ORIGINAL ARTICLE

Pamela Dalton

Odor, irritation and perception of health risk

Received: 12 April 2001 /Accepted: 8 December 2001 / Published online: 19 March 2002� Springer-Verlag 2002

Abstract Objectives: Understanding the potential forvolatile chemicals to elicit chemosensory irritation in theupper respiratory tract is critical to setting occupationalexposure limits that are protective of comfort and well-being for the majority of workers. However, the deter-mination of irritant potency for any volatile chemicalhas been limited by the lack of reliable and non-invasiveassays for studying sensory irritation in humans and afailure to appreciate the many non-sensory factors thatcan influence the reactions to an odor or an irritant inthe workplace. Methods: This paper reviews the issuesinvolved in distinguishing and measuring sensations ofodor and irritation from volatile chemicals, and de-scribes recent developments in psychophysical methodsfor evaluating chemical irritancy in humans, and dis-cusses some of the many non-sensory factors such asexposure history, attitudes and expectations and per-sonality variables that can significantly alter the per-ception of odor, irritation and health risk followingexposure to a volatile chemical. Results: The availabilityof safe, non-invasive assays to measure directly odor andirritant responses in the species of interest, humans, canboth simplify and improve accuracy in the process ofdeveloping appropriate occupational exposure guide-lines. Conclusions: Objective measures of irritation onsetobtained in conjunction with subjective responses canlend valuable input to the decision process for deter-mining occupational exposure limits but should alwaysaccount for other factors (e.g., cognitive or emotional)that may be modulating the subjective response.

Keywords Sensory irritation Æ Olfaction Æ Occupationalexposure Æ Health symptoms Æ Volatile chemicals

Introduction

Although most volatile chemicals are capable of elicitingupper respiratory tract irritation, their odors can oftenbe detected at concentrations far below those that willobjectively evoke sensory irritation. There is little ques-tion that the inherent variability in absolute sensitivitycoupled with the subjectivity in response to odors andsensory irritants can produce significant problems forevaluating the potential for irritation, annoyance orhealth risk from workplace chemical exposures.

Recent research suggests there are at least three im-portant issues to be considered when developing or in-terpreting data for occupational exposure limits basedon chemosensory irritation from volatile chemicals. Thefirst issue involves the extent to which evaluations ofchemical irritancy are confounded by odor sensationand to what extent odor and irritation can be adequatelydistinguished and evaluated. Numerous studies haveshown that the mere presence of odor, at concentrationswell below those that are likely to elicit mucous mem-brane irritation, can nonetheless provoke adverse sen-sory responses and complaints. The second issue forconsideration is the choice of method for measuringsensory irritation. Although data derived from sensoryirritation assays performed in animals (i.e.,RD50) (Alarie1998) have been used to support occupational exposurelimits for many chemicals, the need to extrapolate acrossspecies or to measure the responses of sensitive humansub-populations has limited their utility and widespreadapplication. Fortunately, development and refinementof objective psychophysical and physiological assaysnow permit direct and more precise resolution of achemical’s irritant potential in human observers. Thethird issue concerns the degree to which variables otherthan the properties of the chemical can influence reportsof odor, irritation and perceived health risk during and

Int Arch Occup Environ Health (2002) 75: 283–290DOI 10.1007/s00420-002-0312-x

This paper was presented at the International Workshop onChemosensory Irritation, October 2000, sponsored by the Instituteof Occupational and Social Medicine at the University of Heidel-berg, Germany.

P. DaltonMonell Chemical Senses Center,3500 Market Street, Philadelphia,PA 19104-3308, USAE-mail: pdalton@pobox.upenn.eduTel.: +1-215-8985595Fax: +1-215-8982084

following exposure. As will be discussed, naıve and oc-cupationally exposed individuals often show consider-able variation in the reported levels of perceivedirritation and health complaints from chemical expo-sures. Thus, differences in exposure history have impli-cations for (1) predicting the likelihood of discomfort ina worker population exposed to that chemical and (2)the appropriate selection of a study population forevaluating chemical irritancy. In addition, myriad non-sensory factors, such as expectations, attitudes and be-liefs about the risks from exposure to a specific chemical,can modulate perceived irritancy and symptom reports.

