Tobacco Documents Online

REGULATORY TOXICOLOGY AN[D PH+,R%tACOLOGY 17. !7-8. (1993), Carbon Monoxid'e and Cardiovascular Disease: An Analysis of the Weight of E:: idanc;e ~ 1Oii'N, H. MENNEAR. School of Pharmacy. Campbel! University. Buies Cree7c. North Carolina 27506 Received dugtur 19„l991 The role. if anv, of environmental tobacco smoke (ETS) in the causation and/or exaetttiation of cardiovascular disrase remains to be proven and'defined! Earlier workers suggested thar ETS- ass,xiated carbon monoxide- nicotine. and/or pol}aromauc hydrocsrbons may, be musauve factors. The purpose of this revicw was to assess the weight of evidence supporting a role for ambient carbon monoxide in the etiology of human ischemic c3rdiovascul+sr disase. The findings show thacthere is scant clinical or cxperimental evidence to support a role for carbon monoxide in the causation ofl ischcmic heart disease. Further. the tesulu of field studies of rciative air quality in nonsmoking and smoking homcs offices. and public places show that ETS contributes only minor and toxicologiolly insignificant increments in ambient carbon monoxide conctntntions. These increments are variable and casiiy masked by other common carbon monoxide sources such as internalicombustion engines and the burning of cooking and heating fuels. lt is conclitded that if ETS plays a role in the etiology of cardiovascuiar disGtse. it is most likely not mediated'through carbon monoxide. C ivv3 AooXmK Pn= I,K. INTRODUCTPON Cigarette smok~ing is frequently implicated as a risk factor in the production and/ or exaceroation of cardiovascular disease. Active smoking has been estimated to impart a risk for heart disease of 1.7 relative to nonsmoking (Surgcon General, t983)i Since 1984 a number of epidemiological studies have been conducted to assess the presence or absence of an association between the cohabitation of nonsmokers with smokers and death from cardiovascular disease. Glantz and Parmley (1991) reviewed the results of 13 such studies and' pointed out that in most (9/ 13, 69%) the estimated relati've risk (RR) of c3rdiovascular ddath due to ETS exposure was not significantly different from that of non-ETS-exposed people. In the remaining 4 studies (31% of the studies reviewed)'sma11 elevations in R.R, ranging from 1.2 to 2.0, were considered~ statistically significant. Glantz and Paratley noted that although estimates of cardiovascular death risks were only inconsistently elevated 16 ETS-exposed subjects, risk was not randomly distributed around unisy. The computed RRs and~ 9596 confidence intervals (CI) appear to be skewed toward eievation_ Further, when the results from all! studies were pooled! 77 0273-2300/93 S3.00 Cavrr*r~i c1993 b.? AmOmX r,vm lbc. Au ngac ar M cm,r,ae 10 ,oT . rormwer.aa t

78 JOHN H. MENNEAR analysis revealed a statistically signific3nt 30% increase in risk (RR = 1'_3; 95% CI = 1.2 to L.4). Reviewers of the ETS-cardiowascular death risk issue (Glantz and Parmley, 1'991; NIOSH', 1991; Taylor et al., 11992; and Steenland. 1992),ail noted that the known cardiotoxic compounds identified in mainstream smoke are aiso:presenrin ETS. This has been considered supportive evidence for the thesis of a cause-and-eSecr rzlationship between ETS and cardiovascular disease risk. But in reaching their conclusions ic is obvious that those authors gave little or no thought to one of the most basic principles of toxicology-the concept of dose-response relationships. It is a basic tenet of both clinical and experimental toxicology that there is generally a direct relationship between the amount of chemical to which an organism is exposed and~ the magnitude of the physiological changes produced. This principle of dose- response relationships forms the basis through which the medical profes.sion, industriali hygienists, and fed'eral regulators establish