CHARACTERIZATION OF VOLATILE ORGANIC COMPOUNDSIN SMOKE AT MUNICIPAL STRUCTURAL FIRES

C. C. Austin

Department of Epidemiology, Biostatistics and Occupational Health,Faculty of Medicine, McGill University, Montreal, Quebec, Canada

D. Wang

Air Toxics Section, Analysis and Air Quality Division, EnvironmentCanada, Ottawa, Ontario, Canada

D. J. Ecobichon

Department of Epidemiology, Biostatistics and Occupational Health,Faculty of Medicine, McGill University, Montreal, Quebec, Canada

G. Dussault

Division of Health, Safety and Tactical Strategies, City of MontrealFire Department, Montreal, Quebec, Canada

The objective of this study was to characterize volatile organic compounds (VOCs)found at municipal structural fires in order to identify sources of long-term health risks tofirefighters, which may be contributing factors in heart disease and cancer. Firefighterscollected air into evacuated Summa canisters inside burning buildings at nine municipalstructural fires under conditions where they judged that at least some firefighters mightremove their self-contained breathing apparatus masks. Volatile organic compoundswere identified and quantified for 144 target compounds using cryogenic precon-centration and gas chromatography/mass spectral detection (GC/MSD) methodologyoperating in selected ion monitoring mode. Samples were also analyzed in SCAN modeand examined for the appearance of substances that were not present in the instrumentstandard calibration mixture. The spectra of municipal structural fires were surprisinglysimilar and remarkable for their simplicity, which was largely due to the dominating pres-ence of benzene along with toluene and naphthalene. Propene and 1,3-butadiene werefound in all of the fires, and styrene and other alkyl-substituted benzene compounds werefrequently identified. Similar “fingerprints” of the same 14 substances (propene, benzene,xylenes, 1-butene/2-methylpropene, toluene, propane, 1,2-butadiene, 2-methylbutane,ethylbenzene, naphthalene, styrene, cyclopentene, 1-methylcyclopentene, isopropyl-benzene) previously identified at experimental fires burning various solid combustiblematerials were also found at municipal structural fires, accounting for 76.8% of the total

437

Journal of Toxicology and Environmental Health, Part A, 63:437–458, 2001Copyright© 2001 Taylor & Francis1528-7394 /01 $12.00 + .00

Received 8 September 2000; sent for revision 16 October 2000; accepted 4 January 2001.The authors thank the National Health Research and Development Program (NHRDP), Canada

(grant 6605409858), and the Institut de recherche en santé et sécurité travail du Québec (IRSST) fortheir support during this study.

Address correspondence to C. C. Austin, PhD, CIH, School of Engineering, University ofQuebec at Three Rivers, PO Box 571, Station A, Montreal, Quebec, Canada H3C 2T6. E-mail: caustin@sarec.org

VOCs measured. Statistically significant positive correlations were found between increas-ing levels of benzene and levels of propene, the xylenes, toluene, 1-butene/2-methyl-propene, 1,3-butadiene, and naphthalene. Given the toxicity/carcinogenicity of thoseVOCs that were found in the highest concentrations, particularly benzene, 1,3-buta-diene, and styrene, further investigation of VOC exposures of firefighters is warranted.Benzene, or its metabolic product s-phenylmercapturic acid in urine, was identified as asuitable chemical marker for firefighter exposure to combustion products.

Statistically significant associations have been found between fire-fighters and certain forms of cancer: bladder, brain and nervous system,colon, esophageal, genitourinary (aggregate), kidney, liver, lymphatic/hematopoietic, melanoma, multiple myeloma, non-Hodgkin’s lymphoma,and rectal and ureter cancers (Decoufle et al., 1977; Milham, 1983;Eliopoulos, 1984; Vena & Fiedler, 1987; Howe & Burch, 1990; Sama et al.,1990, Beaumont et al., 1991; Grimes et al., 1991; Demers et al., 1992,Aronson et al., 1993; Guidotti, 1993; Tornling et al., 1994). Small but statis-tically significant excesses of arteriosclerotic, coronary, ischemic, and over-all circulatory disease have been found (Bates, 1987; Feuer & Rosenman,1986; Grimes et al., 1991; Orris et al., 1995; Sardinas et al., 1986). The riskof cancer and heart disease in firefighters has been extensively reviewed(Golden et al., 1995; Guidotti, 1995; Melius, 1995). Although some of theepidemiological evidence suffers from a number of limitations, includinglow statistical power resulting from the small populations studied, possiblemisclassification, and limited documentation of exposure, the preponder-ance of evidence suggests that toxic exposures encountered at the firescene may increase the risk of cancer.

Firefighting consists of two phases, (1) knockdown, during whichthe fire is brought under control, and (2) overhaul, when the fire is extin-guished and cleanup begins. Although previous studies have measured anumber of toxic chemicals at municipal structural fires (Table 1), a com-prehensive assessment of firefighter exposures to substances having chroniceffects has yet to be performed. Austin (1997) verified the use of gas chro-matography and mass spectral detection (GC/MSD) analysis of air collectedin evacuated Summa canisters under the severe conditions (temperature, hu-midity, reactive atmospheres, multiple contaminants) in which firefightersroutinely work for the identification and quantification of volatile organiccompounds (VOCs) found at fires. When this method was used to sampleair at experimental fires burning various combustible material, chromato-graphic fingerprints apparently characteristic of combustion were ob-served (Austin et al., 2001a). Benzene, toluene, 1,3-butadiene, naphtha-lene, and styrene were found at higher concentrations than most otherVOCs, and it was suggested that benzene might be a suitable markercompound for the assessment of exposure to combustion products at fires.The objective of this study was to characterize VOCs found at municipalstructural fires in order to identify sources of long-term health risks to fire-fighters, which may be contributing factors in cancer and heart disease.

