7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 1/8

doi: 10.1136/oem.2010.06406320, 2011

2011 68: 895-901 originally published online AprilOccup Environ Med et al.Lau Caspar Thygesen, Esben Meulengracht Flachs, Kirsten Hanehøj,mercuryassistants occupationally exposed torenal diseases among dentists and dentalHospital admissions for neurological and

 http://oem.bmj.com/content/68/12/895.full.html

Updated information and services can be found at:

These include: 

References http://oem.bmj.com/content/68/12/895.full.html#ref-list-1

This article cites 36 articles, 4 of which can be accessed free at:

serviceEmail alerting

the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in

CollectionsTopic

(682 articles)Other exposures 

Articles on similar topics can be found in the following collections

Notes

 http://group.bmj.com/group/rights-licensing/permissions

To request permissions go to:

 http://journals.bmj.com/cgi/reprintform

To order reprints go to:

 http://group.bmj.com/subscribe/

To subscribe to BMJ go to:

 group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from 

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 2/8

ORIGINAL ARTICLE

Hospital admissions for neurological and renaldiseases among dentists and dental assistantsoccupationally exposed to mercury

Lau Caspar Thygesen,1 Esben Meulengracht Flachs,1 Kirsten Hanehøj,1 Helge Kjuus,2

Knud Juel1

ABSTRACTObjectives For many years an amalgam containingmetallic mercury, which has been associated withneurological and renal diseases, has been used indentistry. In this nationwide study we compared hospitaladmissions due to neurological and renal diseasesamong dentists and dental assistants to admissions incontrols.Methods This register-based cohort study included all

Danish workers employed in dental clinics, generalpractitioners’ clinics or lawyers’ offices between 1964and 2006. We compared dentists with generalpractitioners and lawyers, and dental assistants withmedical secretaries, nurses and legal secretaries. Wealso compared dentists and dental assistants employedduring periods with high occupational mercury exposurewith dentists and dental assistants employed duringperiods with less mercury exposure. We followed allsubjects in a nationwide register of hospital admissions.We analysed risk of neurological diseases, Parkinson’sdisease and renal diseases using a Cox regressionmodel.

Results The cohort consisted of 122 481 workersincluding 5371 dentists and 33 858 dental assistants. Forneurological diseases, no association was observedfor dental assistants, while for dentists an increasing riskfor periods with less mercury exposure was observed.Among dental assistants, a negative associationbetween employment length and risk of neurologicaldisease was observed. Admissions for renal diseaseamong dental assistants were increased during periodswith less mercury exposure compared with controls. Fordentists a non-significant increased risk was observedbetween employment length and renal disease risk.Conclusions Our nationwide study does not indicatethat occupational exposure to mercury increases the risk

of hospital admissions for neurological, Parkinson’s orrenal diseases.

INTRODUCTIONFor many years an amalgam containing metallicmercury has been used by dentists in Denmark.Since 1963 the handling of amalgam fillings hasgradually been more strictly regulated. Previousstudies have shown that exposure to mercury indentistry up to 1970 resulted in average urinary values of 125e200 nmol/l, with individual values

ranging from 0 to 500 nmol/l. Since then, average

urinary values have slowly decreased to about25 nmol/l, with individual values rarely exceeding100 nmol/l.1 There is no evidence of any differencesin urine mercury levels between dentists and dentalassistants.2

Previous studies of the effects of mercury in otheroccupational groups have shown that acutemercury intoxication is characterised by a numberof neurological and neuropsychological symptomsand indicate that urine mercury above 600 nmol/l isassociated with impairment in neuropsychologicaltests.3 The nervous system and the kidneys areespecially sensitive to mercury toxicity.4e6   Very few genuine mercury poisonings have beenreported in dentistry and many of the symptoms of intoxication have been unspecific.7 8 However,some studies have reported that exposure tomercury associated with urine mercury levels of approximately 150 nmol/l results in impairedresults in neuropsychological and psychomotortests,3 9e13 while other studies have shown noeffect14e16  or only slight effects.17 In studies of pastexposure, the effects are smaller and decrease withtime from cessation of exposure,3 although subjectswith previous mercury intoxication still havedecreased performance several years later.18 19

Slight neurophysiological and neuropsychological

effects have been found among chloralkali workers

1National Institute of PublicHealth, University of SouthernDenmark, Copenhagen,Denmark2National Institute ofOccupational Health, Oslo,Norway

Correspondence toDr Lau Caspar Thygesen,National Institute of PublicHealth, University of SouthernDenmark, Øster Farimagsgade5 A, 2nd floor, DK-1353Copenhagen K, Denmark;lct@niph.dk

Accepted 23 March 2011Published Online First20 April 2011

What this paper adds

< Previous studies have usually been cross-sectional and in several the selection ofcontrol groups may have biased the results.

< Our nationwide study of dental workersbetween 1964 and 2006 shows no increasedrisk of neurological, Parkinson’s or renaldiseases among dentists and dental assistantscompared with controls and no increased riskfor high occupational mercury exposurecompared with lower occupational mercuryexposure.