nontoxic doses of drugs, acceptable daily exposure levels to food additives, and no effezt levels of chemicals in the environment. Disregarding the principle of dose-response relationships would necessarily obligate prohibition of human exposure to vinualliy all'chernicals, whether synthetic or natural. Because of the relationship between dose and effect', the detection of a substance in the environment is only the initial step in establishing the presence of a possible human health hazard. When appraising the human health implications of exposure to any environmental factor a thorough assessment of the biological and chemical piausibilities of the purported effect is imperative. Such an asscssment should address three key fa,=ors: (1) Is there a plausible toxicologic mechanism through~ whi& the material could produce the suspected effect? (2) Is the mechanism operative in the human subjects of interest? (3) Are the human subjects exposed to a sufficient quantity of the environmental factor to produce the claimed toxicological consequence? The mechanism(s) though which either active or passive smoking might increase risk of cardiovascular disease have yet to be unequivocally defined. A prominent and frequently mentioned cause or contributor is the production of myocardial ischemia through exposure to ETS-associated carbon monoxide (Glantz and Parmley, 1991;. NIOSH. 1991; Taylor er al:. 1992; and~ Steenlar.d, 1992): The purpose of this review is to weigh the evidence relative to the hypothesis that ETS-related exposures to carbon monoxide (CO) can contribute to either the initiation or exacerbation of ischemic cardiovascular disease in humans. The resulu of this review show that there is llttll: clinical or experimental evidence that is relevant to the issue and that that which is available does not support a role for ETS-associated carbon monoxide in the causation or exacerbation of ischemic hean disease in non/never-smoking humans. MECHANISM OF ACTTON' OF CARBON MONOXIDE Carbon monoxide, produced during the incomplete combustion of all organic ma- terials, is the most extensively studied an& best understood component of either main- stream or sidestr¢am cigarttte smoke. This gas avidlycompetes with oxygen for binding to hemoglobin (Hb)1 The combination of CO with~ HB results in the formation of carbozvhemoglobin (COHB)iand compromises the trartsport of oxygen to the tissues of the body. All consequences of exposure to CO art directly attributable to the production of tissue anoxia. The magnitude of anoxia, and therefore the scveaty of physical svrnp- ~.

CO AND HEART DISEASE 799 toms. is related to the percentage of hemoglobin that is convened to COHb. The production of COHb is proportioncil to the amount of CO present in the imspired air. The remarkable afhnity of hemoglobin for the CO molecule makes the gas decep- tively toxic. If the affinity of hemoglobin for oxygen is assigned a value of 1.0, its affinity for CO is greater than 200. in~ the clinical situation, a few minutes of inhaling air containing as litt;le as 0.1% CO ('i.e., 1000 parts per miilion) results in 5090 of the available hemoglobin being converted to COHb. The presence of a 50% saturation of COHb is physically incapacitating and may even be lethal to the human (Smith, 198b): Toxicological consequences such~ as headache, dyspnea, and visuals disturbanees are associated with lower blood concentrations of COHb and the Amerir.am Conference of Govemmental an6 Industriali Hygienists has indicated its intent to establish 3.5% COHb as its best estimate of a no effect concentration among industrial workerss chronica3ly exposed to~CO:(ACGIH, 1991'). Because of the critical importance of continuous and4dequate oxygenation of heart muscle, it is obvious that a cardiotoxic effect of CO is plausible. Myocardial damage raused CO-induced ischernia