Effects of exposure to vehicle exhaust on health

Effects of exposure to vehicle exhaust on health. Scand J Work Environ Health I3 (1987) 505-512. Exposure to combustion engine exhaust and its effect on crewsof roll-on roll-off ships and car ferriesand on bus garage staff werestudied. The peak concentrations recorded for some of the substances studied were as follows: total particulates (diesel only) 1.0 rng/rn", benzene (diesel) 0.3 mg/rn", formaldehyde (gasoline and diesel) 0.8 mg/rn", and nitrogen dioxide (diesel) 1.2 mg/m'. The highest observed concentration of benzo(a)pyrene was 30 ng/m ' from gasoline and diesel exhaust. In an experimental study volunteers were exposed to diesel exhaust diluted with air to achieve a nitrogen dioxide concentration of 3.8 mg/m '. Pulmonary function was affected during a workday of occupational exposure to engine emissions, but it normalized after a few days with no exposure. The impairment of pulmonary function was judged to have no appreciable, adverse, short-term impact on individual work capacity. In the experimental exposure study, no effect on pulmonary function wasobserved. Analyses of urinary mutagenicity and thioether excretion showed no sign of exposure to genotoxic com pounds among the occupationally exposed workers or among the subjects in the experimental study.

Reprint requests to: Dr R Alexander sson , Department of Occupatio nal Medicine, Karol inska Hospital , S-104 01 Stockholm, Sweden. 3 the complex chemical mixtures contained in engine exhaust, since several earlier studie s have shown that components in engine exhaust, especially the particulate fra ction, are highly mutagenic in experimental systems (28,29). A review of the literature relevant to this investigation has already been publ ished (32).

Subjects and methods
Several workplaces producing comparatively heavy exposure, to diesel exhaust in parti cular , were chosen for the investigation . (See table s I and 2.) Occupational hygiene measurements were carr ied out at these workplaces, along with measurements of pulmonary function and genotoxic exposure. In addition , a small investigation with volunteers exposed to dilute engine emission was carried out in which the same occupational hygiene, medical, and toxicologic measurements were used as at the workplaces.
Workplaces, j ob descriptions and measurement strategies Bus garage. In March-April a bus garage was investigated in which there were large and small dieselpowered vehicles and gasoline-driven buses used for tran sporting disabled persons. The work per formed there, in addition to storage and engine warm-up, included refueling, washing and repair s. In the mornin g, afternoon, and evening elevated concentrations, in particular of diesel exhau st, accumulated in the garage bays when most buses left for or returned from assignments. The occupations represented were mars haler (2 workers), cleaner (I worker), foreman (2 workers) , Table 1. Characterization ofthe workers (men only) exposed to mixed exhaust (gasoline and diesel fumes) or to primarily diesel exhaust in the workplaces chosen for the investigation. _.._ --------a Ten persons took part on both occasions at the ro-ro ships. Table 2. Chemical characterization of the air samples, obtained with a personal sampling technique, from the workplaces and from the experimental study and the occupational exposure limits of the substances determined. The readings from the ro-ro ships are averages for a whole workshift. Other results represent the exposure measured for part of a day. In the latter instances the measurements were carried out during the course of work entailing exposure.

