Downloaded from oem.bmj.com on October 19, 2013 - Published by group.bmj.

com

Hospital admissions for neurological and renal diseases among dentists and dental assistants occupationally exposed to mercury
Lau Caspar Thygesen, Esben Meulengracht Flachs, Kirsten Hanehj, et al. Occup Environ Med 2011 68: 895-901 originally published online April 20, 2011

doi: 10.1136/oem.2010.064063

Updated information and services can be found at:

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

These include:

References Email alerting service

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

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

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

Topic Collections

Articles on similar topics can be found in the following collections Other exposures (682 articles)

Notes

To request permissions go to:

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

To order reprints go to:

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

To subscribe to BMJ go to:

http://group.bmj.com/subscribe/

Downloaded from oem.bmj.com on October 19, 2013 - Published by group.bmj.com

Workplace

ORIGINAL ARTICLE

Hospital admissions for neurological and renal diseases among dentists and dental assistants occupationally exposed to mercury
Lau Caspar Thygesen,1 Esben Meulengracht Flachs,1 Kirsten Hanehj,1 Helge Kjuus,2 Knud Juel1
1

National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark 2 National Institute of Occupational Health, Oslo, Norway Correspondence to Dr Lau Caspar Thygesen, National Institute of Public Health, University of Southern Denmark, ster Farimagsgade 5 A, 2nd oor, DK-1353 Copenhagen K, Denmark; lct@niph.dk Accepted 23 March 2011 Published Online First 20 April 2011

ABSTRACT Objectives For many years an amalgam containing metallic mercury, which has been associated with neurological and renal diseases, has been used in dentistry. In this nationwide study we compared hospital admissions due to neurological and renal diseases among dentists and dental assistants to admissions in controls. Methods This register-based cohort study included all Danish workers employed in dental clinics, general practitioners clinics or lawyers ofces between 1964 and 2006. We compared dentists with general practitioners and lawyers, and dental assistants with medical secretaries, nurses and legal secretaries. We also compared dentists and dental assistants employed during periods with high occupational mercury exposure with dentists and dental assistants employed during periods with less mercury exposure. We followed all subjects in a nationwide register of hospital admissions. We analysed risk of neurological diseases, Parkinsons disease and renal diseases using a Cox regression model. Results The cohort consisted of 122 481 workers including 5371 dentists and 33 858 dental assistants. For neurological diseases, no association was observed for dental assistants, while for dentists an increasing risk for periods with less mercury exposure was observed. Among dental assistants, a negative association between employment length and risk of neurological disease was observed. Admissions for renal disease among dental assistants were increased during periods with less mercury exposure compared with controls. For dentists a non-signicant increased risk was observed between employment length and renal disease risk. Conclusions Our nationwide study does not indicate that occupational exposure to mercury increases the risk of hospital admissions for neurological, Parkinsons or renal diseases.

What this paper adds

< Previous studies have usually been cross-

sectional and in several the selection of control groups may have biased the results. < Our nationwide study of dental workers between 1964 and 2006 shows no increased risk of neurological, Parkinsons or renal diseases among dentists and dental assistants compared with controls and no increased risk for high occupational mercury exposure compared with lower occupational mercury exposure. < We conclude that our nationwide study does not support the suggestion that occupational mercury exposure from dental amalgam increases the risk of hospital admissions for neurological, Parkinsons or renal diseases.

INTRODUCTION
For many years an amalgam containing metallic mercury has been used by dentists in Denmark. Since 1963 the handling of amalgam llings has gradually been more strictly regulated. Previous studies have shown that exposure to mercury in dentistry 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
Occup Environ Med 2011;68:895e901. doi:10.1136/oem.2010.064063

urinary values have slowly decreased to about 25 nmol/l, with individual values rarely exceeding 100 nmol/l.1 There is no evidence of any differences in urine mercury levels between dentists and dental assistants.2 Previous studies of the effects of mercury in other occupational groups have shown that acute mercury intoxication is characterised by a number of neurological and neuropsychological symptoms and indicate that urine mercury above 600 nmol/l is associated with impairment in neuropsychological tests.3 The nervous system and the kidneys are especially sensitive to mercury toxicity.4e6 Very few genuine mercury poisonings have been reported in dentistry and many of the symptoms of intoxication have been unspecic.7 8 However, some studies have reported that exposure to mercury associated with urine mercury levels of approximately 150 nmol/l results in impaired results in neuropsychological and psychomotor tests,3 9e13 while other studies have shown no effect14e16 or only slight effects.17 In studies of past exposure, the effects are smaller and decrease with time from cessation of exposure,3 although subjects with previous mercury intoxication still have decreased performance several years later.18 19 Slight neurophysiological and neuropsychological effects have been found among chloralkali workers