Objective measures of upper respiratory tract irrita-tion onset obtained in conjunction with subjective re-ports of sensory experience and well-being can lendvaluable input to the decision process for determiningoccupational exposure limits based on chemosensory ir-ritation. Subjective reports of irritation at low levels thatcannot be reconciled with objective measures shouldprompt a careful investigation into the other factors (e.g.,cognitive or emotional) that may be modulating thesensory response. For example, anxiety or worry over theconsequences of exposure can elevate symptom reportsand irritation and may not be ameliorated by a reductionin the exposure level if the odor of the chemical is stilldetectable. Distinguishing between the dose that elicitslocal effects of sensory irritation in the upper respiratorytract and the dose that elicits self-reports of irritation is akey component in establishing occupational exposurelimits that are protective of the majority of workers.

This paper focuses on four questions that are ger-mane to understanding and evaluating chemosensoryirritation in occupational settings:

1. What is the relationship between odor and chemo-sensory irritation, and how can these sensations bedistinguished at a sensory level?

2. How can the human response to odors and irritantsbe measured?

3. What are the dominant factors that can affect theperception of odor and irritation?

4. How can this information be used in the process ofrecommending occupational exposure limits for vol-atile chemicals?

What is the relationship between odor and irritation?

Intra-nasal and upper respiratory sensations:odor and irritation

Sensations of odor and upper airway irritation are oftenexperienced as a unitary phenomenon, principally be-cause most volatile chemicals have the potential to ac-tivate two separate, yet interrelated, sensory pathways inthe upper respiratory airways: the olfactory nerve, whichgives rise to sensations of odor, and the trigeminal,glossopharyngeal or vagal nerves, which give rise totemporary burning, stinging, tingling or painful sensa-

tions in the eyes and upper airways (Alarie 1966).Chemical stimulation (known as chemesthesis) of thetrigeminal nerve often combines with stimulation of theolfactory nerve to produce sensations that form anoverall perception of a chemical. For example, lowconcentrations of ammonia produce a distinct odor;however, higher concentrations may also elicit a muco-sal burning or tingling, which is the chemesthetic or ir-ritant component of perception. Most chemicals atsufficiently high concentrations are capable of elicitingupper respiratory tract irritation in addition to odorsensations. Because these two sensory pathways (olfac-tion and chemesthesis) can be activated by a singlechemical stimulus and localized to the nose, workersoften experience and report odor and irritation as aunitary perception. For purposes of evaluating the irri-tant potential of a chemical, this confusion has castdoubts over attempts to rely on self-report methods,such as symptom questionnaires or even scalar ratings ofa chemical’s irritancy.

Relationship between odor, irritation and annoyance

To illustrate the interactions between the sensation ofodor and irritation from any chemical it is instructive toconsider the psychophysical relationships that define thehuman sensory response to a chemical. Psychophysicsrefers to the lawful relationships that exist between thephysical concentration of a stimulus (in this case, achemical) and an observer’s perception of the stimulusintensity across a physical (in this case, concentration)range. Psychophysical functions that relate stimulusconcentration to an observer’s rating of perceived in-tensity can be derived for both the odor and the irritancyof a chemical.

Data obtained from psychophysical scaling studieshave shown that the exponent of the psychophysicalfunction relating concentration to intensity for mostodorants falls well below 1, implying a compression ofsensation magnitude over stimulus magnitude. However,there can be considerable variation in the slope of thepsychophysical function for different odorants, rangingfrom as little as 0.10 to more than 1.0. Apart from dif-ferences in solubility or deposition in the airways, atleast some of this variation arises because, at low con-centrations, most chemicals will stimulate only the ol-factory system, while at higher concentrations thosesame chemicals will stimulate olfactory and trigeminalsensations. Data from a study by Green et al. (shown inFig. 1) (Green et al. 1996) evaluating two scalingmethods for chemosensory stimuli illustrate this differ-ence nicely. In this study, the psychophysical functionfor intensity of phenylethyl alcohol, which in vaporphase does not elicit chemosensory irritation, is quiteshallow, with a slope of approximately 0.3 across thelower concentrations and a slope of zero (implying aninability to resolve differences in concentration) acrossthe higher ones. In contrast, intensity data for acetic

284

acid, which is a trigeminal as well as an olfactory stim-ulant at the higher concentrations, were well describedby a single power function with a slope of approximately0.7.