Range of the substances determined
Acetal-Benzo-Benzene Dust and mechanic (12 workers). Those responsible for the workplace had rated the ventilation as not fully satisfactory.
Car ferry. In June-July two types of car ferries were studied, one serving on a 2-h route and the other on a 20-min route. The job of the upper deck crew on the former was to direct vehicles to and from their parking places on the car deck. Loading and unloading averaged 20 min. In the intervals between the loading and unloading, the deck crew performed other tasks. The first and second mates had duties on the bridge during crossings. On the second type of ferry, work consisted of directing vehicles to the right place as they came aboard (so the maximum number of cars and lorries could be accommodated on the car deck) and assisting in the unloading of vehicles from the deck after the ferry had docked. Their work also included mooring the ferry. The mate's job was to lower the gangway after docking, check the tickets of passengers traveling in vehicles, and supervise operations while vehicles were 506 driven on and off the car deck. Samples were only taken during the loading and unloading of vehicles.
Roll-on roll-off ships. The investigations aboard the roll-on roll-off (ro-ro) ships were carried out in January-February and in June and covered lower deck crew whose exposure to exhaust gas was judged to be the heaviest. During loading, cargo was moved from the quay onto the ship with large diesel-powered trucks. Once the cargo was on deck, smaller trucks took over and carried out the final stowage. The work of each truck driver was directed by one or two men on the deck. A supervisor was in charge of one or more decks. The truck drivers drove their trucks aboard ship and also spent a short time on deck. Drivers of the big trucks transferred cargo between the quay and the ships. They remained in their cabs the whole workday. One group, made up in part of supervisors, placed padding, etc, under cargo to keep it steady. The supervisors directed the loading and unloading of cargo.
Experiments in the exposure chamber. Six workers were placed in an exposure chamber (interior dimensions 4.9 x 2.9 x 2.5 m) into which exhau st gas , dilut ed with air, from a diesel-powered vehicle was piped . The d ilution was adj usted so that the nitrogen d ioxide in th e chamber amounted to about 2 ppm.
The test vehicle was a 1980 Volvo 2440 automobile with a manual , four-speed transmission. The vehicle was powered by a six-cylinder, precombustion chamber diesel engine (2 383 em'). The compression was 23: I, and the engine output 60 kW . Du ring the exper iment the vehicle was run at a con stant speed, equi valent to 60 k/h in th ird gear. Thi s rat e produced an engine speed of about 2 580 revolutions/min. The engine load was calcula ted as 18 kW (maximum engine output about 35 kW at the indicated engine speed). This engine load is about fou r to five times higher than the load during dri ving on a dry, level road. The engine was kept running for 3 hand 40 min without interruption.

Hygienic measurements and chemical analysis
Total du st a nd respirable du st were co llected o n cellulose acetate filters in personal and stat ionary sa mpling equipment. Th e sampling of formaldeh yde and acetaldehyde was performed with person al sampling equipment with chemo sorption , and the samples were analyzed with high-performance liquid chromatography (2,3). H ydro carbons were collected in an aluminum laminate bag and then an alyzed with a nonspecific, direct-reading photoionization detector (HNU PI 101). Some samples were taken with a carbon tube o r Tenax '" tube and then analyzed with gas chr omatography. Hydro carbons were sampled with personal eq uipment only .
Both the person al and th e stati o nary sa mpling was carr ied out fo r ma ss spectrometry determination s of ben zo(a)pyrene and total pol ycyclic aromatic hydroca rbo ns (P AH). The particles were collected on glass fiber filters, whereas gas samples were collected by twostage chillin g of the samples with water an d dr y ice and ethanol. Th e samples were analyzed with a mass spect ro meter employing mult iple ion detecti on (15).
Air was collect ed in an aluminum lam inate bag and an alyzed for carbon monoxide with a direct-reading instrument containing an electro chemi cal cell (Interscan 1144) and for nitric oxide and nitrogen dio xide with a d irect-reading instru ment (Monitor Labs 8440) accordi ng to the chemiluminescence prin ciple. Sulfur dio xide was determined with a dire ct-reading instrument (Interscan) a ccording to an electrochemical measuring principle. Nitrous acid and nitric acid were collected in glass tubes coated with sodium hydroxide. Subsequent analy sis was carried out with ion-excha nge chro ma tography (17) .
Aceto ne extracts of the particulat e co mponent of three air samples from the expo sur e chamber experiment were analyzed for mutagenicit y in the Ames test with two bacterial st ra ins (Salmonella typhimurium, TA 98 and TA 100), both with and witho ut metabolic acti vation (I) .
Biological methods Determination oj urinary mutagenicity and urinary thioether excretion. Workers occupation ally exposed to engine exhausts were asked to pro vide urine specimens (on a random basis) immedi atel y before going to work (unexposed sample) and again at the end of or immediately after wor k (exposed sample). The urine specimens were immediat ely fro zen a nd sto red at -20°C for several months.
Th e subjects in the experimental expos ure supplied urine specimens before and at several times after th e expo sur e (table 3). Since it has been found (5) that diet may have a pow erful effect on the urin ary excretion o f th ioether s and it ca n be assum ed th at certain drugs are ca pable o f influencing thio ether formation, dietar y inta ke was sta nda rdized during the experiment al da y. In addition the subjects refrained from takin g any medication. Th ey care fully avoided certain vegeta bles (th e cru ciferae fa mily suc h as ca bbage and hor seradi sh) which co ntain co mparatively large am ou nts of thio cyan at e (5,37). Th e urine specimens were immediately frozen and sto red at -20°C for several months. Table 3. Thi oet her co nce ntration and mutagenic activity in the urine of si x wo rkers before and after the experi me ntal expos ure t o diese l exhaus t gases. (SE = stan da rd erro r of the mean)