895

Workplace
with mean urine mercury levels of 500 nmol/l examined 12 years after exposure.20e22 Two Norwegian questionnaire studies of dental personnel showed more neurological and general symptoms in dental assistants compared with controls, while dentists had fewer symptoms compared with controls 20e30 years after exposure.23 24 These studies stimulated intense debate in Denmark on 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 for neurological, Parkinsons and renal diseases compared with control groups based on the nationwide registers of employment groups which started in 1964. The employer pays contributions into the employees account in ATP four times each year according to the number of hours worked. For every employment the following information is registered: the employees identication number (CPR number), the employer s identication number (including industrial classication code) and the percentage of employment for each quarter of the year. This information is retained in ATP even after 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 Danish authorisation to work as healthcare personnel including dentists, general practitioners and nurses. This register contains information back to 1982. The Danish Civil Registration System (CRS) contains information on mortality, emigration, immigration and disappearance for all Danish residents since 1968. Furthermore, people can voluntarily register information on their most important job title with the CRS. This information on job title is not dated. The Labour Market Module (LMM) at Statistics Denmark includes information on occupational classication for all Danes for November of each year since 1980. The classication of occupational groups does not include a category for dental assistants. In the cohort we included workers who had paid into a supplementary pension between 1 April 1964 and 31 December 2006. We included employees in dental clinics and compared their morbidity with two control groups (employees in a general practitioner s clinic and employees in a lawyer s

MATERIAL AND METHODS Dening the cohort

This was a historical register-based cohort study. In Denmark, morbidity can be studied in register-based studies, because nationwide registration systems with linkage between registers at the individual level via the unique personal identication number (CPR number) are available for all permanent Danish residents. We obtained information from several registers when constructing the cohort. The Supplementary Pension Fund Register (ATP) was established in 1964 as a compulsory supplementary pension fund (Arbejdsmarkedets Tillgspension) for all workers aged 18e66 years employed in Denmark for more than 8 h per week.

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
ofce). We obtained information from ATP on all workers employed in a dental clinic (n41 328), a general practitioner s clinic (n38 584) or a lawyer s ofce (n53 398) (gure 1). If a person had worked in a dental clinic and a general practitioner s clinic (n1198) or a lawyer s ofce (n492) or had had all three jobs (n28), we only included the employment information from the dental clinic. The three employment groups were divided into seven occupational groups: dentists, dental assistants, general practitioners, nurses, medical secretaries, lawyers and legal secretaries. The study thereby included two occupational groups exposed to mercury (dentists and dental assistants) and ve control groups (gure 1). Information on occupational groups was not available from 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 not authorised to practice as dentists, general practitioners or nurses. As this information was not necessarily obtained at the same time as data from ATP, employees with job titles not associated with the three employment groups had to be categorised. We used the educational level of the specic job titles obtained from the CRS. Workers with job titles requiring four years of education after elementary school were considered to have low and medium education, while those in jobs requiring ve or more years of education after elementary school were considered to be highly educated. Workers in dental clinics with unknown, low or medium education were categorised to the dental assistant group. Highly educated workers were excluded (see gure 1 for the numbers of workers in each category). The same procedure was used for employees in a general practitioner s clinic. To determine whether this method resulted in unbiased occupational groups, we assessed consistency between our occupational groups and the LMM. The distribution of occupational 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 was unbiased. The same was observed for medical secretary . Table 1 Characteristics of the cohort, Denmark, 1964e2006
Dental clinics Dentist Persons Persons Persons Persons employed employed employed employed 1964e2006 1964e1979 1980e1994 1995e2006 5371 2033 3493 3448 2693 50.1 30.2 191 15.9 8 1.2 122 5.7 304 277 3 112 462 Assistant 33 858 12 065 17 514 16789 33 858 100.0 23.8 1227 20.6 14 0.6 625 10.1 1406 832 4 682 682 General practitioners clinics GP 6154 778 3018 3667 2455 39.9 37.3 148 22.6 6 1.6 86 6.7 266 123 0 86 141 Nurse 5872 1433 2447 3805 5872 100.0 37.3 195 18.3 1 0.2 68 6.4 290 66 0 89 476 Secretary 22 785 7626 12 450 11 537 22 785 100.0 35.5 1085 22.2 34 0.8 395 8.4 2254 376 2 425 499 Lawyers ofces Lawyer 11 433 3787 5527 5940 1916 16.8 27.7 355 29.1 16 2.5 237 5.3 1062 397 3 207 059 Secretary 37 717 15659 18 848 16 393 37 717 100.0 27.6 1540 20.3 53 1.1 660 8.9 3537 1017 7 794 055

