Risk Analysis, Vol. 19, No. 6, 1999

Health Risk Assessment of a Modern MunicipalWaste Incinerator

Celine Boudet,1 Denis Zmirou,1 Mauricette Laffond,2 Franck Balducci,1

and Jean-Louis Benoit-Guyod3

During the modernization of the municipal waste incinerator (MWI, maximum capacity of180,000 tons per year) of Metropolitan Grenoble (405,000 inhabitants), in France, a riskassessment was conducted, based on four tracer pollutants: two volatile organic compounds(benzene and 1, 1, 1 trichloroethane) and two heavy metals (nickel and cadmium, measuredin particles). A Gaussian plume dispersion model, applied to maximum emissions measuredat the MWI stacks, was used to estimate the distribution of these pollutants in the atmospherethroughout the metropolitan area. A random sample telephone survey (570 subjects) gathereddata on time-activity patterns, according to demographic characteristics of the population.Life-long exposure was assessed as a time-weighted average of ambient air concentrations.Inhalation alone was considered because, in the Grenoble urban setting, other routes ofexposure are not likely. A Monte Carlo simulation was used to describe probability distribu-tions of exposures and risks. The median of the life-long personal exposures distribution toMWI benzene was 3.2 · 1025 eg/m3 (20th and 80th percentiles 5 1.5 · 1025 and 6.5 · 1025 eg/m3), yielding a 2.6 · 10210 carcinogenic risk (1.2 · 10210–5.4 · 10210). For nickel, the correspondinglife-time exposure and cancer risk were 1.8 · 1024 eg/m3 (0.9.1024 – 3.6 · 1024 eg/m3) and8.6 · 1028 (4.3 · 1028–17.3 · 1028); for cadmium they were respectively 8.3 · 1026 eg/m3 (4.0 · 1026

– 17.6 · 1026) and 1.5 · 1028 (7.2 · 1029 – 3.1 · 1028). Inhalation exposure to cadmium emitted bythe MWI represented less than 1% of the WHO Air Quality Guideline (5 ng/m3), whilethere was a margin of exposure of more than 109 between the NOAEL (150 ppm) andexposure estimates to trichloroethane. Neither dioxins nor mercury, a volatile metal, weremeasured. This could lessen the attributable life-long risks estimated. The minute (VOCsand cadmium) to moderate (nickel) exposure and risk estimates are in accord with otherstudies on modern MWIs meeting recent emission regulations, however.

KEY WORDS: Municipal waste incineration; risk assessment; Monte-Carlo simulation; time activitypatterns.

1. INTRODUCTION

Incineration of municipal wastes is a widely usedmeans to manage solid wastes in large communities.It tends to take the place of landfilling, lessening area

1 Public Health Laboratory, GEDEXE, Grenoble University Med-ical School. F-38700 La Tronche, France, and INERIS.

2 GRECA (Applied Chemistry Research Group), Grenoble Uni-versity.

3 GEDEXE, Grenoble University School of Pharmacy.

12150272-4332/99/1200-1215$16.00/1 1999 Society for Risk Analysis

constraints in highly urbanized areas and growingconcerns about hazards posed to future generationsby landfills.(1) However, incineration also is hotly de-bated.(2–4) While it is a means to extract value fromwastes through energy production, social fear is keptalive by the number of studies reporting emissionsof heavy metals, acid gases and chlorinated organiccompounds. Modern procedures to remove dust inthe stacks have drastically reduced emissions of met-als. Sorting and recycling of plastics in municipalwastes have also contributed to abatement of chlorine

1216 Boudet et al.

emissions. However, both metals and dioxins/furansare still of serious concern among environmentalistsand the public.(5–8)

Assessing health impacts of incineration is diffi-cult because the effects, if any, are delayed, nonspe-cific and weak. Epidemiology has not proved to bevery effective. A retrospective cohort study suggestedan association between residence close to one of 72municipal solid waste incinerators in Great Britainand the incidence of cancer after 10 to 12 years,(9)

but this relationship could well be confounded bysocioeconomic factors since communities living in thevicinity of incinerators tend to belong to poorer socialcategories. A three years prospective follow-up de-sign was used to evaluate the short-term impact ofincinerators of biomedical municipal and hazardouswastes.(10) It failed to show any significant effect onan array of health indices (symptoms, respiratoryfunction or nasal lavage).