Salmonella typh im uri um TA 98
Th ioet her concent ration (mol/mol c reati nine) + S-9 mi x (rever tan ts! mo l c reati ni ne) -S-9 mi x (reverta nts! m ol creati ni ne) 507 Two bacterial strains were used for the determination of urinary mutagenicity, S typhimurium TA 98 and Escherichia coli WP2 uvrA . The analyses were carried out at the Mutagen Laboratory of the Institute of Occupational Health , Helsinki, using the copolymer XAD-2 (for the determination of the concentrations in the urine specimens) and the fluctuation method (16,38).
The urinary concentrations of the thioethers (13) were determined by spectrophotometry, using Ellman's reagent after hydrolysis to thiols according to the method described by van Doorn et al (34,35). All the exposed persons (see table I) were examined with the use of spirometry and the single-breath nitrogen-washout technique. The subjects from the car ferries and bus garage were examined on a Monday after a 2-d break from work . Stevedores on the ro-ro ships were examined after a IO-d break from work. All the exposed subjects were examined before they entered the work premises and after the workday. The exhaust concentrations in the air were determined with a personal sampling technique during the workday . On the basis of these results the study was expanded to include a second examination of the stevedores on the ro-ro ships on a later occasion. Prior to the examination , the stevedores were not exposed to exhaust gases for 5 or 6 d. After expo sure to emission s for 1 or 2 d, the group was monitored during unexposed work so as to determine the time needed for pulmonary function to recover after deterioration.
Forced expired vital capacit y was recorded with a low-re sistance bellows spirometer (Ohio 740). At least two measurements were taken per person. The best result for each variable was cho sen, even if the value had to be taken from different determinations, and the volumes were adjusted to conditions of body temperature and pressure saturated with water vapor (BTPS) (19).
The Reference values for the spirometric variables were taken from Berglund et al (9) and Birath et al (10).
Single-breath nitrogen washout (4) was studied with the use of the aforementioned bellows spirometer and a "bag-in-box" unit. The nitrogen concentration was measured with an analyz er operat ing on the ionization pr inciple (Ohio 720). The gas concentration and volume were documented on an x-y recorder (Bryan s 26000). At least two mea sur ement s were taken at an interval of 7-10 min . The closing volume (CY) was expressed as the percentage of expiratory vital capacity (CY%). The slope of the alveola r plateau (phase III) was expressed as the percentage of nitrogen per liter of exhaled gas . The nitrogen washout data obtained in this manner were compared to reference values from Buist & Ross (II , 12).

508
Linear regression analyses were performed in intraand interindividual comparisons between degree of exposure and lung-function effect. The two-tailed Student's t-test was used (30).