< We conclude that our nationwide study does notsupport the suggestion that occupationalmercury exposure from dental amalgamincreases the risk of hospital admissions forneurological, Parkinson’s or renal diseases.

Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063 895

Workplace

 group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from  group.bmj.comon October 19, 2013 - Published by oem.bmj.comDownloaded from 

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 3/8

with mean urine mercury levels of 500 nmol/l examined12 years after exposure.20e22

Two Norwegian questionnaire studies of dental personnelshowed more neurological and general symptoms in dentalassistants compared with controls, while dentists had fewersymptoms compared with controls 20e30 years after expo-sure.23 24 These studies stimulated intense debate in Denmarkon whether Danish dental assistants have increased morbidity 

due to past mercury exposure.The aim of this study was to investigate whether dentists and

dental assistants have higher hospital admission rates forneurological, Parkinson’s and renal diseases compared withcontrol groups based on the nationwide registers of employmentgroups which started in 1964.

MATERIAL AND METHODSDefining the cohortThis was a historical register-based cohort study. In Denmark,morbidity can be studied in register-based studies, becausenationwide registration systems with linkage between registersat the individual level via the unique personal identificationnumber (CPR number) are available for all permanent Danishresidents. We obtained information from several registers whenconstructing the cohort.

The Supplementary Pension Fund Register (ATP) was estab-lished in 1964 as a compulsory supplementary pension fund(Arbejdsmarkedets Tillægspension) for all workers aged18e66 years employed in Denmark for more than 8 h per week.

The employer pays contributions into the employee’s account in ATP four times each year according to the number of hoursworked. For every employment the following information isregistered: the employee’s identification number (CPR number),the employer ’s identification number (including industrial clas-sification code) and the percentage of employment for eachquarter of the year. This information is retained in ATP evenafter a worker has left the scheme, retired, emigrated or died.

The Register of Authorisation of Health-Care Personnel(RAHP) contains information on all workers with Danishauthorisation to work as healthcare personnel includingdentists, general practitioners and nurses. This register containsinformation back to 1982.

The Danish Civil Registration System (CRS) contains infor-mation on mortality, emigration, immigration and disappear-ance for all Danish residents since 1968. Furthermore, people canvoluntarily register information on their most important jobtitle with the CRS. This information on job title is not dated.

The Labour Market Module (LMM) at Statistics Denmarkincludes information on occupational classification for all Danesfor November of each year since 1980. The classification of occupational groups does not include a category for dentalassistants.

In the cohort we included workers who had paid intoa supplementary pension between 1 April 1964 and 31December 2006. We included employees in dental clinics andcompared their morbidity with two control groups (employeesin a general practitioner ’s clinic and employees in a lawyer ’s

Figure 1 Presentation of the cohort and construction of occupational groups, Denmark, 1964e2006.

896 Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063

Workplace

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 4/8

of fice). We obtained information from ATP on all workersemployed in a dental clinic (n¼41 328), a general practitioner’sclinic (n¼38 584) or a lawyer’s of fice (n¼53398) (figure 1). If a person had worked in a dental clinic and a general practi-tioner’s clinic (n¼1198) or a lawyer ’s of fice (n¼492) or had hadall three jobs (n¼28), we only included the employment infor-mation from the dental clinic.

The three employment groups were divided into seven occu-

pational groups: dentists, dental assistants, general practitioners,nurses, medical secretaries, lawyers and legal secretaries. Thestudy thereby included two occupational groups exposed tomercury (dentists and dental assistants) and five control groups(figure 1). Information on occupational groups was not availablefrom ATP, but was obtained by linkage with RAHP (dentists,general practitioners and nurses) and the CRS.

Information on job title from the CRS was used for those notauthorised to practice as dentists, general practitioners ornurses. As this information was not necessarily obtained at thesame time as data from ATP, employees with job titles notassociated with the three employment groups had to be cate-gorised. We used the educational level of the specific job titlesobtained from the CRS. Workers with job titles requiring four

 years of education after elementary school were considered tohave low and medium education, while those in jobs requiringfive or more years of education after elementary school wereconsidered to be highly educated. Workers in dental clinics withunknown, low or medium education were categorised to thedental assistant group. Highly educated workers were excluded(see figure 1 for the numbers of workers in each category). Thesame procedure was used for employees in a general practi-tioner ’s clinic.

To determine whether this method resulted in unbiasedoccupational groups, we assessed consistency between ouroccupational groups and the LMM. The distribution of occu-pational groups in the LMM was similar for our categories‘unknown

’,

‘low and medium educated

’and

‘dental assistant

’,

which indicated that our method of categorising workers wasunbiased. The same was observed for ‘medical secretary ’.

For workers in lawyers’ of fices, the group with no informationon job title in the CRS was divided into lawyer and legalsecretary in that men were classified as lawyers and women aslegal secretaries since linkage with Statistics Denmark showedconsiderable sex differences consistent with this classification.