438 C. C. AUSTIN ET AL.

439

TAB

LE 1

.Po

tential F

irefighter Exp

osures to Com

bustion Prod

ucts

Num

ber

Com

poun

dCon

centration

SDSa

mples

aof fires

Type

of fire

Sampling metho

dAutho

r

Ace

taldeh

yde

n.db–8

.1 ppm

c—

d21

22Kno

ckdo

wn

Polymer tub

eJank

ovic et al. (19

92)

n.d.–0

.9 ppm

—4

22Inside

mask

Polymer tub

eJank

ovic et al. (19

92)

n.d–

1.6 pp

m—

522

Ove

rhau

lePo

lymer tub

eJank

ovic et al. (19

92)

0.34

ppm

0.41

9625

Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Acrolein

0.2–

15 ppm

—29

200

Mun

icipal

13× Siev

eBurge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)0.1–

4 pp

m—

1222

Kno

ckdo

wn

Polymer tub

eJank

ovic et al. (19

92)

0.2 pp

m—

122

Ove

rhau

lPo

lymer tub

eJank

ovic et al. (19

92)

1 pp

m—

122

Inside

mask

Polymer tub

eJank

ovic et al. (19

92)

0.12

3 pp

m0.13

396

25Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Ben

zene

0.2–

180 pp

m—

181/19

7f20

0Mun

icipal

Cha

rcoa

lBurge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)8–

250 pp

m—

18/25

14Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

9)n.d.–2

2 pp

m—

1522

Kno

ckdo

wn

Cha

rcoa

lJank

ovic et al. (19

92)

n.d.–0

.3 ppm

—2

22Ove

rhau

lCha

rcoa

lJank

ovic et al. (19

92)

n.d.–2

1 pp

m—

422

Inside

mask

Cha

rcoa

lJank

ovic et al. (19

92)

0.38

3 pp

m0.42

595

25Ove

rhau

lCha

rcoa

lBolstad

-Joh

nson

et a

l.

(200

0)

Ben

zaldeh

yde

0.05

7 pp

m0.05

796

25Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

(Tab

le con

tinue

s on

next pa

ge)

440

TAB

LE 1

.Po

tential F

irefighter Exp

osures to Com

bustion Prod

ucts (C

ontin

ued)

Num

ber

Com

poun

dCon

centration

SDSa

mples

aof fires

Type

of fire

Sampling metho

dAutho

r

Carbo

n diox

ide

1000

–75,00

0 pp

m—

8920

0Mun

icipal

Bag

—de

tector

Burge

ss et a

l. (197

9);

tube

Treitm

an et al

(198

0)35

0–54

10 ppm

—20

22Kno

ckdo

wn

Bag

—FT

IRJank

ovic et al. (19

92)

130–

1420

ppm

—7

22Ove

rhau

lBag

—FT

IRJank

ovic et al. (19

92)

460–

21,300

ppm

—28

22Inside

mask

Bag

—FT

IRJank

ovic et al. (19

92)

Carbo

n mon

oxide

3–10

00 ppm

—90

90Mun

icipal

Bag

—Ec

olyz

erGold et al. (197

8)15

–500

0 pp

m—

4920

0Mun

icipal

Bag

—Ec

olyz

erBurge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)n.d.–1

5,00

0 pp

m—

7575

Mun

icipal

Detec

tor tube

Lowry et a

l. (198

5)11

–108

7 pp

m—

2614

Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

9)5–

1900

ppm

—33

22Kno

ckdo

wn

Bag

—FT

IRJank

ovic et al. (19

92)

5–82

ppm

—7

22Ove

rhau

lBag

—FT

IRJank

ovic et al. (19

92)

<1–

105 pp

m—

622

Inside

mask

Bag

—FT

IRJank

ovic et al. (19

92)

n.d.–1

7 pp

m—

601

Forest fire

Detec

tor tube

Kelly (19

91)

52 6 ppm

6665

25Ove

rhau

lMetroso

nics

Bolstad

-Joh

nson

et a

l.4-ga

s meter

(200

0)

Carbo

xyhe

mog

lobin

2.76

–7.3%

—8

5-mo

Mun

icipal

Non

smok

ing

Sammon

s an

d mea

nsColem

an (1

974)

3.3%

(3.4–

13.2%)

1.6

33

Mun

icipal

Non

smok

ing

Loke

et a

l. (1

976)

2.45

%2.5

519

124

Mun

icipal

Non

smok

ing +

Rad

ford and

Lev

ine

mask

(197

6)1.4–

9.1%

—55

—Mun

icipal

Non

smok

ing

Stew

art e

t al. (1

976)

Form

alde

hyde

0.4–

8.3 pp

m—

6/24

14Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

8)n.d.–8

ppm

—16

22Kno

ckdo

wn

Polymer tub

eJank

ovic et al. (19

92)

n.d.–0

.4 ppm

—5

22Ove

rhau

lPo

lymer tub

eJank

ovic et al. (19

92)

n.d.–0

.3 ppm

—5

22Inside

mask

Polymer tub

eJank

ovic et al. (19

92)

0.25

ppm

0.25

296

25Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

441

Glutaraldeh

yde

0.04

6 pp

m0.04

9625

Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Hyd

roge

n ch

loride

18–1

50 ppm

—5/90

90Mun

icipal

13× Siev

eGold et al. (197

8)1–

200 pp

m—

6920

0Mun

icipal

13× Siev

eBurge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)0–

40 ppm

—7/75

75Mun

icipal

Detec

tor tube

Lowry et a

l. (198

5)2.17

–13.3 pp

m—

2/19

14Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

8)n.d.–8

.5 ppm

—2

22Kno

ckdo

wn

Silic

a ge

l tub

eJank

ovic et al. (19

92)

0.99

mg/m

31.1

9525

Ove

rhau

lORBO 53 tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Hyd

roge

n cy

anide

0.02

–0.89 pp

m—

43/90

90Mun

icipal

Ascarite

Gold et al. (197

8)0.1–

4 pp

m—

1320

0Mun

icipal

Ascarite

Burge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)n.d.–4

0 pp

m—

9/75

75Mun

icipal

Detec

tor tube

Lowry et a

l. (198

5)0.8-75

ppm

—10

/26

14Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

8)n d.–2

3 pp

m—

1222

Kno

ckdo

wn

Soda

lime tube

Jank

ovic et al. (19

92)

n.d.–0

.4 ppm

—3

22Ove

rhau

lSo

da lime tube

Jank

ovic et al. (19

92)

<1.0 mg/m

3—

2525

Soda

lime tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Hyd

roge

n flu

oride

0.2–

7 mg/m

3—

822

Kno

ckdo

wn

Silic

a ge

l tub

eJank

ovic et al. (19

92)

Isov

aleralde

hyde

0.07

ppm

0.03

896

25Ove

rhau

lDNPH

tube

Bolstad

-Joh

nson

et a

l.

