Occupation and the risk of bothersome tinnitus: results from a prospective cohort study (HUNT).

0.17 MB·English

by  Engdahl, Bo

Open Access Research
Occupation and the risk of bothersome
tinnitus: results from a prospective
cohort study (HUNT)
Bo Engdahl,1 Norun Hjertager Krog,1 Ellen Kvestad,1
Howard J Hoffman,2 Kristian Tambs1,3
Tocite:EngdahlB,KrogNH, ABSTRACT ARTICLE SUMMARY
KvestadE,etal.Occupation Objectives: Estimates ofoccupation-specific tinnitus
andtheriskofbothersome prevalence may helpidentifyhigh-risk occupations Article focus
tinnitus:resultsfrom
whereinterventions are warranted. Theauthors - We studied the effect of occupation on preva-
aprospectivecohortstudy
(HUNT).BMJOpen2012;2: studiedtheeffect ofoccupation onprevalence of lence of bothersome tinnitus and estimated the
bothersome tinnitus andestimated the attributable attributable fraction due tooccupation.
e000512.doi:10.1136/
bmjopen-2011-000512 fractiondue tooccupation. The authors also studied - We also studied if the effect remained after
howmuchoftheeffect remained after adjustingfor adjustment for noise exposure, education and
<Prepublicationhistoryfor noiseexposure,educationincome,hearingthresholds income,otherriskfactorsandhearingthreshold.
thispaperisavailableonline. andother risk factors.
Toviewthesefilesplease Design: Aprospective cohort study. Key messages
visitthejournalonline(http:// Setting:AhealthsurveyoftheNord-Trøndelagcounty - There are very few epidemiological studies
bmjopen.bmj.com). quantifying work-related tinnitus risk, and our
ofNorway.
large and representative sample gives precise
Received21October2011 Participants:Asampleofthegeneraladultpopulation
estimates of occupational risk.
Accepted28November2011 (n¼49948).
- Thisstudyshowsmoderateeffectsofoccupation
Primary outcome measure: The primary outcome
onbothersome tinnitus andpresents prevalence
Thisfinalarticleisavailable measure isbothersome tinnitus.
foruseunderthetermsof estimatesof122differentoccupationsin49948
Results: Occupation had amarked effectontinnitus
theCreativeCommons subjects.
prevalence. The effect ofoccupationontinnitus was
AttributionNon-Commercial - Noisy occupations were associated with an
2.0Licence;see reduced inmenby controlling forself-reported increased risk of bothersome tinnitus in men,
http://bmjopen.bmj.com occupational noise exposure andin womenby butinwomen,occupationswiththehighestrisk
controllingforeducation and income.Addinghearing
fortinnituswerenottypicallynoisyones,andthe
loss asa predictor increased theeffect ofoccupation
attributable fraction was determined mainly by
somewhat.Inmen, age-adjusted prevalence ratios of
the groupofoccupationally inactive.
tinnitusrangedfrom1.5(workshopmechanics)to2.1
(craneandhoistoperators)inthe10occupationswith Strengths and limitations of this study
highesttinnitus prevalence. Inwomen,themost - The major advantages of the present study are
importantcontribution tothetinnitus prevalence was the prospective design and that the study
fromthelarge group ofoccupationally inactive population is representative of the general
persons,with aprevalence ratio of1.5. working population.
Conclusion: Thisstudyfoundamoderateassociation - Theoccupationsarenotclassifiedbyriskfactors
between occupationand bothersome tinnitus. fortinnitus,butaccordingtothetasksandduties
1NorwegianInstituteof undertaken in the job. Heterogeneity regarding
PublicHealth,Divisionof exposure within occupational categories implies
MentalHealth,Oslo,Norway thatoccupation,asanexplanatoryvariable,does
2EpidemiologyandStatistics notcapturealleffectsofoccupationalexposures
INTRODUCTION
Program,NationalInstitute ontinnitus.
onDeafnessandOther Tinnitus,ortheperceptionofsoundwithout
CommunicationDisorders an external acoustic stimulus, is a common
(NIDCD),NationalInstitutes health complaint in the adult population. In higher in men and increases with age up to
ofHealth(NIH),Bethesda, addition to general irritation and annoyance a certain point, after which it declines.3e5
Maryland,USA
3DepartmentsofPsychiatry withthesound,tinnituscancausedifficulties Tinnitus frequently occurs together with
andHumanGenetics,Virginia with sleep and concentration, reduced permanent hearing loss,6 7 suggesting that
CommonwealthUniversity, speech intelligibility and various psychoso- tinnitus may be associated with cochlear
Richmond,Virginia,USA matic, emotional and interpersonal prob- damage. Tinnitus shares many of the same
lems.1 The prevalence of chronic tinnitus in risk factors as hearing loss, including occu-
Correspondenceto
DrBoEngdahl; the adult population is estimated at 8%e pational noise, work-related diseases, expo-
[email protected] 15%, depending on the definition.2 3 It is sure to toxins, non-occupational noise
EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512 1
Occupation and the risk of tinnitus
exposure, drugs or medications, otological diseases, Nord-Trøndelag county in Norway was invited to partic-
dizziness, head injury and socioeconomic and general ipate in HUNT-2, which was conducted from January
health status.4 7e11 1996 to February 1998. Screening included several types
However, tinnitus is not always secondary to hearing of examinations and two questionnaires (HUNT-2 Q1
loss and may occur in individuals with normal hearing. and Q2). Seventeen of the 24 municipalities were
Some have suggested that tinnitus is an early sign of offeredandacceptedhearingexamination,consistingof
hearing loss, in particular noise-induced hearing loss,12 pure-tone audiometry and the completion of two ques-
although there are studies opposing this hypothesis.13 tionnaires (Hearing Q1 and Q2), as part of the
Central nervous system mechanisms are believed to play screening program.
an important role in the pathology of tinnitus.14 There- The subjects ranged in age from 20 to 101years
fore, risk factors related to neural plasticity and sensiti- (median 48.0years; mean (SD) 50.2 (17.0)years). The
sationmaybeofimportance.Stressseemstoplayarole: participation rate for all municipalities was 69% except
patients often report worsening of tinnitus with stress. one (Levanger), 65% among male subjects and 73%
Workersperceivinghighjobstresshaveanincreasedrisk among female subjects. The corresponding rates for
of tinnitus,9 15 and tinnitus may be induced by stressful Levanger (where the HUNT-2 participants had to be
life events and trauma.16 Work-related stress such as low re-invited to have their hearing examined) were 42%,
degree of control, conflicting work demands, conflict 39% and 45% overall and for male subjects and female
between work and family life and lack of support from subjects, respectively. The participation rates varied with
superiors may therefore be risk factors for tinnitus. age, from about 40% for subjects younger than 30years
Tinnitus has been associated with mental health and or older than 80years to 82% for subjects from 60 to
well-being,17 factors that might themselves be work 69years.Thelowparticipationrateamongyoungpeople
related, thus mediating the association between occu- islikelydueinparttotheabsenceofstudentsandyoung
pation and tinnitus. The direction of influence between adults serving their (compulsory) military service who,
tinnitus and many of these factors is, however, unclear, while formally keeping their childhood home address,
and there may even be bidirectional relationships.18 had moved to other parts of the country.