For a chemical such as acetic acid, there is a fairlylarge window between the concentration necessary tostimulate olfaction and that necessary to stimulatetrigeminal or other upper respiratory irritant sensations.However, once the threshold for irritation is exceeded,doubling the concentration tends to produce largerchanges in the perception of irritation than in that ofodor. It is important to point out, however, that inpractice, the window between odor and irritationthresholds can be altered either because the irritationthreshold varies or because the odor threshold does.However, for a number of chemicals for which odor andirritancy thresholds have been established, the distancebetween these two thresholds appears to be fairly large(Cometto-Muniz and Cain 1995).

Given that a number of studies have established re-liable estimates of odor and irritation thresholds for avariety of chemicals, how can we explain subjective re-ports of irritation at relatively low concentrations? Atintermediate concentrations, i.e., those that lie above theodor detection threshold but well below the irritant

threshold, reports of perceived irritation most likelysignify odor intolerance or annoyance. As the concen-tration of a volatile chemical increases, such responsesare often initiated after the threshold for odor recogni-tion is crossed but often before the concentration ex-ceeds the threshold for trigeminal stimulation. To thebest of our knowledge, such intolerance or annoyancereactions are not mediated by stimulation of cranialnerves V, IX or X, but rather represent a psychologicalresponse to the presence of an unfamiliar, uncontrolla-ble or unpleasant odor. At low or moderate concentra-tions, the reaction appears to be related more closely tothe perception of an odor than to the presence of truesensory irritation (Dalton et al. 1997a).

How can odor and sensory irritation be measuredin humans?

Although it may appear to be a straightforward exerciseto expose individuals to a volatile chemical and ask themto report on the level of irritation they experience, thepotential for confusion between olfactory and irritantmodalities has produced extreme variability in directscaling of upper airway irritant sensations by individualswith intact olfactory and trigeminal systems. This isexacerbated when chemicals present in the air at levelsthat stimulate only odor sensation can prompt exposedindividuals, such as workers, to report ‘‘irritation’’, evenif the perception is largely mediated through the psy-chological discomfort incurred by smelling the odor ofan unfamiliar or unpleasant chemical, or even frommisattributions of unrelated symptoms that just happento coincide with chemical exposure. Not surprisingly,considerable interest has centered on developing meth-ods of assessing irritancy that either do not rely solely onself-report or that measure or otherwise control for thevariance that is produced by the perception of odor andother factors.

Two approaches have been taken to address thisproblem. The first embodies the use, validation andcontinued development of a number of objective assaysfor evaluating irritant potency across a wide range ofchemicals. The second approach involves the design ofstudies to identify the mechanisms of sensory irritationand perceived irritation, with a focus on systematicallystudying variables that may impact on the perception ofirritancy in humans.

Two psychophysical methods have proven to be quitevaluable in ongoing efforts to relate the subject’s per-ception of irritation to the physical concentration of thechemical. Sensory scaling, a technique that has been inuse for decades, relies upon the relationship betweensystematic changes in physical concentration and cor-responding changes in the subjects’ report of perceivedirritation intensity. For scaling sensory stimuli in otherdomains (e.g., vision, audition), many different types ofintensity scales have been developed, of which most canbe adapted for use in the scaling of sensory irritation,

Fig. 1. The psychophysical functions for two odorants: acetic acid,which stimulates both the olfactory and trigeminal system (leadingto a combined percept of odor and irritation), and phenylethylalcohol (PEA) (rose smell) which in vapor phase stimulates onlyolfactory sensations. Reprinted with permission from Green et al.(1996)