In this context, risk assessment is a reasonablealternative. Although it might be viewed as a puretheoretical exercise, it offers the advantage of antici-pating long-term putative consequences of MWIemissions and provides the basis for informed riskmanagement decisions and risk communication.(11)

This approach was followed in a number of set-tings.(12–16)

In France, many of the 300 municipal waste in-cinerators (MWI) currently operating are out of date.It is expected that more than 100 new MWI will bebuilt within 5 to 10 years, in substitution of about thesame number of old, smaller units. The Grenoblemetropolitan area local authorities decided in 1990to transform the furnaces and stacks of the old (1971)incinerator, in order to meet the growing demand forhouse-hold waste destruction, and for environmentpreservation. A risk assessment was conducted inorder to inform both local and national authoritiesabout risks associated with the new incineratorequipped with modern pollutant-removal devices.

This paper presents the results of this risk assess-ment and discusses its sensitivity to improvementsin exposure assessment by considering time-activitypatterns of the population within the study area.

2. MATERIAL AND METHOD

The Grenoble MWI was originally built in 1971in the periphery of Metropolitan Grenoble. Due topopulation expansion, it is now located within adensely populated zone. The total population of the

metropolitan area is 405,000 inhabitants, of whomabout half live within 5 km of the incinerator. Domi-nant winds direct stack emissions towards the citycenter, due to the mountainous topography of thenearby area (mountain levels of about 800 metersoccur above the urban area). In 1993–1995, the plantwas extended and extensively transformed. House-hold refuse is, in part, sorted by families (about 30%of the total mass) and also in the plant (paper, paste-board, aluminum, iron, plastics and glass are re-claimed for the market). Three incinerators (withthermal energy reclamation) burn the remainingwaste (102,000 tons in 1995, maximum capacity 5180,000 tons per year). The 50 m high stacks areequipped with electrofilters (99.6% maximum effi-ciency on particles mass) and liquid lime and sodascrubbers (for acid gas removal).

Stack emissions were characterized in the springof 1995. Table I also presents data from one stackbefore renovation of the process (1994), showingemissions abatement ranging from 1 to 3 orders ofmagnitude for the criteria pollutants that were mea-sured.

Among the specific pollutants quantified at theMWI stacks in 1995, four pollutants were selectedfor the present risk assessment: two VOCs (benzeneand 1,1,1 trichloroethane) and two heavy metals(nickel and cadmium). The criteria for choosing theserisk tracers were: the amount emitted (in favor ofbenzene), the specificity of MWI emissions relative tourban background levels (in favor of trichloroethaneand cadmium), their well characterized toxicity (car-cinogenic [benzene, nickel, cadmium], or/andnoncarcinogenic [trichoroethane, cadmium]), withpublished unit risks or NOAEL/reference doses.Other important risk tracers, such as dioxins andfurans were not measured, due to the complexityof their sampling and analysis (only a very limitednumber of laboratories had mastered the analysistechniques in France, at the time of the study); be-cause metals had only been measured in the particle

Table I. Typical concentrations of criteria pollutantsa of theGrenoble MWI (mg/Nm3)

Before modernization After modernization

Particles 35 4.7HCL 1100 3.5SO2 220 15.5Metalsb 250 .15

a French standards: Arrete on MWI, 25/1/1991.b Pb 1 Cr 1 Cu 1 Mn.

Incineration Risk Assessment 1217

Table II. Concentrations of Some Specific Pollutants Emitted bythe Grenoble MWI in 1995 (Average, [Maximum])

VOCs (pptv) Metals (eg/m3)

Trichloroethane 420 [615] Zn 24.8 [25]Dichloromethane 230 [442] Ni 5.3 [6.1]Xylene 169 [286] Cr 4.2 [5.0]Benzene 150 [350] Pb 5.2 [8.9]Dichlorobenzene 142 [206] Cu .33 [1.0]Ethylbenzene 140 [497] Cd .22 [.3]

phase, mercury was not selected either, since it ismostly emitted in the gaseous phase.(17) VOCs wereanalyzed by thermodesorption cold trap and gas spec-trometry; metals were analyzed by atomic adsorptionspectrometry. Table II summarizes emission concen-trations for some VOCs and metals, including ourtracer pollutants.