For workers in lawyers ofces, the group with no information on job title in the CRS was divided into lawyer and legal secretary in that men were classied as lawyers and women as legal secretaries since linkage with Statistics Denmark showed considerable sex differences consistent with this classication. For dental assistants, 33 858 workers (95.0%) were women. The same proportions were observed for nurses (99.4%), medical secretaries (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 practitioners and lawyers, the proportions of women were lower (see table 1) and we therefore included both sexes. Finally, we excluded 70 subjects who emigrated from Denmark before they were employed. The nal cohort consisted of 122 481 persons employed in seven occupational groups (rst row, table 1).

Exposure
Using information from ATP on the annual percentage of employment for all years, we constructed a rst date of employment and a date of termination of employment. We compared these dates with information from the CRS on vital and emigration status to determine the rst and last day of employment. Dentists, general practitioners and lawyers normally changed from being employees early in their careers to being self-employment later on. Since ATP only contains information on employees, the date of the subjects 66th birthday or the date of termination after age 60 if available was considered to be the last date of employment. In analyses where dental assistants and dentists were compared with the control groups, we used the rst and last date of employment to construct the period of employment, which we then used as the measurement for exposure. Since occupational mercury exposure decreased with calendar time, we included the period effect using four categories (1964e1969, 1970e1979, 1980e1989 and 1990e2006). A proxy measure for excess cumulative mercury exposure among dental personnel compared with the general population

Female (n) Female (%) Age at employment (mean) Neurological disease cases (n) Neurological disease IR* Parkinsons disease cases (n) Parkinsons disease IR* Renal disease cases (n) Renal disease IR* Deaths by 31 Dec 2006 (n) Emigrations by 31 Dec 2006 (n) Disappearances by 31 Dec 2006 (n) Follow-up (years)

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
was calculated as the weighted amount of occupational mercury exposure during their working career multiplied by the annual percentage of employment. The weights were based on urine samples from Norwegian dental personnel collected since the late 1950s.2 Mercury exposure was highest in the 1960s (weight of 4.8 for each year) and decreased until 1989. The weights were 2.2 for each year in 1970e1974, 1.2 in 1975e1979, 0.8 in 1980e1984 and 0.2 in 1985e1989. The weight from 1990 and onwards was set to 0 as dental personnel were not thought to be exposed to excess mercury after this date. The maximum weighted exposure was 50.8, which was given to workers with full time employment in a dental clinic for the entire period 1964e1989. In analyses of cumulative exposure, weighted exposure was divided into ve categories (0, 0.1e0.9, 1.0e4.7, 4.8e14.9 and 15.0e50.8). A Cox regression analysis was performed with R software v 2.9.025 to estimate the HR of employment as a dental assistant or dentist compared with the control groups in the betweengroup analyses, and the HR of cumulative exposure among dental assistants and dentists in the within-group analyses, while taking potential confounding variables into account. We tested the between-group effect by conducting a test of interaction between period and occupational group, that is, a test of whether the HR for the exposed occupational groups compared with the control groups varied by period. Because age was a strong predictor of the diseases of interest, age was used as the time scale with delayed entry, thereby adjusting for the effect of age on the risk estimates. For the analyses of neurological and renal diseases, we related time-dependent occupational mercury exposure to risk of disease the following year. For the analysis of Parkinsons disease, we included a latency period of 5 years between the time-dependent exposure and disease. We thereby assumed that Parkinsons disease occurring during the latency period was not related to the exposure. To evaluate the assumption of no latency period for neurological and renal diseases, we repeated the analyses assuming latencies of 5 and 10 years. For Parkinsons disease too few cases were observed for such sensitivity analysis. We examined the proportional hazards assumption by testing whether the correlation between the scaled Schoenfeld residuals and underlying time was different from zero and examined plots of underlying time by Schoenfeld residuals. These tests and plots did not indicate violation of the assumption. We followed the cohort from their rst employment or 1 January 1977 to the date of disease diagnosis, date of death, emigration or disappearance, or end of follow-up (31 December 2006), whichever came rst. For each outcome, we followed the cohort to rst admission for that specic disease.