285

including category scales, ratio scales (Stevens andGalanter 1957) and more recent hybrids of these twotypes, known as category-ratio scales (Green et al. 1993,1996). Although a detailed description of the differencesbetween scaling methods are beyond the scope of thisreview, one common feature among scaling methodsthat is often a cause for concern is the inherent subjec-tivity of the response. Self-report, scalar ratings of irri-tation intensity reflect an integrated response to avolatile chemical that is comprised of (1) the sensory andphysiological signals that the chemical elicits and (2) theinterpretation of those signals as influenced by experi-ence, expectations, personality factors and other psy-chosocial or situational variables. For this very reason,however, scalar ratings can provide quite valuable in-formation in the investigation of chemical irritancy asthey are often most analogous to the subjective reportsof irritation and symptoms that occur in the workplace.Moreover, progress has been made in refining the utilityof self-report information. For example, asking exposedindividuals to provide both an affective rating (e.g.,perceived annoyance or bother) in addition to the moretypical sensory ones (e.g., irritation quality or intensity)can usefully partition the response to the volatile stim-ulus into its sensory and affective components, e.g.,(Fernandez and Turk 1992). Such a partition has beenfound to increase associations between the sensorycomponent of the response and more objective assays ofsensory irritation, such as the lateralization threshold orother techniques described below.

The second group of psychophysical methods that areused to distinguish odor and irritation accomplish thisby measuring absolute thresholds for irritation. Histor-ically, irritation thresholds were estimated by measuringnasal detection thresholds to odorants in normosmics –subjects with a normal sense of smell – and comparingthem with detection thresholds obtained from anosmics– individuals without a functional sense of smell(Cometto-Muniz and Cain 1990, 1991). The concentra-tion at which a normosmic can detect a chemical isassumed to correspond to the odor threshold, whereasthe concentration at which an anosmic can detect it isassumed to correspond to the nasal irritation threshold.Detection thresholds collected using this technique for adiverse set of chemicals have revealed that odorthresholds are typically orders of magnitude below irri-tation thresholds. However, recently, the degree towhich anosmic detection thresholds are equivalent to theirritation thresholds of normosmic individuals has beenquestioned, as evidence from electrophysiological andanatomical studies has emerged, suggesting that chem-esthetic perception is diminished among individuals wholack a functional sense of smell (Roscher et al. 1996;(Dunnegan 1993).

For this reason, a novel psychophysical method thatallows one to obtain independent thresholds for odorand irritation in individuals with normal olfactoryability has led to rapid progress in this area (Cometto-Muniz and Cain 1998;Wysocki et al. 1997). This

technique, called nasal lateralization, relies upon the factthat irritants, but not pure olfactory stimuli, can be lo-calized in the nasal mucosa. Thus, a chemical vapor (at aconcentration above odor threshold) presented to onenostril while clean air is presented to the other nostrilcan elicit a smell, but cannot be localized to the stimu-lated nostril, as long as the concentration does not ex-ceed the threshold for activating nasal-trigeminalreceptors (the functional definition of sensory irritation).In contrast, when chemical concentrations exceed thethreshold for trigeminal activation, the subject can reli-ably discriminate or lateralize which nostril has beenstimulated. By varying the concentrations presented tothe subject in a stepwise fashion throughout the test, onecan determine the concentration at which the individualcan reliably identify the stimulated nostril, which istaken to represent the nasal irritancy threshold, e.g.,(Cometto-Muniz and Cain 1998;Wysocki et al. 1997).One important advantage of this method is that incombination with an odor threshold test it providesobjective measures of both odor and intra-nasal irrita-tion within the same individual.

What factors can influence odor, irritantand health perception?

Individuals can vary in their sensitivity or response toirritants just as they vary in their sensitivity and responseto odors. Inter-individual variation in sensitivity to anirritant can be due to physiological factors, age or ex-posure history. Alternatively, extreme variations insubjective evaluations of irritancy can result from cog-nitive and emotional factors that can influence percep-tion in all sensory modalities. It is critical that thesesources of variation are well understood when designingor interpreting the outcome of research.