Exposure assessment combined dispersion mod-eling and evaluation of time-activity patterns.(18) Inha-lation was the only route of exposure that was consid-ered. A Gaussian plume model(19) (POLAIR), whichis a simplified version of ICAIR 3 C built by theFrench Atomic Energy Agency CEA(20) was used toestimate annual ambient air concentrations attribut-able to the MWI throughout the metropolitan area(a grid of 2.5 3 2.5 km was selected, with unit cellsof 500 3 500 m). Meteorology parameters that wereinput in the model came from a local station 500meters away from the MWI. In order to computeconservative risk estimates, the dispersion model wasapplied to maximum stack emission concentrationsfrom Table II.

People commute within the city. Therefore, air-borne exposure is not accurately characterizedthrough place of residence. A random sample tele-phone survey was conducted with 570 subjects inorder to describe activity patterns of the populationand the time spent in different locations of the Metro-politan area, according to demographic characteris-tics (age [5 categories: ,10 years; 10–19; 20–39;40–59; 601] and gender). Four main types of activi-ties were investigated: at home, at work (includingschool/university), recreation, other key activity last-ing more than 1 hour a day. While the dispersionmodel provides data for areas defined with regulargeometric ordinates (squares), administrative bor-ders within Metropolitan Grenoble are more com-plex. Resident population sizes were estimated ineach exposure grid cell, based on census data. Whengrid cells overlapped different administrative units,populations were assigned according to the relative

unit areas. For other activities than residence, thepopulation was apportioned in the different grid cellsbased on the geographical distribution described bythe telephone survey. Since an individual’s exposureis a time-weighted average of attributable ambient airconcentrations prorated by the time spent in differentgrid cells for his/her daily activities, a time-weightedaverage was computed for the assessment of the indi-vidual life-long exposure. Due to the great variabilityof penetration rates of outdoor pollutants, indoorattributable concentrations (at work or at home)were assumed equal to the corresponding grid cell’sestimated outdoor air concentration, for each activ-ity; this tends to maximize indoor attributable expo-sures.

Let - Cijk be ambient (and assumed indoor) air con-centration of a given tracer pollutant duringactivity i, of a subject of age category j, in gridcell k;

- Nijk be the number of corresponding subjects;- Tijk be the relative time (daily hours) spent in

cell k for activity i by a subject of age category j;

For the age category j, the average daily populationexposure across activities i and cells k is:

Ej (eg/m3 · person) 5 1/(24 3 100) ·OiO

kNijk · Cijk · Tijk

4

The life-long individual exposure due to the MWIwithin Metropolitan Grenoble is:

E (eg/m3) 5 365 ·Oj

Ej · rj/N

where rj is the breadth (in years) of age category j(i.e. 0–9 years or 4–59 years; for the upper openended age category, the age limit was set at 70 years),and N is the Metropolitan Grenoble total population.

In order to incorporate sources of variability,these life-long individual exposure estimates werecomputed using a Monte Carlo simulation (@Risk,Palisade Corp.) where age and activity-specific statis-tical distributions of time spent across MetropolitanGrenoble were input (general distributions from thepopulation sample survey, 100 iterations, Latin hy-percube simulation).(21) Results are expressed interms of percentiles (20th, 50th and 80th), meaningthat, under the assumptions that were made, there isa 60% probability that the estimated exposure liesbetween the 20th and 80th percentiles.

The risk characterization steps followed a differ-ent rationale according to the nature of the health

4 24 stands for 24 hours; 100 stands for 100 grid cells.