Outcomes
The Danish National Patient Register contains information on all hospital discharges, including diagnoses and surgical procedures, in Denmark since 1977. This register contains information on mode and date of admission and discharge and ICD codes for primary and secondary diagnoses for all hospital admissions. We followed the cohort of hospital admissions due to neurological disease except eye and ear disease (ICD-8, 320e359; ICD-10, G00-G99), Parkinsons disease (ICD-8, 342.9; ICD-10, G20) and renal disease except nephritis and obstructive uropathy (ICD-8, 580e584 and 590e593; ICD-10, N00eN08 and N14eN29). We included primary and secondary diagnoses in 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 rates for dentists and dental assistants were compared with control groups and (2) a within-group analysis where dentists and dental assistants with employment during periods with high mercury exposure were compared with dentists and dental assistants employed during periods with less mercury exposure. In the between-group analysis, we examined whether relative rates among dentists and dental assistants compared with control groups were higher during periods with occupational mercury exposure compared with periods with no excess occupational mercury exposure. In the within-group analyses, we evaluated the effect of cumulative weighted exposure among dental assistants and dentists on each health outcome. As ATP was established on 1 April 1964, information on employment was not available before this date. To account for exposure before 1964, we adjusted for age in 1964 for those employed in 1964 (4.2% of the cohort) using four categories: 17e24 years (n2381), 25e39 years (n1458), 40e65 years (n1370) and not employed in 1964 (n117 272). As a sensitivity analysis, we repeated the analyses excluding those employed in 1964. We also adjusted for sex in the analyses of dentists, general practitioners and lawyers. In the between-group analyses of dental assistants, we furthermore adjusted for length of employment using four categories (1e154 days, 155e822 days, 823e2477 days and 2478+ days). In the within-group analyses, we adjusted for follow-up period using ve categories (1977e1982, 1983e1988, 1989e1994, 1995e2000 and 2001e2006).
898

RESULTS
The number of workers employed in each of the seven occupational groups is shown in table 1. The proportion of women was highest for dentists and lowest for lawyers. The mean age at employment was highest for general practitioners, nurses and medical secretaries and lowest for dental assistants. The numbers of cases and age-standardised incidence rates showed that dentists had a lower incidence of neurological and Parkinsons disease compared with the control groups (general practitioners and lawyers) (table 1). Dental assistants had rates comparable with medical and legal secretaries for all diseases. Nurses had lower rates. By the end of follow-up, 304 dentists and 1406 dental assistants had died.

Neurological disease
Admission rates due to neurological disease among dental assistants compared with the three control groups for the four periods did not show any consistent pattern (table 2). Tests for different period effects were insignicant in all control groups (eg, p0.49 for dental assistants compared with medical secretaries). For dentists there was a signicant difference in period effects with increasing risk for dentists compared with general practitioners and lawyers over time (table 2). Among dental assistants, a negative association between cumulative exposure and neurological disease was observed, although the association was not signicant (p0.21) (table 3). Among dentists there was a non-signicant U-shaped association (p0.51).
Occup Environ Med 2011;68:895e901. doi:10.1136/oem.2010.064063

