Fig 1.
Flow chart of the search strategy and selection of studies in order to evaluate the relationship between the hand-arm vibration and health outcomes in accordance with the PRISMA (Raynaud’s phenomenon, neurosensory injury, carpal tunnel syndrome).
Table 1.
Includes studies of Raynaud’s phenomenon and their assessed risk of bias (quality score) regarding the diagnosis “Raynaud’s phenomenon” (Diagnosis sum) and the total sum when the quality score for assessing diagnosis’ Raynaud’s phenomenon, “study method” and “exposure” has been added (Total score).
The studies are presented in descending order based on the total score. Higher scores indicate higher “quality”, indicating a smaller possible risk of bias. Furthermore, the study design is given for each study.
Fig 2.
The prevalence of Raynaud’s phenomenon in the included studies sorted by year of publication.
Fig 3.
Statistics and forest plot of a “random - effect” meta-analysis of the prevalence of Raynaud’s phenomenon between the groups exposed to HAV and non-exposed reference groups.
The size of the square of the individual studies is proportional to the study’s weight in the analysis. The red diamond indicates the overall risk. The studies have been ranked in order from highest to lowest quality score points according to quality criteria in appendix (S1 Table) and Table 1. The asterisk indicates that the study data presented made it possible to calculate the unadjusted odds ratio.
Fig 4.
Statistics and forest plot of the weighting for each dose studies in Raynaud’s phenomenon sorted by various measuring cup.
The size of the square of the individual studies is proportional to the study’s importance in the analysis. The blue diamond’s (rhomboids) shows the combined effect of the subgroups of measuring cup; Dose 1 = Number of exposure years (year); Dose 2 = Number of exposure hours (h); Dose 3 = Daily vibration exposure, A (8) (m/s2); Dose 4 = Cumulative vibration exposure (mh/s2, m2h/s4 or m2h3/s4). The studies have been ranked in order from highest to lowest quality score points according to Table 1. The asterisk indicates that the study data presented made it possible to calculate the unadjusted odds ratio.
Fig 5.
Meta-Regression analysis of the relationship between the logarithm of the prevalence of Raynaud’s phenomenon in vibration-exposed and vibration exposure A(8) (random effect; n = 14).
The figure also shows the regression line’s 95% confidence intervals intervals (red lines). Blue lines show the prediction interval. The size of the circles represents the importance of the study results in the estimates of the regression.
Fig 6.
Funnel plot with pseudo 95% confidence interval for publication bias in studies of the association between the occurrence of Raynaud’s phenomenon among groups exposed to HAV and non-exposed reference groups.
Beggs test shows no evidence of publication bias (p = 0.14), while Eggers test indicates such an effect (p <0.01). The trim and fill method imputed three missing studies to the left of the mean (random-effects model).
Fig 7.
Funnel l plot with pseudo 95% confidence interval for publication bias in studies of the association between the occurrences of Raynaud’s phenomenon among groups exposed to different levels of HAV.
Beggs test (p = 0.04), but not Eggers test (p = 0.36), showed evidence of publication bias and trim and fill method imputed seven studies lacked the left of the mean (random-effect model).
Table 2.
Included studies of neuro-sensory injury and their assessed risk of bias (quality score) regarding the diagnosis “neuro-sensory injury “(Diagnosis sum) and the total sum when the quality score for assessing diagnosis’ neuro-sensory injury, “study method” and “exposure” has been added (Total score).
The studies are presented in descending order based on the total score. Higher scores indicate higher “quality”, indicating less possible risk of bias. Furthermore, the study design is given for each study.
Fig 8.
The prevalence of neuro-sensory injury in the studies included sorted by year of publication.
Fig 9.
Statistics and “forest plot” with aggregate from “random - effect” meta-analysis of the incidence of neurosensory injury between groups exposed to HAV and non-exposed reference groups.
The size of the squares of the individual studies is proportional to the importance of the study in the analysis. The red diamond represents the weighted risk for all studies. The studies have been sorted in order from highest to lowest quality score points according to Table 2. The asterisk indicates that the study data presented made it possible to calculate the unadjusted odds ratio.
Fig 10.
Statistics and Forest plot of the weighting for each dose studies in neuro-sensory injury sorted by the various dose measures.
The size of the square of the individual studies is proportional to the study’s importance in the analysis. The blue diamonds (rhomboids) shows the combined effect of the subgroups of dose measures. Dose 2 = Number of exposure hours (h); Dose 4 = Cumulative vibration exposure (mh/s2, m2h/s4 or m2h3/s4). The studies have been ranked in order from highest to lowest quality score points according to Table 2. The asterisk indicates that the study data presented made it possible to calculate the unadjusted odds ratio.
Fig 11.
Meta-Regression analysis of the relationship between the logarithm of the prevalence of neurosensory injury in vibration-exposed and vibration exposure A(8) (random effect; n = 15).
The figure also shows the regression line 95% confidence limits (red lines). Blue lines show the prediction interval. The size of the circles represents the importance of the study results in the estimates of the regression.
Fig 12.
Funnel plot with pseudo 95% confidence interval for publication bias in studies of the association between the occurrences of neuro-sensory injury among groups exposed to HAV and non-exposed reference groups.
A Beggs test shows evidence of publication bias (p = 0.04), while Eggers test indicates no evidence the effect (p = 0.07). The trim and fill method imputed no missing study (random-effects model).
Fig 13.
Funnel l plot with pseudo 95% confidence interval for publication bias in studies of the association between the occurrences of neuro-sensory impairment among groups exposed to different levels of HAV.
A Beggs test (p = 0.04) and Eggers test (p = 0.02) showed evidence of publication bias while the trim and fill method imputed no missing study (random-effects model).
Table 3.
Studies included of CTS and their estimated risk of bias (quality score) regarding the diagnosis “Carpal tunnel syndrome” (Diagnosis sum) and the total sum when the quality score for assessing diagnosis’ CTS, “study method” and “exposure” has been added (Total score).
The studies are presented in descending order based on the total score. Higher scores indicate higher “quality” indicating less possible risk of bias. Furthermore, the study design is given for each study.
Fig 14.
The prevalence of CTS in the studies included sorted by year of publication.
Fig 15.
Statistics and “forest plot” with aggregate from “random - effect” meta-analysis of the prevalence of CTS among groups exposed to HAV and non-exposed reference groups.
The size of the squares of the individual studies is proportional to the study’s importance in the analysis. The red diamond represents the weighted risk for all studies. The studies have been sorted in order from highest to lowest quality score points according to Table 3. The asterisk indicates that the study data presented made it possible to calculate the unadjusted odds ratio.
Fig 16.
Funnel plot with pseudo 95% confidence interval for publication bias in studies of the association between the occurrences of CTS among groups exposed to HAV and non-exposed reference groups.
A Beggs and Eggers test showed no evidence of publication bias (p = 0.50; p = 0.21) and the trim and fill method imputed one missing study to the right of the mean (random-effects model).
Fig 17.
Calculated 10% correlation between the prevalence of Raynaud phenomenon (25 studies; 40 values) and neurosensory injury (17 studies; 21 values) as a function of the 8-hour equivalent frequency-weighted acceleration and number of years of exposure.
In the figure shown, the linear regression line for the two outcomes, and the corresponding curve of ISO 5349–1 [(Equations: Prevalence Raynaud phenomenon (%) = 10 ^ (1.35+ log10 (A (8) + - 0.53)) r = 12.39; Prevalence neurosensory damage (%) = 10 ^ (0.9+ log10 (A (8) + - 0.54)), r = 12.55)].