower normal limit for FSBP%10° (mean – 2SD) was 59.6% at baseline (n = 143) and 60.1% at the end of the follow-up (n = 99). Over the entire follow-up period, 498 cold tests were performed in the controls and an overall lower normal limit of 61% was estimated for FSBP%10°. As a result, a value of ⩽60% for FSBP%10° was assumed as a discriminating threshold to differentiate between normal and pathological findings at the cold test in the HTV workers.
Measurement and assessment of vibration exposure
Vibration was measured on the brush saws, chain saws and pneumatic tools used by the forestry workers, and on the grinders, polishers and inline hammers used by the stone workers. Vibration measurements were made in the field during real operating conditions according to the ISO 5349-1 recommendations.3
Vibration magnitudes were expressed as root-mean-square (r.m.s.) acceleration, frequency-weighted using frequency weighting Wh in accordance with ISO 5349-1.3 In addition, unweighted acceleration magnitudes were obtained over the same one-third octave band frequency range (6.3–1250 Hz). The root-sum-of-squares (also called “vibration total value”) of the frequency-weighted (ah,wv) and unweighted (ah,uwv) r.m.s. acceleration values in three orthogonal directions was calculated:
The mean values (range) of ah,wv and ah,uwv for the various tools are reported in online supplementary material.
Questionnaire data, information obtained by interviewing employees and employers, and company records were used to estimate the duration of exposure to vibration for each individual. To assess daily exposure duration to vibration, direct observation of exposure patterns at the workplace was made by supervisors over an entire week period. They used a stopwatch method and recorded the contact time the hands of the operator were actually exposed to the vibration from the tools.
Tool-operating durations were obtained in hours per day, days per year and total number of years, separately for each period of use of each tool type. The estimated total (ie, lifetime) vibration exposure duration in hours for each subject was obtained by addition of the operating durations for the different tools. Total operating hours and years of usage of vibratory tools were calculated as those at the time of the study (for the workers without vascular disturbances) or those at the time of the onset of symptoms (for the subjects who reported white finger).
Using the vibration magnitudes and total exposure durations, various alternative measures of cumulative vibration doses were constructed for each subject, according to the following general form9:
where ai is the acceleration magnitude (ah,wv or ah,uwv in ms−2 r.m.s.) of tool i, ti is the total exposure duration (hours) for tool i, and m = 0, 1, 2 or 4.
In these doses, the relative importance of the acceleration, ai (weighted (ie, ah,wv) or unweighted (ie, ah,uwv)) and the total exposure duration, ti, depends on the value of m. If m has the value 2, the relationship between a and t is that assumed in r.m.s. averaging as suggested by the international standard ISO 5349-1 to evaluate vibration exposure over a working day.3 Assigning values of 1 or 4 to m decreases or increases, respectively, the “importance” of the vibration magnitude, ai, relative to that of exposure duration, ti. With m = 0, the dose takes no account of vibration magnitude. Doses with m = 0, 1, 2 and 4 were computed for each worker, with both frequency-weighted acceleration and unweighted acceleration.9
In addition to the above measures of cumulative vibration dose, which all increase with increasing exposure duration, total equivalent frequency-weighted acceleration r.m.s. magnitudes and unweighted r.m.s. acceleration magnitudes (ah,w(eq,T) and ah,uw(eq,T), respectively) were computed for each subject:
where ai is the acceleration magnitude (ah,wv or ah,uwv in ms−2 r.m.s.) of tool i, and ti is the total exposure duration (hours) for tool i.
Data analysis
The statistical analysis of data was performed with the Stata software, V.10.1 (Stata Corporation, 2008).
Continuous variables were summarised with the mean or the median as measures of central tendency and quartiles or range as measures of dispersion.
Comparisons between independent groups were made with non-parametric statistics.
Point prevalence, period prevalence, cumulative incidence, crude relative risk (RR) and 95% CI were estimated by either conventional or exact epidemiological methods when appropriate.