Workplace
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* HR (95% CI) Neurological disease 1964e1969 1970e1979 1980e1989 1990e2006 p Value p Value (trend) Parkinsons disease 1964e1969 1970e1979 1980e1989 1990e2006 p Value p Value (trend) Renal disease 1964e1969 1970e1979 1980e1989 1990e2006 p Value p Value (trend) 0.92 0.83 0.97 0.85 0.49 0.95 0.78 0.83 0.68 0.96 1.00 0.96 1.04 1.19 1.06 1.38 0.58 0.45 (0.76 (0.73 (0.82 (0.68 to to to to 1.12) 0.95) 1.15) 1.06) DA/NU* HR (95% CI) 1.10 0.99 0.97 0.99 0.96 0.58 e e e e (0.76 (0.77 (0.70 (0.70 to to to to 1.60) 1.28) 1.32) 1.39) DA/LS* HR (95% CI) 1.11 1.08 1.08 1.22 0.82 0.65 (0.96 (0.95 (0.92 (0.97 to to to to 1.30) 1.22) 1.26) 1.53) DE/GPy HR (95% CI) 0.46 1.02 1.10 1.50 0.07 0.03 (0.23 (0.70 (0.73 (0.94 to to to to 0.91) 1.49) 1.66) 2.38) DE/LAy HR (95% CI) 0.64 0.86 0.98 1.16 0.21 0.03 (0.44 (0.65 (0.67 (0.72 to to to to 0.93) 1.14) 1.43) 1.86)

(0.26 (0.29 (0.14 (0.10

to to to to

2.31) 2.35) 3.25) 9.24)

0.74 (0.31 to 1.80) 0.92 (0.33 to 2.56) 0.67 (0.14 to 3.34) e 0.57 0.55 (0.54 (0.89 (0.89 (0.71 to to to to 1.53) 2.06) 2.91) 2.23) 1.24 1.18 1.19 1.69 0.28 0.29 (0.99 (0.99 (0.95 (1.21 to to to to 1.56) 1.42) 1.49) 2.36)

0.40 (0.07 to 2.20) 0.21 (0.02 to 2.04) 2.84 (0.25 to 32.6) e 0.43 0.28 1.27 1.27 0.88 0.83 0.68 0.26 (0.50 (0.76 (0.52 (0.39 to to to to 3.25) 2.11) 1.49) 1.76)

1.00 (0.32 to 3.12) 0.33 (0.04 to 3.03) 2.29 (0.21 to 25.4) e 0.35 0.61 1.38 0.89 0.88 0.98 0.37 0.22 (0.93 (0.63 (0.54 (0.46 to to to to 2.05) 1.28) 1.42) 2.09)

(0.77 (0.96 (0.83 (0.96

to to to to

1.40) 1.46) 1.37) 1.98)

0.90 1.36 1.61 1.26 0.52 0.40

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.

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

Among dental assistants, a negative association was shown between cumulative exposure and risk of Parkinsons disease, although the association was not signicant (p0.45) (table 3). Too few cases were observed among dentists to perform a meaningful analysis of the association between cumulative exposure and Parkinsons disease.

Parkinsons disease
Admission rates for Parkinsons disease among dental assistants compared with medical and legal secretaries did not show any consistent pattern and no signicant difference for period effects was observed (table 2). This analysis was not performed for nurses as only one nurse was admitted to hospital due to Parkinsons disease. There was no consistent pattern for dentists (table 2).

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

Table 3 Within-group analyses of cumulative exposure and hospital admission rates among dental assistants and dentists, Denmark, employment 1964e2006, follow-up 1977e2006
Neurological disease Cumulative exposure Dental assistant* 0 0.1e0.9 1.0e4.7 4.8e14.9 15.0e50.8 p Value p Value (trend) Dentisty 0 0.1e0.9 1.0e4.7 4.8e14.9 15.0e50.8 p Value p Value (trend) Cases, n 163 380 348 249 87 HR (95% CI) 0.96 (0.78 1 (ref) 0.93 (0.80 0.83 (0.70 0.79 (0.61 0.21 0.31 1.36 (0.82 1 (ref) 0.94 (0.56 1.03 (0.63 1.28 (0.76 0.51 0.29 to 1.17) to 1.08) to 0.99) to 1.02) Parkinsons disease Cases, n 1 5 3 4 1 HR (95% CI) 1.68 (0.17 1 (ref) 0.46 (0.10 0.57 (0.13 0.15 (0.01 0.45 0.22 e e e e e to 16.2) to 2.01) to 2.43) to 1.70) Renal disease Cases, n 94 196 183 113 39 HR (95% CI) 1.05 (0.79 1 (ref) 0.98 (0.80 0.81 (0.64 0.82 (0.56 0.41 0.21 0.97 (0.43 1 (ref) 1.31 (0.68 1.30 (0.68 1.74 (0.88 0.52 0.18 to 1.39) to 1.21) to 1.04) to 1.20)