Although tinnitus has been associated with a few A total of 51574 persons arrived for their hearing
occupationalriskfactorssuchasnoiseexposure,458911 examination and provided written informed consent.
there are very few studies quantifying occupational- Participantscompletedaquestionnaire(HearingQ1)on
specific tinnitus risk.4 Epidemiological studies of work- hearing-related information at the examination site.
relatedtinnitusareneededinordertoidentifyhigh-risk Audiometric data were missing for 774 persons (1.5%).
occupations with specific types of harmful exposure so Questionnaire data were missing or incomplete for 815
that protective measures can be implemented. persons (1.6%). The sample is described in greater
Previous analyses of data from the Nord-Trøndelag detail elsewhere.20
Hearing Loss Study (NTHLS) showed effects of self- Information on occupation, education and income
reported occupational and impulse noise exposure on wasobtainedforall,but37subjectsfromthepopulation
tinnitus.4 Analyses also revealed effects of education, register information from Statistics Norway. In total, the
income, general health status, recurrent ear infections, sample consisted of 49948 subjects with complete data.
headinjuryandcigarettesmoking.Frequentexposureto
loudmusicandhavingplayedinabandwere,incontrast, Measures
morefrequentamongsubjectswithouttinnitus.Detailed The Hearing Q1 included questions about bothersome
information on occupation type was not included in the tinnitus.Tinnituswasheredefinedbya‘yes’responseto
previous analyses. However, information from the the single general question: ‘Are you bothered by
nationwideoccupationregisterhasrecentlybeenusedto ringing in the ears?’ Missing values and a ‘no’ response
study the effect of occupation-specific hearing loss.19 were considered equivalent to ‘not bothered’. The
Theprimaryaimofthepresentstudywastodetermine Hearing Q2 included a slightly differently phrased
the effect of occupation on bothersome tinnitus. question about the degree to which the respondent is
Second,weestimatedthefractionoftinnitusthatcanbe botheredbytinnitus(responsecategories:notbothered,
attributedtooccupation-associatedrisks.Wealsowanted a little bothered and strongly bothered). In the present
to examine the extent to which differences in tinnitus study, the question and data from Q1 were used.
between the various occupations remained after adjust- Previous analysis has shown a testeretest polychoric
ment for self-reported occupational noise exposure, correlation for 27792 persons tested twice on both Q1
non-occupational noise exposure, other risk factors, andQ2,withthemajorityoftimelagsrangingfrom3to
education, income and hearing loss. 6months, of 0.65 (95% CI 0.63 to 0.66) indicating
relatively high reliability for our tinnitus measure.21
MATERIALS AND METHODS Data on occupation were obtained using census
Study population records from 1970, 1980 and 1990. The most recent
The NTHLS is part of the Nord-Trøndelag Health occupation information was used. For example, if
Study (HUNT-2). The entire adult population of asubjectwasnotworkingin1990,hisorheroccupation
2 EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512
Occupation and the risk of tinnitus
statusfrom1980wasused.About22%ofthepopulation occupation group and age in 5-year groups as fixed
(14% of male subjects and 30% of female subjects) had factors. The analyses were stratified by sex and age
no registered occupation (ie, were occupationally inac- groups (20e44, 45e64 and $65years). The model,
tive)duringallthecensusregistrationyears.Occupation a generalised linear model in which the link function is
was coded according to the Nordic Classification of the logarithm of the proportion under study and the
Occupations22 using a three-digit code. The digits distribution of the error is binomial, was estimated by
represent the major class (‘felt’), the sector (‘omra˚de’) maximum likelihood. The occupation groups 060e069
and the occupation group. The codes consist of 13, 86 were aggregated into one occupation group 06 ‘peda-
and 412 groups, respectively.23 For supplementary anal- gogical work’ and used as a reference with a sufficient
yses, the group of occupationally inactive was further number of subjects for estimating prevalence ratios
distinguished into subclasses based on questions on (PRs). Occupations with fewer than 40 subjects were
working situation in the HUNT-2 Q1. The subgroups collapsedintoonegroup.DirectestimatesofPRsbylog-
were(1)full-timehouseholdworkers,(2)militaryservice binomial regression have some advantages over ORs
or student, (3) unemployed and (4) receiving social estimatedwithlogisticregressionanalysis,25andthehigh
security or disability pension. These latter groups were prevalenceoftinnitusinoursamplemakesPRseasierto
not mutually exclusive, so that individuals could belong interpret.
to more than one of these groups. Overall model fit was determined by the residual
Education data were available for 1980, 1985, 1990, deviance, the lack of fit that remains after modelling
1995 and 1998. We used the most recent education with m predictors, as well as the McFadden pseudo R2
information. Education was classified into nine levels, defined as:
from elementary school to tertiary studies leading to
advanced professional degrees. R2 ¼ 1(cid:2)lnðLmÞ; (1)
Income data from 1980, 1985, 1990, 1995 and 1998 l lnðL Þ
0
were calculated as the mean income over the years
available, corrected for an increase in the general where Lm is the likelihood function of the model
population income during the period 1980e1998. containingmpredictorsandL isthelikelihoodfunction
0
Self-reportednoiseexposureandotherriskfactorsfor of the model containing only the intercept. As R2 does
l
hearing loss were obtained from the Hearing Q1. not reach 1, a rule of thumb is that the model has an
Occupational noise exposure was measured by ques- excellent fit with R2 being between 0.2 and 0.4.26
l
tionnaire items on the duration of exposure to loud The overall effect of occupation in the model was
noise at work in general (scored 0e3) and from specific estimated by the partial R2 (the difference in R2 values
l l
noise sources: staple gun/hammering, metal betweenamodelwithandwithoutoccupation).Changes
hammering/riveting, circular saw/machine planing, in the overall effect of occupation by controlling for
chain saw operation, tractor/construction machines, hearing loss, self-reported noise exposure, other
sledgehammer operation, blasting, machine room risk factors, education and income were measured
noise and other factory noise (scored as ‘yes’ or ‘no’). by changes in partial R2 after subsequently entering
l
Non-occupational risk factors were measured by ques- additional control variables in the model.