The relations of vascular outcomes to alternative measures of cumulative vibration exposure, while controlling for potential confounders, were assessed by the generalised estimating equations (GEE) method with an exchangeable correlation structure to account for the within-subject dependency of the observations over time.10 Continuous or binary response variables were modelled using identity or logit link functions, respectively. For binary outcomes, odds ratios (ORs) and 95% CI were estimated from the logistic regression coefficients and their robust standard errors.
GEE analysis was performed with a time-lag model to investigate the temporal sequence of the relationship between the occurrence of vascular outcomes and the predictor variables: the response variable for subject i at time point t (Yit) was related to independent variable(s) k for subject i measured at time point t −1 (Xikt −1), that is, at one time point earlier (1-year lag period in this study).
Both exposure variables and confounding factors entered the linear or logistic models as time-dependent covariates, except for age at entry which was included as a time-independent continuous variable. All models included a linear term for time effect. The technique of fractional polynomials was used to check for non-linear relationships between outcomes and alternative measures of cumulative vibration exposure (program fracpoly in Stata). The contribution of covariates to the fit of the regression models was tested by the Wald χ2 statistic assuming a p value <0.05 as the limit of statistical significance.
The “quasi-likelihood under the independence model criterion” (QIC), a modification of the Akaike’s Information Criterion (AIC), was used to compare the fit of GEE longitudinal models including alternative measures of cumulative vibration dose.11 12 The model with the smallest QIC value was chosen as the best-fitting model for the relation between vascular outcomes and measures of vibration dose, while adjusting for covariates. To aid comparison, a Δ QIC was calculated as the difference between the QIC values for a specific dose model and the model including exposure duration in years. By analogy with a suggested rule of thumb for the AIC method, a Δ QIC ⩾2 was assumed to support evidence for one model to fit the data better than another model.13
Results
Characteristics of the study populations
At the cross-sectional survey, the controls and the HTV workers were comparable for various individual characteristics (age, anthropometric variables, drinking) with the exception for smoking habit which was more frequent in the HTV workers than in the controls (table 1). Vibration exposure, in terms of either vibration magnitude (total equivalent frequency-weighted and unweighted acceleration magnitudes in ms−2 r.m.s.) or duration of exposure (total operating hours with vibratory tools), was significantly greater in the stone workers than in the forestry operators, while there was no difference for job seniority (years of employment) between the two groups.
Prevalence and incidence of VWF
Table 2 reports the point prevalence at baseline (2003–2004), the period prevalence (2003–2004 to 2006–2007), and the cumulative incidence (2004–2005 to 2006–2007) of VWF according to the aforementioned diagnostic criteria (VWF(H) or VWF(HCT), see methods).
At the initial investigation, the point prevalence of VWF(H) among the HTV workers who participated in the follow-up was 17.3% (14.0% in the forestry workers, 38.2% in the stone workers), while the prevalence of subjective symptoms of Raynaud’s phenomenon was 5.8% in the controls (crude RR 2.97 (95% CI 1.44 to 6.15)). During the follow-up period, there were 11 new cases of VWF(H) in the HTV workers and two new cases of Raynaud’s phenomenon in the controls, giving a 3-year incidence of 5.3% and 1.5%, respectively (crude RR 3.47 (95% CI 0.78 to 15.4)). Over the study period, the overall prevalence of VWF(H) and Raynaud’s phenomenon was 21.7% and 7.3% in the HTV workers and the controls, respectively (crude RR 2.99 (95% CI 1.58 to 5.69)).