33 30 30 47 51

to 2.26) to 1.58) to 1.68) to 2.16)

0 2 0 1 5

10 16 23 31 42

to 2.18) to 2.51) to 2.48) to 3.43)

*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
There was no association between cumulative exposure and risk of renal disease for dental assistants, but for dentists a non-signicant increasing risk was observed (p0.52) (table 3). We repeated the analyses with 5 and 10 years of latency, which showed the same relationships as described in the main analyses (not shown). We also repeated the analysis after excluding those employed in 1964, which also did not change the main results (not shown). One limitation of ATP is that it does not include selfemployed subjects. We think this problem only affects dentists, general practitioners and lawyers since they normally change from being employees early in their careers to being selfemployment later on. Analytically, we handled this by dening last date of employment in these three occupational groups as the date of their 66th birthday or date of termination of employment after age 60 if that was available. As the CPR number was rst introduced April 1968, workers who died before this date were not included. It is unlikely that these exclusions inuenced the results. Since ATP does not include information on occupational groups, misclassication of occupational groups may have occurred. We categorised workers in lawyers ofces with unknown occupational group based on sex, which may have caused misclassication. Furthermore, we were not able to differentiate between those working with dental amalgam for only a few hours per day and those working more extensively with amalgam. We consider all these misclassications nondifferential since they hardly depend on the outcomes of interest. Information on outcomes was retrieved from administrative data based on medical diagnosis, which may have resulted in misclassication. We also consider this non-differential as it is probably independent of occupational group. These misclassications will therefore, in general, underestimate the association between exposure and outcome and could therefore partly explain our negative ndings. It is important to emphasise that several of the reported associations are in the opposite direction to that expected if occupational exposure has detrimental effect on the studied outcomes. Furthermore, we only included people employed for more than 8 h per week in dental clinics, which means that the misclassication would be between dentists and dental assistants with actual employment exposed to mercury. The same reasoning applies to those employed in the control employment groups. Since both dentists and dental assistants were exposed to occupational mercury and all control groups were not exposed, we do not think this misclassication can be the main explanation for our negative ndings. We used urine mercury samples among Norwegian dental personnel2 to estimate the weights in our within-group analyses. The few samples collected in Denmark agree with the average mercury 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 exposure conditions were similar in Norway and Denmark, while relative weights only assume a parallel development over calendar time. Exposure to amalgam llings may also have inuenced exposure levels among dental workers and control groups. Compared with the historical mercury dose from occupational exposure, the contribution of mercury from personal dental llings is regarded as limited, and we further assume that the numbers of llings are equally distributed between the groups. Thus, we nd it unlikely that mercury exposure from personal dental llings would have confounded the associations observed. Our study has several strengths. We used a nationwide cohort extending over many years, which is independent of any of the outcomes studied. By using ATP, we were able to fully identify all employees in dental clinics and the control groups. Furthermore, contributions have been made to ATP each quarter of the year since April 1964, which allows each employees course of employment to be followed for the entire study period. For dentists and dental assistants this is an advantage since we have
Occup Environ Med 2011;68:895e901. doi:10.1136/oem.2010.064063