tionnaire items about impulse noise (ie, explosions, Inordertoestimatetheportionoftinnituscasesinthe
shootings); playing in a band or going to discotheques, population that can be attributed to an occupation, the
rock concerts or similar loud events; recurrent ear occupation-specific adjusted attributable fraction (AF)
infections (in childhood or later); hospitalisation (ever) was calculated by the following formula:27
for a head injury (scored as ‘no’, ‘perhaps or I don’t
k‘nnoo’w,’‘yaens,df‘oyres0’)toan<d5smyeoakrsin’,g‘yceigs,arfoetrte$s5datioly<(1sc5oyreeadrsa’s, pdiPRPiR(cid:2)i 1; (2)
‘yes,for$15years’).Theitemsonthequestionnaireare
described in detail elsewhere.4 24 where PRi is the adjusted prevalence ratio for the ith
Air conduction hearing thresholds were obtained by occupation relative to occupation group 063 and pdi
represents the proportion of cases in the ith occupation
pure-tone audiometry as described in an earlier publi-
cation.20 The hearing scores were computed as pure- to the total population. The sum of the occupation-
specific adjusted AF is thus:
tone average on the worse ear for three independent
mean values: (1) low-frequency hearing level (250 and
500Hz),(2)medium-frequencyhearinglevel(1000and 1(cid:2) +k pdi; (3)
2000Hz) and (3) high-frequency hearing level (3000, i¼0PRi
4000, 6000 and 8000Hz).
where k is the total number of occupations.
Statistical methods The 95% CIs of the occupation-specific AF where
The effects of occupation on the prevalence of estimated by non-linear bootstrapping with the percen-
tinnitus were estimated using alog-binomial model with tile method and 1000 replications.Allstatisticalanalyses
EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512 3
Occupation and the risk of tinnitus
Table1 Tinnitus prevalence
Men Women
Tinnitus Tinnitus
Age group Samplesize prevalence(%) Samplesize prevalence(%)
All 23374 16.4 26574 12.1
20e44 years 9359 10.6 10920 8.4
45e64 years 8618 18.5 9246 12.0
>64 years 5397 23.0 6408 18.7
were performed using the computer program R, hadthehighestPR,1.9.Thelargegroupofsubjectswith
V.2.11.1. no reported occupation had the highest AFs both in
menandinwomen,althoughtheirPRwasmoderate,1.2
RESULTS and 1.5, respectively. The sum of all occupation-specific
The tinnitus prevalence is higher in men and increases age-adjusted AFs was estimated to be 13.3% (95% CI
by age (table 1). 9.1% to 17.0%) and 21.4% (95%CI 13.9% to 24.9%) in
The goodness of fit (viz., pseudo R2) for models of men and women, respectively. In women, the overall AF
tinnitus prevalence by age, occupation, self-reported was to a great extent determined by the contribution
noise exposure, other risk factors, education, income from the group not reporting an occupation.
andhearingloss,enteredstep-andblock-wise,areshown In order to further investigate the nature of tinnitus
in tables 2 and 3 for men and women, respectively. prevalence in the group of occupationally inactive
Occupation contributed significantly to the prediction women, we restricted our sample to subjects younger
of tinnitus after adjusting for age in all age groups and than 65years (N¼5850). Significant effects after adjust-
forbothsexes.DifferencesinpseudoR2values,DR2,for mentfor age were estimated for receivingsocialsecurity
l
each model showing specifically the additional effect of ordisabilitypension(PR2.0;95%CI1.6to2.5,N¼567),
occupation compared with the same model without for being unemployed (PR 1.6; 95% CI 1.2 to 2.0,
occupation, ranged from 0.9% to 2.5%. The effects of N¼503) and for being full-time household workers (PR
occupationasobservedbyDR2 werehigherinmenthan 1.2; 95% CI 1.0 to 1.5, N¼1713). There was no effect of
l
inwomenandhighestinmenolderthan64years.Inall being in the military service or a student (PR 0.9; 95%
strata,thebest-fitmodelwasobtainedwiththecomplete CI 0.7 to 1.1, N¼936). The effect of receiving social
set of predictors, with a significant increase in pseudo security or disability pension was slightly reduced
R2 value for each step. by controlling for noise exposure and other risk factors
Controlling for self-reported occupational noise (PR 1.7; 95% CI 1.4 to 2.1) and further reduced by
exposure reduced the effect of occupation only in men controlling for education and income (PR 1.4; 95% CI
aged 45years or older. The effect of occupation was still 1.2 to 1.8). Controlling for hearing loss resulted in
statistically significant. Additional control for leisure- a negligible additional change (PR 1.3; 95% CI 1.1
relatednoise,recurrentearinfections,headinjuriesand to 1.6).
smokinghad littleinfluenceon theeffectofoccupation
at all; this was also true when controlling for education DISCUSSION
and income. Controlling for hearing loss, however, Our results showed that occupation has moderate but
increased the effect of occupation somewhat. significant effects on the prevalence of bothersome
Inwomen,theeffectofoccupationwasnotreducedby tinnitus.Theeffectofoccupationontinnitusprevalence
controlling for self-reported occupational noise expo- was smaller in women than in men, but the gender
sureorbycontrollingforleisure-relatednoise,recurrent difference was not as marked as the previously reported
ear infections, head injuries and smoking. However, effect of occupation on hearing loss in this sample.19
controlling for education and income considerably Controlling for self-reported occupational noise expo-
reduced the estimated effect of occupation, so that the sure reduced the occupation effect in men but had no
effect was no longer statistically significant. effectinwomen; thisislikely aconfirmation ofprevious
Tables 4 and 5 report the predicted age-adjusted reports that women are in general exposed to less
prevalence estimates for tinnitus according to various occupational noise. However, additional control for
occupational groups. The tables also show the AFsdthe education and income only affected the occupation
fraction of tinnitus cases in the population attributed to effect in women.
