Eye/airway irritation and odor are important components included in the classic ‘sick building syndrome’ in non-industrialized buildings (Burge, 2004; Hodgson, 2002; Redlich et al., 1997). They are common symptoms (complaints) that may be experienced simultaneously and thus may interact with each other. Thus independent evaluation may be difficult, if not impossible (Dalton, 2003). For this reason, it is important to understand the characteristics and contributions of sensory irritation and odor to the overall perception and reporting of the indoor air quality. Perceived indoor air quality, in this paper, refers to the overall perception of sensory irritation symptoms and odor that accumulates during a working day. This is to be opposed to the ‘immediately’ perceived experience when entering a building or room from the outside (or short-term sensory perception of material emissions in climate chambers). The distinction is relevant for the development of practical guidelines for sensory irritation and odor annoyance in the indoor climate (cf. World Health Organization, 1989).
Eye and upper airway sensory irritation
Both eye and upper airway irritation are common symptoms in indoor environments; however, data collection is problematic (Brightman and Moss, 2000). Differences in design of the questionnaires including symptom type, the use of different recall periods and frequency categories of symptoms may explain the large differences from study to study. For example, in 56 European buildings in nine countries, the mean prevalence of dry eyes was 39% expressed as at least once the preceding month, this dropped to 26% by asking as ‘experienced at work at this moment’ (Bluyssen et al., 1996). Two studies have administered sequential questionnaires over an extended period. One such study showed a remarkable reduction of the prevalence of eye symptoms within a period of 4–12 months (Chao et al., 2003). A similar time trend has been found within a period of 6 weeks (Tamblyn et al., 1992). The possible causes of this decline are not entirely clear. The mean prevalence of eye-related symptoms is considerably lower if the complaint frequency is often or constant (Doughty et al., 2002), but substantially above an estimated background prevalence of 5% for eye irritation (Wolkoff et al., 2003).
Different terminology has been used to characterize the sensory perception evoked by airborne chemicals, especially as they refer to exposures close to TLV levels (cf. Doty et al., 2004). For example, ‘the common chemical sense’, which describes mucosal sensitivity to chemicals and more recently pungency and chemesthesis (‘chemical irritation’), both encompass mucosal and dermal sensations, but not odor. Pungency refers to nasal and oral chemosensation responses mediated through the trigeminal nerve (fifth cranial nerve). ‘Sensory irritation’ is a general term, comprising specifically eye and upper airway irritation, used by indoor air scientists and airway toxicologists.
Many symptoms and signs have been used for the characterization of sensory irritation (Doty et al., 2004). Some of these are conceptually overlapping in questionnaires, which adds to the overall confusion (Doughty et al., 2002; Rolando et al., 1998; Wolkoff et al., 2003). For example, one of the most common symptoms in indoor environments, ‘dry eyes’ has been equated to and associated with complaints of irritated eyes, and in some cases, the combination ‘dry, itching, or irritated eyes’ is used (cf. Guillon, 2002). The large number of different symptoms for eye irritation (e.g. dry or smarting) or clusters thereof may reflect different ocular mechanisms that are difficult to differentiate (Wolkoff et al., 2005a). For example, the symptoms itching, irritating, grating and sandy have been found to cluster together (Lundin, 1991). In another study, the three symptoms ‘dryness, smarting and itching’ were found to be occupationally related (Aronsson and Strömberg, 1995).
Sensory irritation symptoms have been reported with intensity from severe, such as pain, smarting, burning or irritating, to less severe, such as itchy eyes, dry eyes or discomfort in the eye (cf. Hedge et al., 1996; Norn, 1992). However, descriptors of eye irritation have so far not included details about its location (i.e. the inner and exterior eyelids vs. the eyeball itself), diurnal variation, onset, duration, or alleviating factors (cf. Gilbard, 1999). In addition, the same symptom(s) may arise from different diseases, for example dry eyes and inflammation of the Meibomian glands (Meibomian dysfunctions) both result in sandy-gritty irritation, and the symptoms are insidious in onset (Gilbard, 1999); similarly, dry eyes and allergic conjunctivitis are difficult to differentiate. Further, individuals with perfume contact allergy or allergic rhinitis are likely to report sensory irritation more frequently and more severely following VOC exposure than those without (e.g. Elberling et al., 2004; Shusterman et al., 2003).