DISCUSSION
In this nationwide study we investigated hospital admission rates for neurological, Parkinsons and renal diseases among dentists 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 for dentists and dental assistants employed during periods with high mercury exposure compared with dentists and dental assistants employed during periods with less mercury exposure. We conclude that our study does not support the suggestion that occupational mercury exposure from dental amalgam increases the hospital admission rates for these diseases. Mercury exposure has been associated with central nervous system effects and kidney impairment.5 Several cross-sectional studies among active employed dentists and dental assistants with current exposure have reported associations between low mercury exposure and decreased performance in neuropsychological tests, memory disturbance and inferior logical memory tests,9e11 16 26e29 while others report no effects.14 15 In employees in other occupations, for example, chloralkali workers20 30e32 and workers from a zincemercury amalgamation plant33 or thermometer manufacturing facility,34 mercury exposure has been associated with risk of glomerular proteinuria,32 33 glomerular dysfunction,20 30 psychomotor function impairment,33 mild sensory polyneuropathy31 and static tremor.34 No signicant excess mortality due to nephritis and nephrosis or non-malignant diseases of the nervous system in a cohort of Norwegian chloralkali workers has been reported.35 As regards possible persistent effects after cessation of exposure, studies among Norwegian chloralkali workers suggest more slight persistent effects at higher exposure levels on the central nervous system20 21 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 assistants previously exposed to amalgam compared with assistant nurses.23 In a larger cross-sectional Norwegian study including 608 dental assistants (56% participation rate) and 456 population controls (43% participation rate), dental assistants report more cognitive symptoms than controls.24 However, female dentists with corresponding exposure levels report less neurological symptoms compared with the same population controls. A small Australian study has shown that school dental nurses have a higher prevalence of occupational overuse syndrome and are more anxious.38 The cross-sectional design, low participation rates and subjective reporting of exposure and symptoms may limit the validity of these studies. We present the only study that explores the association between employment as a dentist or dental assistant and risk of hospital admission for neurological and renal diseases. Previous studies 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 crosssectional questionnaire-based studies with low participation rates.
900

Workplace
information on annual percentage of employment for every year during follow-up. We have complete follow-up until death or emigration for all Danish citizens and information on hospital admissions is independent of exposure. The results are therefore not inuenced by missing outcome information on, for example, those who leave the work force before retirement, or by different outcome information between dentists and dental assistants and the 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 from subjective reporting of exposure and outcome. Finally, since all Danish employees are included, the representativeness of the results is ensured.40 In this long-term and nationwide study, we followed dentists and dental assistants for neurological, Parkinsons and renal diseases. We found no increased risk of these diseases among dentists and dental assistants in periods of occupational mercury exposure compared with control groups and no increased risk for dentists and dental assistants employed during periods with high mercury exposure compared with dentists and dental assistants employed during periods with less mercury exposure. We conclude that our study does not show that occupational mercury exposure from dental amalgam increases the hospital admission rates of neurological, Parkinsons or renal diseases.
Funding This study was supported by a research grant from Arbejdsmiljforskningsfonden, Denmark. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed.
14. 15. 16. 17. 18. Nilsson B, Gerhardsson L, Nordberg GF. Urine mercury levels and associated symptoms in dental personnel. Sci Total Environ 1990;94:179e85. Langworth S, Almkvist O, Soderman E, et al. Effects of occupational exposure to mercury vapour on the central nervous system. Br J Ind Med 1992;49:545e55. Ritchie KA, Gilmour WH, Macdonald EB, et al. Health and neuropsychological functioning of dentists exposed to mercury. Occup Environ Med 2002;59: 287e93. Bittner AC Jr, Echeverria D, Woods JS, et al. Behavioral effects of low-level exposure to Hg0 among dental professionals: a cross-study evaluation of psychomotor effects. Neurotoxicol Teratol 1998;20:429e39. Kishi R, Doi R, Fukuchi Y, et al. Subjective symptoms and neurobehavioral performances of ex-mercury miners at an average of 18 years after the cessation of chronic exposure to mercury vapor. Mercury Workers Study Group. Environ Res 1993;62:289e302. Zachi EC, Faria MA, Taub A. Neuropsychological dysfunction related to earlier occupational exposure to mercury vapor. Braz J Med Biol Res 2007;40: 425e33. Ellingsen DG, Barregard L, Gaarder PI, et al. Assessment of renal dysfunction in workers previously exposed to mercury vapour at a chloralkali plant. Br J Ind Med 1993;50:881e7. Mathiesen T, Ellingsen DG, Kjuus H. Neuropsychological effects associated with exposure to mercury vapor among former chloralkali workers. Scand J Work Environ Health 1999;25:342e50. 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. Moen B, Hollund B, Riise T. Neurological symptoms among dental assistants: a cross-sectional study. J Occup Med Toxicol 2008;3:10. 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. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austri: R Foundation for Statistical Computing, 2009. Aydin N, Karaoglanoglu S, Yigit A, et al. Neuropsychological effects of low mercury exposure in dental staff in Erzurum, Turkey. Int Dent J 2003;53:85e91. Echeverria D, Heyer NJ, Martin MD, et al. Behavioral effects of low-level exposure to elemental Hg among dentists. Neurotoxicol Teratol 1995;17:161e8. Echeverria D, Aposhian HV, Woods JS, et al. Neurobehavioral effects from exposure to dental amalgam Hg(o): new distinctions between recent exposure and Hg body burden. FASEB J 1998;12:971e80. 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. Buchet JP, Roels H, Bernard A, et al. Assessment of renal function of workers exposed to inorganic lead, calcium or mercury vapor. J Occup Med 1980;22:741e50. Albers JW, Kallenbach LR, Fine LJ, et al. Neurological abnormalities associated with remote occupational elemental mercury exposure. Ann Neurol 1988;24:651e9. Foa V, Caimi L, Amante L, et al. Patterns of some lysosomal enzymes in the plasma and of proteins in urine of workers exposed to inorganic mercury. Int Arch Occup Environ Health 1976;37:115e24. 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. Ehrenberg RL, Vogt RL, Smith AB, et al. Effects of elemental mercury exposure at a thermometer plant. Am J Ind Med 1991;19:495e507. Ellingsen DG, Andersen A, Nordhagen HP, et al. Incidence of cancer and mortality among workers exposed to mercury vapour in the Norwegian chloralkali industry. Br J Ind Med 1993;50:875e80. 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 Health 2000;26:427e35. He FS, Zhow XR, Lin BX, et al. Prognosis of mercury poisoning in mercury renery workers. Ann Acad Med Singapore 1984;13(Suppl 2):389e93. 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. n M. National Health Data Registers: a Nordic heritage to public health. Scand Rose J Public Health 2002;30:81e5. Sorensen HT, Schulze S. Danish health registries. A valuable tool in medical research. Dan Med Bull 1996;43:463.