Exposure conditions
An overvie w of the workplaces visited , the subjects, and the measured concentrations of substa nces can be found in tables I and 2.
Th e carbon monoxide peaks were lower in the bus garage (where diesel-powered vehicles were predominant) than on the ferries. The carbon monoxide concentrations were also low on the ro-ro ships on which only diesel trucks were used. The blood of exposed nonsmokers before and after a workshift displayed no increase in carboxyhemoglobin. The concentrations of nitrogen oxides were not higher than at the other workplaces where both gasoline-and diesel-powered vehicles were used .
At the concentrations tested, none of the three extract s from the filter samples was toxic to either strain of S typhimurium (TA 98 or TA 1(0), but extracts were mutagenic and produced a linear dose-response relation ship . The maximal number of ob served revertants per cubic meter of air sample was 643 for strain TA 100-S9. Mutagenicity was ob served both with and without a rat -liver metabolic system (S-9 mix), but the respon se was slightly lower in both strains when metabolic acti vation was used .
Urinary mutagenicity and urinary excretion oj th ioethers Occupat ionally exposed work ers. Comparison s between the urinary mutagenicity of samples from an expo sed and an unexposed period failed to disclose any significant differences between the number of revertants per mole of creatinine, ie, either in the occupationally exposed group as a whole or after the group had been divided into smokers and nonsmokers. A few sample s displa yed values that were barely above the limits regarded as mutagenic (>600 for S typhimurium and > 300 for E coliv (16). All but one of the se urine specimens had been submitted by smokers. For the mut agenicit y of the S typhimurium stra in TA 98, the mean value for the smo kers was higher than that of th e non smokers [mean 231 (SE 58) a nd 197 (SE 54) revertants/mol of creatinine, respectively]. No significant differences in thioether excretion were found between the samples from the exposed and unexposed period s, ie, either in the occupationally expo sed group as a whole or after the group had been divided into smokers and nonsmokers [5.8 (SE 0.5) and 5.5 (SE 0.5) mmo llmol , respectively ]. If , on the other hand , all the samples from the smo kers are compared to all the samples from the non smokers, the value of the smo kers is significantl y higher than that of the nonsmokers [7.2 (SE 0.6) and 4.7 (SE 0.4) mmollmol, respectively, P = 0.001.] Experimental exposure. Neither the thioether findings nor any of the values obtained in the tests for urinary mutagenicity were increased after the experimental exposure to the diesel exhaust-air mixture. (See table 3

.)
Pulmonary function studies For each person, a comparison was made with the reference material on a Monday before work and after a break from work lasting at least 2 d (table 4). The exposed group displayed an average reduction of 0.33 I in FYC (P < 0.001) and of 0.23 I in FEYl.o (P < 0.005). There were no differences between the smokers and nonsmokers. There were no corresponding significant changes in the MMF, CYOJo, or phase III variables . As a further check for a possible correlation between exhaust exposure and pulmonary function, various linear regression anal yses were used, but they failed to disclose any significant correlation s.
The pulmonary function of all the exposed subjects was studied before and after a workday. The mean exposure on this day was as follows: nitric oxide 0.6 mg/m ', nitrogen dioxide 0.54 mg/m', carbon monoxide 1.1 mg/rn", and formaldehyde 0.15 mg/m'. Significant impairment in pulmonary function was found (table 4). The FYC declined by an average of 0.15 I and the FEVl.o by a corresponding value of 0.11 I. No changes were found in the MMF, the CYl1fo, or the alveolar plateau gradient (phase III). There was no significant difference between the smokers and nonsmokers in pulmonary function on Monday before work.
The presence of any correlation between the exposure to various investigated components in exhaust gas and the investigated pulmonary function variables was studied. However , no correlation was found between the changes in pulmonary function and exposure to nitri c oxide, nitrogen dioxide, or formaldehyde in comparison with either peak values or the mean values for the various components in the exhaust gases during one workshift. Nor was any correlation found between the changes in pulmonary function and the carbon monoxide concentration.
The group of stevedores who had only been exposed to diesel emissions on ro-ro ships had a lO-d break from exposure before the medical investigation. There was no difference between the values of this group and the reference values before the start of the workday (table 4). However, pulmonary function did display deterioration during a workshift. The deterioration averaged 0.441 for FYC and 0.30 I for FEYJ.o' There was no difference between the smokers and nonsmokers .
A new study was carried out on another group of stevedores (N = 24) to ascertain whether the Table 4. Spirometric and nitrogen wash -out data for the exposed subjects (according to type of exposure) on a Monday be· fo re work and after periods of 2, 5, 10 and 13 days of no exposure and after 3 days of exposure. Reference values were matched with regard to sex, age and height (8,10,11   "acute" deterioration in pulmonary func tion found in the stevedores during one wo rkshift was reproducible and, if so , how long an y normalization would take. On this occasion exposure conditions were about the same as in the first investigation series (table 2). In this case, the break in exposure prior to the investigation was 5 to 6 d. As in the first investigation, no significantly impaired pulmonary function was found before the workshift. But acute deterioration was recorded for FVC (which averaged 0 .161) during one workday. This decline was accompanied by an increase in FEVOJo . As table 4 shows, this decl ine in pulmonary function normalized after a 3-d break from exposure. Ferry and garage per sonnel expo sed to both gasoline and diesel exhaust had a 2-d break from exposure before the measurement of pulmonary function began before work on Monday. A decline in pulmonary function was found in this group (0.45 I for FVC and 0.34 I for FEV 1.0) in comparison to the reference values (table 4). No additional deterioration was found in the pulmonary function during a workshift.