1218 Boudet et al.

effect of the tracer pollutant. For carcinogens (ben-zene, nickel, cadmium), life-long exposures weremultiplied by the corresponding unit risk.(22) The can-cer unit risk for inhalation of benzene(23) is 8.3 · 1026

(eg/m3)21; for nickel(24) it is 4.8 · 1024 (eg/m3)21 andfor cadmium(25) it is 1.8 · 1023 (eg/m3)21; exposure totrichloroethane will be compared to the ATSDRNOAEL(26) (150 ppm) which was based on cardiovas-cular, neurological, kidney and hematological effects.Although cadmium is a carcinogen,(25) this risk assess-ment also compares its attributable exposure to therecent WHO Office for Europe Air Quality Guide-lines(27) of 5 ng/m3.

3. RESULTS

Typical time-activity patterns of the Grenoblepopulation are presented in Table III showing slightdifferences according to gender, and important dis-crepancies with age. Thereafter, gender was not con-sidered in the analysis. An example of the exposurecell apportionment of the population, based on thesetime-activity patterns, is presented in Table IV (cellk/E8).

Figure 1 displays the spatial distribution of nickelambient air concentrations attributable to the MWI,after the dispersion model (based on local autumnaverage meteorology parameters in 1990–94), show-ing that the influence of the MWI is negligible beyond2.5 km. Attributable concentrations are very small(maximum 2 · 1024 eg/m3 for nickel and 1 · 1025 eg/m3 for cadmium), especially for benzene and trichlo-roethane (maximum 4.1025 eg/m3 and 1 · 1024 eg/m,3 respectively).

Therefore, life-long exposure and risk estimatesare minute to moderate (Table V). The 80th percen-tile of excess risk for benzene is 5.4.10210; it is 3.1 · 1028

Table III. Time Activity Patterns in the Study Population: Average [Standard Error] Time Spent During 24 Hours in DifferentActivities (In Minutes)

Total Female Male 0–9 10–19 20–39 40–59 601 year

At home 989 1019 952 1028 888 899 963 1216[283] [285] [277] [186] [254] [289] [263] [242]

At worka 252 251 253 260 369 307 300 0[233] [228] [240] [211] [153] [340] [234] [0]

Leisure 88 77 99 69 90 89 106 80[54] [43] [61] [31] [41] [46] [91] [60]

Other 56 66 43 84 28 58 40 60[104] [115] [89] [160] [36] [104] [61] [77]

a Including at school or university.

for cadmium: time-weighted life-long exposures tocadmium attributable to the MWI represent no morethan 1 % of the proposed Air Quality Guideline ofthe European Office of WHO.(27) There is a at leasta 5 · 109 ratio between the ATSDR NOAEL and theattributable ambient air concentrations for 1,1,1 tri-chloroethane. The 80th percentile of excess risk fornickel is a bit greater: 1.7 · 1027; 5% of the populationincur a life-long cancer risk greater than 2.2 · 1027

(95th percentile). As an illustration, Fig. 2 displaysthe probability distribution of the life-long individualexposure due to nickel attributable to the MWI.

To illustrate the gain due to incorporating time-activity patterns into the process of airborne expo-sure assessment, individual exposures to nickel, andthe associated risks, were also estimated assumingthat residents were exposed to MWI’s pollutants onlywhile residing in their houses where they were sup-posed to stay 24 hours a day: on average, over thewhole percentile distribution, life-long exposure andrisk are 35% greater than the corresponding estimateswhen mobility is taken into account. In such a simplis-tic scenario, no Monte-Carlo simulation can be runsince exposure is fully determined by the place of resi-dence.

4. DISCUSSION

The main results of this study are the following:(1) in the setting of a MWI equipped with modernpollutants removal devices, stacks emissions contrib-ute to a very small extent to the general populationexposure to the tracers that were evaluated. Mostof the corresponding risks are minute, if any, andcertainly indistinguishable by the way of conven-tional epidemiological investigations; (2) the expo-sure estimates based on time-weighted average ac-

Incineration Risk Assessment 1219

Table IV. Distribution of the Study Population in an Exposure Cell (Example) According to Age and Activity Categories

Grid cell E8 Total 0–9 Years 10–19 Years 20–39 Years 40–59 Years 160 Years

Home 3330 310 391 1217 763 649Work place 4339 357 496 1278 2208 0Other 4091 233 233 976 1395 1254Leisure 684 342 0 0 0 342

cording to daily activities were not dramaticallyaltered when the more simplistic approach consider-ing only place of residence was followed.