would be no less significant than ischernic damage sec- ondary to coronary thrombosis or atherosclerosis. Since tobacco smoking may incTt= the concentration of CO in certain environments it is reasonable to assess the sensitivity of humans to CO-induced~ cardiotoxicity and d'eter7nine the quantitative impact of indoor smoking on the CO concentration in air. CARDIOVASCULAR EFFECTS OF CARBON MONOXIDE IN HUMANS Stern et al: (1988) presented evidence ofa possible CO-induced: risk of cardiovascular disease in humans expose& to automobile exhaust. These workers reponed a 35F0 excLSS in ischemic heartd%sease deaths among male traffic officers employed in tunnels in New York City. Additional evidence of probable occupationall association of the deaths was the fact that elevated risk promptly declined upon cessation of the occu- pationall exposure. These officers were occupationally exposed to environments containing about 5& ppm of CO. Although direct measures of COHb were not reported. it has been estimated that 8 hr of exposure to 50 ppm of CO will produce a COHb concentration of 6?790 (Singh et aL. 1991)a This indicates that the traffic ofinctrs may have had' blood' con- centrations of COHb approximately twofold greater than the ACGIH no effect eon- centration. - SeveraU investigators have studied' the effects of controlled tobacco smoke or CO inhalation on exercise tolerance and cardiac rhythms. Elevated serum carboxyhe- moglobin levels have been associated with decreased exercise tolerance in healthy subjects (McMurray et al:. 1985) and; decreased exercise toieranct and increased sus- ceptibility to exercise-induced~ cardiac arrhythmias in patients with coronary artery disease (Allired et al.. 1989; Sheps et aL. 1990a;b): Other workers, however, have re- ported~ the absence of effects of exposure to low concentrations of CO in patients with known coronary artery disease (Hind'erliter et aL 1989). E'ffects in healthy human. McMurray et al, (1985) exposed healthy smokers and nonsmokers to cigarette smoke during strenuous exercise. These workers reported tliat the exposure decreased the amount of exercise required to produce exhaustion in both groups. In addition, exercise-associated changes in biochemical measurements indicated tthatexposure to smoke caused an incrta,sed reliance on anaerobic meabolisrn. evidence ~

8& 1OHN' H. MENNEAR of decreased tissue oxygenation. The authors attributed these changes to systcmic anoxia secondary to the formation of COHb. McMurray et aL stated a belief that the exposure of their subjects was similar to a typical exposure of humans to environmental tobacco smoke. They did not, however,, present quantitative information to support'this contenrion and it is possible that their subjecLs were exposed to unrealistically high levels of smoke. During the exercise portions of the experiment cZgarettes were mechanically smoked, at' a rate of one every 3 tnin, and the smoke was mixed with air and delivered directly to exercising subjects via a mouthpiece and an intaiation tube. The minimum duration of exercise was 20, min. Consequently, subjecu were exposed to the smoke from ap- proximately seven cigarettes during their exercise session: Preexercise COHb concen- trauon in nonsmoking subjects was 1i.1'mo and, at the conclusion of the experimental session it had riscn to 2:20/'a. Similar data for subjects who were smokers was not pttsented. While the smoke exposure regimen in the McMurray et aL stud',v may have caused the slight decrement in exercise performance, the relevance of the data to the exposure of humans to ETS is difficult t& assess because the authors failed to report cither the smoke:a~r ratios in the mixtures delivered to their subjects or the CO~concentr-ations to which they were exposed Since subjects were exposed to some portion of the smoke from approximately