Exposure history

Repetitive exposure to a volatile chemical, such as oc-curs in occupational environments, can result in a de-crease in sensitivity and a reduction in the perceivedintensity of that chemical both during and followingexposure. This phenomenon, known as chemosensoryadaptation, can lead to dramatic reductions in both theodor and irritant sensations that an individual perceives.Although chemosensory adaptation is often quite spe-cific to the exposure chemical and appears to represent atransient shift in responsiveness that reverses upon re-moval of the individual from the chemical exposure, itnonetheless has implications for understanding andevaluating sensory irritancy from a workplace chemical.As can be seen in Fig. 2, absolute sensitivity to both theodor and irritancy of acetone is significantly reducedamong acetone-exposed textile workers when comparedwith matched, unexposed controls. This decreased sen-sitivity can also be observed in the much lower ratings of

286

odor, irritation and health symptoms by workers whoare experimentally exposed to concentrations at orslightly above their normal workplace levels.

Expectation and beliefs

Although symptoms and somatic sensations from vola-tile chemical exposures are determined, in part, byphysiological activity, they can also be influenced bygeneral beliefs about exposure risks. Thus, one impor-tant source of influence on the response to a chemical isthe individual’s accumulated knowledge or ‘‘mentalmodel’’ of exposure effects. Such a schema can guide theinterpretation of everyday odor experiences (Reiser et al.1985) and delimit the symptoms a person monitors andultimately perceives, e.g., (Leventhal et al. 1980; Shus-terman 2001). For example, individuals living nearhazardous waste sites who report smelling an odor oftenreport much higher symptom rates than those who donot perceive an odor (Dayal et al. 1994; Kilburn andWarshaw 1995; Shusterman et al. 1991), indicating thatpeople can be cued to monitor and report symptomsthat are related to and activated by a mental schema.

Evidence in support of these field studies comes froma series of laboratory studies that explicitly manipulatedthe perceived risk from chemical exposure. In thosestudies, individuals who were informed that the chemicalto which they were being exposed was an industrialsolvent reported significantly more odor, irritation andhealth symptoms from exposure than did individualswho were exposed to the exact same chemical but whowere told that it was a natural extract (Fig. 3) (Dalton1996; Dalton et al. 1997b;Dalton 1999). Moreover, thefrequency of spontaneous (Dalton 1996) and surveyed(Dalton et al. 1997a, 1997b) symptom reports variedsignificantly with perceived odor intensity, suggestingthat symptom perception was correlated with, or per-haps triggered by, the awareness of an odor.

Bias from odor perception

The role of response bias has emerged as a criticalvariable in studies of sensory irritation, particularlywhere odors are also present. Numerous studies of irr-itancy from airborne chemicals have contrasted re-sponses during exposure to the chemical with responsesmade during exposures to clean air. However, for mostpeople, non-odorized air signals the absence of a volatilechemical. For this reason individuals may adopt verydifferent criteria for reporting irritation in the presenceof an odor than in the presence of relatively odorless air.For example, subjects who were successively exposed toclean air and the chemical phenylethyl alcohol (PEA,used as a positive control for odor because it has anodor, but is not an irritant) reported more irritation inthe presence of PEA than in the presence of clean air(Smeets et al. 2002). Accordingly, measuring bias andsensitivity in studies that use positive odor controls canshed considerable illumination on the factors that gov-ern the reports of irritancy in the workplace.

Social factors

Reports of sensory irritation in the workplace may alsobe significantly influenced by the reactions of co-workersand other bystanders. A study that employed a confed-erate subject, (e.g., an actor who was trained to follow acarefully scripted set of verbal or behavioral responses tothe ambient odor in the chamber) showed that the per-ception of chemical odor and irritancy can be signifi-cantly altered by the reactions of other people (Daltonet al. 1999). In these studies, the confederate (who pre-sented him or herself as a research participant to the realsubject) delivered a set of positive, negative or neutralcomments/behaviors about the ambient odor. Themeasure of this effect was on the responses of the realsubjects in each of the three conditions. Subjects who

Fig. 2. Median odor andirritation detection thresholdsfor acetone and butanol (partsper million vapor concentration±SEM). Data collected fromworkers who were occupation-ally exposed to acetone andcontrol subjects who were notregularly exposed to eithercompound. Reprinted withpermission from Wysocki et al.(1997)

287

were exposed to a solvent odor while being given nega-tive cues were much more likely to report irritation andhealth symptoms (70%) than were individuals who weregiven positive cues (12%) or neutral cues (34%).