Inhalation was the only route of exposure thatwas studied. This can be seen as a flaw in our riskassessment. The land usage and the wind directionin the Grenoble metropolitan area (surrounded bymountains) are not in favor of the contamination ofthe food chain; agriculture constitutes less than 2% ofthe regional human activities and takes place upwind,relative to the MWI. Also, the Isere river that flowsthrough Grenoble is hardly used for fishing in thisarea. Indirect routes of exposure should be consid-ered in settings where cattle would graze, crops wouldbe grown on fields or fish would be captured on lakesor rivers influenced by atmospheric deposition of pol-lutants emitted by a MWI.(16,28–30) However, such emis-sions were not shown to influence significantly theburden of pollutants in the food chain in the vicinityof modern facilities.(30–35) Moreover, the origin of foodsupplies in large cities is wide-spread, only a fractioncoming from surrounding stock farms or crops fields.The exposure through food goods, even in a long-term transport perspective, would thus be seriouslydiluted. Conventional ‘‘worst cases’’ of subjects

Fig. 1. Modeling of maximum ambient air concentrations of nickel(eg/m3 3 1025) attributable to the Grenoble MWI* (based on theautumn wind-rose from 1990 to 1994).

whose diet would be entirely based on local productsis highly unrealistic in the specific setting of the Gre-noble MWI. However, soil ingestion by children hasnot been taken into account in this assessment, thusunder-estimating the exposures and risks.

In the absence of environmental monitoring orbiological exposure data, only statistical models allowthe assessment of air pollutants environmental con-centrations.(36) The Gaussian type model that wasused in this case gives accurate results for passivegases and fine particles as well as for VOCs in anideal plain area. Grenoble topography is somewhatunusual because of its mountainous environment;hence, the true spatial distribution of pollutants dif-fers from the model estimates. Still, among the avail-able models, the Doury model that was used seemsmore relevant than Pasquill-Turner(19) in a mild windsituation like in Grenoble.

The choice of the tracer pollutants was based onseveral criteria: the intrinsic pollutant toxicity, theMWI source specificity, and the emitted quantity.Among the quantified components (VOCs and met-als in particulate phase), 2 organics (one carcinogen,one systemic toxicant) and 2 inorganics (both carcino-gens, with one showing also a systemic toxicity) werechosen. Trichoroethane seemed to be the most rele-

Fig. 2. Monte-Carlo probability distribution of the time-weighteddaily average of life-long individual exposure to nickel attributableto the MWI.

1220 Boudet et al.

Table V. Exposure Assessment and Risk Characterization Associated with the MWI Emissions

Benzene Nickel Cadmium Trichloroethaneunit 5 1025 eg/m3 unit 5 1024 eg/m3 unit 5 1026 eg/m3 unit 5 1025 eg/m3

Life-long exposure20th percentile 1.5 .9 4.0 4.050th percentile 3.2 1.8 8.3 8.380th percentile 6.5 3.6 17.6 17.2

Life-long cancer riska (1026) Ratio to guideline or NOAELb

20th percentile 1.2 · 1024 4.3 · 1022 12.5 · 102 2 · 1010

50th percentile 2.6 · 1024 8.6 · 1022 6.0 · 102 109

80th percentile 5.4 · 1024 17.3 · 1022 2.8 · 102 4.7 · 109

a 20th, 50th, and 80th percentiles of life-long cancer risk for cadmium are respectively (1026): 0.72 · 1022, 1.5 · 1022, and3.1 · 1022.

b Computed as the ratio of the guideline (or NOAEL) to the exposure level: NOAEL/exposure.

vant among chlorinated monoaromatic components;its toxicity has been well described in the literature.(26)

However, in spite of the fact that it is one of the mostMWI-specific and -emitted VOCs, it is also found invehicle traffic exhaust (due to organic lead additives).Benzene is the most well known of monocyclic aro-matics. It is a carcinogen(23) and is one of the mostemitted VOCs at the MWI stack. However, benzeneurban air concentrations are often greater than thosemeasured directly at the incinerator stack.