an occupation. The aggregated occupational group Several occupations recognised to be associated with
‘pedagogical work’ was specified as a reference group. loud noise exposure were associated with an increased
For men, the occupations with the highest PRs were riskoftinnitusinmenandcontributedtotheoverallAF
crane and hoist operators and miners, with PRs of 2.1 of tinnitus, that is, the fraction of tinnitus cases due to
and 1.9, respectively. For women, laboratory assistants occupation. This is in agreement with previous findings
4 EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512
Occupation and the risk of tinnitus
all n %) * ** all n %) * **
Age,occupation,yriskfactors,income,educatioandhearinglossD22RR(%)(ll 8.31.16*10.11.8011.71.97*14.62.48* Age,occupation,yriskfactors,income,educatioandhearinglossD22RR(%)(ll 7.80.45*7.40.977.31.08*7.00.85*
n
e
gmen on,y, 2R(%)l 78**7374**26** gwom on,y, 2R(%)l 38010277
ccupationamon Age,occupatiallriskfactorsincomeandeducationD2R(%)l 6.80.6.61.7.01.12.92. ccupationamon Age,occupatiallriskfactorsincomeandeducationD2R(%)l 5.40.5.41.4.11.4.10.
o o
ut n. ut n.
o o o o
handwith ationyfactorsD2R(%)l 0.76**1.741.76**2.21** outoccupati handwith ationyfactorsD2R(%)l 0.59**1.201.19*1.04 outoccupati
wit upsk with wit upsk with
nmodels Age,occandallri 2R(%)l 6.76.36.812.6 withand nmodels Age,occandallri 2R(%)l 5.14.63.93.7 withand
e el e el
e d e d
etw mo ng. etw mo ng.
22doRanddifferencesinpseudoRb Age,occupationandoccupationalupationnoiseexposureDD222RRR(%)(%)(%)lll 1.06**5.00.82**1.83*3.91.822.04**4.41.73**2.49**11.82.20**2R2Risthedifferenceinvaluesbetweenal entearinfections,headinjuriesandsmoki 22doRanddifferencesinpseudoRb Age,occupationandoccupationalupationnoiseexposureDD222RRR(%)(%)(%)lll 0.59**3.40.62**1.31*2.71.38**1.15*1.81.17*0.91*2.20.992R2Risthedifferenceinvaluesbetweenal entearinfections,headinjuriesandsmoki
mialregressionmodels.Pseu Modelvariables AgeandoccAge22RR(%)(%)ll 2.43.40.11.90.52.50.12.5D2Rdden,1979);,partialpseudolkelihoodratioteston89df.ationalnoise,leisurenoise,recurr mialregressionmodels.Pseu Modelvariables AgeandoccAge22RR(%)(%)ll 2.32.80.01.30.31.40.51.4D2Rdden,1979);,partialpseudolkelihoodratioteston54df.ationalnoise,leisurenoise,recurr
bino McFa01,liccup bino McFa01,liccup
Table2Log- Agegroup Alle2044yearse4564years>64years2R2,pseudoR(l<<*p0.05;**p0.ySelf-reportedo Table3Log- Agegroup Alle2044yearse4564years>64years2R2,pseudoR(l<<*p0.05;**p0.ySelf-reportedo
EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512 5
Occupation and the risk of tinnitus
Table4 Predictedage-adjustedPR and AFsof tinnitusamong men*
Samplesize
Nordic Classificationof Occupational Codes PR(95%CI) AF (%)y(95%CI) (23374)
872 Craneand hoistoperators, etc 2.1(1.4 to3.1) 0.23(0.05 to0.42) 53
501 Miners (inunderground mines,quarrymen, 1.9(1.5 to2.5) 0.69(0.36 to1.03) 171
shotfirers)
754 Sheet-metal workers 1.8(1.2 to2.8) 0.19(0.01 to0.38) 68
871 Stationary engine operators 1.8(1.1 to2.8) 0.15((cid:2)0.01to0.33) 40
827 Dairyworkers 1.7(1.2 to2.3) 0.35(0.11 to0.61) 125
A30 Military (senior officers) 1.7(1.1 to2.5) 0.19(0.01 to0.39) 58
881 Longshoremenand vehicle loaders 1.7(1.1 to2.6) 0.17((cid:2)0.01to0.37) 64
912 Cooks 1.7(1.0 to2.9) 0.12((cid:2)0.02to0.28) 55
909 Othersin 90public safetyand 1.6(1.0 to2.7) 0.11((cid:2)0.02to0.26) 44
protection work
751 Workshop mechanics 1.5(1.1 to2.0) 0.33(0.04 to0.64) 183
757 Metalplate and steelstructural workers 1.5(1.1 to2.0) 0.32(0.04 to0.62) 174
77 Wood work 1.5(1.0 to2.5) 0.12((cid:2)0.04to0.29) 46
821 Millers 1.5(0.9 to2.5) 0.12((cid:2)0.04to0.30) 51
331 Salesmen operatingfroman office 1.4(1.0 to1.9) 0.30((cid:2)0.02to0.58) 182
836 Papermakers 1.4(1.0 to2.0) 0.21((cid:2)0.01to0.44) 124
875 Material-handling equipmentoperators 1.4(0.9 to2.1) 0.15((cid:2)0.04to0.36) 96
0X6 Personnelspecialists 1.4(0.8 to2.3) 0.08((cid:2)0.06to0.24) 50
876 Oilers andgreasers,etc 1.4(0.8 to2.5) 0.06((cid:2)0.07to0.22) 57
753 Machine andmotorrepairmen 1.3(1.0 to1.5) 0.71(0.08 to1.31) 695
761 Electricians and electrical fitters 1.3(1.0 to1.6) 0.39((cid:2)0.08to0.83) 410
772 Sawmill andplaning millworkers 1.3(1.0 to1.6) 0.36((cid:2)0.06to0.72) 282
756 Welders and flamecutters 1.3(1.0 to1.7) 0.27((cid:2)0.07to0.63) 256
A10 Military (corporals andprivates) 1.3(0.9 to1.9) 0.21((cid:2)0.07to0.51) 247
755 Plumbersand pipefitters 1.3(0.9 to1.8) 0.20((cid:2)0.10to0.47) 177
75 Iron andmetalware work 1.3(0.9 to1.9) 0.14((cid:2)0.07to0.36) 106
7 Manufacturing and constructionwork 1.3(0.9 to2.0) 0.13((cid:2)0.06to0.33) 86
769 Othersin 76electrical work 1.3(0.9 to2.1) 0.11((cid:2)0.08to0.30) 82
765 Linemen andcable jointers 1.3(0.8 to1.9) 0.10((cid:2)0.10to0.31) 87
759 Othersin 75iron andmetal warework 1.3(0.7 to2.2) 0.06((cid:2)0.08to0.22) 56
X Occupationnot reported 1.2(1.0 to1.5) 1.88(0.16 to3.33) 3216
111 Directors, managers and workingproprietors 1.2(1.0 to1.5) 0.46((cid:2)0.10to0.98) 469
003 Otherengineers, engineertechnicians, 1.2(1.0 to1.6) 0.43((cid:2)0.07to0.91) 459
industrial designers