For dental assistants, 33 858 workers (95.0%) were women.The same proportions were observed for nurses (99.4%), medicalsecretaries (92.2%) and legal secretaries (94.2%). Therefore, in

the analyses we only included women from these groups,thereby excluding 6053 men. For dentists, general practitionersand lawyers, the proportions of women were lower (see table 1)and we therefore included both sexes.

Finally, we excluded 70 subjects who emigrated fromDenmark before they were employed. The final cohort consistedof 122 481 persons employed in seven occupational groups (firstrow, table 1).

ExposureUsing information from ATP on the annual percentage of employment for all years, we constructed a first date of employment and a date of termination of employment. Wecompared these dates with information from the CRS on vitaland emigration status to determine the first and last day of employment. Dentists, general practitioners and lawyersnormally changed from being employees early in their careers tobeing self-employment later on. Since ATP only contains infor-mation on employees, the date of the subject’s 66th birthday orthe date of termination after age 60 if available was consideredto be the last date of employment.

In analyses where dental assistants and dentists werecompared with the control groups, we used the first and lastdate of employment to construct the period of employment,which we then used as the measurement for exposure. Sinceoccupational mercury exposure decreased with calendar time,we included the period effect using four categories (1964e1969,

1970e

1979, 1980e

1989 and 1990e

2006). A proxy measure for excess cumulative mercury exposure

among dental personnel compared with the general population

Table 1 Characteristics of the cohort, Denmark, 1964e2006

Dental clinics General practitioner’s clinics Lawyer’s offices

Dentist Assistant GP Nurse Secretary Lawyer Secretary  

Persons employed 1964e2006 5371 33 858 6154 5872 22 785 11 433 37 717

Persons employed 1964e1979 2033 12 065 778 1433 7626 3787 15659

Persons employed 1980e1994 3493 17 514 3018 2447 12 450 5527 18 848

Persons employed 1995e2006 3448 16789 3667 3805 11 537 5940 16 393

Female (n) 2693 33 858 2455 5872 22 785 1916 37 717

Female (%) 50.1 100.0 39.9 100.0 100.0 16.8 100.0Age at employment (mean) 30.2 23.8 37.3 37.3 35.5 27.7 27.6

Neurological disease cases (n) 191 1227 148 195 1085 355 1540

Neurological disease IR* 15.9 20.6 22.6 18.3 22.2 29.1 20.3

Parkinson’s disease cases (n) 8 14 6 1 34 16 53

Parkinson’s disease IR* 1.2 0.6 1.6 0.2 0.8 2.5 1.1

Renal disease cases (n) 122 625 86 68 395 237 660

Renal disease IR* 5.7 10.1 6.7 6.4 8.4 5.3 8.9

Deaths by 31 Dec 2006 (n) 304 1406 266 290 2254 1062 3537

Emigrations by 31 Dec 2006 (n) 277 832 123 66 376 397 1017

Disappearances by 31 Dec 2006 (n) 3 4 0 0 2 3 7

Follow-up (years) 112 462 682 682 86 141 89 476 425 499 207 059 794 055

GP, general practitioner; IR, incidence rate.*Age-standardised (European standard population) incidence rate per 100 000 person-years. To make the rates comparable we only included women in all occupational groups.

Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063 897

Workplace

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 5/8

was calculated as the weighted amount of occupational mercury exposure during their working career multiplied by the annualpercentage of employment. The weights were based on urinesamples from Norwegian dental personnel collected since thelate 1950s.2 Mercury exposure was highest in the 1960s (weightof 4.8 for each year) and decreased until 1989. The weights were2.2 for each year in 1970e1974, 1.2 in 1975e1979, 0.8 in1980e1984 and 0.2 in 1985e1989. The weight from 1990 and

onwards was set to 0 as dental personnel were not thought to beexposed to excess mercury after this date. The maximumweighted exposure was 50.8, which was given to workers withfull time employment in a dental clinic for the entire period1964e1989.

In analyses of cumulative exposure, weighted exposure wasdivided into five categories (0, 0.1e0.9, 1.0e4.7, 4.8e14.9 and15.0e50.8).

OutcomesThe Danish National Patient Register contains information onall hospital discharges, including diagnoses and surgical proce-dures, in Denmark since 1977. This register contains informa-tion on mode and date of admission and discharge and ICDcodes for primary and secondary diagnoses for all hospitaladmissions.

 We followed the cohort of hospital admissions due to neuro-logical disease except eye and ear disease (ICD-8, 320e359;ICD-10, G00-G99), Parkinson’s disease (ICD-8, 342.9; ICD-10,G20) and renal disease except nephritis and obstructiveuropathy (ICD-8, 580e584 and 590e593; ICD-10, N00eN08and N14eN29). We included primary and secondary diagnosesin the analyses and we used the date of admission as date of diagnosis.

Statistical analyses We performed two types of analyses for each health outcome:

(1) a between-group analysis where the hospital admission ratesfor dentists and dental assistants were compared with controlgroups and (2) a within-group analysis where dentists anddental assistants with employment during periods with highmercury exposure were compared with dentists and dentalassistants employed during periods with less mercury exposure.