(200

0)

Methy

lene

chloride

0.28

ppm

—1

14Mun

icipal

Cha

rcoa

lBrand

t-Rau

f et al.

(198

8)

Nitrog

en dioxide

0.02

–0.89 pp

m—

8/90

90Mun

icipal

13× siev

eGold et al. (197

8)0.2–

10 ppm

—33

200

Mun

icipal

13× siev

eBurge

ss et a

l. (197

9);

Treitm

an et al.

(198

0)0.13

ppm

0.21

6525

Ove

rhau

lMetroso

nics

Bolstad

-Joh

nson

et a

l.

4-ga

s meter

(200

0)

(Tab

le con

tinue

s on

next pa

ge)

442

TAB

LE 1

.Po

tential F

irefighter Exp

osures to Com

bustion Prod

ucts (C

ontin

ued)

Num

ber

Com

poun

dCon

centration

SDSa

mples

aof fires

Type

of fire

Sampling metho

dAutho

r

PAHs

Mixture

g0.26

2 mg/m

3—

322

Kno

ckdo

wn

Polymer tub

es +

Jank

ovic et al. (19

92)

filter

0.00

6 mg/m

3—

322

Ove

rhau

lpo

lymer tub

es +

Jank

ovic et al. (19

92)

filter

0.54

8 mg/m

30.01

9 88

25Ove

rhau

lORBO 43/PT

FE

Bolstad

-Joh

nson

et a

l.

(avg

)filter

(200

0)Ben

zo(e)pyren

e0.02

2 mg/m

3—

322

Kno

ckdo

wn

Polymer tub

es +

Jank

ovic et al. (19

92)

filter

0.00

1 mg/m

3—

322

Ove

rhau

lPo

lymer tub

es +

Jank

ovic et al. (19

92)

filter

Ace

naph

thylen

e0.45

1 mg/m

30.53

688

25Ove

rhau

lORBO 43/PT

FE

Bolstad

-Joh

nson

et a

l.

filter

(200

0)Nap

htha

lene

0.22

3 mg/m

30.10

188

25Ove

rhau

lORBO 43/PT

FE

Bolstad

-Joh

nson

et a

l.

filter

(200

0)

Particulates

4–75

0 mg/m

3—

2090

Mun

icipal

25 m

m glass filte

rGold et al. (197

8)20

–20,00

0 mg/m

364

200

Mun

icipal

Fibe

rglass filte

rBurge

ss et a

l. (197

9);

Trietm

an et al.

(198

0)10

.1–3

44 m

g/m

3—

514

Mun

icipal

Glass fibe

rBrand

t-Rau

f et al.

(198

8)n.d.–5

60 m

g/m

3—

422

Kno

ckdo

wn

Casca

de impa

ctor

Jank

ovic et al. (19

92)

n.d.–4

5 mg/m

3—

422

Ove

rhau

lCasca

de impa

ctor

Jank

ovic et al. (19

92)

443

Particulates (respirable)

370 mg/m

3—

251

Forest fire

PVC filter

Kelly (19

91)

8.01

mg/m

38.02

9325

Ove

rhau

lPV

C filter

Bolstad

-Joh

nson

et a

l.

(200

0)

Perchloroe

thylen

e0.06

4–0.13

8 pp

m—

314

Mun

icipal

Cha

rcoa

lBrand

t-Rau

f et al.

(198

8)

Sulphu

r diox

ide

0.4–

41.7 ppm

—12

/26

14Mun

icipal

Detec

tor tube

Brand

t-Rau

f et al.

(198

8)0.6–

3.0 pp

m—

261

Forest fire

Detec

tor tube

Kelly (19

91)

1.6 pp

m2.06

6525

Ove

rhau

lMetroso

nics

Bolstad

-Joh

nson

et a

l.

4-ga

s meter

(200

0)

Toluen

e0.16

–0.28 pp

m—

314

Mun

icipal

Cha

rcoa

lBrand

t-Rau

f et al.

(198

8)

Trichloroe

thylen

e0.11

2–0.18

1 pp

m—

214

Mun

icipal

Cha

rcoa

lBrand

t-Rau

f et al.

(198

8)

Trichlorop

heno

l0.1 pp

m—

114

Mun

icipal

Cha

rcoa

lBrand

t-Rau

f et al.

(198

8)

Xylen

e0.06

ppm

—3

22Kno

ckdo

wn

Cha

rcoa

lJank

ovic et al. (19

92)

Note.

Some of th

e stud

ies cited also sam

pled

other sub

stan

ces such

as metals an

d asbe

stos fibe

rs not m

entio

ned in th

e table.

a Num

ber of usable samples.

b Not detec

ted.

c Parts per m

illion.

d Information no

t given

in th

e original article.

e Not all fires invo

lved

ove

rhau

l.f (N

umbe

r of sam

ples w

here the

che

mical w

as detec

ted)/(total num

ber of sam

ples).

g The

sum

of 13

polya

romatic hyd

roca

rbon

s (ace

naph

thalen

e, anthrac

ene, ben

z[a]an

thrace

ne, be

nzo[a]py

rene

, be

nzo[b]flu

oran

then

e, ben

zo[ ghi] -

perylene

, ben

zo[ k] fluoran

then

e, chrysen

e, diben

z[a,h]an

thrace

ne, flu

oran

then

e, ind

eno[1,2,3-cd

] pyren

e, phe

nanthren

e, pyren

e).