874 Operators ofearth-moving and construction 1.2(1.0 to1.6) 0.36((cid:2)0.09to0.76) 383
machinery
826 Butchers, sausagemakers, etc 1.2(0.9 to1.8) 0.15((cid:2)0.10to0.42) 157
299 Othersin 29other clerical work 1.2(0.9 to1.7) 0.14((cid:2)0.11to0.36) 149
853 Plasticproduct makers 1.2(0.8 to1.7) 0.13((cid:2)0.13to0.40) 163
002 Chiefengineers 1.2(0.8 to1.7) 0.11((cid:2)0.16to0.35) 153
781 Building and furniture painters 1.2(0.8 to1.7) 0.10((cid:2)0.14to0.37) 146
105 Senior administratorsand executiveofficials, 1.2(0.8 to1.7) 0.08((cid:2)0.12to0.30) 99
municipal administration
113 Administration secretaries 1.2(0.8 to2.0) 0.07((cid:2)0.08to0.23) 57
834 Mechanical pulpworkers 1.2(0.7 to2.0) 0.06((cid:2)0.09to0.23) 57
106 Otheradministrators and executiveofficials, 1.2(0.7 to2.1) 0.05((cid:2)0.09to0.22) 54
municipal administration
Yz Occupationswith <40subjects 1.1(1.0 to1.3) 1.48((cid:2)0.31to3.04) 2519
774 Constructioncarpenters andworkers 1.1(0.9 to1.4) 0.53((cid:2)0.28to1.28) 911
882 Warehouseworkers 1.1(0.9 to1.5) 0.16((cid:2)0.22to0.53) 290
441 Forestry workersand loggers 1.1(0.8 to1.5) 0.12((cid:2)0.24to0.49) 276
641 Bus drivers 1.1(0.8 to1.7) 0.08((cid:2)0.15to0.32) 133
302 Working proprietors, retail trade 1.1(0.8 to1.6) 0.06((cid:2)0.19to0.32) 131
0X2 Social workers 1.1(0.7 to1.8) 0.05((cid:2)0.13to0.25) 92
822 Bakers andpastry cooks 1.1(0.7 to1.9) 0.05((cid:2)0.12to0.22) 78
612 Able andordinaryseamen 1.1(0.6 to2.1) 0.03((cid:2)0.11to0.17) 49
Continued
6 EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512
Occupation and the risk of tinnitus
Table4 Continued
Samplesize
Nordic Classificationof Occupational Codes PR(95%CI) AF (%)y(95%CI) (23374)
119 Othersin 11, administrationof private 1.1(0.6 to1.9) 0.02((cid:2)0.13to0.18) 61
enterprises and organisations.
404 Managers and supervisors(farms) 1.1(0.6 to1.8) 0.02((cid:2)0.14to0.20) 84
031 Otherphysicians 1.1(0.6 to2.0) 0.01((cid:2)0.12to0.15) 49
793 Cement finishers,excavators,etc 1.0(0.8 to1.3) 0.04((cid:2)0.48to0.58) 509
104 Otheradministration governmental 1.0(0.6 to1.7) 0.01((cid:2)0.15to0.20) 80
servantsdlocalstate administration
0X1 Auditors 1.0(0.6 to1.8) 0.01((cid:2)0.14to0.17) 64
06 Pedagogical work 1.0(ref) 0.00(ref) 1095
764 Installers, fitters, repairmen (radio,TV, 1.0(0.6 to1.6) 0.00((cid:2)0.20to0.21) 123
phone, telegraph)
0X9 Othersin technical,physicalscience, 1.0(0.5 to2.1) (cid:2)0.01((cid:2)0.12to0.14) 48
humanistic, artistic work
311 Salesmen ofinsurance 1.0(0.5 to1.9) (cid:2)0.01((cid:2)0.13to0.12) 45
911 Housekeepers, etc (notprivate orpublic 1.0(0.5 to1.8) (cid:2)0.01((cid:2)0.14to0.13) 51
service)
931 Janitors,vergers, etc 1.0(0.7 to1.2) (cid:2)0.07((cid:2)0.49to0.32) 333
644 Lorry and vandrivers 1.0(0.8 to1.2) (cid:2)0.10((cid:2)0.75to0.55) 782
024 Silviculturists andforestry consultants 0.9(0.5 to1.7) (cid:2)0.02((cid:2)0.16to0.12) 61
681 Postmen 0.9(0.5 to1.6) (cid:2)0.03((cid:2)0.18to0.14) 74
76 Electrical work 0.9(0.4 to1.8) (cid:2)0.03((cid:2)0.14to0.10) 48
671 Localpostmasters, postalassistance 0.9(0.6 to1.5) (cid:2)0.03((cid:2)0.22to0.17) 117
A20 Non-commissioned officers and subalterns 0.9(0.6 to1.4) (cid:2)0.04((cid:2)0.30to0.22) 164
643 Taxi drivers 0.9(0.5 to1.5) (cid:2)0.04((cid:2)0.19to0.11) 70
791 Masons,bricklayers andplasterers 0.9(0.5 to1.4) (cid:2)0.05((cid:2)0.24to0.14) 101
333 Shopassistants 0.9(0.6 to1.2) (cid:2)0.17((cid:2)0.53to0.22) 373
663 Railway supervisors 0.8(0.4 to1.6) (cid:2)0.04((cid:2)0.17to0.09) 43
851 Concrete productmakers, etc 0.8(0.5 to1.5) (cid:2)0.05((cid:2)0.20to0.11) 80
201 Accountants and bookkeepers 0.8(0.5 to1.3) (cid:2)0.08((cid:2)0.28to0.11) 119
903 Policemen anddetectives 0.8(0.5 to1.3) (cid:2)0.10((cid:2)0.29to0.10) 126
777 Wood workingmachine settersand 0.8(0.5 to1.2) (cid:2)0.13((cid:2)0.36to0.09) 159
operators
332 Shopmanagers 0.8(0.6 to1.1) (cid:2)0.20((cid:2)0.49to0.09) 236
412 Livestockworkers (general) 0.8(0.6 to1.1) (cid:2)0.23((cid:2)0.59to0.13) 436
411 Farm helpers(general) 0.8(0.6 to1.0) (cid:2)0.35((cid:2)0.76to0.06) 429
401 General farmers,livestock farmers 0.8(0.7 to1.0) (cid:2)2.20((cid:2)4.33to(cid:2)0.42) 2763
(working onownbehalf)
095 Editors andjournalists, etc 0.7(0.3 to1.6) (cid:2)0.06((cid:2)0.17to0.06) 44
297 Real estatemanagers, store-room 0.7(0.4 to1.4) (cid:2)0.07((cid:2)0.21to0.08) 64
keepers,etc
103 Leadingadministrators and executive 0.7(0.3 to1.4) (cid:2)0.09((cid:2)0.22to0.04) 57
officialsdlocalstate administration
403 Gardeners, horticulturalfarmersand 0.6(0.2 to1.5) (cid:2)0.08((cid:2)0.17to0.02) 43
fruitgrowers
021 Veterinarians 0.5(0.2 to1.5) (cid:2)0.08((cid:2)0.17to0.02) 40
432 Fish hatchers 0.5(0.2 to1.4) (cid:2)0.09((cid:2)0.19to0.02) 61
023 Agronomists and horticulturists, 0.5(0.2 to1.1) (cid:2)0.15((cid:2)0.28to0.00) 69
agricultural consultants
292 Clerks(bank) 0.4(0.2 to1.2) (cid:2)0.13((cid:2)0.23to0.00) 67
699 Othersin 69other transportand 0.3(0.1 to1.0) (cid:2)0.15((cid:2)0.22to0.00) 40
communication work
Occupationalgroupsaresortedbyprevalenceratio.