In the between-group analysis, we examined whether relativerates among dentists and dental assistants compared withcontrol groups were higher during periods with occupationalmercury exposure compared with periods with no excessoccupational mercury exposure.

In the within-group analyses, we evaluated the effect of cumulative weighted exposure among dental assistants anddentists on each health outcome.

 As ATP was established on 1 April 1964, information onemployment was not available before this date. To account forexposure before 1964, we adjusted for age in 1964 for thoseemployed in 1964 (4.2% of the cohort) using four categories:17e24 years (n¼2381), 25e39 years (n¼1458), 40e65 years(n¼1370) and not employed in 1964 (n¼117272). As a sensi-tivity analysis, we repeated the analyses excluding thoseemployed in 1964. We also adjusted for sex in the analyses of dentists, general practitioners and lawyers.

In the between-group analyses of dental assistants, wefurthermore adjusted for length of employment using fourcategories (1e154 days, 155e822 days, 823e2477 days and2478+ days). In the within-group analyses, we adjusted for

follow-up period usingfi

ve categories (1977e

1982, 1983e

1988,1989e1994, 1995e2000 and 2001e2006).

 A Cox regression analysis was performed with R softwarev 2.9.025 to estimate the HR of employment as a dental assistantor dentist compared with the control groups in the between-group analyses, and the HR of cumulative exposure amongdental assistants and dentists in the within-group analyses,while taking potential confounding variables into account. Wetested the between-group effect by conducting a test of inter-action between period and occupational group, that is, a test of 

whether the HR for the exposed occupational groups comparedwith the control groups varied by period. Because age wasa strong predictor of the diseases of interest, age was used as thetime scale with delayed entry, thereby adjusting for the effect of age on the risk estimates.

For the analyses of neurological and renal diseases, we relatedtime-dependent occupational mercury exposure to risk of disease the following year. For the analysis of Parkinson’sdisease, we included a latency period of 5 years between thetime-dependent exposure and disease. We thereby assumed thatParkinson’s disease occurring during the latency period was notrelated to the exposure. To evaluate the assumption of nolatency period for neurological and renal diseases, we repeatedthe analyses assuming latencies of 5 and 10 years. For Parkin-son’s disease too few cases were observed for such sensitivity analysis.

 We examined the proportional hazards assumption by testingwhether the correlation between the scaled Schoenfeld residualsand underlying time was different from zero and examined plotsof underlying time by Schoenfeld residuals. These tests and plotsdid not indicate violation of the assumption.

 We followed the cohort from their first employment or 1January 1977 to the date of disease diagnosis, date of death,emigration or disappearance, or end of follow-up (31 December2006), whichever came first. For each outcome, we followed thecohort to first admission for that specific disease.

RESULTSThe number of workers employed in each of the seven occupa-tional groups is shown in table 1. The proportion of women washighest for dentists and lowest for lawyers. The mean age atemployment was highest for general practitioners, nurses andmedical secretaries and lowest for dental assistants.

The numbers of cases and age-standardised incidence ratesshowed that dentists had a lower incidence of neurological andParkinson’s disease compared with the control groups (generalpractitioners and lawyers) (table 1). Dental assistants had ratescomparable with medical and legal secretaries for all diseases.Nurses had lower rates. By the end of follow-up, 304 dentistsand 1406 dental assistants had died.

Neurological disease Admission rates due to neurological disease among dentalassistants compared with the three control groups for the fourperiods did not show any consistent pattern (table 2). Tests fordifferent period effects were insignificant in all control groups(eg, p¼0.49 for dental assistants compared with medical secre-taries). For dentists there was a significant difference in periodeffects with increasing risk for dentists compared with generalpractitioners and lawyers over time (table 2).

 Among dental assistants, a negative association betweencumulative exposure and neurological disease was observed,although the association was not significant (p¼0.21) (table 3).

 Among dentists there was a non-signifi

cant U-shaped association(p¼0.51).

898 Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063

Workplace

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 6/8

 We repeated the analyses with 5 and 10 years of latency,which showed the same relationships as described in the mainanalyses with no latency period (not shown). We also repeatedthe analysis after excluding those employed in 1964, which alsodid not change the main results (not shown).

Parkinson’s disease

 Admission rates for Parkinson’s disease among dental assistants

compared with medical and legal secretaries did not show any consistent pattern and no significant difference for period effectswasobserved (table2). This analysis wasnot performedfor nursesas only one nurse was admitted to hospital due to Parkinson ’sdisease. There was no consistent pattern for dentists (table 2).

 Among dental assistants, a negative association was shownbetween cumulative exposure and risk of Parkinson’s disease,although the association was not significant (p¼0.45) (table 3).Too few cases were observed among dentists to performa meaningful analysis of the association between cumulativeexposure and Parkinson’s disease.

Renal diseaseThe admission rate among dental assistants compared with thethree control groups showed an increasing rate for dental assistantsduring the four periods compared with the three control groups,although the differences were not statistically significant (table 2).There was no significant difference in period effects for dentists.