METHODS

The same U.S. Environmental Protection Agency (EPA) TO14A methodpreviously modified and verified for the sampling and analysis of nonpolarVOCs in smoke that had been used to characterize VOCs at experimentalfires was used in this study to characterize VOCs at municipal structuralfires, with the average sampling and analytical error being ±23% (U.S. EPA,1997a, 1997b; Austin, 1997; Austin et al., 2001a). Firefighters inside burn-ing buildings at 9 municipal structural fires collected air into evacuated 1-Land 3-L Summa electropolished stainless-steel canisters fitted with a stain-less-steel 2-µm Nupro prefilter. Samples were collected at seven mixed-occupancy fires, one electronics industry fire, and one structural fire thathad been smoldering for 9 d. Samples were collected without regard tolocation, except at one mixed-occupancy fire where they were collectedon both the first and second floors. Firefighters were instructed to collectsamples when they judged that at least some firefighters might removetheir self-contained breathing apparatus (SCBA) masks. Nonpolar VOCspresent in the air and smoke samples were identified and quantified usingcryogenic preconcentration at –183°C and a Hewlett-Packard model 5890series II gas chromatograph (GC), capable of subatmospheric temperatureprogramming, equipped with an HP model 5971 quadrupole mass-selec-tive detector (MSD). The instrument standard calibration mixture, pre-pared by Environment Canada to monitor outdoor ambient air, contained144 target compounds (Table 2) and 4 internal standards (1,4-difluoro-benzene, bromochloromethane, chlorobenzene-d5 and 1-bromo-4-fluoro-benzene). Use of a Nafion drier to remove water vapor from the airstreamprior to cryogenic preconcentration precluded the analysis of polar VOCsand one- and two-carbon compounds. The volatiles were separated on aHewlett-Packard 50-m, 0.32-mm ID fused silica capillary column with a1.0-µm-thick film of HP-1 bonded liquid phase using helium as a carriergas. The revolatilized sample from the preconcentrator was cryofocusedon the column using liquid nitrogen. The initial temperature of the column,which was –60°C, was held for 3 min, then raised to 250°C at a rate of8°C/min. The rate was then increased to 20°C/min, and the temperaturewas held at 280°C for 8 min, after which it was lowered to 150°C. Thetotal chromatographic separation time was 57.75 min, including a 5-minsolvent delay time. The mass-selective detector was operated in selectedion monitoring (SIM) mode, eluting compounds being monitored for char-acteristic and qualifier ions in 27 acquisition windows, with each windowmonitoring 3 to 17 ions. Overnight and between runs, the GC was left onstandby at 150°C. Samples were also analyzed in SCAN mode and exam-ined for the appearance of substances that were not present in the instru-ment standard calibration mixture. All molecular ion fragments present inthe mass spectrum were monitored from 35 amu to 500 amu. Tentativeidentification of major nontarget compounds was made by comparison ofthe mass spectrum obtained with appropriate reference spectra found by

444 C. C. AUSTIN ET AL.

445

TAB

LE 2

.Ta

rget Com

poun

ds (1

44) in Air Sam

ples O

btaine

d at M

unicipal Struc

tural Fires

Ben

zene

1,1-Dichloroe

than

eFreo

n 11

4 (1,2-dichlorotetraflu

oroe

than

e)3-Methy

lpen

tane

Ben

zyl ch

loride

(alpha

-chlorotolue

ne)

1 2-Dichloroe

than

eFreo

n 12

(dichlorod

ifluorom

etha

ne)

Nap

htha

lene

Bromod

ichlorom

etha

ne1,1-Dichloroe

then

eFreo

n 22

(chlorod

ifluo

rometha

ne)

Non

ane

Bromoform

c-1,2-Dichloroe

then

eHep

tane

1-Non

ene

Bromom

etha

net-1,2-Dichloroe

then

e1-Hep

tene

Octan

eBromotrich

lorometha

neDichlorom

etha

nec-2-Hep

tene

1-Octen

e1,3-Butad

iene

1,2-Dichlorop

ropa

nec-3-Hep

tene

c-2-Octen

eButan

ec/t-1,3-Dichlorop

rope

net-2-Hep

tene

t-2-Octen

eIsob

utan

e (2-m

ethy

lpropa

ne)

c-1,3-Dichlorop

rope

net-3-Hep

tene

Pentan

ec-2-Buten

e1,2-Diethylbe

nzen

eHex

achlorbu

tadien

e1-Pe

nten

et-2-Buten

e1,3-Diethylbe

nzen

eHex

ane

c-2-Pe

nten

e1-Buten

e/2-methy

lprope

ne1,4-Diethylbe

nzen

ec-2-Hex

ene

t-2-Pe

nten

etert-B

utylbe

nzen

e2,2-Dim

ethy

lbutan

et-2-Hex

ene

Prop

ane

iso-Butylbe

nzen

e2,3-Dim

ethy

lbutan

e1-Hex

ene/2-methy

l-1-pe

nten

eProp

ene

n-Butylbe

nzen

ec-1,2-Dim

ethy

lcyc

lohe

xane

Hex

ylbe

nzen

eIsop

ropy

lben

zene

sec-Butylbe

nzen

ec-1,3-Dim

ethy

lcyc

lohe

xane

Inda

nen-Prop

ylbe

nzen

e1-Butyn

ec-1,4/t-1,3-Dim

ethy

lcyc

lohe

xane

Isop

rene

(2-methy

l-1,3-bu

tadien

e)Prop

yne

Carbo

ntetrach

loride

t-1,2-Dim

ethy

lcyc

lohe

xane

2-Methy

l-1-bu

tene

Styren

eChlorob

enze

net-1,4-Dim

ethy

lcyc

lohe

xane

3-Methy

l-1-pe

nten

e1,1,2,2-Te

trac

hloroe

than

eChloroe

than

e2,2-Dim

ethy

lhex

ane

4-Methy

l-1-pe

nten

eTe

trac

hloroe

then

eChloroform

2,4-Dim

ethy

lhex

ane

2-Methy

l-2-bu

tene

Toluen

eChlorom

etha

ne2,5-Dim

ethy

lhex

ane

c-3-Methy

l-2-pe

nten

e1,2,4-Trichlorob

enze

neCyc

lohe

xane

3,6-Dim

ethy

loctan

ec-4-Methy

l-2-pe

nten

e1,1,1 -Trich

loroetha

neCyc

lohe

xene

2,2-Dim

ethy

lpen

tane

t-3-Methy

l-2-pe

nten

e1,1,2-Trichloroe

than

eCyc

lope

ntan

e2,3-Dim

ethy

lpen

tane

t-4-Methy

l-2-pe

nten

eTrichloroe

then

eCyc

lope

nten

e2,4-Dim

ethy

lpen

tane

2-Methy

lbutan

e1,2,3-Trim

ethy

lben

zene

p-C

ymen

e2,2-Dim

ethy

lpropa

neMethy

lcyc

lohe

xane

1,2,4-Trim

ethy

lben

zene

Dec

ane

Dod

ecan

e1-Methy

lcyc

lohe

xene

1,3,5-Trim

ethy

lben

zene

1-Dec

ene

2-Ethy

l-1-Buten

eMethy

lcyc

lope

ntan

e2,2,3-Trim

ethy

lbutan

eDibromoc

hlorom

etha

neEthy

lben

zene

1-Methy

lcyc

lope

nten

e2,2,5-Trim

ethy

lhex

ane

1,2-Dibromoe

than

e (EDB)