During exposure, sensory irritation symptoms may be persistent or transient. Their development in the indoor environment is characterized by latency, i.e. the symptom is experienced with delay in contrast to odor perception. This has been reported from studies of city halls and libraries where reported ‘irritation symptoms’ increased during a working day (Baird et al., 1994; Skov et al., 1989). In a climate chamber study, subjects exposed over the period of hours to butanol and formaldehyde emitted from an acid-curing lacquer reported sensory irritation with considerable delay. In contrast, a naive panel perceived the odor immediately, but no sensory irritation (Wolkoff et al., 1991b). These and similar observations indicate the role of time for the development and perception of irritative symptoms (Bender et al., 1983; Hempel-Jørgensen et al., 1999; Hudnell et al., 1993). The odor masking effect of butanol in the above case is another possibility (cf. Cain and Murphy, 1980; van Thriel et al., 2003). For this reason, published irritation thresholds that are based on subjective evaluation of short-term exposure, are less suited for the evaluation of indoor concentrations, because of longer exposure durations and lower concentrations. In addition, adaptation (physiological process) and habituation (mainly a psychological process, i.e. familiarity with the sensation) are other important confounding factors that may result in overestimated thresholds (Arts et al., 2002). Overall, the sensory irritation symptoms are often reversible after cessation of exposure.
Observations and cautions for the assessment of sensory irritation from organic compounds in the indoor environment:
Table 1. Estimated threshold limit values (based on sensory irritation), indoor air norm values for sensory irritation, human sensory irritation threshold, odor threshold and reported concentration of selected common organic compounds in indoor air
|Organic compound|| 0.03 × RD50 TLV (mg/m3)||Estimated indoor air guideline1,2 [0.03 × RD50/40 (TLV/40)mg/m3]||Human sensory irritation threshold [ref 3, if not otherwise stated (mg/m3)]|| Odor threshold (μg/m3)|| Indoor concentrations reported after year 2000 [mean–max (μg/m3)]|
|Decane||>129 (based on 0.2 × RD0)4||>3|| ||43,7005 30876||9–297; 3–23708|
|Toluene||3899||10|| ||60010 12396 60005||30–447; 28–95008; 2211|
|p-xylene||1769||4|| ||14105 mixt. of m-/-p 7810 p-isomer1616||10–597; 1511|
|(+)-α-pinene||35012||9||20||1006||23–447; 13–29528; 711|
|(+)-limonene||18013||4||44014||2116||33–657; 8–249018; 1611|
|2-butoxyethoxyethanol|| || ||Disregarded||910||<3–6218|
|Butanone||7959||60|| ||13006 87010||3–2438|
|Acrolein||0.07||0.002|| ||4075 86||Oxidation product|
|Methacrolein||0.917||0.03|| ||246||Oxidation product of isoprene|
|2-decenal|| || || ||∼418||Emitted from linoleum19|
|Acetic acid||2320||0.6||2.5||3635 4310 156||Emitted from wood products|
|Hexanoic acid|| || || ||6010 36||Emitted from linseed-based products|
|Tetrachloethylene|| || || ||427005 52176||5–157; <1–55408; 311|
|Ozone||No sensory irritation21|| || ||66||10–30022|
|Nitrogen dioxide|| || || ||3555 2256||300–300022|
Sensory irritants formed during terpene oxidation reactions
The sensory irritation of the monoterpene oxidation products has been evaluated by a mouse bioassay and a human eye exposure model. The results from the mouse bioassay, which estimates airway irritation from reduction in the respiratory rate, suggested that the R-limonene/ozone (LO; Clausen et al., 2001), α-pinene/ozone (PO; Wolkoff et al., 1999) and isoprene/ozone (IO) reactions generate sensory irritants of known and unknown structures (Wilkins et al., 2001). The sensory irritation effect is significantly higher than that exhibited by the identified reaction products and residual concentration of the reactants. The identified sensory irritants inter alia include formaldehyde, methacrolein, methyl vinylketone and formic and acetic acid (Wolkoff et al., 2000). In a study, male subjects have been exposed in one eye for 20 min with LO, IO, the nitrate radical, methacrolein and residual reactants. The eye blink frequencies of the subjects were recorded as a physiological measure of trigeminal stimulation (Klenø and Wolkoff, 2004; Nøjgaard et al., 2005). Mean blink frequencies increased significantly only during exposure to LOs and methacrolein compared with that of clean air, and the findings coincided with qualitative reporting of weak eye irritation symptoms. The blink frequency showed a decreasing trend with increase of the relative humidity from 20% to 50% for LO mixtures (Nøjgaard et al., 2005). A similar effect was observed in the mouse bioassay in which sensory irritation was highest at low relative humidity (Wilkins et al., 2003). The observed effects may be ascribed to the formation of less irritation species, a more stable mucous membrane, a more stable eye tear film or a combination.