Table 2 Between-group analyses of hospital admission rates among dental assistants and dentists compared with control groups, Denmark,employment 1964e2006, follow-up 1977e2006

DA/MS* DA/NU* DA/LS* DE/GPy DE/LAyHR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI)

Neurological disease

1964e1969 0.92 (0.76 to 1.12) 1.10 (0.76 to 1.60) 1.11 (0.96 to 1.30) 0.46 (0.23 to 0.91) 0.64 (0.44 to 0.93)

1970e1979 0.83 (0.73 to 0.95) 0.99 (0.77 to 1.28) 1.08 (0.95 to 1.22) 1.02 (0.70 to 1.49) 0.86 (0.65 to 1.14)

1980e1989 0.97 (0.82 to 1.15) 0.97 (0.70 to 1.32) 1.08 (0.92 to 1.26) 1.10 (0.73 to 1.66) 0.98 (0.67 to 1.43)

1990e

2006 0.85 (0.68 to 1.06) 0.99 (0.70 to 1.39) 1.22 (0.97 to 1.53) 1.50 (0.94 to 2.38) 1.16 (0.72 to 1.86)p Value 0.49 0.96 0.82 0.07 0.21

p Value (trend) 0.95 0.58 0.65 0.03 0.03

Parkinson’s disease

1964e1969 0.78 (0.26 to 2.31) e 0.74 (0.31 to 1.80 ) 0.40 (0.07 to 2.2 0) 1.00 (0.32 to 3.12)

1970e1979 0.83 (0.29 to 2.35) e 0.92 (0.33 to 2.56 ) 0.21 (0.02 to 2.0 4) 0.33 (0.04 to 3.03)

1980e1989 0.68 (0.14 to 3.25) e 0.67 (0.14 to 3.34 ) 2.84 (0.25 to 32.6) 2.29 (0.21 to 25.4)

1990e2006 0.96 (0.10 to 9.24) e e e e

p Value 1.00 0.57 0.43 0.35

p Value (trend) 0.96 0.55 0.28 0.61

Renal disease

1964e1969 1.04 (0.77 to 1.40) 0.90 (0.54 to 1.53) 1.24 (0.99 to 1.56) 1.27 (0.50 to 3.25) 1.38 (0.93 to 2.05)

1970e1979 1.19 (0.96 to 1.46) 1.36 (0.89 to 2.06) 1.18 (0.99 to 1.42) 1.27 (0.76 to 2.11) 0.89 (0.63 to 1.28)

1980e1989 1.06 (0.83 to 1.37) 1.61 (0.89 to 2.91) 1.19 (0.95 to 1.49) 0.88 (0.52 to 1.49) 0.88 (0.54 to 1.42)

1990e

2006 1.38 (0.96 to 1.98) 1.26 (0.71 to 2.23) 1.69 (1.21 to 2.36) 0.83 (0.39 to 1.76) 0.98 (0.46 to 2.09)p Value 0.58 0.52 0.28 0.68 0.37

p Value (trend) 0.45 0.40 0.29 0.26 0.22

DA, dental assistant; DE, dentist; GP, general practitioner; LA, lawyer; LS, legal secretary; MS, medical secretary; NU, nurse; e, too few cases in control group.*Cox regression model adjusted for age (underlying time scale), age in 1964 and length of employment. Only women were included.yCox regression model adjusted for age (underlying time scale), sex and age in 1964.

Table 3 Within-group analyses of cumulative exposure and hospital admission rates among dental assistants and dentists, Denmark, employment1964e2006, follow-up 1977e2006

Cumulative exposure

Neurological disease Parkinson’s disease Renal disease

Cases, n HR (95% CI) Cases, n HR (95% CI) Cases, n HR (95% CI)

Dental assistant*

0 163 0.96 (0.78 to 1.17) 1 1.68 (0.17 to 16.2) 94 1.05 (0.79 to 1.39)

0.1e

0.9 380 1 (ref) 5 1 (ref) 196 1 (ref)1.0e4.7 348 0.93 (0.80 to 1.08) 3 0.46 (0.10 to 2.01) 183 0.98 (0.80 to 1.21)

4.8e14.9 249 0.83 (0.70 to 0.99) 4 0.57 (0.13 to 2.43) 113 0.81 (0.64 to 1.04)

15.0e50.8 87 0.79 (0.61 to 1.02) 1 0.15 (0.01 to 1.70) 39 0.82 (0.56 to 1.20)

p Value 0.21 0.45 0.41

p Value (trend) 0.31 0.22 0.21

Dentisty

0 33 1.36 (0.82 to 2.26) 0 e 10 0.97 (0.43 to 2.18)

0.1e0.9 30 1 (ref) 2 e 16 1 (ref)

1.0e4.7 30 0.94 (0.56 to 1.58) 0 e 23 1.31 (0.68 to 2.51)

4.8e14.9 47 1.03 (0.63 to 1.68) 1 e 31 1.30 (0.68 to 2.48)

15.0e50.8 51 1.28 (0.76 to 2.16) 5 e 42 1.74 (0.88 to 3.43)

p Value 0.51 0.52

p Value (trend) 0.29 0.18

*Cox regression model adjusted for age (underlying time scale), age in 1964 and calendar time. Only women were included.yCox regression model adjusted for age (underlying time scale), sex, age in 1964 and calendar time.

Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063 899

Workplace

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 7/8

There was no association between cumulative exposureand risk of renal disease for dental assistants, but for dentistsa non-significant increasing risk was observed (p¼0.52) (table 3).

 We repeated the analyses with 5 and 10 years of latency,which showed the same relationships as described in the mainanalyses (not shown). We also repeated the analysis afterexcluding those employed in 1964, which also did not changethe main results (not shown).

DISCUSSIONIn this nationwide study we investigated hospital admissionrates for neurological, Parkinson’s and renal diseases amongdentists and dental assistants. We found no increased risk of these diseases among dentists and dental assistants in periods of occupational mercury exposure compared with control groups.Neither did we observe any increased risk of these diseases fordentists and dental assistants employed during periods withhigh mercury exposure compared with dentists and dentalassistants employed during periods with less mercury exposure.

 We conclude that our study does not support the suggestionthat occupational mercury exposure from dental amalgamincreases the hospital admission rates for these diseases.

Mercury exposure has been associated with central nervoussystem effects and kidney impairment.5 Several cross-sectionalstudies among active employed dentists and dental assistantswith current exposure have reported associations between lowmercury exposure and decreased performance in neuro-psychological tests, memory disturbance and inferior logicalmemory tests,9e11 16 26 e29 while others report no effects.14 15 Inemployees in other occupations, for example, chloralkaliworkers2 0 3 0e32 and workers from a zincemercury amalgam-ation plant33 or thermometer manufacturing facility,34 mercury exposure has been associated with risk of glomerular protein-uria,3 2 3 3 glomerular dysfunction,2 0 3 0 psychomotor functionimpairment,33 mild sensory polyneuropathy 31 and static

tremor.34

No significant excess mortality due to nephritis andnephrosis or non-malignant diseases of the nervous system ina cohort of Norwegian chloralkali workers has been reported.35

 As regards possible persistent effects after cessation of exposure, studies among Norwegian chloralkali workers suggestmore slight persistent effects at higher exposure levels on thecentral nervous system2 0 2 1 than on the kidney.36  However,mercury-related effects seem to a large extent to be reversible. 37

Moen and colleagues have recently reported a higher rate of several neurological symptoms among 41 dental assistantspreviously exposed to amalgam compared with assistantnurses.23 In a larger cross-sectional Norwegian study including608 dental assistants (56% participation rate) and 456 popula-tion controls (43% participation rate), dental assistants report

more cognitive symptoms than controls.24 However, femaledentists with corresponding exposure levels report less neuro-logical symptoms compared with the same population controls.

 A small Australian study has shown that school dental nurseshave a higher prevalence of occupational overuse syndrome andare more anxious.38 The cross-sectional design, low participationrates and subjective reporting of exposure and symptoms may limit the validity of these studies.

 We present the only study that explores the associationbetween employment as a dentist or dental assistant and risk of hospital admission for neurological and renal diseases. Previousstudies have to a large extent focused on pre-clinical outcomes.Furthermore, our prospective study design using register-based

information may be more appropriate compared with cross-sectional questionnaire-based studies with low participation rates.

One limitation of ATP is that it does not include self-employed subjects. We think this problem only affects dentists,general practitioners and lawyers since they normally changefrom being employees early in their careers to being self-employment later on. Analytically, we handled this by defininglast date of employment in these three occupational groups asthe date of their 66th birthday or date of termination of employment after age 60 if that was available. As the CPR 

number was first introduced April 1968, workers who diedbefore this date were not included. It is unlikely that theseexclusions influenced the results.

Since ATP does not include information on occupationalgroups, misclassification of occupational groups may haveoccurred. We categorised workers in lawyers’ of fices withunknown occupational group based on sex, which may havecaused misclassification. Furthermore, we were not able todifferentiate between those working with dental amalgam foronly a few hours per day and those working more extensively with amalgam. We consider all these misclassifications non-differential since they hardly depend on the outcomes of interest. Information on outcomes was retrieved from adminis-trative data based on medical diagnosis, which may haveresulted in misclassification. We also consider this non-differen-tial as it is probably independent of occupational group. Thesemisclassifications will therefore, in general, underestimate theassociation between exposure and outcome and could thereforepartly explain our negative findings. It is important to empha-sise that several of the reported associations are in the oppositedirection to that expected if occupational exposure has detri-mental effect on the studied outcomes. Furthermore, we only included people employed for more than 8 h per week in dentalclinics, which means that the misclassification would bebetween dentists and dental assistants with actual employmentexposed to mercury. The same reasoning applies to thoseemployed in the control employment groups. Since both

dentists and dental assistants were exposed to occupationalmercury and all control groups were not exposed, we do notthink this misclassification can be the main explanation for ournegative findings.