Exposure conditions
In the experimental study the nitrogen oxide concentrations were generally higher than any peak concentration found at the workplaces at which only diesel engines were used (table 2). The concentrations of carbon monoxide were comparable with the peak concentrations found at the workplaces using diesel engines.
The total concentration of particulate-bound and gaseous P AH was 638 ng/ " in the exposure chamber.
The low molecular-weight compounds (methyl phenanthrenes, anthracenes, and phenanthrene) accounted for 76 070 of the total P AH content.

Genotoxicity
The particulate extract of the air samples from the exposure chamber displayed clear genotoxicity with a linear dose-response in the Ames test. Both test strains used , ie, S typhimurium T A 98 and TA 100, yielded a higher respo nse when no exogenous metabolic system wa s used. This finding po int s to the pre sence of direct-acting mutagenic substa nces reg, nitroarenes such as l-nitropyrene (28,29)] charac teristically found in diesel exhaust. The ambient air samples taken from the exposure chamber during the volunta ry exposure experiment and several previous studies performed on engine exhau st ha ve all shown these emissions to be mutagenic, irres pective of the fuel type (14,15,22,23,26,28,29). However no mutagens were found in the urine specimen s of our subjects occupat ion all y or experimentally exposed to vehicle emi ssions.
The negati ve results of the mutagenicity ana lysis of th e ur ine do not preclude the absorption of mutagenic 510 substa nces fro m engine exhaust. Several earli er studies (31,38) ha ve shown the test strai n S typhimurium TA 98, with metabolic activation (S-9 mix) to be sensitive to tobacco smoke. The smokers in the pre sent study also displayed greater mutagenic activity than th e non smokers.
The excretion of thioethers in urine was similar before and afte r exposure to engine exh au st. On th e other hand, the values of the smo kers were significantly higher (P = 0.001 ) than those o f th e nonsmo kers. Th is findi ng agrees with results reported earlier (5,33). The individu al results for the occupationally exposed subjects (some va lues exceeding 10 mmol/mol of creatinine in samples taken before expo sure but declining in those taken a fter expo sure) indicate that factors unrelated to the work environment, ie, probably dietary factors (5,37) or pos sibly medication , had aff ected the results obtained for certain individuals in this group. The lower thioether levels and the lower vari ation in the values obtained in the experimental stu dy, compared to the values o f the occupationally exposed persons, might be due to the diet standardization emp loyed in the experimental study .
Th e nega tive results o f the urinary assays do not preclude th e po ssibility that exhaust exposures may cause local mutagenic effects in, eg, lungs and airways, the regional lymph system, or the gastro intestinal tract, aft er exhaust particles are swallowed . Compounds excreted in bile were not detectable with the urinary assay methods used.