19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

REFERENCES
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Baelum J, Po ckel 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. Lenvik K, Woldbk T, Halgard K. Kvikkslveksponering blant tannhelsepersonell. Nor Tannlaegeforen Tid 2006;116:350e6. Meyer-Baron M, Schaeper M, Seeber A. A meta-analysis for neurobehavioural results due to occupational mercury exposure. Arch Toxicol 2002;76:127e36. The World Health Organisation. Inorganic Mercury. The World Health Organisation Environmental Health Criteria Series. Geneva: The World health Organisation, 1991. American Conference of Governmental Industrial Hygienists (ACGIH). Mercury, All Forms Except Alkyl. Cincinatti: ACGIH, 2001. Brownawell AM, Berent S, Brent RL, et al. The potential adverse health effects of dental amalgam. Toxicol Rev 2005;24:1e10. Jacobsen P. Dansk Selskab for Arbejds- og Miljmedicin. [Mercury poisoning at dental clinics? The Danish Society of Occupational Medicine] (In Danish). Ugeskr Laeger 2007;169:1097. Iyer K, Goodgold J, Eberstein A, et al. Mercury poisoning in a dentist. Arch Neurol 1976;33:788e90. Ngim CH, Foo SC, Boey KW, et al. Chronic neurobehavioural effects of elemental mercury in dentists. Br J Ind Med 1992;49:782e90. 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;28:39e48. 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. Meyer-Baron M, Schaeper M, van Thriel C, et al. Neurobehavioural test results and exposure to inorganic mercury: in search of dose-response relations. Arch Toxicol 2004;78:207e11. Rohling ML, Demakis GJ. A meta-analysis of the neuropsychological effects of occupational exposure to mercury. Clin Neuropsychol 2006;20:108e32.

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

901