 We used urine mercury samples among Norwegian dentalpersonnel2 to estimate the weights in our within-group analyses.The few samples collected in Denmark agree with the averagemercury level of approximately 200 nmol/l found in the 1960s,1

which is in close agreement with values reported in Norway.2

 We did not use absolute mercury exposure for the calculation of cumulative exposure, because this presupposes that exposureconditions were similar in Norway and Denmark, while relativeweights only assume a parallel development over calendar time.

Exposure to amalgam fillings may also have influenced expo-

sure levels among dental workers and control groups. Comparedwith the historical mercury dose from occupational exposure,the contribution of mercury from personal dental fillings isregarded as limited, and we further assume that the numbers of fillings are equally distributed between the groups. Thus, we findit unlikely that mercury exposure from personal dental fillingswould have confounded the associations observed.

Our study has several strengths. We used a nationwide cohortextending over many years, which is independent of any of theoutcomes studied. By using ATP, we were able to fully identify all employees in dental clinics and the control groups. Further-more, contributions have been made to ATP each quarter of the

 year since April 1964, which allows each employee’s course of 

employment to be followed for the entire study period. Fordentists and dental assistants this is an advantage since we have

900 Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063

Workplace

7/27/2019 Occup Environ Med 2011 Thygesen 895 901

http://slidepdf.com/reader/full/occup-environ-med-2011-thygesen-895-901 8/8

information on annual percentage of employment for every yearduring follow-up. We have complete follow-up until death oremigration for all Danish citizens and information on hospitaladmissions is independent of exposure. The results are thereforenot influenced by missing outcome information on, for example,those who leave the work force before retirement, or by differentoutcome information between dentists and dental assistants andthe control groups, which is a very positive feature of using

registers for research in Denmark and other Nordic countries.39

The objective outcome registration minimises bias fromsubjective reporting of exposure and outcome. Finally, since allDanish employees are included, the representativeness of theresults is ensured.40

In this long-term and nationwide study, we followed dentistsand dental assistants for neurological, Parkinson’s and renaldiseases. We found no increased risk of these diseases amongdentists and dental assistants in periods of occupational mercury exposure compared with control groups and no increased risk fordentists and dental assistants employed during periods withhigh mercury exposure compared with dentists and dentalassistants employed during periods with less mercury exposure.

 We conclude that our study does not show that occupationalmercury exposure from dental amalgam increases the hospitaladmission rates of neurological, Parkinson’s or renal diseases.

Funding This study was supported by a research grant fromArbejdsmiljøforskningsfonden, Denmark.

Competing interests None.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES1. Baelum J, Pockel H. Reference Document on Exposure to Metallic Mercury and the

 Development of Symptoms with Emphasis on Neurological and Neuropsychological  Diseases or Complaints. Odense: Odense University Hospital, 2007.

2. Lenvik K, Woldbæk T, Halgard K. Kvikksølveksponering blant tannhelsepersonell. Nor Tannlaegeforen Tid  2006;116:350e6.

3. Meyer-Baron M, Schaeper M, Seeber A. A meta-analysis for neurobehaviouralresults due to occupational mercury exposure. Arch Toxicol  2002;76:127

e36.

4. The World Health Organisation. Inorganic Mercury. The World HealthOrganisation Environmental Health Criteria Series. Geneva: The World healthOrganisation, 1991.

5. American Conference of Governmental Industrial Hygienists (ACGIH). Mercury, All Forms Except Alkyl . Cincinatti: ACGIH, 2001.

6. Brownawell AM, Berent S, Brent RL, et al . The potential adverse health effects ofdental amalgam. Toxicol Rev  2005;24:1e10.

7. Jacobsen P. Dansk Selskab for Arbejds- og Miljømedicin. [Mercury poisoning atdental clinics? The Danish Society of Occupational Medicine] (In Danish). Ugeskr 

 Laeger  2007;169:1097.8. Iyer K, Goodgold J, Eberstein A, et al . Mercury poisoning in a dentist. Arch Neurol 

1976;33:788e90.9. Ngim CH, Foo SC, Boey KW, et al . Chronic neurobehavioural effects of elemental

mercury in dentists. Br J Ind Med  1992;49:782e90.10. Echeverria D, Woods JS, Heyer NJ, et al . The association between a genetic

polymorphism of coproporphyrinogen oxidase, dental mercury exposure and

neurobehavioral response in humans. Neurotoxicol Teratol  2006;28:39e

48.11. Echeverria D, Woods JS, Heyer NJ, et al . Chronic low-level mercury exposure,

BDNF polymorphism, and associations with cognitive and motor function. Neurotoxicol Teratol  2005;27:781e96.

12. Meyer-Baron M, Schaeper M, van Thriel C, et al . Neurobehavioural test results andexposure to inorganic mercury: in search of dose-response relations. Arch Toxicol 2004;78:207e11.

13. Rohling ML, Demakis GJ. A meta-analysis of the neuropsychological effects ofoccupational exposure to mercury. Clin Neuropsychol  2006;20:108e32.

14. Nilsson B, Gerhardsson L, Nordberg GF. Urine mercury levels and associatedsymptoms in dental personnel. Sci Total Environ 1990;94:179e85.