Ethy

lbromide

2-Methy

lhep

tane

2,2,4-Trim

ethy

lpen

tane

Dibromom

etha

ne2-Ethy

ltoluen

e3-Methy

lhep

tane

2,3,4-Trim

ethy

lpen

tane

1,3-Dichlorob

enze

ne3-Ethy

ltoluen

e4-Methy

lhep

tane

Und

ecan

e1,2-Dichlorob

enze

ne4-Ethy

ltoluen

e2-Methy

lhex

ane

Vinyl chloride (chloroe

then

e)1,4-Dichlorob

enze

neFreo

n 11

(Trich

loroflu

orom

etha

ne)

3-Methy

lhex

ane

m/p-Xylen

e1,4-Dichlorob

utan

eFreo

n 11

3 (1 1 2-Trich

lorotrifluoroe

than

e)2-Methy

lpen

tane

o-Xylen

e

searching NBS75K computerized mass spectral and molecular structureslibraries (NIST, 1992). The probability that a correct match was found wasrequired to be greater than 90%. The internal standard method was usedto estimate the concentrations of tentatively identified unknowns. A Bruel& Kjaer model 1302 multigas analyzer was used to measure CO and CO2

in four of the canister samples.The ratio of selected VOCs to benzene was determined by linear cor-

relation using results from all nine fires and Excel spreadsheet software,with the critical value for a statistically significant correlation coefficientbeing .798 (n = 9, p £ .01).

RESULTS

The spectra of the seven mixed-occupancy, municipal structural fireswere surprisingly similar and remarkable for their simplicity, which waslargely due to the dominating presence of the benzene peak along withtoluene and naphthalene (Figures 1 and 2). Propene and 1,3-butadienewere found in all of the fires, and styrene and other alkyl-substituted ben-zene compounds were frequently identified. Other substances found in-cluded methyl-substituted butanes, pentanes, propanes and hexanes, andcyclopentane, all measured at ppb levels. The air at one fire, which hadbeen smoldering for 9 d, contained the same VOC combustion products ashad been seen previously (Figure 3A), while an electronics industry firecontained proportionately higher levels of ethylbenzene and isopropylben-zene (Figure 3B). GC/MSD analysis of municipal fire samples in SCANmode tentatively identified six substances not already present in the instru-ment standard calibration mixture: furan, benzaldehyde, benzofuran, ben-zonitrile, 2,3-dihydrofuran, and 2-methylfuran. Estimates for these sub-stances ranged from 0.2 ppm to 2 ppm, with furan and benzaldehyde beingpresent in the greatest concentration.

Of the 144 possible VOCs measured, 14 substances (propene, benzene,xylenes, 1-butene/2-methylpropene, toluene, propane, 1,2-butadiene, 2-methylbutane, ethylbenzene, naphthalene, styrene, cyclopentene, 1-methyl-cyclopentene, and isopropylbenzene) were found in proportionately higherconcentrations, accounting for 76.8% (SD = ±10.4%) of the 123 VOCsfound. “Fingerprints” of these 14 substances were found to be similar fromfire to fire (Figure 4). Twenty-one substances, many of them chlorinatedcompounds present at levels less than 1 ppb (bromoform, bromotrichloro-methane, tert-butylbenzene, carbon tetrachloride, dibromochloro-methane, 1,2-dibromoethane, 1,2-dichlorobenzene, 1,3-dichloroben-zene, 1,1-dichloroethane, c-1,2-dichloroethene, t-1,2-dichloroethene,1,1-dichloroethene, c/ t-1,3-dichloropropene, c-1,3-dichloropropene, 2,2-dimethylhexane, hexachlorobutadiene, c-2-pentene, tetrachloroethene,1,1,2-trichloroethane, trichloroethene, and 2,2,5-trimethylhexane), wereexcluded from the database.

446 C. C. AUSTIN ET AL.

The mean level of 123 VOCs found was 22.8 ppm (n = 9,SD = ±25.5 ppm). Levels of benzene, toluene, 1,3-butadiene, naphtha-lene, and styrene were in the ranges 0.12–10.76 ppm, 0.05–5.52 ppm,0.03–4.84 ppm, 0.01–2.14, ppm and 0.003–2.01 ppm, respectively(Table 3). These toxicologically important substances accounted for 31.1%(SD = ±12.7%) of the total VOCs found. In addition to the usual com-bustion products, unusually high levels of ethylbenzene (6 ppm) and ofisopropylbenzene (0.6 ppm) were found at an electronics factory, firenumber 2, the only industrial fire sampled (Figures 3B and 4). In

VOCs FOUND AT MUNICIPAL FIRES 447

FIGURE 1. Typical chromatograms from two of seven municipal structural fires sampled exhibitingthree major peaks: benzene (1), toluene (2), and naphthalene (3). Asterisks indicate the internal stan-dards normally added to samples prior to analysis.

448 C. C. AUSTIN ET AL.

FIGURE 2. Derived chromatograms (CorelDraw v4.0) for samples collected at seven different munic-ipal structural fires. The predominant combustion products were propene, 1,3-butadiene, benzene,toluene, styrene, and naphthalene.

samples collected simultaneously at one fire, the concentration of ben-zene was found to be 100-fold higher on the second floor than on thefirst floor of a burning building. Mean carbon monoxide and carbon di-oxide levels were 160 ppm (n = 4, SD = ±28 ppm) and 3376 ppm (n =4, SD = ±1593 ppm), respectively. The differences in chemical concen-

VOCs FOUND AT MUNICIPAL FIRES 449

FIGURE 3. Derived chromatograms (CorelDraw v4.0) for samples collected at (A) a 9-d, smoldering,municipal structural fire, and (B) an electronics industry structural fire. The usual combustion prod-ucts seen at municipal structural fires were observed to be present (primarily propene, 1,3-butadiene,benzene, toluene, styrene, and naphthalene) in both cases. In addition, unusually high levels of ethyl-benzene and isopropylbenzene were observed at the electronics industry fire (B).

450

FIG

UR

E 4.

Cha

racteristic

“fing

erprints” of the

predo

minan

t VOCs from

nine mun

icipal struc

tural fires sho

wing the simila

rity

in com

position

of the VOCs prod

uced

as well as the

variability fou

nd betwee

n fires. Fire 2 is an

electronics ind

ustry fire, fire

4 is a 9-d smolde

ring

fire, an

d the othe

rs are m

ixed

-occ

upan

cy fires.