Respiratory effects
Exp osu re to exha ust fum es appeared to imp a ir pulmon ary fun ction . These effects were found on a Monday morning , before work began , after cessation of exposure for at least 2 d and were accentua ted during a workday with exposure to exhaust fum es.
The subjects comprised stevedores expos ed only to exhaust fro m diesel-powered truck s and ferry and garage per sonnel exposed to both diesel and gasoline exhaust fumes. The stevedores had had a IO-d break from expo sure before the investigat ion started and displayed normal pulmonary fun ction on the Monday before go ing to work as compared with th e refe rence material. Ho wever , both the ir FVC and FE Vl.o values det eriorated during a wo rkday with exp osur e to exhau st. Similar but less pronounced effects were fo und in a seco nd, subsequent study. P ulmonar y funct ion returned to no rmal after 3 d witho ut occ upationa l exposur e to exha ust fumes.
Th e reco very to normal fun ction afte r th ree exposu re-free da ys may explain why the steve dore group did not display poorer pulmonary function than th e refe rence gro up before expo su re. Th is find ing is co nsistent with th e fact that no chro nic lun g ef fect s so lely att ributa ble to exhaust gases and particles hav e been discernible in va rious cross-s ectio na l st udies of per son s expo sed to diesel exha ust (6, 7, 8, 18, 24).
After a 2-d break in exposure, the ferry and garage personnel exposed to mixed exhausts displayed a deterioration in pulmonary function of the same nature and magnit ude as the stevedores after work . The functional deterioration showed no progression after one workday. Thu s th e same degree of dysfunction was ob served in both the stevedores exposed solely to diesel exhaust and in the ferry and garage workers expo sed to mixed exhaust emissions. These findings suggest that subjects who display functional deterioration after previous exposure to exhaust gases may not display any additional deterioration after renewed acute exposure and that subjects who have time to recover during a break from exposure may display acute deterioration. How ever, the interpretation of these findings is complicated by the differences in the expo sure of the stevedores and that of the exposure of ferry and garage personnel and by the fact that no dose-related correlation was found between any irritating substances in the exhaust gases (nitric oxide, nitrogen dioxide, and formaldeh yde) and an effect on pulmonary function. Nor was an y correlation found with the carbonmono xide exposure, investigated as a possible indicator of expo sure to all exhaust compounds.
A decline of up to 0.4 I in the FVC du ring a workday corresponded to a deviation of less than 10 % from prework values. Smaller differences were fou nd for the other variables . Presumably, these changes do not produce any discernible impairment in physical wo rk capacity during moderate physical exertion, but they are a disturbing finding in a cross-sectional study since their pro gno stic implications are unknown.
The occupational exhaust exposure levels measured are among the highest in Sweden, and they appear to affect pulmonary funct ion during a workday but cause relatively few subjective symptom s. This decline, which is reproducible in repeated measurements, normali zes after a few days without exposure to exhaust emissions. The chan ges are relatively slight and need not have any signi ficant effe ct on physical work cap acity. Becau se the changes are small, the nature of the dysfunction is difficult to assess. The decrease in FVC and FEY1.0' but not in FEV% or MMF, like the virtu ally normal, single-breath nitro gen washout findin gs, may suggest a restri ctive rathe r than an obstructive dysfunction , or a combination of both.

Possible indicators of exhaust exposure
In group studies, formaldehyde at concentrations of DA D mg/m' in the air has been show n to be capable of affecting pulmonary function , ie, causing mainly obstruct ion. Formaldehyde alon e, at co ncentratio ns do wn to 0.05 ppm (0.06 rng/rn ' ), irr itat es the nose, eyes, a nd throat of sensitive people a nd produces discomfort at somewhat lower concentrations (25,39).
It is possible that aldehydes could explain some of the airway symptoms found in the present study. The levels were relatively low, and a higher incidence of eye, nose, and thr oat sympto ms might have been expected.
High concentratio ns of nitrogen dio xide are known to cau se acute effe cts on the lungs (20). Nitrogen dioxide pro duced increased respiratory resistance after a IS-min exposure in the 1.6-2.0 ppm (3.6 -2.9 mg/m' ) concentration range. Lower concentrations produce nasal irritation and lar yngitic sympto ms but no changes in lung phy siolog y (21).
It should be pointed out that the alveo lar co ncentr ation of several of the gases may be even higher th an concentrations in the breathing zone, since additional input of injurious agents could con ceivably occur by adsorption on to the soot particles present in exhaust emissions. The particulate phase in diesel exhaust resembles carbon black , since the particulate matter consists of more than 85 % elementary carbon in the form of nearly spherical particles formed in the partial combustion of hydrocarbons. Diesel particles and carbon black particles are almost always respirable (36). A calculation assuming adsorption of a monomo lecular layer of nitrogen dioxide on a diesel aerosol ind icates that the sur face area of aerosol particles is sufficient to permit the particle-bonded nitro gen dioxide to constitute a substantial proportion of the total nitrogen dio xide concentration.
In summary publi shed data on th e effects of the sub stances studied at different conc entrations and on measured concentratio ns of these substances in occupational exhaust exposure do not rule out the possibility that nitrogen dioxide and formaldehyde, or both , may give rise to effects on lung and mucou s membranes at the concentrations found in practice.