Personality variables

A significant amount of the variation in irritant andsymptom perception in healthy individuals can be at-tributed to differences in personality orientations. Ingeneral, positive affective orientations appear to lowerindividuals’ expectancies of becoming ill, while negativeorientations appear to heighten those same expectancies.Negative affectivity (NA) (Tellegen 1985; Watson andClark 1984) is a personality dimension that reflects sta-ble and pervasive differences in emotional processing,negative mood and self-concept. Individuals who arehigh in NA are more likely to experience distress in theabsence of overt stressors, exhibit hyper-vigilance inscanning their environment, interpret ambiguous stimuliin a negative manner, and report more subjective healthcomplaints (Watson and Pennebaker 1989). These ten-dencies may allow environmental stimuli, such as odors,to trigger detection of baseline levels of physiologicalactivity that would otherwise go unnoticed (Moyle 1995;Wickramasekera 1995).

Smeets and Dalton (2000) conducted a study aimedat understanding the relationship between NA and re-sponses to odor exposure. Healthy volunteers whoscored at the low and high extremes on the NA items ofthe positive and negative affectivity scale (PANAS) wereexposed to isopropyl alcohol at 400 ppm for 30 min,during which time a variety of subjective (e.g., reportedsymptoms) and objective (e.g., eye redness) adverse ef-fects was measured. High NA individuals gave signifi-cantly higher ratings of irritation during exposure thandid low NA individuals; following exposure, they alsoreported more frequent and more severe symptoms, such

as eye irritation, headache, and dizziness (Fig. 4). De-spite reports of high levels of eye irritation among thehigh NA group, no exposure-related differences in con-junctival hyperemia (redness) were observed.

The value in studying chemosensory irritationin humans

Although animal models of sensory irritation have beenwidely used to evaluate the irritant potency for manychemicals, differences in breathing patterns and theanatomy and biochemistry of the upper respiratory tract

Fig. 3. Average intensity rat-ings of the perceived odor andirritation from a 20-min cham-ber exposure to isobornylacetate for participants given apositive or negative bias aboutthe nature of the odorant.Intensity ratings were madeusing the labeled magnitudescale (BD barely detectable,W weak,M moderate, S strong,VS very strong), modified fromDalton (1999)

Fig. 4. Reported ocular irritation following a 30-min exposure to400 ppm isopropanol for individuals scoring high and low innegative affectivity (NA). Although there was no objective evidenceof ocular irritation, the high NA group rated more sensoryirritation from exposure throughout the entire duration ofexposure than did the low NA group. From Smeets and Dalton(2000)

288

in humans and rodents necessitates sometimes signifi-cant adjustments to values derived from animal studiesin order to set human exposure guidelines. Thus, theavailability of safe, non-invasive assays to measure odorand irritant responses in the species of interest, humans,can both simplify and improve accuracy in the process ofdeveloping appropriate occupational exposure guide-lines.

It is also important to acknowledge, however, thatfactors that produce variation and complicate the mea-surement of human sensory responses in experimentalsituations (such as non-specific responses to odors andgeneralized tendencies to report irritation) are likely tobe critical components of people’s reactions to irritantsin occupational environments as well. Fortunately, thegeneral nature of this problem for measuring humanresponses in other domains has prompted researchers incognitive and sensory psychology to develop methodsthat separate and quantify measures of generalized re-sponse tendencies (bias) separately from measures ofperceptual sensitivity. Thus, rather than adopting a strictfocus on designing studies that eliminate the influence ofbias, a more comprehensive strategy may be to evaluateand quantify the degree and types of bias that influencesubject reports of sensory irritation, and identify ex-perimental analogues of real-world conditions underwhich biases are likely to modify reports of irritationfollowing chemical exposure.