Chromium VI, mercury, nickel and cadmium areheavy metals classified as carcinogens, that are emit-ted by MWIs. In Grenoble, only total Cr was mea-sured (no speciation). The fraction of total Cr beingCr VI varies from 10 to 0.007% in the literature.(12)

Moreover, metals were only measured in particles.Mercury being mainly emitted in the gaseous phase,this explains the low concentrations found at thestacks (from 0.1 to 0.3 eg/m3), values that are notrelevant for this risk assessment. By contrast, nickelis mainly found in the particulate phase. Concentra-tions measured at the MWI stacks are great relativeto the other heavy metals, and MWIs are a prepon-derant Ni source in the global environment.(24) Whilethe most specific MWI metal is Zn, emitted in thegreatest amount (2.5 e/m3), it is less relevant in termsof public health. Among the systemic inorganic toxi-cants, lead is not specific of MWIs. Cadmium is abetter choice as a MWI tracer, yet emitted in muchsmaller amounts than Zn; it’s volatility is lower thanthat of Hg.(37) Its renal systemic effects are well de-scribed.(25) On the other hand, several studies alsodesignated Cd as the second largest source of carcino-genic risk in terms of incineration, after dioxins andfurans.(22) It is also classified as a carcinogen (class 1)by the International Agency for Research on Can-

cer(25) (IARC). Its estimated unit risk is 1.8 · 1023 (eg/m3)21, i.e. about five times greater than Ni.(38) Becausethe measured concentrations at the MWI stack were20 times smaller than for Ni, the associated cancerrisk for Cd is only one fifth of that attributable to Ni.

Our results cannot be extended to chlorinatedorganics such as PCDD/Fs or volatile metals becausethey were not measured at the time of our study.Dioxins, furans and mercury are seen as the mostrelevant MWI-associated pollutants in terms of bi-oaccumulation through the food chain.(33–34) However,most studies showed low risks, as modern technolo-gies drastically reduce these emissions.(39) Direct inha-lation exposure in the vicinity of a modern MWIrepresents less than 1% of the total dioxin-estimateddaily absorption, which is essentially associated withdiet.(16,40) Compared to background levels in the gen-eral population, in Germany, concentrations of diox-ins in human blood and milk in the proximity of aMWI did not indicate an enhanced body burden ofdioxins.(32) While 2,3,7,8 TCDD is classified as a car-cinogen by the IARC,(40) the issue of a reference doseor the value of its unit risk is still controversial. Re-cently, additional measurements were done to assessemissions of dioxins and furans at the GrenobleMWI; they showed values ranging from 0.6 to 4.8I-TEQ ng/m,3 i.e., concentrations that exceed by 6to 50 times the limit values set by the Ministry ofEnvironment in 1997 (0.1 ng/m3 for new MWIs).Based on the dilution factors observed for nickel,these emissions could result in a daily intake throughinhalation near 16 · 1022 I-TEQ pg/kg for an adultmost exposed (percentile 95), a value which is about10 times lower than the WHO allowable daily intake.Again, ingestion through contaminated soil dust isoverlooked by this computation.

Incineration Risk Assessment 1221

For this risk assessment, the focus was placedon the variability of direct exposure due to time-activity patterns of the population. Sources of vari-ability were accommodated using a Monte Carloanalysis. Little difference was found between expo-sure estimates based solely on the place of residence,or weighted according to time spent in other activitylocations (work, school, leisure practice . . .). Thisresult might not hold true in other urban settingsbecause it depends largely upon the geographical dis-tribution of land usage. In the Grenoble metropolitanarea, as in most French cities, there is a close overlapbetween residential buildings, work locations (plants,offices, shops and alike) and recreative areas. Hence,people commute between activity locations in a sortof random-like movement that results in a somehowhomogeneous geographical distribution. This is likelyto be different in regions or countries where the landusage is more sharply partitioned between residentialand work areas. This issue warrants further studies.

To conclude, this study, in accord with manyothers, shows that the relative contribution of mod-ern MWIs to population exposure to significanthealth-related pollutants is small. It provides dataand illustrates an approach that is useful for riskcommunication and management.