*Inrelationtothereferenceoccupationgroup06‘pedagogicalwork’.
yAFestimatedbybootstrapwith1000replications.
zOccupationswith<40arecollapsedintoonegroup.
AF,attributablefraction;PR,prevalenceratio.
EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512 7
Occupation and the risk of tinnitus
Table5 Predictedage-adjustedPR and AFsof tinnitusamong women*
Sample
Nordic Classificationof Occupational Codes PR(95%CI) AF (%)y(95%CI) size(26574)
013 Laboratory assistants 1.9(1.1 to3.3) 0.18(0.01 to0.41) 73
294 Clerks(public healthinsurance) 1.6(0.9 to3.0) 0.12((cid:2)0.05to0.32) 68
681 Postmen 1.6(0.8 to3.4) 0.07((cid:2)0.05to0.23) 41
X Occupationnot reported 1.5(1.3 to1.8) 11.30(6.96 to15.62) 7946
915 Housekeepers (publicservice) 1.5(1.2 to1.8) 1.42(0.59 to2.30) 842
049 Others in04nursing care 1.5(1.1 to2.0) 0.52(0.11 to0.96) 355
921 Headwaiters, waiters 1.5(1.1 to2.1) 0.43(0.05 to0.83) 295
914 Housekeepers, maids (privateservice) 1.5(1.0 to2.2) 0.29((cid:2)0.01to0.63) 212
911 Housekeepers, etc (notprivate or 1.5(1.0 to2.2) 0.20((cid:2)0.07to0.50) 129
public service)
861 Packers, labellersand relatedworkers 1.5(0.9 to2.4) 0.18((cid:2)0.06to0.44) 111
853 Plastic productmakers 1.4(0.8 to2.4) 0.11((cid:2)0.09to0.35) 102
0X3 Librarians, archivists andscientific 1.4(0.7 to2.8) 0.06((cid:2)0.08to0.23) 47
personnel in museums
003 Otherengineers, engineer technicians, 1.4(0.6 to3.3) 0.04((cid:2)0.07to0.18) 48
industrial designers,draughtsman
932 Char workers andcleaners 1.3(1.1 to1.6) 2.03(0.64 to3.43) 1888
913 Kitchen assistants 1.3(0.9 to1.7) 0.32((cid:2)0.15to0.80) 404
716 Sewers andembroiderers(textileproducts, 1.3(0.7 to2.2) 0.09((cid:2)0.11to0.29) 94
leather garments)
825 Canning andother preservationworkers 1.3(0.7 to2.5) 0.06((cid:2)0.09to0.25) 80
104 Otheradministration governmental 1.3(0.7 to2.6) 0.06((cid:2)0.09to0.23) 58
servantsdlocalstate administration
822 Bakers and pastrycooks 1.3(0.6 to2.6) 0.04((cid:2)0.10to0.18) 45
Yz Occupationswith <40subjects 1.2(1.0 to1.5) 0.82((cid:2)0.15to1.87) 1449
045 Otherpractical nurses 1.2(0.9 to1.5) 0.60((cid:2)0.15to1.35) 1063
411 Farm helpers(general) 1.2(0.9 to1.5) 0.53((cid:2)0.23to1.29) 843
912 Cooks 1.2(0.9 to1.8) 0.20((cid:2)0.15to0.57) 247
919 Others in91public safety andprotection 1.2(0.8 to1.8) 0.17((cid:2)0.17to0.53) 310
work
671 Local postmasters,postalassistance 1.2(0.8 to1.9) 0.12((cid:2)0.17to0.40) 203
941 Barbers,hairdressers andbeauticians 1.2(0.8 to1.9) 0.10((cid:2)0.16to0.37) 186
292 Clerks(bank) 1.2(0.7 to1.9) 0.09((cid:2)0.15to0.37) 202
413 Nursery workersand gardeners 1.2(0.7 to2.1) 0.08((cid:2)0.14to0.30) 106
046 Dental assistance 1.2(0.7 to2.2) 0.07((cid:2)0.10to0.26) 93
675 Telegraph dispatchers 1.2(0.6 to2.4) 0.05((cid:2)0.10to0.22) 62
059 Others in05other professional health 1.2(0.5 to2.7) 0.02((cid:2)0.09to0.17) 53
and medical work
333 Shopassistants 1.1(0.9 to1.4) 0.59((cid:2)0.73to1.91) 2042
401 General farmers, livestockfarmers 1.1(0.8 to1.4) 0.24((cid:2)0.47to1.05) 905
(working onownbehalf)
04 Nursing care 1.1(0.7 to1.8) 0.04((cid:2)0.18to0.30) 120
211 Secretaries and stenographers 1.1(0.6 to1.8) 0.02((cid:2)0.20to0.29) 161
302 Working proprietors, retail trade 1.1(0.6 to1.9) 0.02((cid:2)0.16to0.23) 94
043 Practical nursesin psychiatricinstitutions 1.1(0.5 to2.4) 0.02((cid:2)0.12to0.18) 67
111 Directors, managers, andworking proprietors 1.1(0.5 to2.6) 0.02((cid:2)0.10to0.14) 42
412 Livestockworkers (general) 1.0(0.8 to1.4) 0.05((cid:2)0.51to0.60) 439
299 Others in29other clerical work 1.0(0.8 to1.3) 0.03((cid:2)1.05to1.11) 1636
673 Telephone switchboard operators 1.0(0.5 to2.2) 0.01((cid:2)0.12to0.16) 47
(public service)
951 Laundryand dry-cleaningworkers 1.0(0.5 to1.9) 0.00((cid:2)0.17to0.19) 81
06 Pedagogical work 1.0(ref) 0.00(ref) 1299
0X2 Social workers 1.0(0.6 to1.5) (cid:2)0.01((cid:2)0.29to0.29) 263
922 Otherwaiting personnel 1.0(0.6 to1.7) (cid:2)0.01((cid:2)0.22to0.21) 136
201 Accountants and book-keepers 0.9(0.5 to1.8) (cid:2)0.02((cid:2)0.20to0.19) 121
674 Telephone switchboard operators 0.9(0.3 to2.3) (cid:2)0.02((cid:2)0.11to0.11) 48
(private exchange)
41 Farm work andlivestockwork 0.9(0.4 to2.0) (cid:2)0.02((cid:2)0.16to0.14) 56
Continued
8 EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512
Occupation and the risk of tinnitus
Table5 Continued
Sample
Nordic Classificationof Occupational Codes PR(95%CI) AF (%)y(95%CI) size(26574)
332 Shopmanagers 0.9(0.4 to1.7) (cid:2)0.04((cid:2)0.19to0.14) 89
041 Professional nurses 0.9(0.7 to1.3) (cid:2)0.13((cid:2)0.69to0.46) 745
052 Physio- andoccupational therapists 0.8(0.4 to1.5) (cid:2)0.07((cid:2)0.26to0.15) 134
203 Othercashiers 0.8(0.5 to1.4) (cid:2)0.09((cid:2)0.31to0.15) 176
916 Concierge (hotels) 0.7(0.2 to2.0) (cid:2)0.05((cid:2)0.13to0.07) 51
047 Nursemaids in hospitalsand other 0.3(0.1 to1.0) (cid:2)0.20((cid:2)0.31to0.00) 119
institutions