Discussion

Although exposure to chemicals can elicit physiologicaland biochemical changes in the ocular and upper re-spiratory mucosa, these changes may be insufficient toproduce discomfort or to compel workers to complain ofadverse effects. Alternatively, workers frequently reportirritation and health symptoms in the absence of anyobjective markers of sensory irritation and at concen-trations that can be well below those thought to producetoxicity and adverse effects. The frequent lack of corre-spondence between exposure concentration, objectivesigns of exposure-related symptoms and adverse reportshas led to problems for setting occupational exposurelimits (OELs). For this reason, there are a number ofcaveats to be considered in the design, conduct or in-terpretation of investigations of sensory irritation inhumans.

1. Selection of irritancy assays: objective and subjectivemethods can be profitably used to establish irritantpotency; the emphasis should be on well-controlledstudies of both types, in which exposures are carefullyquantified, response instruments are designed tominimize bias and the confounds of other psychoso-cial variables, and response bias in the subjectivedependent measures is either controlled or measured.Combining subjective and objective techniques with-in a group of individuals can provide important

information about irritancy mechanisms and inter-and intra-group variation. Fortunately, researcherscontinue to develop and refine a number of promisingobjective assays to measure sensory irritation at thebiochemical, psychophysiological and electrophysio-logical levels of analysis. These assays can then becoupled with well-designed subjective assessments inorder to articulate fully the human response to sen-sory irritation from volatile chemical exposure.

2. Selection of test populations: populations (e.g., oc-cupationally exposed workers and naıve individuals)differ with respect to both their exposure history andtheir expectations about chemicals. Because perceivedirritancy can be altered by sensory adaptation, ha-bituation and perceived risk, evaluations of irritancyin repetitively exposed populations may yield sub-stantively different results from those in naıve indi-viduals.

3. The evaluation of bias: response bias is a ubiquitousinfluence on people’s assessments of stimuli in allsensory domains. Nonetheless, factors that can con-tribute to response bias in experimental investigationsare equally likely to be present in real-world, occu-pational settings. For this reason, it is likely thatcorrelations between responses in the laboratory andresponses in the real world could increase signifi-cantly if investigators were to design studies thatmanipulate and quantify the degree of bias present inreports of chemosensory irritancy.

Objective measures of irritation onset obtained inconjunction with subjective responses can lend valuableinput to the decision process for determining occupa-tional exposure limits. Negative findings on objectivemeasures of irritation that cannot be reconciled withsubjective reports occurring at much lower levels of ex-posure should prompt a careful investigation into theother factors (e.g., cognitive or emotional) that may bemodulating the sensory response. For example, anxietyor worry over the consequences of exposure can elevatesymptom reports and irritation and may not be ame-liorated by a reduction in the exposure level if the odorof the chemical is still detectable. Distinguishing betweenthe exposure concentration that elicits local effects ofsensory irritation in the upper respiratory tract and theconcentration that elicits self-reports of irritation is akey component in establishing OELs that are protectiveof the majority of workers.

Acknowledgements This paper was supported by a grant from theNIH (RO1-DC0374).

References

Alarie Y (1966) Irritating properties of airborne materials to theupper respiratory tract. Arch Environ Health 13:433–449

Alarie Y (1998) Computer-based bioassay for evaluation of sensoryirritation of airborne chemicals and its limit of detection. ArchToxicol 72:277–282

289

Cometto-Muniz JE, Cain WS (1990) Thresholds for odor and nasalpungency. Physiol Behav 48:719–725

Cometto-Muniz JE, Cain WS (1991) Nasal pungency, odor, andeye irritation thresholds for homologous acetates. PharmacolBiochem Behav 39:983–989

Cometto-Muniz JE, Cain WS (1995) Relative sensitivity of theocular trigeminal, nasal trigeminal and olfactory systems toairborne chemicals. Chem Senses 20:191–198

Cometto-Muniz JE, Cain WS (1998) Trigeminal and olfactorysensitivity: comparison of modalities and methods of mea-surement. Int Arch Occup Environ Health 71:105–110

Dalton P (1996) Odor perception and beliefs about risk. ChemSenses 21:447–458

Dalton P (1999) Cognitive influences on health symptoms fromacute chemical exposure. Health Psychol 18:579–590