ACKNOWLEDGMENTS

This work was supported by a grant fromADEME (Agency for Environment and Energy,Paris, France); it was conducted within the collabora-tive group GRIDEC (Groupe de Recherche Interdis-ciplinaire sur les Dechets). The authors are gratefulto Prof. J. Evans (Harvard School of Public Health)for his thoughtful advice.

REFERENCES

1. Nat. Research Council, Environmental. Public Health andHazardous Wastes, Nat. Ac. Press, Washington D.C. (1993).

2. C. C. Travis and H. A. Hattemer-Frey, Health Effects of Munic-ipal Waste Incineration. CRC Press, Boca Raton, Fla (1991).

3. C. R. Dempsey and E. T. Oppelt, Incineration of HazardousWaste: A Critical Review Update, J. Air Waste Manage. Assoc.43, 25–73 (1993).

4. M. Gochfeld, Incineration: Health and Environmental Conse-quences, Mount Sinai J. Med. 62, 365–374 (1995).

5. L. Morselli, S. Zappoli, and S. Militerno, The Presence andDistribution of Heavy Metals in Municipal Solid Waste Incin-erators, Toxicol. Environ. Chem. 37, 139–145 (1993).

6. R. Sedman and J. Esperza, Evaluation of the Public HealthRisks Associated with Semi Volatile Metals and Dioxin Emis-

sions from Hazardous Waste Incinerators, Environ. HealthPerspect. 94, 181–187 (1991).

7. C. Rappe, Sources of Exposure, Environmental Concentra-tions and Exposure Assessment of PCDD and PCDF, Chemo-sphere 27, 211–225 (1993).

8. H. Fiedler, Sources of PCDD/PCDF and Impact on the Envi-ronment, Chemosphere 32, 55–64 (1996).

9. P. Elliot, G. Shaddick, I. Kleinschmidt et al., Cancer IncidenceNear Municipal Solid Waste Incinerators in Great Britain, Br.J. Cancer 73, 702–710 (1996).

10. C. M. Shy, D. Degnan, D. C. Fox et al., Do Waste IncineratorsInduce Adverse Respiratory Effects? An Air Quality and Epi-demiological Study in Six Communities, Environ. Health Per-spect. 103, 714–724 (1995).

11. V. T. Covello and M. W. Merkhofer, Risk Assessment Methods.Approaches for Assessing Health and Environmental Risks,Plenum Press, New York (1993).

12. A. Levin, D. B. Fratt, A. Leonard, R. J. F. Bruins, and L.Fradkin, Comparative Analysis of Health Risk Assessmentsfor Municipal Waste Combustors, J. Air Waste Manage. Assoc.41, 20–31 (1991).

13. W. H. Hallenbeck, S. P. Breen, and G. R. Brenniman, CancerRisk Assessment for the Inhalation of Metals from MunicipalSolid Waste Incinerators Impacting Chicago, Bull. Environ.Contam. Toxicol. 51, 165–170 (1993).

14. W. H. Hallenbeck, Health Impact of a Proposed Waste-to-Energy Facility in Illinois, Bull. Environ. Contam. Toxicol. 54,342–348 (1995).

15. T. Eickman, Environmental Toxicological Assessment ofEmissions from Waste Incinerators, Fresenius Environ. Bull.3, 244–249 (1993).

16. A. C. Cullen, The Sensitivity of Probabilistic Risk AssessmentResults to Alternative Model Structures: A Case Study ofMunicipal Waste Incineration, J. Air Waste Manage. Assoc.45, 538–546 (1995).

17. W. M. Schaub, Mercury Emissions from MSW Incinerators:An Assessment of the Current Situation in the United Statesand Forecast of Future Emissions, Resources, Conserv. Recycl.9, 31–59 (1993).

18. EPA, Guidelines for Exposure Assessment; Notice. Part II,Federal Register, 57, 104, 22888–22938 (1992).

19. F. Balducci, O. Grandamas, and D. Zmirou, Polair: Logicielet guide methodologique pour l’investigation epidemiologiqued’une pollution atmospherique ponctuelle, Rev. Epidem. SantePub. 43, 594–603 (1995).

20. B. Crabol and M. Montfort, Presentation du code de calculICAIR3C. de dispersion passive de polluants dans l’atmosph-ere. Note technique, IPSN, Fontenay aux Roses, France(1992).