451

TAB

LE 3

.Fo

urteen

“Fing

erprint” VOCs Fo

und in A

ir at Nine Mun

icipal Struc

tural F

ires Ran

geMea

nSD

________________

TLV

aST

ELb

n(ppm

)(ppm

)Minim

umMax

imum

(ppm

)(ppm

)BEI

c

“Finge

rprint” VOCsd

Prop

ene

94.96

6.8

0.22

21.64

——

—Ben

zene

e9

3.38

3.45

0.12

10.76

0.5

2.5

Yes

Xylen

es (o

, m, a

nd p

isom

ers)

91.74

2.9

0.06

9.19

100

150

Yes

1-Buten

e/2-methy

lprope

ne9

1.38

1.52

0.03

4.08

——

—To

luen

ee9

1.57

2.2

0.05

5.52

50—

Yes

Prop

ane

90.71

1.14

0.03

3.63

2500

——

1,3-Butad

iene

e9

1.03

1.49

0.03

4.84

2—

—2-Methy

lbutan

e9

0.07

0.13

0.00

40.43

——

—Ethy

lben

zene

90.86

1.94

0.01

5.97

100

125

Yes

Nap

htha

lene

e9

0.62

0.68

0.01

2.14

1015

—Styren

ee9

0.5

0.68

0.00

32.01

2040

Yes

Cyc

lope

nten

e9

0.41

1.08

0.00

23.29

——

—1-Methy

lcyc

lope

nten

e9

0.22

0.59

0.00

11.79

——

—Isop

ropy

lben

zene

90.07

0.18

0.00

040.55

——

—

Total “fin

gerprint” VOCs

17.5

19.9

0.6

75.8

Percen

t of total VOCs foun

d76

.810

.446

.193

.7

Total selected toxic VOCse

97.1

7.4

0.2

25.3

Percen

t of total VOCs foun

d9

31.1

12.7

17.3

31.2

Total V

OCs foun

d (123

che

micals)

922

.825

.51.3

80.9

Carbo

n diox

ide

433

7615

9320

8556

2150

0030

,000

—Carbo

n mon

oxide

416

028

134

199

25—

Yes

a Thresho

ld limit va

lues (A

CGIH

, 20

01).

bSh

ort-term

exp

osure lim

it (ACGIH

, 200

1).

c Biologica

l ex

posure in

dice

s (ACGIH

, 200

1).

dTh

e 14

sub

stan

ces co

mprising the VOC “fin

gerprint” foun

d to be simila

r from

fire to

fire.

e Five toxico

logica

lly im

portan

t VOCs (ben

zene

, toluen

e, 1,3-butad

iene

, nap

htha

lene

, styrene

).

trations between fires are reflective of different sites and conditions at dif-ferent fires.

Statistically significant positive linear correlations (p < .01) were foundbetween increasing levels of benzene and levels of propene, the xylenes,toluene, coeluants 1-butene/2-methylpropene, 1,3-butadiene, and naph-thalene (Figures 5 and 6). The ratios of benzene to these substances were0.6, 1.6, 2, 2.5, 2.9, and 5.5, respectively (Table 4).

DISCUSSION

One might have expected that firefighters would be exposed to quitedifferent substances from one fire to another. However, similarity in thenature of the combustion products from a variety of sources, demonstratedby the characteristic prevalence of benzene, toluene, and naphthalenefound in air at experimental fires, suggested that similar patterns might befound in municipal structural fires (Austin et al., 2001a). That this is indeedthe case was demonstrated by chromatograms obtained from municipalstructural fires. Although it is true that fires emit a myriad of combustionproducts, the same predominant substances (benzene, toluene, 1,3-buta-diene, naphthalene, and styrene) found by Austin et al. (2001a) at experi-mental fires were prevalent also at municipal structural fires. A number ofsubstances, including the six chemicals found by SCAN analysis, werefound at municipal fires that had not been seen previously at experimen-tal fires. Similar “fingerprints” of the same 14 substances identified at firesburning various solid combustible materials were also found at municipalstructural fires, accounting for 76.8% of the total VOCs measured. Eightof these substances have established ACGIH (2001) threshold limit values(TLVs), and five of them have biological exposure indices (BEIs). Five ofthem (benzene, toluene, 1,3-butadiene, naphthalene, and styrene) accountfor 31.1% of the total VOC concentrations. The maximum levels of ben-zene (a proven human carcinogen), 1,3-butadiene (a probable human car-cinogen) and styrene (a possible human carcinogen) found were 11 ppm, 5ppm, and 2 ppm, respectively, with benzene exceeding the current short-term exposure limit (STEL) 4.3-fold (ACGIH, 2001; IARC 1985, 1987, 1992,1994). Bearing in mind that the analytical method used precluded theanalysis of C1 and C2 hydrocarbons likely to be produced in large quantitiesat fires, the total levels of VOCs quantified (1.3–80.9 ppm), where it waspresumed that at least some of the firefighters would not have been wearingtheir masks, were consistent with the results of previous studies.

Given that the cohorts studied in epidemiological studies to date didnot, as a general rule, use respiratory protection and that firefighters todaydo use SCBAs, it could be argued that the positive associations betweenfirefighting and certain forms of cancer identified in these studies are notrelevant to the modern firefighter. However, one should also consider thatin previous decades when firefighters did not wear respiratory protection

452 C. C. AUSTIN ET AL.

VOCs FOUND AT MUNICIPAL FIRES 453

FIGURE 5. Linear correlations observed for the levels of (A) propene, (B) xylenes, and (C) toluene vs.the levels of benzene found at nine municipal structural fires (p < .01).

454 C. C. AUSTIN ET AL.

FIGURE 6. Linear correlations observed for the levels of (A) 1-butene + 2-methylpropene, (B) 1,3-butadiene, and (C) naphthalene, vs. the levels of benzene found at nine municipal structural fires(p < .01).

they naturally attempted to avoid entering smoke-filled areas that mightbe more readily penetrated by firefighters today who do wear SCBAs. It isnot clear, then, to what extent their true exposures are different particularlyin view of the fact that firefighters today do not, in fact, wear their SCBAsall of the time (Burgess et al., 1977; Brandt-Rauf et al., 1988; Jankovic etal., 1992; Austin et al., 2001b).