A different picture of the prevalence and the incidence of white finger was observed in both the controls and the HTV workers when the more restrictive diagnostic criterion VWF(HTC) was adopted for the definition of the outcome (see above). At baseline, the point prevalence of VWF(HTC) was 5.2% in the HTV workers and that of Raynaud’s phenomenon with abnormal cold test results was 0.7% in the controls (crude RR 7.20 (95% CI 0.95 to 54.5)). During the follow-up, the cumulative incidence of VWF(HTC) was 5.9% in the HTV workers, while no new cases of Raynaud’s phenomenon were observed in the controls (crude RR +∞ (exact 95% CI 1.98 to +∞)). The overall prevalence of VWF(HTC) and Raynaud’s phenomenon over the study period was 10.8% and 0.7% in the HTV workers and the controls, respectively (crude RR 14.9 (95% CI 2.06 to 109)).
Exposure-response relationship for VWF in the HTV workers
Tables 3–5 report the results of GEE regression analyses aimed at investigating possible exposure-response relationships in the HTV workers. To avoid spurious findings, the control subjects were excluded from data analysis. After adjustment for potential confounders (age, body mass index, smoking and drinking habits, leisure activity with vibratory tools, survey time), logistic analysis with a GEE time-lag model showed that most of the alternative measures of cumulative vibration dose were significantly related to the occurrence of either VWF(H) (table 3) or VWF(HCT) (table 4) (0.0001<p<0.05), with the exception for the incidence of VWF(H) (third column in table 3). The risk of VWF over time increased with the increase of lifetime vibration doses, linearly for VWF(H) or after square root transformation of dose measures for VWF(HCT). Total operating hours with vibratory tools (Σti) was more significantly associated with VWF than years of employment. The magnitude of the Wald statistic suggested that cumulative vibration doses computed from combining duration of exposure and frequency-weighted or unweighted accelerations () performed better for the prediction of VWF than measures of lifetime exposure duration alone (years of employment or total operating hours). In general, the Δ QIC values suggested better fits to the data when dose measures with high powers of unweighted acceleration (m>1) were included in the models, mainly for the prediction of VWF(H) and VWF(HCT) incidences.
These findings were confirmed when a measure of vibration magnitude (total equivalent r.m.s. frequency-weighted (ah,w(eq,T)) or unweighted (ah,uw(eq,T)) acceleration), and a measure of lifetime exposure duration (years of exposure or total operating hours) were included together as independent variables in the GEE logistic models (table 5). Unweighted r.m.s. acceleration ah,uw(eq,T) was a stronger predictor of both VWF(H) and VWF(HCT) than frequency-weighted r.m.s. acceleration ah,w(eq,T). The QIC value for the model including ah,uw(eq,T) and total operating hours was smaller than the QIC values for the other combinations of vibration magnitude and duration.
The relations between the results of the cold test expressed in terms of FSBP%10° and the various measures of cumulative vibration dose are reported in online supplementary material. The vasoconstrictor response to cold was investigated separately in three subgroups: (1) all HTV workers; (2) the HTV workers without VWF(H) at the cross-sectional survey; and (3) the HTV workers with normal cold test results at baseline (ie, FSBP%10°>60%). After adjustment for several covariates, cold-induced digital vasoconstriction over time (ie, reduction of FSBP%10°) in the three samples was significantly associated with the increase in almost all of the measures of vibration dose. Similarly to the findings observed for the binary outcomes VWF(H) and VWF(HCT), dose measures with high powers of unweighted r.m.s. acceleration (m>1) performed better for the prediction of cold test results than other alternative measures of vibration dose.
In addition to vibration exposure, multivariable regression analysis of repeated measures over time showed significantly inverse associations between vascular outcomes and body mass index (OR 0.78 (95% CI 0.62 to 0.97) to 0.86 (95% CI 0.76 to 0.98)). In some logistic models including measures of vibration dose with high powers of unweighted acceleration (, with m>1), the occurrence of VWF(H) or VWF(HCT) over the follow-up period was positively associated with age at entry (OR 1.72 (95% CI 1.02 to 2.90) to 1.87 (95% CI 1.08 to 3.24)) and smoking habit (OR 3.56 (95% CI 1.04 to 12.2) to 4.07 (95% CI 1.09 to 15.2)). There were no significant interactions between these variables and measures of vibration dose.