seven cigarettes it is possible that unrtalistically, high ETS an& CO concentrations were uscd: Levesque et a!: (1991) studied the relationship between CO in ambient air and the formation of COHb in hockey players under game conditions. These workers found that for every 10 ppm of CO in environmental air, COHb saturation increases by 0.76%. Ifa similar relatio nship ~ holds for Mclvlunay's exercising subjects itis estimated that the nonsmoker's experimental exposure was to 15 ppm of CO in excess of their normal background concentrations. Effects in humans with coronary artery disease. Studies in which coronary-anery- discased subjects were exposed to.CO prior to exercise have yieldeda variety of result.s. "Ihese variable results are doubtless due to differences ia experimental designs and measured endpoints and subject selections. zClcin¢nan et al: (1989) reported' that exposure of male subjects with stable angina to 100 ppm of CO for 1 hr increased COHb saturation from a preexposure 1.5% to 2.9.°0. The 2.9% COHb concentration causcd~ a more rapid onset of exercise-induced anginal pain: thaa was experienced during the control exercise period without CO exposure. Hinderliter er a1: (1989) exposed coronary-artery-ciiseased patients, with low baseline ltvels of ventricular arrtiythmias, to cither 1100 or 200 ppm of CO for sulncienr durations to increase COHb levels to as high as 5.8%. Subjects then performed symptom-limited exercise. Continuous ambulatory EKG monitoring revealed that this level of COHb saturation was nonarrhythmogenic in these cardiac-diseased patients. Unfortunately; these workers did not compare pre- and' postexposure susceptibiliry to anginal pain. Using the same protocol with eoronarry-artery-dise2xd patients who had ventricular arrhythmias Sheps er al: (1990a,b) found that 5.7Fo COHb saturation caused an in- creased frequency and complexity of postexercise ventricular arrhythmias. Carboxy- hemoglobin saturation of 3.9%, however, was without' effecr. Allred' et aL (1989) reported the results of a multicenter study of the effects of CO exposure on exer6se performance in coronary artery disease patients. Subjects were

CO A1+D HEART DISEASE 81 exposed toeither 117 or 253 ppm of CO for periods sufficient to elevate COHb con- centrauons to values of 2.0 or 3.9°'0. Under the conditions of these experiments control COHb concentrations were unusually low (0.6-4.7%). These workers reported that both the 2.0 and the 3.9% COHb concentrations exacertiated exerezse-induced myocardial ischemia as evidenced by EKG changes and decreased time of onset of anginal pain., The Allred study has been criticized (Katzenstein, 1990) because of the low control values reported for pretest blood~COHb concentrations. Levels of COHb in nonexpose6 nonsmokers are generally found to be from~ two to three times higher than those reported by the Allred group. For example, Hinderliter et al. (1989) reported a preex- posure leveliof 1.8%; Sheps et al. (1990a) reported 1.82%; and McMurray et al: (1985) reported 1.1%. The Allred~ gToup~ explained that their low levels were due to their use of a gass chromatography assay of COHb rather than the more frequently us.^& optically based assay (Dahms et al:, 1990). They stated that, the commercial instruments generally provide inaccurately high COHb readimgs when concentrations of less than 5% are assayed'. For this reason the relevance of the Allred'data to:other contemporary studies is open to questaon. Overall, the results of studies in humans afford some evidence that exposure to extremely high concentrations of CO may elevate risk of ischemic heart disease and decrease the exercise tolerance of people with coronary artery disease. Such effects are consistent with the production of systemic anoxia and' impaired myocardial oxygen- ation. However, it remains to be established