It has also been suggested that modern-day building materials (poly-mers) might give rise to new supertoxicants, either during the phase ofrapid combustion (knockdown) or during the smoldering conditions follow-ing extinction of the fire (overhaul). Neither this study of municipal struc-tural fires, nor the previous one by Austin et al. (2001a) of experimentalfires burning various combustible materials, found any such new toxic non-polar VOCs that would have not been present in the past. The analysis ofsamples obtained at experimental fires and at municipal structural firesfound the same substances (propene, isoprene, benzene, toluene, ethyl-benzene, styrene, propene, and 1,3-butadiene) at high concentrationsrelative to other combustion products. Along with naphthalene, thesedegradation products of polymeric material were also the principal com-bustion products of wood, the predominant construction material in thepast (Austin et al., 2001a). Burgess et al. (1979) found that benzene levelswere highest at fires involving wood structures. Jankovic et al. (1992) andBolstad-Johnson et al. (2000) found many of the same contaminants duringoverhaul as had previously been found during knockdown but at muchlower levels, with the exception of aerosolized building materials. Theresults of the present study also suggest that the lower combustion tem-peratures characteristic of the latter stages of a fire do not result in higherlevels of toxic combustion products. Furthermore, the spectra of combus-tion products were similar at mixed-occupancy, municipal structural fires,an electronics industry fire and, a 9-d smoldering fire. The presence andconcentration of contaminants, such as benzene, styrene, and 1,3-buta-diene, having known long-term effects may not have been less in previousdecades (where wood, cotton, and wool dominated the fire scene) than inmodern fires where there is a prevalence of polymers.

VOCs FOUND AT MUNICIPAL FIRES 455

TABLE 4. Ratio of Benzene to Selected VOCs in Air at Nine Municipal Structural Fires

95% Confidence interval Correlation_________________ coefficient (r)VOCs Ratio Minimum Maximum (p < .01)

Benzene/propene 0.6 0.5 0.8 0.94Benzene/xylenes 1.6 1.2 2.5 0.89Benzene/toluene 2.0 1.4 3.2 0.85Benzene/(1-butene + 2-methylpropene) 2.5 1.9 3.5 0.89Benzene/1,3-butadiene 2.9 2.2 4.1 0.91Benzene/naphthalene 5.5 4.4 7.1 0.92

In the present study, firefighters were asked to sample air at times whenthey judged that at least some firefighters might remove their masks. Al-though the overall levels of VOCs were not high by industrial standards, theresults show that even minimal exposure to smoke may result in importantexposure to highly toxic and carcinogenic benzene. The concentration ofbenzene at fires was also among the highest of the 144 possible VOCs sam-pled. In addition, it had the interesting and unexpected property of correlat-ing directly with the levels of a number of other important combustionproducts, in particular propene, toluene, 1,3-butadiene, and 1-butene/2-methylpropene. These results are comparable to those observed by Austinet al. (2001a) at experimental fires burning solids where the ratios of ben-zene to these chemicals were 0.5, 3.2, 2.4, and 1.4, respectively. The ratioof benzene to naphthalene was comparable to that found at experimentalfires burning liquids, 5.3 (the result obtained at fires burning solids was notsignificant). These observations point to benzene, or its metabolic products-phenylmercapturic acid in urine, as a suitable chemical marker for fire-fighter exposure to combustion products. This requires further study.

Those VOCs likely to be responsible for many of the toxic effects ofsmoke do not appear to be new, they appear to be relatively few in num-ber, their levels are considerably higher than the remaining numerous com-bustion products, and they appear also in the combustion of wood, a tra-ditional building material. In spite of the small number of fire samplescollected, the consistency of the results obtained indicates that there maybe less variability in VOC exposures between fires than had been previ-ously thought. Given the toxicity/carcinogenicity of those VOCs that werefound in the highest concentrations, particularly benzene, 1,3-butadiene,and styrene, investigation of time-integrated personal exposures of fire-fighters to VOCs is warranted. Future studies should be expanded to in-clude industrial fires and sampling of polar VOCs.

REFERENCESAmerican Conference of Government Industrial Hygienists. 2001. Threshold Limit Values for Chemical

Substances and Physical Agents and Biological Exposure Indices. ACGIH, Cincinnati, Ohio. Aronson, K. J., Tomlinson, G. A., and Smith, L. 1993. Mortality among firefighters in metropolitan

Toronto. Am. J. Ind. Med. 25:1–13.Austin, C. C. 1997. Municipal firefighter exposures to toxic gases and vapours. PhD thesis, McGill

University, Montreal, Quebec, Canada.Austin, C. C., Wang, D., Dussault, G., and Ecobichon, D. J. 2001a. Characterization of volatile

organic compounds in smoke at experimental fires. J. Toxicol. Environ. Health A 63:101–116.Austin, C. C., Wang, D., Dussault, G., and Ecobichon, D. J. 2001b. Municipal firefighter exposure

groups, time spent at fires and use of self-contained-breathing-apparatus (SCBAs). Am. J. Ind.Med., in press.

Bates, J. T. 1987. Coronary artery disease in the Toronto fire department. J. Occup. Med. 29:132–135.Beaumont, J. J., Chu, G. S. T., Jones, J. R., Schenker, M. B., Singleton, J. A., Piantanida, L. G., and

Reiterman, M. 1991. An epidemiological study of cancer and other causes of mortality in SanFrancisco firefighters. Am. J. Ind. Med. 19:3517–372.

456 C. C. AUSTIN ET AL.

Bolstad-Johnson, D. M., Burgess, J. L., Crutchfield, C. D., Storment, S., Gerkin, R., and Wilson, J. R.2000. Characterisation of firefighter exposures during fire overhaul. Am. Ind. Hyg. Assoc. J. 61:636–641.

Brandt-Rauf, P. W., Fallon, L. F., Tarantini, T., Idema, C., and Andrews, L. 1988. Health hazards offirefighters: exposure assessment. Br. J. Ind. Med. 45:606–612.

Brandt-Rauf, P. W., Cosman, B., Fallon, L. F., Tarantini, T., and Idema, C. 1989. Health hazards offirefighters: exposure assessment. Br. J. Ind. Med. 46:209–211.

Burgess, W. A., Sidor, R., Lynch, J. J., Buchanan, P., and Clougherty, E. 1977. Minimum respiratoryprotection factors for respiratory protective devices for firefighters. Am. Ind. Hyg. Assoc. J.38:18–23.