21. EPA, Guiding Principles for Monte-Carlo Analysis. EPA/630/R-97/001, Washington D.C. (March 1997).

22. EPA, Proposed Guidelines for Carcinogenic Risk Assessment;Notice; Part II. Federal Register, 61, 79, 17960–18010 (1996).

23. ATSDR, Toxicological Profile for Benzene (update). U.S. De-partment of Health and Human Services Public Health Ser-vice, ATSDR, Atlanta (1996).

24. ATSDR, Toxicological Profile for Nickel (update). U.S. De-partment of Health and Human Services Public Health Ser-vice, ATSDR, Atlanta (1996).

25. IARC, Cadmium and Certain Cadmium Compounds. In:IARC Monographs on the Evaluation of the CarcinogenicRisk of Chemicals to Humans. Beryllium, Cadmium, Mercuryand Exposures in the Glass Manufacturing Industry, IARCMonographs, vol 58, Lyon, France p 119–146, 210–236 (1993).

26. ATSDR, Toxicological Profile for 1.1.1 Trichloroethane (up-date). US. Department of Health and Human Services PublicHealth Service, ATSDR, Atlanta (1995).

27. WHO, Air Quality Guidelines for Europe. European Epi-

1222 Boudet et al.

Marker, Kracow, Poland. (April 1997, the final publication ofthe WHO-Euro guidelines is pending).

28. C. C. Travis and B. P. Blaylock, Validation of a TerrestrialFood Chain Model, J. Expos. Anal. Environ. Epidemiol. 2,221–239 (1992).

29. R. M. Sedman, J. M. Polisini, and J. R. Esperza, The Evaluationof Stack Metal Emissions from Hazardous Waste Incinerators:Assessing Human Exposure Through Noninhalation Path-ways, Environ. Health Perspect. 102 (sup2), 105–112 (1994).

30. M. Schummacher, S. Granero, M. Belles, J. M. Llobet, andJ. L. Domingo, Levels of Metals in Soils and Vegetation inthe Vicinity of a Municipal Solid Waste Incinerator, Toxicol.Environ. Chem., 119–132 (1996).

31. B. D. Eitzer, Polychlorinated Dibenzo-p-dioxins and Dibenzo-furans in Raw Milk Samples from Farms Located Near a NewResource Recovery Incinerator, Chemosphere 30, 1237–1248 (1995).

32. E. Demi, I. Mamgelsfdorf, and H. Greim, Chlorinated Diben-zodioxins and Dibenzofurans (PCDD/F) in Blood and HumanMilk of Non Occupationally Exposed Persons Living in theVicinity of a Municipal Waste Incinerator, Chemosphere 33,1941–1950 (1996).

33. K. L. Hentz, S. W. Pirages, and K. A. Williams, Differences

in Indirect Risk Methodologies: Dioxin Emissions from Com-bustion Facilities, Regulatory Toxicology and Pharmacology20, 83–95 (1994).

34. H. Greim, Toxicological Evaluation of Emissions from Mod-ern municipal Waste Incinerators, Chemosphere 20(3/4), 317–331 (1990).

35. C. S. Krivanek, Mercury Control Technologies for MWCS—The Unanswered Questions, Journal of Hazardous Material47, 119–136 (1996).

36. NRC, Human Exposure Assessment for Airborne Pollutants:Advances and Opportunities, National Academy of Sciences,321 pages, Washington D.C. (1991).

37. E. T. Oppelt, Air Emissions from the Incineration of Hazard-ous Waste, Toxicol. Indust. Health 6, 23–51 (1990).

38. IRIS, Integrated Risk Information System, US-EPA, Wash-ington D.C. (1996).

39. J. D. Kilgroe, Control of Dioxin, Furan, and Mercury Emis-sions from Municipal Waste Combustor, Journal of HazardousMaterial 47, 163–194 (1996).

40. IARC Monographs on the Evaluation of Carcinogenic Risk ofChemicals to Humans, Volume 69: Polychlorinated Dibenzo-para-Dioxins and Polychlorinated Dibenzofurans, IARC,Lyon, France (1997).