whether ETS can contribute sufficient environmental CO to impact on: the cardiovascular status of either healthy or com. promised~ humans. THE CLINICAL SIGNIFICANCE OF ETS-ASSOCIATED CARBON MONOXIDE To assess the potential cardiac risk of exposure to ETS-associated CO, it is necessary to~esvmate a maximal COHb saturation that would produce no physiological changes in exposed humans. Concentrations of CO in excess of that value should be considered potentially dangerous to~human health. A COHb concentration of 2.596 is proposed as the no effect leYel. This level of saturation is far below that which was associated~with increased ischemic heart disease risk in trahc tunnel workers (estimated to be 6.27% COHb) (Stern et aL. 1988). It is aiso:well below the 3.9°0 level, a levelithat did not result in exercise-induced arrhythmias in; patients with preexisting coronary artery disease (Sheps er aL. 1990a.b) and it is less than the 2.9,°0 level that was associated with decreased exercise tolerance in cor- onary-araery-diseased' patients (Kleinman er aL. 1989): The proposed value is also lower than the 3.5% COHb saturation that the ACGIH intends to establish~ as its best estimate of a no effcct concentration among industrial workers (ACGIH, 1991). The ACGIH value repczsents that body of health scientists' best estimate of a chronic, no effect level in workers expose6 to: CO~ 8 hr per day, 40 hr per we:ek. The 2.2% COHb concentration reported by Ivfc:ySurray er aC (1985) to produce an 8% decrement inthe performance of strenuous exercise was not considered because ~

82 1OHN' H., MENNEAR the effezt, which was minimal, was noted in only a small numbtr of strcnuouslv exercising subjects. Similarly, the 2.0% COHb saturation~ reported to reduce exercltr toltrance in~ patients with coronary artery disease (Allred'~ et al:, 1989) was not incot- porated into the estimation of a no effect level'because of uncertainty about the com~ parability of the COHb analyses in that study with those of the more nutneroils roi„ temporary'studies. In view of the available data relative to potential cardiov.~<culai effccts of CO exposure in, humans, the 2.5% COHb concentration reprrscnts a coo- servau've estimate of a probable no effezt level. The likelihood of a human achieving a serum concentra`uon of 2.5% COHb dctx:'nds upon the ambient concentrauon of CO and the duration of exposure. Singh et al: (]i991) reviewed', experimentally achieved COHb concentrations after exposures of varying durations to different concentrations of the gas. With exposure to 100 ppm, a serum concentration of 2.5% COHb was reached after between 30 and 45 min of exposure. At aniambient concentration of 50 ppm CO, longer than 60'rnin wasrequ1ired;and two, hours exposure to 45 ppm causes a COHb concentration of 2.48%. With the exception of accidents, employment in occupations involving itutcrnal combustion engines,, and intentional self inflicted exposures, humans are seldom ex- posed, even for brief periods, to CO concentrations in the range of 45 to I CK> pFm,. At lower, more probable levels of CO exposure still longer periods are required to produce the 2.5% COHb saturation. For example, exposure to 15 ppm of CO requires continuous exposure for 10 hr to produce a serum concentration of 2.5% COHb (Guerin et al:. 