Burgess, W. A., Treitman, R. D., and Gold, A. 1979. Air Contaminants in Structural Firefighting.Harvard School of Public Health, prepared for the National Fire Prevention and control Admin-istration. NTIS order no. PB 299017, Springfield, VA.

Decoufle, P., Lloyd, J. W., and Salvin, L. G. 1977. Mortality by cause among stationary engineers andstationary firemen. J. Occup. Med. 19:679–670.

Demers, P. A., Heyer, N. J., and Rosenstock, L. 1992. Mortality among firefighters from threeNorthwestern United States cities. Br. J. Ind. Med. 49:664–670.

Eliopoulos, E. 1984. Mortality of firefighters in Western Australia. Br. J. Ind. Med. 41:183–187.Feuer, E., and Rosenman, K. 1986. Mortality in police and firefighters in New Jersey. Am. J. Ind.

Med. 9:517–527.Gold, A., Burgess, W. A., and Clougherty, E. V. 1978. Exposure of firefighters to toxic air contami-

nants. Am. Ind. Hyg. Assoc. J. 39:534–539.Golden, A. L., Markowitz, S. B., and Landrigan, P. J. 1995. The risk of cancer in firefighters. Occup.

Med. State Art Rev. 10:803–819.Grimes, G., Hirsch, D., and Borgeson, D. 1991. Risk of death among Honolulu firefighters. Hawaii

Med. J. 50:82–85.Guidotti, T. L. 1993. Mortality of urban firefighters in Alberta, 1927–1987. Am. J. Ind. Med. 23:921–

940.Guidotti, T. L. 1995. Occupational mortality among firefighters: Assessing the association. J. Occup.

Environ. Med. 37:1348–1356.Howe, G. R., and Burch, J. D. 1990. Firefighters at risk of cancer: An assessment and overview of the

epidemiological evidence. Am. J. Epidemiol. 132:1039–1050.IARC. 1985. 1,3-Butadiene. IARC Monogr. Eval. Carcinogen. Risks Hum. 39.IARC. 1987. Overall evaluations of carcinogenicity: An updating of IARC monographs volumes 1 to

42. Supplement 7 (benzene; ethylene dibromide; vinyl chloride) . Lyon: IARC.IARC. 1992. 1,3-Butadiene. IARC Monogr. Eval. Carcinogen. Risks Hum. 54.IARC. 1994. Isoprene; propene; propylene oxide; styrene. IARC Monogr. Eval. Carcinogen. Risks

Hum. 60.Jankovic, J., Jones, W., Burkhart, J., and Noonan, G. 1992. Environmental study of firefighters. Ann.

Occup. Hyg. 35:581–602.Kelly, J. E. 1991. Firefighters’ exposure to smoke during fire suppression activities at wildland fires.

Health hazard evaluation report HETA 91-312-2185. U.S. Department of the Interior, NationalPark Service, Gallatin National Forest, Montana.

Loke, J., Farmer, W. C., Matthay, R. A., Virgulto, J. A., and Bouhuys, A. 1976. Carboxyhemoglobinlevels in firefighters. Lung 154:35–39.

Lowry, W. T., Juarez, L., Petty, C. S., and Roberts, B. 1985. Studies of toxic gas production duringactual structural fires in the Dallas area. J. Forens. Sci. 30:59–71.

Melius, J. M. 1995. Cardiovascular disease among firefighters. Occup. Med. State Art Rev. 10:821–827.

Milham, S. 1983. Occupational Mortality in Washington State, 1950–1979. DHHS (NIOSH) Publica-tion 83-115. Washington, DC: U.S. Government Printing Office.

National Institute of Science and Technology. 1992. NBS75K Mass Spectral Library. NIST. NationalBureau of Standards, Gaithersburg, Md.

VOCs FOUND AT MUNICIPAL FIRES 457

Orris, P., Melius, J., and Duffy, R. M. 1995. Firefighters’ safety and health. Occup. Med. State Art Rev.10:11–13.

Radford, E. P., and Levine, M. S. 1976. Occupational exposures to carbon monoxide in Baltimorefirefighters. J. Occup. Med. 18:628–632.

Sama, S. R., Martin, T. R., Davis, L. K., and Kriebel, D. 1990. Cancer incidence among Massachusettsfirefighters, 1982–1986. Am. J. Ind. Med. 18:47–54.

Sammons, J. H., and Coleman, R. L. 1974. Firefighters’ occupational exposure to carbon monoxide.J. Occup. Med. 16:543–546.

Sardinas, A., Miller, J. W., and Hansen, H. 1986. Ischemic heart disease mortality of firemen andpolicemen. Am. J. Public Health 76:1140–1141.

Stewart, R. D., Stewart, R. S., Stamm, W., and Seelen, R. P. 1976. Rapid estimation of carboxyhemo-globin level in firefighters. J. Am. Med. Assoc. 235:390–392.

Tornling, G., Gustavsson, P., and Hogstedt, C. 1994. Mortality and cancer incidence in Stockholmfirefighters. Am. J. Ind. Med. 25:219–228.

Treitman, R. D., Burgess, W. A., and Gold, A. 1980. Air contaminants encountered by firefighters.Am. Ind. Hyg. Assoc. J. 41:796–802.

U.S. Environmental Protection Agency. 1997a. Compendium of Methods for the Determination ofToxic Organic Compounds in Ambient Air—Compendium Method TO-14A. Determination ofVolatile Organic Compounds (VOCs) in Ambient Air Using Specially Prepared Canisters WithSubsequent Analysis by Gas Chromatography. Environmental Protection Agency, QualifyAssurance Division, Environmental Monitoring Systems Laboratory, Research Triangle Park, NC.

U.S. Environmental Protection Agency. 1997b. Method TO-15: The Determination of VolatileOrganic Compounds (VOCs) in Air Collected in SUMMAÒ Canisters and Analyzed by GasChromatography/Mass Spectrometry (GC/MSD). U.S. Environmental Protection Agency, QualityAssurance Division, Environmental, Monitoring Systems Laboratory, Research Triangle Park,NC.

Vera, J. E., and Fiedler, R. D. 1987. Mortality of a municipal-worker cohort: IV. Firefighters. Am. J.Ind. Med. 11:671–684.

458 C. C. AUSTIN ET AL.