15. Langworth S, Almkvist O, Soderman E, et al . Effects of occupationalexposure to mercury vapour on the central nervous system. Br J Ind Med 1992;49:545e55.

16. Ritchie KA, Gilmour WH, Macdonald EB, et al . Health and neuropsychologicalfunctioning of dentists exposed to mercury. Occup Environ Med  2002;59:287e93.

17. Bittner AC Jr, Echeverria D, Woods JS, et al . Behavioral effects of low-levelexposure to Hg0 among dental professionals: a cross-study evaluation ofpsychomotor effects. Neurotoxicol Teratol  1998;20:429e39.

18. Kishi R, Doi R, Fukuchi Y, et al . Subjective symptoms and neurobehavioralperformances of ex-mercury miners at an average of 18 years after the cessation ofchronic exposure to mercury vapor. Mercury Workers Study Group. Environ Res1993;62:289e302.

19. Zachi EC, Faria MA, Taub A. Neuropsychological dysfunction related to earlieroccupational exposure to mercury vapor. Braz J Med Biol Res 2007;40:425e33.

20. Ellingsen DG, Barregard L, Gaarder PI, et al . Assessment of renal dysfunction inworkers previously exposed to mercury vapour at a chloralkali plant. Br J Ind Med 1993;50:881e7.

21. Mathiesen T, Ellingsen DG, Kjuus H. Neuropsychological effects associated withexposure to mercury vapor among former chloralkali workers. Scand J Work Environ

 Health 1999;25:342e50.22. Andersen A, Ellingsen DG, Morland T, et al . A neurological and neurophysiological

study of chloralkali workers previously exposed to mercury vapour. Acta Neurol Scand  1993;88:427e33.

23. Moen B, Hollund B, Riise T. Neurological symptoms among dental assistants:

a cross-sectional study. J Occup Med Toxicol  2008;3:10.24. Hilt B, Svendsen K, Syversen T, et al . Occurrence of cognitive symptoms in dental

assistants with previous occupational exposure to metallic mercury. Neurotoxicology 2009;30:1202e6.

25. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austri: R Foundation for Statistical Computing, 2009.

26. Aydin N, Karaoglanoglu S, Yigit A, et al . Neuropsychological effects of low mercuryexposure in dental staff in Erzurum, Turkey. Int Dent J 2003;53:85e91.

27. Echeverria D, Heyer NJ, Martin MD, et al . Behavioral effects of low-level exposureto elemental Hg among dentists. Neurotoxicol Teratol  1995;17:161e8.

28. Echeverria D, Aposhian HV, Woods JS, et al . Neurobehavioral effects fromexposure to dental amalgam Hg(o): new distinctions between recent exposure andHg body burden. FASEB J 1998;12:971e80.

29. Heyer NJ, Echeverria D, Bittner AC Jr, et al . Chronic low-level mercury exposure,BDNF polymorphism, and associations with self-reported symptoms and mood.Toxicol Sci  2004;81:354e63.

30. Buchet JP, Roels H, Bernard A, et al . Assessment of renal function of workersexposed to inorganic lead, calcium or mercury vapor. J Occup Med 1980;22:741

e50.

31. Albers JW, Kallenbach LR, Fine LJ, et al . Neurological abnormalities associated withremote occupational elemental mercury exposure. Ann Neurol  1988;24:651e9.

32. Foa V, Caimi L, Amante L, et al . Patterns of some lysosomal enzymes in the plasmaand of proteins in urine of workers exposed to inorganic mercury. Int Arch Occup

 Environ Health 1976;37:115e24.33. Roels H, Lauwerys R, Buchet JP, et al . Comparison of renal function and

psychomotor performance in workers exposed to elemental mercury. Int Arch Occup Environ Health 1982;50:77e93.

34. Ehrenberg RL, Vogt RL, Smith AB, et al . Effects of elemental mercury exposure ata thermometer plant. Am J Ind Med  1991;19:495e507.

35. Ellingsen DG, Andersen A, Nordhagen HP, et al . Incidence of cancer and mortalityamong workers exposed to mercury vapour in the Norwegian chloralkali industry.

 Br J Ind Med  1993;50:875e80.36. Ellingsen DG, Efskind J, Berg KJ, et al . Renal and immunologic markers for

chloralkali workers with low exposure to mercury vapor. Scand J Work Environ Health2000;26:427e35.

37. He FS, Zhow XR, Lin BX, et al . Prognosis of mercury poisoning in mercury refineryworkers. Ann Acad Med Singapore 1984;13(Suppl 2):389e93.38. Jones L, Bunnell J, Stillman J. A 30-year follow-up of residual effects on New

Zealand School Dental Nurses, from occupational mercury exposure. Hum Exp Toxicol 2007;26:367e74.

39. Rosen M. National Health Data Registers: a Nordic heritage to public health. Scand  J Public Health 2002;30:81e5.

40. Sorensen HT, Schulze S. Danish health registries. A valuable tool in medicalresearch. Dan Med Bull  1996;43:463.

Occup Environ Med  2011;68:895e901. doi:10.1136/oem.2010.064063 901

Workplace