1992). IMPACT OF ETS ON AMBIENT CARBON MONOX'IDE CONCENTRATIONS It has be:n frequently and correctly noted that sidestrearn tobacco smoke contains a higher concentration of CO than does mainstream smoke. Sidestrcam smoke is prod'uced at a lower temperature at which the combustion of carbonaceous materials is less complete. American cig3rettes are recognized to deliver approximately 15 mg/ ci3arette of CO via mainstream smoke and 50 mgJcigarette via sidestream smoke (Guerin er aL. 1992); This rtlativel',v high doncentration in sidesutam smoke has led many to conclude that ETS is a major contributor to environmental CO concentrations. Such a conclusion is not supported by the results generated in ficid studies during whi& the air in resi- dences, work places, andpublic places has been analyzed under both smoking and' nonsmoking conditions. Gucrin er al. (1992) reviewed the data: generated during field studies of CO con- centrauons in a variety ofsmoking and nonsmoking areas. The results of the reviewed studies indicated that in general, smoking contributes only small increments in en- vironmental CO. For example, mean concentrations of CO in the air of offices in which smoking was permitted rangcd from 1.2 to~ 2.8 ppm, whereas values in non- smoking areas ranged from 1.2 to 2:5 ppmL In restautants and cafeterias pcrzzitting smoking, the environmental CO concentrations ranged from 1.2'to 9.9 ppm as con. tsasted' against nonsmoking control areas where concentrations ranged from 0.5 to 7.1 ppm. On the basis of the available data: obtained from field studies, it is clear that ETS contributes CO to the environment. However, the increment of environmental CO ~

CO AND~ HEART DISE.aSE 83 attributable to tobacco smoking is exceedingly small. Further. this small increase is easily masked by normal day-to-day va,riations in ambient concentrations which are attributable to the presence of other CO sources such as automobiles and the com- bustion of heating and cooking fuels. More importantiy; however, the results of the field studies also show that whether or not tobacco smoking is permitted, CO concentrations to wiuch humans are exposed~ seldom exceed the 9 ppm indoor standard that has been recommended by the American Society for Heating; Refrigerating, and Air Conditioning Engineers (ASHRAE, 1989): Since 10 hr of exposure to 15 ppm of CO is rrquired to produce a 2.5% level of COHb saturation in humans, and' since this is a no effcct ]evei, few Americans are ever exposed:, even for brief periods, to cardiotoxic concentrations of COHb. The small increment in ambient CO concentrations contributed by ETS is insignifieant. While conducting this analysis no attempt was made to directly address the issue of whether or not exposure to ETS per se causes or exacerbates cardiovascular disease. The results of this review have established, however, that if the purported impact of ETS on cardiovascular disea_se is real, it can be neither explained nor mediate& through ETS-assoc.-iated~ increases in ambient concentrations of carbon monoxide. There is scant evidence to support a role for carbon monoxide in the causation of ischemic heart disease. Further, the results of field studies of air quality in nonsmoking and smoking homes, offices, and public places demonstrate that ETS contributes only minor and toxicologically insignifieant increments in ambient carbon monoxide con- centrations. These increments are variable and easily masked by other commonly encountered carbon monoxide sources such as internal combustion engines and the burning of cooking,and heating fuels. Earlier workers have suggested that inhalatjon exposure to environmental tobacco smoke-associated nicotine and/or polycyclic aromatic hydrocarbons may also cause cardiovascular disease in humans (Glantz el al:. 1991; NIOSH, 1991: Taylor er al:. 1992; and~ Stecnland, 1992). Such claims cannot be taken seriously at this time since critaical' reviews of the experimental and clinical evidence claimed to~support the hy- potheses have yet to be conducted. CONCLUSION If ETS is an etiological factor in cardiovascular disease, its effect is most likely not mediated through carbon monoxide. REFERENCES A11RED. E. N.. Bt.F_ECKax. E. R.. Qt.,rrraA,4, B. R.. D..Hms, T. E_, GoTrues. S. 0.. HAcxxEY. J. D.. P,.Gr+o, M.. SELVFSrEtt- R. H.. WALDE*r. S. M.. At+rn WARRfr+. J. (1989): Shon-term effees oF nmon monoxide cxposurt on the exercise pcriormancz of subjects with coronary artery diuase. N. Sngl. L.N?d: 321., 1426-1432. Amencan Confertnce of Governmental lndustrial Hygienisu (ACGIH) '(1991): Threshold Limii Yalues jor Cntmical Sudstcnct and PhlysualAgrnu and BioFogical E_rposurt Indices 1991-1992. ACGiH. Cncinnan. OH: Americin Soeittv for Hesdng. Refrigeranng, and Air Conditioning Engineers (ASHRAE) (1989)..aSH1Ufi Siandard: k'eraiiation for Accepiable Indoor.4ir Quality: ASHR,IE: Inc.. Atlanta GA. DAHms, T. E.. WAxaEr+. J.. Bl_EaCxEx- E. R.. PAGAr+o, M.. CHAtTM..N. B. R.. H..