769 Others in76electrical work 0.2(0.0 to1.6) (cid:2)0.10((cid:2)0.14to0.00) 48
Occupationalgroupsaresortedbyprevalenceratio.
*Inrelationtothereferenceoccupationgroup06‘pedagogicalwork’.
yAFestimatedbybootstrapwith1000replications.
zOccupationswith<40arecollapsedintoonegroup.
AF,attributablefraction;PR,prevalenceratio.
in regard to the effects of occupation on hearing loss hear speech’) compared with unexposed participants.7
based on this same study sample.19 Among 3753 older adults in Beaver Dam, Wisconsin, no
In women, occupations with the highest risk for associationwasfoundbetweenmajoroccupation,history
tinnitus were not typically noisy ones, and the AF was of occupational noise exposure or hunting history,
determined mainly by the group of occupationally with the likelihood of having tinnitus (prevalence) or
inactive.Thisisdifferentfromtheanalysisofoccupation developing tinnitus (incidence).5
effects on hearing loss19dthere was no increased risk of We found adding hearing loss as a predictor did not
hearing loss in occupationally inactive women. decrease the effect of occupation; when controlling for
Only afewprevious studies have reportedoccupation- some other risk factors, a small increase in occupational
specific tinnitus prevalence. The odds for tinnitus effect was found. Thus, hearing loss does not seem to
according to different occupational groups were mediate,butrathersuppress,theeffectofoccupationon
reported based on the 1994e1995 US National Health tinnitus. This points to the importance of occupational
Interview Survey Disability Supplement data set.4 This risk factors not related to hearing loss, such as work
studyshowedamarginalelevationintinnitusprevalence demands, level of control, social support and other
in skilled and unskilled workers compared with profes- psychosocial factors.
sionals (OR 1.18; 95% CI 1.00 to 1.39). This study also Thenegativeeffectforwomenofbeingoccupationally
found a reduced prevalence in the two occupational non-active is partly in agreement with results from
groups of managerial or administrative (OR 0.82; 95% a study of self-reported hearing problems in a Swedish
CI0.68to0.99)andtechnicalorsales(OR0.83;95%CI working and non-working population.11 A higher prev-
0.70 to 0.98). alenceoffrequentorconstanttinnituswasfoundamong
Some studies evaluating the effect of self-reported non-workers(15%)thanworkers(11%).Whilewefound
occupational noise exposure found results comparable this effect mainly among women, the previous study
to the present study. In a British general population reported highereffects formen(25%and15% fornon-
sampleof12907subjects,9age-adjustedPRsfor working workersandworkers,respectively)thanforwomen(10%
in noisy environment for >10years was estimated to be and 8%). These numbers were not age-adjusted,
2.6 (95% CI 2.0 to 3.4) in men and 1.9 (95% CI 1.0 to however, and the age distribution of workers was quite
3.7) in women in comparison with those with no occu- different from non-workers.
pational exposure to noise. The overall tinnitus preva- Receiving social security or disability pension was
lence of the sample was 6% in men and 3% in women. associated with the highest risk of tinnitus in the group
Previous British data also showed tinnitus to be about of occupationally non-active women. It may be that
twiceascommon inthosewithahistoryofoccupational reduced functional ability or poor general health
exposure to noise.28 decreases the ability to cope with tinnitus, although the
In an earlier analysis of tinnitus in the NTHLS, being causaldirectionisnotclear,astinnitusmaywellbeapart
exposed to loud noise at work for >15hours per week ofthedisabilityinthefirstplace.Also,theincreasedrisk
resultedinanORof1.7(95%CI1.5to1.9)inmenand oftinnitusmayberelatedtopsychosocialfactors,suchas
1.6 (95% CI 1.4 to 1.9) in women compared with those loss of social status and self-esteem, social support,
who had not been exposed to loud noise at work.4 personal economy and lifestyle factors such as physical
Among 2015 older Australians, the RR of tinnitus was inactivity. Regardless of the underlying cause, the
1.4 (95% CI 1.1 to 1.7) for participants exposed to elevated occurrence of tinnitus among unemployed
‘tolerable noise’ and 1.5 (95% CI 1.11 to 2.1) for those women is hardly a real occupational effect, so the true
exposed tohighlevelsofoccupational noise(‘unable to fraction of tinnitus attributable to type of occupation is
EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512 9
Occupation and the risk of tinnitus
much lower than the 21.4% estimated by including CONCLUSIONS
unemployed women. This study found a moderate association between occu-
pation and tinnitus. Estimates of occupation-specific
Strengths and limitations tinnitus prevalence may help identify high-risk occupa-
The major advantages of the present study are the tions in which interventions are needed.
