Analysis and statistics
For each study population, the ‘normal’ cut-off was esti-
mated using the 95th percentile of the sensory thresholds
of each digit. As the bend forces chosen for this study are
not truly continuous, it would not be appropriate to assume
a particular distribution so the 95th percentile was taken
directly from the data. The 95% confidence intervals (95%
CI) were bootstrapped with the subject as the resampling
unit, thereby taking into account any clustering due to mul-
tiple digits per person. Results were analysed by sex, age,
hand and digit for each study population. Because of the
method used to estimate the 95th percentiles and CIs, no
formal statistical tests were undertaken to compare the 95th
percentiles across groups. Instead it was assumed that if the
CIs did not overlap, then the two values were significantly
different at the
p
< 0.05 level. The
t
test was used to compare
group ages.
To determine intra-digit variability, 20 office workers and
10 heavy manual workers were re-tested in the same way and
by the same investigator two to four weeks later. The degree
of agreement between the two measurement occasions was
assessed using the percentage agreement and linear weighted
kappa. The 95% CIs were bootstrapped with the subject as
the resampling unit.
For workers with HAVS, the mean sensory perception
threshold of the two digits with the highest thresholds was
ascertained. This was compared with the number of abnor-
mal thermal (hot and cold) and vibration perception (31.5
and 125 Hz) thresholds in the index and little fingers of the
same hand. On the same day, thermal aesthesiometry was
undertaken according to the method described by Lindsell
and Griffin
2003
and vibration perception according to
ISO 13091–2 (
2003
) and as described (Poole et al.
2016
).
Thresholds > 48.5 °C for hot or < 19.0 °C for cold; > 0.4 m/s
2
for 31.5 Hz or > 1.0 m/s
2
for 125 Hz were taken as abnormal
(HSE
2005
; Lindsell and Griffin
2003
), making the maxi-
mum QST score per finger four.
The relationship between the mean SWM threshold of
the two digits with the highest thresholds and the number of
QST abnormalities in the same hand of subjects with HAVS
was modelled using mixed effects Poisson regression with
the subject as the random effect and using a restricted cubic
spline with three knots. The small number of results ≥ 10 g-f
were excluded. The utility of SWM for identifying abnor-
malities of QST was assessed using Receiver Operating
Characteristic (ROC) curves and sensitivity and specificity
calculated at different sensory thresholds.
Results
There were 300 office workers of whom 155 were male,
144 female and one unstated. Their mean age was 42 (range
19–68) years. There were 272 who declared their right-hand
and 27 their left-hand to be dominant, with one declaring
ambidexterity. The ambient temperature of the laboratory
varied between 22–24 °C and humidity between 30 and
35%. Table
1
shows sensory perception broken down by
sex, age, digit and hand-dominance. The 95th percentile
was 0.16 g-f and did not vary by sex, hand-dominance or
digit. Women < 30 had a significantly lower 95th percentile
for sensory perception than women ≥ 30 years. Men had a
higher 95th percentile for the dominant thumb (0.40 g-f)
compared with the other digits, but as the CIs overlapped the
difference did not reach statistical significance.
The frequency distribution of sensory perception for the
office workers showed the median to be 0.07 g-f with no
digit with a threshold greater than 0.6 g-f (Table
2
). The
data were positively skewed towards the lower sensory
thresholds.
There were 115 heavy manual workers all of whom
were male. Their mean age was 40 (range 18–66) years,
which was not significantly different from the office workers
(
p
= 0.114). There were 99 who declared their right-hand and
14 their left-hand to be dominant, with one declaring ambi-
dexterity and one unstated. There was no difference in sen-
sory perception by hand dominance. The 95th percentile for
the dominant thumb was 1.00 g-f (95% CI 1.00–1.40), which
was significantly higher than the middle 0.60 g-f (95% CI
0.60–0.60) and the ring fingers 0.60 g-f (95% CI 0.40–0.60)
on the dominant hand.
Table
2
shows the frequency distribution of the sensory
perception thresholds for the office and heavy manual work-
ers. The median threshold for heavy manual workers was
0.16 g-f and there was no threshold > 2.00 g-f. The 95th
percentile was 1.00 g-f (95% CI 0.60–1.00), which was
significantly greater than for office workers 0.16 g-f (95%
CI 0.16–0.16). Heavy manual workers ≥ 50 years had the
highest 95th percentile of 1.4 g-f (95% CI 1.00–2.00),
but there was no consistent trend with age; 40–49 years
0.60 g-f (95% CI 0.40–1.00); 30–39 years 0.40 g-f (95% CI
0.28–0.60); < 30 years 0.60 g-f (95% CI 0.60–1.00).
Table
3
shows the degree of agreement between
repeated measurements on the same digits. There was per-
fect agreement for 198/300 (66%) of digits; 92/300 (31%)
differed by one filament; 10/300 (3%) differed by two
International Archives of Occupational and Environmental Health
1 3
filaments. No finger differed by more than two filaments.
The weighted kappa statistic for intra-subject reliability
was 0.63 (95% CI 0.53–0.70). The degree of agreement
for the office workers was not substantially different from
that of the heavy manual workers.
There were 62 cases of neurological HAVS all of whom
were male. Their mean age was 51 (23–69) years. They
were older than the office (
p
< 0.001) and the heavy man-
ual workers (
p
< 0.001). There were 55 who declared their
right-hand and 7 their left-hand to be dominant. There
were no significant differences in the 95th percentiles of
sensory perception by age, digit or hand dominance. Of the
124 hands that were classified and neurologically staged
according to the ICC (Poole et al.
2019
); 4 were stage 0;
66 were stage 1; 36 were stage 2 and 18 were stage 3.
For workers with HAVS the median threshold of per-
ception by SWM was 0.6 g-f and 11 digits had a thresh-
old ≥ 10 g-f. The 95th percentile was 4.0 g-f (95% CI
2.0–6.0), which was significantly greater than the office
workers and the heavy manual workers (confidence intervals
not overlapping). Table
4
shows the frequency distribution
of sensory perception thresholds for workers with HAVS.
Figure
1
shows the spread of sensory perception in the
digits by SWM for the three study populations. The range
increased from office workers (0.04–0.60 g-f), to heavy man-
ual workers (0.04–2.0 g-f), to HAVS cases (0.04 to ≥ 10 g-f).
Dostları ilə paylaş: |