Difference between revisions of "SETURI: National estimates of DALY of environmental risks"

Revision as of 11:55, 4 April 2008

<accesscontrol>seturi_group</accesscontrol>

The purpose of this page is to serve as a forum for estimating DALYs due to exposure to environmental and other risks.

General

General procedure

The aim is to analyze current chemical and physical exposures in Finland and their health consequences. The procedures used a similar to a project done in the Netherlands (de Hollander 1999) and the Global burden of disease project of WHO (http://www.who.int/topics/global_burden_of_disease/en/). The work is expected to contribute to a comparative study between Finland, the Netherlands, and Norway (Jantunen: Study plan)

The work will start with selected exposures after which more exposures will be analyzed. The emphasis is on comparability, not comprehensiveness. The selected procedure requires good data on exposure and dose-response, which means that all exposures and all outcomes can not be included.

First, the current exposure distribution is estimated for all Finns (or only those exposed). Based on this exposure distribution and the uncertainties in its estimations, the best guess for the average exposure in Finland (or among those exposed) is estimated together with its uncertainty. The uncertainty in the average exposure is expressed by almost lowest possible (5th percentile) and the almost highest possible (95th percentile) value for the average exposure (i.e. as a distribution, more details below).

To be able to calculate attributable cases, in addition to current average exposure, one needs to determine, which is the lowest feasibly achievable average exposure in Finland. For several substances this is not zero, e.g. there is a natural background for particulate air pollution.

Second, exposure/dose-response functions and their uncertainties (5th and 95th percentiles, as above) are derived for all outcome with sufficient data. It is important that the exposure/dose-response function uses exactly the same exposure/dose marker that was used in the exposure estimation above (more details below).

Third, attributable number of cases is estimated by multiplying the exposure difference with exposure/dose-response and number of exposed using probabilistic methods (Monte-Carlo)

List of potential exposures that could be considered in the evaluation

Criteria 1) Public health effects 2) Concern 3) High exposures in specific groups, e.g. occupational exposures

• Alcohol, metanol
• Particulate air pollution by source
• Ozone
• CO and NO2 (probably included in the above)
• PAHs* Environmental tobacco smoke
• CO indoors
• Benzene* Formaldehyde, naphthalene, hexane, asetaldehyde
• 1,3-butadiene
• Lead
• Damp housing
• Noise* Foodborn epidemics
• Waterborn epidemics
• Chlorination by-products
• Arsenic
• Fluoride
• Dioxin, PCBs, phtalates
• Methyl mercury, mercury
• Radon
• UV radiation
• EMF
• Man-made radiation
• Chemicals in food (acrylamide, pesticides, food additives)
• Accidents (traffic, occupational, domestic/other)

List of selected exposures and responsible persons

• Indoor radon, STUK/Päivi Kurttio and ?
• Fine particles, KTL/Juha Pekkanen, Olli Leino?
• Formaldehyde, TTL, Eero Priha
• Alcohol, STAKES/Timo Ståhl
• Acrylamid, EVIRA/Tero Hirvonen
• Occupational disease, wood dust, TTL/Timo Kauppila

• Dioxins, KTL/Juha Pekkanen, Olli Leino?
•  ?Damp housing, KTL/Aino Nevalainen, Ulla Haverinen
• Arsenic, KTL/Hannu Komulainen
•  ?Methyl mercury, KTL/Juha Pekkanen, Olli Leino?

Methods

Exposure estimation

Below link to the exposure assessment done for the earlier exposure seminar of our group http://www.ttl.fi/Internet/Suomi/Aihesivut/Kemikaaliturvallisuus/Valittua+kemikaalitietoa/suomaltkemsateil.htm

Occupational exposure data is based on environmental measurements, modelling and other information on exposure factors (e.g. exposure times, skin contact, consumption data etc.). In workplaces exposure varies considerably depending on the sector of industry and even within a factory depending on work task. Also exposure level also changes often during a work day.