=Nty. J. D.. GoTrt.tES, ~

84 JOHN H'. MENNEAR REGUL S. 0.. WAt_DEr:; S. M.. AND SELVESTER. R. H. (1990). Carbon monoxide and myocvdial ischemiy X Engl.' J: Ned, 322. 1087. GLnNTZ S. A.. AND PARMtEY.,W: W:,('1991): Passive smoking and heart discase. epidcmiology, physioiosy 1' and biochemistry: Circvlation 83; 1-12. GUERIN, M. R'.- JENK]NS, R: A.. AND ToMKSNS, B. A. (1992). The Chemistry of Ertvironmental Tobacco Smoke: Composuion and Measuremnu: Lewis. Baa Raton, Ann Attior, London. Tokyo. HINDERLI7ER, A. L, PRt[T, C. J., HERBSr, M. C.. Kom G.. AND SHEPS, D. S. (1989). Effxs of low level earbon monoxide exposure on rrsring and cxercise-induoed ventricular arrhythmias in patients with cor_ otsary artery dise:a u and no baseline ectopy. Arch. Environ. Haalth 44, 89-93. KA7ZENSTEtN, A. W. (1990). Carban monoxide and myocardial'ischemia. N. Engl: l:,rt;fed: 322, 1086. KtE3NMAr+, M. T., Dr,vtnsort, D. M., VANDAGruFF, R B., CAtoaO, V. J!, AND WHtTPENBERGFJt, 1; L ()989): Effects of shon-term exposure to mrbon monoxide in subjects with coronary artery diwa se..lrrlt' Environ. Health 44, 361-369. [EvFSQUE. B., LaVO(E, R-, DEQAlLLY, E., PRtdD'HOMME, D., AND ALLAIRE, S. (1991): An expCrimenrto evaluate cartion monoxide absorpoon by hockey players in iee skaung rinks, Ver. Hum. Toxicoi 33, 5-8. MCMuRR..Y; R. G., H,iCxs, L L. AND THOMPSOt:, D. L(1985): The effects of'passivc inhalhtion of cigarene smoke on exercise performance. Eur. l. appC Ph'ysioC 54, 196-200. NationaJ Ihsatute for Oecupational Salcty and Health (NIOSH) (1991).,Environmental Tobacco Smoke in the Workplace. Lung Cancer and Other Health Ejjeas: U.1 Dept- HHS. PHS. CDC. SHEPS, D. S., HERSST, M. C_ Ht:wDERUrER. A. L, ADAMS, K. F_ Ex.Et.UND: L G.. O'NEtL,1.1:, GoLDsr~, G. M., BROMeERG, P. A.. Du.TDr:, 1. L, BAuF* GER, M. N.. DHVts. S. M,- AND KOCH. G. G. (1990a). Production of arrhytbmias by eltvated carboxyhemoglobin in patients with coronary artery disease. Ann. lntern. Med.,113, 343-351.. SHEj-, D. S.. HER.BST, M. C.. HtNDERUrER, A. L. ADAMS K. F., Ex€LarxD, L G., O'NEIL 1,1., GOUMTEN, G: M.. BROMBERG: P. A.. BAt1.ENGER. M. N.. DAvii S. M.. AND KOcii G: G. (1990b). Effe,^.s of 4 Percent and 6 Percent Carooxyhemoglobi,r on drrh,vrhmia Production in Pattents wnh Coronary Artery Disease. Health Effects lhstitutc Rescsrt:h Reporr41. Health Effccu lnstitute, Cambridge. MA. StNGH. M: P.. MArmtL7, S., AND SELV,UCUntwR. S. (1991)j A mathemaund model' for the computation of cartioxyhemoglobin in human blood as a funcuon of'ezposure time. Philos. Trans. R. Sac. London B 334, 13 5- i 47. SMmH, R. P. (1986), Toxic Raponses ofrhe Blood. in Casarat and Douil's Toxicology. The Basic Scimce of Poisoru (C. D. KJaaasen. M. 0. Atndun and 1. Doull. Eds.); 3rd ed1, pp. 223-244. MacMillan Ncw York. STEENLUND. K. (1992). Passive smoking and the risk of hean disease. JAMA 267, 94-99. STERN; F.. HALPERIN. W:, AND HortNUNG. R. (1988). Heart disease mortality among bridge and'tunnel ot$cers oxposed to earbon monoxide. Am. 1. EpidemioL 128, 1276-1288. Surgeon General (1983). The Health Co+uequences of SmoktRg: Cardiovascular Disease: A Report of the Surgeon General: U.S. PHS. Depti HHS. RockviUe, MD. TAYLOR. A. E.. JOHNSOrv, D. C.. AND KAzE*.tt. H: (1992), Tobacco smoke and mrdiovascular disease. Circulation 86, II-4,