prospective design and that the study population is
representative of the general working population. A AcknowledgementsTheNord-TrøndelagHealthStudy(TheHUNTStudy)is
acollaborationbetweentheHUNTResearchCentre(FacultyofMedicine,the
substantial selection bias is unlikely since occupational
NorwegianUniversityofScienceandTechnology),theNord-Trøndelag
data were complete for all participants, and the partici- CountyCouncilandtheNorwegianInstituteofPublicHealth.
pation rate in this population survey was relatively high
FundingThisworkwassupportedbytheNationalInstituteonDeafnessand
(69% for the vast majority of the county). OtherCommunicationDisorders(NIDCD),NationalInstitutesofHealth(NIH)
A recent survey of sufferers from tinnitus showed that viaresearchcontractNo.N01-DC-6-2104.
they report excessive noise in the work environment as CompetinginterestsNone.
the single most important factor for developing
PatientconsentObtained.
tinnitus.29 Thus, there might be a serious problem with
EthicsapprovalThestudywasapprovedbytheNorwegianRegional
recall bias that tends to exaggerate an association when
CommitteeofMedicalEthicsandbytheNorwegianDataInspectorate.
boththeexposureandtheoutcomeareself-reported.In
the present study, occupation data were obtained from ContributorsThefollowingauthorshavecontributedtothepaper:BE:Design,
analysisandinterpretationofdata,writing.NHK:Preparingdata,draftingthe
highly valid prospective registry data, thus, we consider
manuscript,criticalrevisionofmanuscript.EK:Preparingdata,draftingthe
recall bias on this factor not to be a problem.
manuscript,criticalrevisionofmanuscript.HJH:Obtainingfunding,data
Testeretest of the question ‘are you bothered by acquisition,draftingthemanuscript,criticalrevisionofmanuscript.KT:
tinnitus?’ indicated a relatively high reliability. Tinnitus Conceptionanddesign,obtainingfunding,dataacquisition,draftingthe
is a personal, subjective experience that cannot be manuscript,criticalrevisionofmanuscript.
measured objectively and is thus per definition ProvenanceandpeerreviewNotcommissioned;externallypeerreviewed.
described by self report. The clinical validity of the
DatasharingstatementsThereisnoadditionaldataavailable.
measure is unknown as we have no data on the correla-
tion between being bothered by tinnitus and seeking
REFERENCES
medical help. Subjective need for treatment has been
1. ErlandssonSI.Psychologicalprofilesoftinnituspatients.In:TylerRS,
reported in as many as 2/3 of the subjects who found ed.TinnitusHandbook.SanDiego:ThomsonLearning2000:25e57.
them self-suffering from tinnitus often or always.30 The 2. HenryJA,DennisKC,SchechterMA.Generalreviewoftinnitus:
prevalence,mechanisms,effects,andmanagement.JSpeechLang
prevalence of suffering from tinnitus in that study was HearRes2005;48:1204e35.
reported to be 14%, which is very similar to the preva- 3. ShargorodskyJ,CurhanGC,FarwellWR.Prevalenceand
characteristicsoftinnitusamongUSadults.AmJMed
lence of bothered by tinnitus in our study. But probably 2010;123:711e18.
only a few per cent of the subjects with bothersome 4. HoffmanHJ,ReedGW.Epidemiologyoftinnitus.In:SnowJrJB,ed.
Tinnitus:TheoryandManagement.HamiltonOntario:BCDeckerInc,
tinnitus have actually been seeking help for this condi-
2004:16e41.
tion, a number that will depend on factors such as 5. NondahlDM,CruickshanksKJ,WileyTL,etal.Prevalenceand
whethertreatmentsaredeterminedtobeeffectiveandif 5-yearincidenceoftinnitusamongolderadults:theepidemiologyof
hearinglossstudy.JAmAcadAudiol2002;13:323e31.
they are known or available to the broader public.
6. KonigO,SchaetteR,KempterR,etal.Courseofhearinglossand
A weakness of the study is the lack of information occurrenceoftinnitus.HearRes2006;221:59e64.
7. SindhusakeD,GoldingM,NewallP,etal.Riskfactorsfortinnitusin
about the duration of employment and exposure.
apopulationofolderadults:theBlueMountainsHearingStudy.Ear
However, the estimated average tinnitus prevalence for Hear2003;24:501e7.
each occupation group applies to workers whose age is 8. SulkowskiW,KowalskaS,LipowczanA,etal.Tinnitusandimpulse
noise-inducedhearinglossindrop-forgeoperators.IntJOccupMed
equal to the sample mean. Accordingly, we have essen- EnvironHealth1999;12:177e82.
tially adjusted for exposure duration. 9. PalmerKT,GriffinMJ,SyddallHE,etal.Occupationalexposureto
noiseandtheattributableburdenofhearingdifficultiesinGreat
The Nordic Classification of Occupations does not Britain.OccupEnvironMed2002;59:634e9.
classifyoccupationsonthebasisofnoiseexposurelevels 10. GopinathB,McMahonCM,RochtchinaE,etal.Riskfactorsand
or other risk factors for tinnitus, but according to the impactsofincidenttinnitusinolderadults.AnnEpidemiol
2010;20:129e35.
tasks and duties undertaken in the job. Heterogeneity 11. HassonD,TheorellT,WesterlundH,etal.Prevalenceand
regardingnoiseandotherexposurewithinoccupational characteristicsofhearingproblemsinaworkingandnon-working
Swedishpopulation.JEpidemiolCommunityHealth
categories implies that occupation, as an explanatory 2010;64:453e60.
variable, does not capture all effects of occupational 12. GriestSE,BishopPM.Tinnitusasanearlyindicatorofpermanent
hearingloss.A15yearlongitudinalstudyofnoiseexposedworkers.
exposures on tinnitus. Our results showed that adding
AAOHNJ1998;46:325e9.
information on self-reported occupational noise expo- 13. RubakT,KockS,Koefoed-NielsenB,etal.Theriskoftinnitus
sure improved the prediction of tinnitus somewhat. followingoccupationalnoiseexposureinworkerswithhearinglossor
normalhearing.IntJAudiol2008;47:109e14.
Selection for good hearing function in some occupa-
14. SaundersJC.Theroleofcentralnervoussystemplasticityintinnitus.
tions could in principle bias the results, but we think it JCommunDisord2007;40:313e34.
15. LinYH,ChenCY,LuSY.Physicaldiscomfortandpsychosocialjob
unlikely that this type of selection had a major effect on
stressamongmaleandfemaleoperatorsattelecommunicationcall
our results. centersinTaiwan.ApplErgon2009;40:561e8.
10 EngdahlB,KrogNH,KvestadE,etal.BMJOpen2012;2:e000512.doi:10.1136/bmjopen-2011-000512