The following table outlines some occupationally important exposures on Finnish workplaces:

Dose-response assessment

Dose-response information can based on epidemiological studies or in toxicological tests. Epidemiological data is normally preferred to animal tests. However, in most cases dose-response data (e.g. IRIS database) is based animal tests and consequent calculations. In some cases epidemilogical data is suitable for quantitative estimates of risk. In majority of cases, the available epidemiological studies are inadequate and then cancer risks assessment is based on animal bioassays.

Estimation of DALYs

Probabilistic risk assessment

Probabilistic exposure/risk assessment methods can be used to analyze the exposure and risk more closely. The needed parameters of the used formula/algoritm are given as distributions (e.g. log-normal, normal, uniform, triangular) and the result of simulation are also distributions. Thus different recentiles of risk are obtained.

Exposures and dose-responses for selected exposures

Indoor radon

Target population: All Finnish subjects
Average radon concentration in Finland (mean; 5th; 95th percentile): 123; 20; 320 Bq/m3
Number of exposed persons
Table 1. Distribution of indoor radon concentration in homes in Finland (Mäkeläinen et al. 2005).

Radon concentration Bq/m3   

% of population     Category Mean Bq/m3
0-99 63 58
100-199 23 138
200-399 10 274
400 - 4 842

There are differences in Rn concentrations in different parts of the country (figure):

xxx

Dose-response of Rn and lung cancer
Based on Darby et al. 2004 and 2006, the relative risk estimate for 100 Bq/m3 changes in Rn concentration increases the lung cancer incidence by 16% (95% CI 5%–31%).
ERR 0.16; 95%CI 0.05-0.31
The risk of a non-smoker developing lung cancer before their 75th birthday increases from 0.5 % to 0.7 % if the radon concentration of the indoor air is in the range of 100 Bq/m3 to 400 Bq/m3. The corresponding figures for those who smoke are an increase from 12 to 16 %. Living for 30 years in a place which has a radon concentration of 800 to 1 400 Bq/m3, doubles the chance of developing lung cancer, compared with a dwelling which has a concentration of less than 100 Bq/m3 (Figure below).

Lung-cancer incidence in Finland(year 2006) [1]: (M 31.2 + F 10.0=) 41.2 cases/100 000 pyrs, (M 1522 + F 611=) 2133 new lung cancer cases in 2006 in Finland.

Number of excess deaths'''' in Finnish population: Indoor Rn is associated with 200-400 lung cancer deaths per year in Finland.
Of 1000 non-smokers who live in a dwelling which has radon concentration ranging from 100 to 400 Bq/m3, about 5 will develop lung cancer. Of 1000 smokers who live in a dwelling which has a high concentration of radon, about 160 will develop lung cancer. (Darby et al. 2006).

Table 2. Number of lung cancer deaths caused by radon in homes in the year 2003 in Finland (Mäkeläinen et al. 2005).
Lifetime indoor radon concentration (Bq/m3) Males Females Total
0-99 102 32 134
100-199 69 24 93
200-399 58 23 81
400- 64 28 92
Total 293 107 400

References
Darby S, Hill D, Auvinen A, Barros-Dios JM, Baysson H, Bochicchio F, Deo H, Falk R, Forastiere F, Hakama M, Heid I, Kreienbrock L, Kreutzer M, Lagarde F, Mäkeläinen I, Muirhead C, Obereigner W, Pershagen G, Ruano-Ravina A, Ruosteenoja E, Schaffrath-Rosario A, Tirmarche M, Tomasek L, Whitley E, Wichmann H-E, Doll R. Radon in homes and lung cancer risk: collaborative analysis of individual data from 13 European case-control studies. British Medical Journal 2005; 330: 223–226.