Dalton P, Wysocki CJ, Brody MJ, Lawley HJ (1997a) Perceivedodor, irritation and health symptoms following short-term ex-posure to acetone. Am J Ind Med 31:558–569

Dalton P, Wysocki CJ, Brody MJ, Lawley HJ (1997b) The influ-ence of cognitive bias on the perceived odor, irritation andhealth symptoms from chemical exposure. Int Arch OccupEnviron Health 69:407–417

Dalton P, Dilks D, Ruberte J (1999) Effects of social cues onperceived odor, irritation and health symptoms from solventexposure. Proceedings and abstracts of the Annual Meeting ofthe Eastern Psychological Association 70:135

Dayal HH, Baranowski T, Li Y, Morris R (1994) Hazardouschemicals: psychological dimensions of the health sequelae of acommunity exposure in Texas. J Epidemiol Community Health48:560–568

Dunnegan KJ (1993) Framing, cognitive modes and image theory:toward an understanding of a glass half full. J Appl Soc Psychol78: 491–503

Fernandez E, Turk DC (1992) Sensory and affective components ofpain: separation and synthesis. Psychol Bull 112:205–217

Green BG, Shaffer GS, Gilmore MM (1993) Derivation and eval-uation of a semantic scale of oral sensation magnitude withapparent ratio properties. Chem Senses 18:683–702

Green BG, Dalton P, Cowart BJ, Shaffer G, Rankin KR, Hig-gins J (1996) Evaluating the ‘‘Labeled Magnitude Scale’’ formeasuring sensations of taste and smell. Chem Senses21:323–334

Kilburn KH, Warshaw RH (1995) Neurotoxic effects from resi-dential exposure to chemicals from an oil reprocessing facilityand superfund site. Neurotoxicol Teratol 17:89–102

Leventhal H, Meyer D, Nerenz D (1980) The commonsense rep-resentation of illness danger. In: Rachman S (ed) Medicalpsychology. Pergamon, New York, pp 7–30

Moyle P (1995) The role of negative affectivity in the stress process:Tests of alternative models. J Org Behav 16:647–68

Reiser BJ, Black JB, Abelson RP (1985) Knowledge structures inthe organization and retrieval of autobiographical memories.Cognit Psychol 17:89–137

Roscher S, Mohammadian P, Schneider A, Wendler J, Hummel T,Kobal G (1996) Olfactory input facilitates trigeminal chemo-sensitivity. ECRO Abstracts 21:24

Shusterman D, Lipscomb J, Neutra R, Satin K (1991) Symptomprevalence and odor-worry interaction near hazardous wastesites. Environ Health Perspect 94:25–30

Shusterman D (2001) Odor-associated health complaints: compet-ing explanatory models. Chem Senses 26:339–343

Smeets MA, Dalton P (2000) The nose of the beholder. AromacholRev 8:1–10

Smeets MA, Maute CM, Dalton P (2002) Acute sensory irritationfrom exposure to isopropanol in workers and controls: objec-tive versus subjective effects. Ann Occup Hyg (In press)

Stevens SS, Galanter EH (1957) Ratio scales and category scalesfor a dozen perceptual continua. J Exp Psychol 54:377–411

Tellegen A (1985) Structures of Mood and Personality and TheirRelevance to Assessing Anxiety, with an Emphasis on Self-re-port. In: Tuma AH, Maser JD (eds) Anxiety and anxiety dis-orders. Erlbaum, Hillsdale, NJ, USA pp 681–706

Watson D, Clark LA (1984) Negative affectivity: the disposition toexperience aversive emotional states. Psychol Bull 96:465–490

Watson D, Pennebaker JW (1989) Health complaints, stress, anddistress: exploring the central role of negative affectivity. Psy-chol Rev 96:234–254

Wickramasekera IE (1995) Somatization: concepts, data, and pre-dictions from the high risk model of threat perception. J NervMent Dis 183:15–23

Wysocki CJ, Dalton P, Brody MJ, Lawley HJ (1997) Acetone odorand irritation thresholds obtained from acetone-exposed fac-tory workers and from control (occupationally non-exposed)subjects. Am Ind Hyg Assoc J 58:704–712

290