Discussion
This prospective cohort study showed significant associations between the occurrence of VWF over time and measures of cumulative vibration exposure in a population of professional users of vibratory tools. The strength of the associations varied according to the job title of the exposed workers, the diagnostic criteria for VWF, and the metrics used to construct the alternative measures of vibration dose.
Comparisons with previous studies of forestry and stone workers
In this cohort study, the incidence of VWF outcomes was greater in the stone workers than in the forestry operators. The results of previous epidemiological surveys seem to be consistent with the findings of the present study. In a personal review of the cross-sectional studies of stone workers carried out from 1918 to 1993, the prevalence of VWF symptoms ranged from 20.1 to 89.5%.14 VWF prevalence weighted across 20 surveys carried out in Europe, the USA, and Japan averaged 35%, a figure comparable with that observed in the present investigation.
The occurrence of VWF symptoms among forestry workers has gradually decreased in the last three decades. It has been reported that the prevalence of VWF in Finnish chain saw operators lowered from 40% in 1972 to 4% in 1995.15 A retrospective cohort study of 1551 Japanese forestry workers showed a reduction of the prevalence of VWF from 31% in 1973 to 17% in 1988.16 The changes in the occurrence of VWF have been ascribed to the use of lighter anti-vibration (AV) chain saws and the introduction of administrative measures curtailing saw usage time and improving the organisation of forest work. Even though the introduction of AV chain saws in forest work has had beneficial effects in terms of reduction of the occurrence of VWF, our findings and those of other investigators suggest that health surveillance should be maintained for saw operators who handle modern chain saws as the risk of developing VWF is still present in forest work.
Dose-response patterns for VWF
It is believed that the physical characteristics of vibration (magnitude, frequency, direction) and exposure duration are the major determinants of the adverse effects of HTV on the upper limbs of tool users.1 In this study, frequency-weighted or unweighted acceleration magnitudes and total hours of tool usage were combined to construct alternative measures of lifetime vibration dose with four different time dependencies (dose = , where m = 0, 1, 2 or 4). After adjusting for potential confounders, in the HTV workers of this cohort study measures of cumulative (ie, lifetime) vibration dose derived from total operating hours and high powers of unweighted acceleration (, with m>1) gave better predictions of the prevalence and incidence of VWF (VWF(H) or VWF(HCT)) than measures of total exposure duration alone or in combination with vibration acceleration frequency-weighted according to current standards.3 Similar findings were obtained when the two components of dose measures (acceleration magnitude and exposure duration) were included together, while adjusting for each other, in logistic time-lag models. Total equivalent unweighted acceleration of vibration performed better than frequency-weighted acceleration for the prediction of either VWF(H) or VWF(HCT). The findings of this cohort study are broadly consistent with those of our previous cross-sectional survey of 1557 users of powered vibratory tools (stone workers, dockyard workers, and forestry operators), in which vibration doses constructed with unweighted acceleration were more significant predictors of the occurrence of VWF than doses calculated from frequency-weighted acceleration magnitude.9
Our findings seem to be in agreement with the results of experimental and epidemiological studies which investigated the (adverse) effects of HTV in either healthy subjects or users of vibratory tools. In an experiment of the acute vascular responses to the frequency of HTV, it was found that vibration frequencies of 31.5–250 Hz with constant weighted or unweighted acceleration magnitude provoked a greater reduction of finger blood flow than did vibration of 16 Hz in both a vibrated finger and a non-vibrated finger, suggesting that the frequency weighting given in the international standard ISO 5349-1 overestimates the acute effects of low vibration frequency on the digital circulation.17 This finding is also supported by animal experiments in which arterial damage in vibrated rat tails was found to be frequency dependent in the range 30–800 Hz.18 It should be noted that the frequency weighting defined in ISO 5349-1 is basically derived from laboratory investigations of equal sensation contours produced by vibration in the hands of a small number of healthy subjects, and as such it may be unsuitable for the