The above findings substantiate that gaseous LO reaction products cause trigeminal stimulation and possibly eye irritation at ozone- and limonene concentrations that are close to high-end values measured in indoor settings (Wolkoff et al., 2000). The etiological fraction, explained by such alkene oxidation products, however, remains to be evaluated in the context of other occupational factors, e.g. demanding computer work in combination with low relative humidity (cf. Wolkoff et al., 2005a). The impact of ultrafine particles on short-term symptoms such as eye and upper airway irritation is unknown and their possible role in the development of effects in the lower airways is at present speculative (cf. Rohr et al., 2002, 2003).
It is clear that oxidation reactions between certain unsaturated VOCs and oxidants like ozone produce sensory irritants. This is referred to as ‘the reactive chemistry’–hypothesis (Weschler and Shields, 1997b; Wolkoff and Nielsen, 2001). Some epidemiological studies have indicated that the sum of detectable VOCs may be lower in an office building classified as ‘sick’ as compared with a similar building classified as ‘healthy’ (cf. Berglund et al., 1993; Groes et al., 1996; Höppe et al., 1995; Lundin, 1993; Subramanian et al., 2000; Sundell et al., 1993; Willers et al., 1996). In a study of buildings in California, it was found that cleaning products and water-based paints accounted for a significant proportion of the observed association of irritation symptoms (Ten Brinke et al., 1998). Citrus and pine oils, in which terpenes (unsaturated VOCs) are major constituents, are common ingredients in such US products (Nazaroff and Weschler, 2004). In addition, one study has shown an association between terpene concentrations and deteriorated lung functions; however, an interpretation is hampered, because of other risk factors (Norbäck et al., 1995).
Observations and cautions for the assessment of formation of sensory irritants in the indoor environment:
Unidentified species for which sampling techniques are unavailable may be partly responsible for sensory irritation effects (Weschler and Shields, 1997b
; Wolkoff et al., 1997
The ozone (outdoor and indoor) and formaldehyde concentrations should be measured in environments suspected to have strong sources of terpenes.
High relative humidity may alleviate sensory irritation effects of alkene oxidation products.
Odor perception is omnipresent in our daily life, including work. Odor as opposed to sensory irritation is immediate with steep time-response curves (Berglund and Lindvall, 1992). The character of odors represents a large variety from pleasant (e.g. perfumes, flowers) to unpleasant (malodors; Distel et al., 1999; Duffee and O'Brien, 2000), but the interaction between odor and a person's psychological state (e.g. emotion/mood) is complex, and cultural differences exist (Ayabe-Kanamura et al., 1998).
The step from odor to the cognitive evaluation of the odor (e.g. annoyance, Berglund et al., 1999) is influenced by a number of personal factors including adaptation, habituation, exposure history, expectation and beliefs about health risk (i.e. informational bias), personal psychological variables and social factors (e.g. personal bias) (Dalton, 2002), and environmental factors (Sucker et al., 2001). In particular, belief concerning health risk has a strong influence, because ‘it creates a context through which perception is filtered’ (Bell and Paton, 2001). However, a major limitation is that most of the research on odors has been carried out at industrial concentrations close to threshold limit values, and its relevance to indoor air settings is questionable. In any case, habituation would be expected to diminish any concern about health risk (cf. Distel et al., 1999).