Darby S, Hill D, Deo H, Auvinen A, Barros-Dios JM, Baysson H, Bochicchio F, Falk R, Farchi S, Figueiras A, Hakama M, Heid I, Hunter N, Kreienbrock L, Kreuzer M, Lagarde FC, Mäkeläinen I, Muirhead C, Oberaigner W, Pershagen G, Ruosteenoja E, Schaffrath Rosario A, Tirmarche M, Tomášek L, Whitley E, Wichmann H-E, Doll R. Residential radon and lung cancer – detailed results of a collaborative analysis of individual data on 7148 persons with lung cancer and 14 208 persons without lung cancer from 13 epidemiologic studies in Europe. Scandinavian Journal of Work, Environment & Health 2006; 32 Suppl 1: 1–84.

Mäkeläinen I, Arvela, H, Kurttio P, Auvinen A. Number of lung cancer deaths caused by radon in Finland. In: Valentin J, Cederlund T, Drake P, Finne IE, Glansholm A, Jaworska A, Paile W, Rahola T (eds). Radiological Protection in Transition – Proceedings of the XIV Regular Meeting of the Nordic Society for Radiation Protection, NSFS – Rättvik, Sweden, 27–31 August 2005. SSI Report 2005:15. Stockholm: Swedish Radiation Protection Authority; 2005. p. 203–206.

Alcohol (ethanol)

Fine particles

Target population: All Finnish subjects aged 35 and above

Average, population weighted, PM2.5 concentration: So far we don’t have any available data on fine particles concentrations in Finland. The population average concentration can be estimated to be somewhere between 5 and 10 µg/m3 based on measurements done in Helsinki.

• Population average concentration (mean; min-max estimate mean): 7.0; 6.0-8.0
• Distribution of population PM exposure 5%; 50%; 95%;: 2; 6; 12;

All non-accidental mortality in Finland (year 2000): Mortality counts can be estimated from WHO-mortality database (http://www.who.int/healthinfo/morttables/en/index.html). The number of non-accidental deaths in Finnish subjects aged 35 and above in year 2000 was ?.

Plausibility of the causal association between fine particles and all-cause mortality:Estimate here 100%

Concentration-response function for PM2.5 (RR/(10ug/m3)): Based on Pope et al. 2002 study, the relative risk estimate for 10 µg/m3 changes in PM2.5 concentration is 1.06 and 95% confidence interval 1.02 – 1.11.

• Relative risk for PM2.5 (mean; min; max): 1.06; 1.02; 1.11

Low exposure/threshold: Although current scientific knowledge has not revealed any threshold value for fine particles, for the analysis we use threshold value 5 µg/m3.

• Threshold value: 5 µg/m3

Exposure above threshold: This is the average fine particles exposure above the threshold limit. Estimated by subtracting the threshold from the average exposure.

• Population average concentration above threshold (mean; min; max): 2.0; 1;0; 3.0

Relative risk above threshold: Relative risk estimate for the given exposure. This is estimated from RR-values with formula exp(ln(RR)/10*exposure), that is, for mean estimate exp(ln(1.02)/10*2.0.

• Relative risk above threshold (mean; min; max): 1.01; 1.00; 1:03

Attributable risk above threshold: Percentage change in all-cause mortality due to exposure. Estimate from Relative Risk above threshold with formula -> (RR-1)/RR

• Attributable death above threshold (mean; min; max): 1.2%; 0.2%; 3.1%

Number of excess deaths in Finnish population: Mortality with the current information.

• Number of excess deaths in Finnish population (mean; min; max): 877; 150; 2333

The excel sheet with the calculations: http://heande.pyrkilo.fi/heande/images/5/5a/Tainio_Finland.xls

Formaldehyde

Formaldehyde has been classified as a human carcinogen by WHO/IARC (2004). Formaldehyde causes sinonasal cancer in rats (Kerns 1983) already at exposure levels (2-6 ppm) close those occurring at workplaces in the 1970s. In the epidemiological studies, there was sufficient evidence thatformaldehyde causes nasopharyngeal cancer and limited evidence that formaldehyde causes sinonasal cancer in humans.