assessment of chronic vascular disorders provoked by HTV.19 Several epidemiological studies have reported a poor agreement between the occurrence of VWF in various occupational groups and that predicted by the ISO 5349 model based on the currently recommended frequency weighting (Wh) of vibration acceleration.2 Both overestimation and underestimation of the risk for VWF have been observed by investigators. Risk overestimation has been mainly found in worker groups using tools with a predominantly low-frequency percussive action such as road breakers, stone hammers, and sand rammers.2 14 20 Since the ISO frequency-weighting curve gives greater weight to low-frequency vibration, it might be argued that the evaluation of such vibration according to the current standard does not reflect adequately the risk of vascular disorders. Conversely, other biodynamic, physiological, and epidemiological investigations have pointed out that the ISO weighting underestimates the vascular effects of vibration containing intermediate- and high-frequency components.2 17 21 22 23 Overall, these findings indicate that there are poor experimental and epidemiological foundations for selecting the ISO frequency weighting to assess VWF. As suggested by this and other studies, unweighted acceleration (ie, a flat weighting curve) may be more appropriate for predicting the occurrence of VWF.9 23 24
Potential sources of bias
In this study there are some potential sources of bias that deserve attention. Vibration measurements were made on the tools currently used by the forestry and stone workers and this may be a source of uncertainty for the estimation of lifetime vibration doses. It should be noted, however, that all forestry operators had work experience limited to AV chain saws, and the vibration emission from the pneumatic tools used by the stone workers has been found fairly similar over time.14 Consistently, the weighted r.m.s. acceleration magnitude of vibration measured in the tools of the present study are broadly comparable with those reported in recent and past investigations and in vibration guidelines.1 25
Quantification of duration of exposure to HTV is a difficult task because recall bias cannot be ruled out when daily exposure time is estimated by means of questionnaire or direct interview of employees and employers. Several investigations have reported that HTV workers tend to overestimate their actual daily duration of exposure to the vibration from tools.26 27 To reduce this bias, a survey was conducted in the field over an entire week period by supervisors who used a stopwatch method recommended by the EU guide on HTV to measure the time the tool operators are exposed to the vibration.22
In this study, about 20% of the HTV workers were lost to follow-up, and 23% did not participate in all follow-up investigations. The main reasons for dropping out of the study were retirement (21%) and change of residence (9%). Only a minority of the HTV workers refused to participate in the follow-up (3%) or could not be identified after the cross-sectional survey (5%). We assumed that the pattern of missing data on the incidence of VWF during the follow-up was random, even though this assumption could not be completely verified. However, there was no significant difference in the incidence of VWF at the various follow-up periods for the HTV workers lost to subsequent follow-up compared with those with complete follow-up.
Conclusions
The results of this prospective cohort study suggest that measures of cumulative vibration doses calculated from unweighted r.m.s. acceleration over the frequency range 6.3–1250 Hz perform better for the prediction of VWF and the cold response of digital arteries in users of vibratory tools than dose measures derived from r.m.s. acceleration frequency-weighted according to the recommendation of current standards. There is experimental and epidemiological evidence that more weight should be given to intermediate- and high-frequency vibration for evaluating the severity of HTV. These findings should contribute to the improvement of the vibration frequency weighting and the time dependency for VWF currently recommended by international standard ISO 5349-1.
Acknowledgments
The author would like to acknowledge the valuable contribution of the occupational health personnel of the NHS and the University of Trieste, as well as of the employers and employees of the surveyed companies, who made it possible to carry out this epidemiological study.
REFERENCES
Footnotes
Funding This research was supported by the European Commission under the Quality of Life and Management of Living Resources programme - Project No. QLK4-2002-02650 (VIBRISKS).
Competing interests None.
Provenance and Peer review Not commissioned; externally peer reviewed.