There is no evidence that malodors per se are associated with objective adverse health effects (Cavalini et al., 1991; Rosenkranz and Cunningham, 2003). However, malodors (as perceived in industrialized cultures, cf. Ayabe-Kanamura et al., 1998) are generally undesirable in the indoor environment. Generally, odor perception provides an adequate warning for the onset of eye/airway irritation (Cometto-Muñiz and Cain, 1995). Some odors appear to influence the pattern of reporting symptoms, for example self-reported health, productivity and mood (cf. Gilbert et al., 1997; Gijsbers van Wijk and Kolk, 2001; Knasko, 1996). For example, it has been found that visual contact to the odor source, e.g. the smoker, enhances the intensity of reporting tobacco smoke (Moschandreas and Relwani, 1992). Exposure to a malodor resulted in startle potentiation, which may be interpreted that the odor triggers a negative emotion (Miltner et al., 1994). When the odor source is unidentified, the level of negative emotion could increase and this would decrease the hedonic quality (pleasantness or acceptability) and increase the arousal level. Provision of information about the odor source could thus decrease the level of negative emotion and increase the hedonic quality; this may reduce the general arousal level.
Certain ‘vulnerable’ subjects may experience health effects in form of somatic symptoms (Segala et al., 2003; Steinheider, 1999). Combined mechanisms of panic disorder and cognitive mediated fear response have been proposed for explanation (Staudenmayer et al., 2004). Environmental awareness and belief through warning about both pleasant odors and malodors facilitates learning about subjective health symptoms such as airway irritation, i.e. learned aversions (Devries et al., 2004; Van den Bergh et al., 2002; Winters et al., 2003). For example, self-selected healthy subjects reported six times more eye irritation when exposed for 1 h to diluted air from a swine confinement than from clean air, this despite measured compounds were well below known irritation thresholds (Schiffman et al., 2005). The authors suggest combined effects of pollutants or learned aversions are responsible for this. Co-pollutants that are part of an odorant mixture (e.g. microbiological species from water-damaged materials) could also cause health effects.
Results from experiments that involve the presentation of certain odors, like lavender and rosemary, under controlled laboratory conditions suggest that their effects are mainly psychological (Ilmberger et al., 2001; see below). These exposures may alter a number of psychological conditions such as mood, alertness and performance (associated with alertness/arousal/vigilance) relative to clean air conditions, but the effects differ depending on concentration, repetition of odor stimuli, type of task and possibly the individual arousal (motivation towards a change) level prior to exposure. The complexity is reflected in the Yerkes–Dodson law, which describes the association between arousal (e.g. stress reflected as odor) and performance as an inverted U-curve (Yerkes and Dodson, 1908). This relationship predicts that as arousal increases, performance of a task improves, but only to a point where it then starts to decrease. Exposure to a sedative odor (e.g. lavender) or an ‘alerting’ odor (e.g. peppermint or jasmine) will cause a decrease of the performance (see e.g. Degel et al., 2001). It is quite clear that mood and alertness (attention) influence the mental and cognitive state, and perhaps mental creativity; however, for how long and what concentration(s) are required is unknown. Based on an extensive review, it was concluded that ‘weak and even unnoticed concentrations of odors often exert a stronger influence on human behaviour than stronger and explicitly perceived ones’ (Köster and Degel, 2001), but the effects under laboratory conditions appear to be modest (e.g. Baron, 1990); quantified implications for the performance at the actual workplace is difficult to predict.
Reported odor detection thresholds of VOCs are generally one to four orders of magnitude lower than estimated thresholds for irritation effects of the upper airways (Cometto-Muñiz et al., 2004; Wolkoff, 1999), see also Table 1. In addition, many reported odor thresholds are too high, sometimes by orders of magnitude. A likely explanation is errors associated with the olfactometric measurements; only recently, an international standard procedure has been developed for odor threshold determination by dynamic olfactometry and use of butanol as a reference odor (CEN, 2003). A comparison of compiled data of older odor thresholds (Devos et al., 1990) with newer data from (Woodfield and Hall, 1994) and (Nagata, 2003) shows the trend of lower thresholds for a number of chemical classes; see also examples in Table 1 (note that the odor threshold of butanol in the two latter compilations is about the same, 40 ppb). The possible influence of temperature and relative humidity on odor threshold determination appears not to be known.