Dose-response data

US EPA has established a unit risk for formaldehyde which is 1.3x10-5 per ug/m3. This unit risk (for sinonasal cancer) is based on animal inhalation tests and linearized multistage (LMS) model (upper bound extrapolation from 10% response level). Unit risk means lifetime excess risk at exposure level of 1 ug/m3. The unit determined for formaldehyde can be applied to exposure level under 800 ug/m3.

The most important epidemiological study (Hauptman 2004) found SMR 2.1 (1.05-4.21)to nasopharyngeal cancer in workers exposed in formaldehyde/resin industry. For sinonasal cancer excess was less clear, SMR=1.19 (0.38-3.29).

It is less clear, if  residential formaldehyde exposures cause cancer in humans.  Vaughan et al. (1986) found an OR=2.1 (0.7-6.6) of nasopharyngeal cancer for the residents of mobile homes. The formaldehyde concentrations of the  American mobile homes were commonly between 100-300 ug/m3 in the 1980s.

Exposure

People are widely exposed to formaldehyde in homes and at workplaces in Finland. Exposure levels have decreased to less formaldehyde emitting products (particle board, MDF board, textiles, wallpaper etc.). It has been evaluated that about 10000-12000 Finns are exposed to formaldehyde at workplaces (exposure level over background). Of these workers about 2000 has been evaluated (in 2003) to be more heavily exposed so formaldehyde concentration in air may at least occasionally reach the current occupational exposure limit which is 0.3 ppm as 8.h TWA. Background exposure level in Finland is today less than 1 ug/m3 in countryside but it may exceed this (max about 5 ug/m3) in cities due to traffic etc. The highest average exposure levels are in formaldehyde and resin glue industry, in woodboard factories, in woodworking and furniture industries as well as in foundries using furane resin.

In Finnish family dwellings formaldehyde levels are currently mainly between 20-80 ug/m3 The level is higher in single family houses being (average about 45 ug/m3)and than in block houses (20 ug/m3 ). The concentrations have considerably decreased since the beginning of the 1980s. The current  Finnish limit value for indoor air is 100 ug/m3.

Risk calculations

Cancer risk (lifetime, excess risk) wascalculated according to the following formula (deterministic point estimate):

Risk = C x EDx'UR, where

C=average exposure ug/m3, ED= exposure duration as proportion of life-time ( working life = 8h/day, 220 days/year, 40 years)

UR= unit risk,  lifetime excess risk of specific cancer at 1 ug/m3 exposure level

Cancer cases = Risk x Number of e

xposed

'Results

The following risks were obtained by combining exposure and unit risks:

Table 1.Calculated excess risk of occupational sinonasal and number of cases using KymCAREX-database and IRIS unit risk estimates (1.3x10^-5 per µg/m³)

Table 2. Calculated risk levels and cancer cases from formaldehyde among non-occupationally exposed populations based on median exposure levels in the beginning of the 1980s and in the end of the 1990s.

Supposed proportion of time spent indoors 65%

Comparison with cancer statistics

According to Finnish cancer registry statistics the annual number of new pharynx cancer cases has been between 79-98 (1997-2003) and sinonasal cancers 35-45 (1997-2003). Most sinonasal and pharyngeal cancers are diagnosed in the age groups of 50-65 years.

The attributable fraction (AF) of nasopharyngeal cancer is about 0.55 (RR=2.1, Hauptman et al., 2004 and Vaughan, 1987). This suggests that there should be 43-54 npharynx cancers and 19-25 sinonasal cancers annually in Finland. These figures are relatively close to the above exposure based calculations.

References

US EPA,IRIS (Integrated Risk Information System) database. Formaldehyde.

Hauptman et al. (2004) Mortality from workers in formaldehyde industries. Am J Epidemiol 159, 1117-1130.

Vaughan et al. (1986)

Formaldehyde and cancers of the pharynx, sinus and nasal cavity: II Residential exposures. Int J Cancer 38, 685-688.

Chemical agents causing occupational diseases

 In rare cases, it is possible to obtain data on attributable cases directly without any information on exposure distributions or dose-response relationships. One such example is the Finnish Register of Occupational Diseases (FROD).