At VOC concentrations that are well below their irritation thresholds, but above their corresponding odor thresholds, reports of perceived irritation most likely is a result of odor annoyance, and possibly accompanied by concern for toxicity (Dalton, 2003). These reactions are probably psychological in nature, possibly a reaction to an ‘unknown’ airborne chemical. It is plausible that a similar mechanism also exists for indoor levels. However, it is unlikely that individual odors emitted from building materials, office equipment or the ventilation system can be differentiated from odors originated from other sources. Exceptions, however, include ozone and nitrogen oxides emitted from photocopiers and certain odors from mold growth.
Clearly, mold odor is a sign of moisture damage of building materials. This might be interpreted as uncontrolled risk of exposure to elevated concentrations of indoor pollutants, e.g. VOCs and particles, thus possibly triggering a psychological process towards more negative reporting of the indoor air quality.
Reduced air quality because of emission of organic compounds from an old carpet or office equipment in field laboratories has been associated with productivity deterioration, for example slower text typing speed and more typing errors (Bakó-Biró et al., 2004; Wargocki et al., 1999). Two different explanations have been proposed: (i) perception of poor air quality caused headache (carpet study, only), which reduced the effort exerted by the subjects, thus lowering the speed of typing. (ii) Unidentified organic compounds caused the decrements (office equipment study). Headache itself can be the result of depression of breathing caused by perceived odors (cf. Danuser et al., 2003; Schiffman and Williams, 2005). However, a more general explanation could be that nearly perceptible odors of the emitted organic compounds cause mental and cognitive distraction of the subjects (e.g. by extension of the reaction time), which results in reduced performance, especially if the odor were perceived as unpleasant or unrecognizable (Danuser et al., 2003; see also Herz, 2002). The etiological fraction of performance alteration that is caused by odors (e.g. material emissions) needs analysis in context of other important occupational factors in the office environment.
Observations and cautions for the assessment of odors from organic compounds in the indoor environment:
The degree of annoyance of just perceptible odors greatly depends on personally related differences, i.e. different coping strategies and possibly strong individual associations with a given odor (Dalton, 2002
). For example, odors that are perceived as pleasant have a lower annoyance potential than unpleasant ones (Both et al., 2004
), and possibly also stronger physiological changes (cf. Danuser et al., 2003
A fraction of the population may perceive a given odor at least one order lower than the majority of the population according to the definition of an odor threshold (50% median response fraction).
Sensory irritation and odor perception may be confused.
Information, experience and habituation may alter the association of perceived health effects (Devries et al., 2004
; Opiekun et al., 2003
; Van den Bergh et al., 2002
; Winters et al., 2003
), and reported odor intensity in some cases (Distel et al., 1999
Concentration may not be the best indicator of the impact of an odor, partly because persistence and hedonic value influence its impact (Nicell, 2003
), and the relationship between concentration and pleasantness or intensity may be inverse for some compounds (Cocheo et al., 1991
; Whelton and Dietrich, 2004
Short-term evaluation of the perceived air quality is probably not relevant for the evaluation of symptoms built up during the working day (Bluyssen et al., 1996
; Wolkoff et al., 1991b
). One exception could be the presence of moldy odor.
Many reported odor thresholds are too high (comparison of data from Devos et al., 1990
, with the data from Nagata, 2003
; Woodfield and Hall, 1994
Some odor thresholds may be too low because of trace amounts of impurities (e.g. formed by oxidation of the compound) that have much lower thresholds than the unaltered compounds.
The immediately perceived odor of VOCs emitted from some building materials is influenced by thermal factors, and in particular high relative humidity may deteriorate the sensory perception (Cain et al., 2002
; Fang et al., 1998
Productivity reduction caused by organic compounds (e.g. from building materials) in climate chambers is encumbered by the complexity and influence of odors on human behaviour. The effect may be only temporary (Danuser et al., 2003