Volatile Organic Compounds, Indoor Air Quality and Health

Lars Mfl)lhave1

AbstraC:t lntrodudion 'I'his publication summarizes field investigations and Volatile Organic Compounds controlled experiments on the relation between low le vels ofindoor air pollution with volatile organic com- and Health pounds (VOC) and human health and comfort. 1'he Volatile organic compounds (VOC's): VOC's Henle-Kock criteria from epidemiology are reoised for are :frequent air-pollutants-in non-industrial the dose-response relation between VOC's and health environments. A WHO working group cate as comfort effects and existing e'lJidence for each criter gorized the entire range of organic indoor ion are discussed. A biological model for human re pollutants into four groups, as indicated in sponses is suggested, based on three mechanisms: sen Table 1. The four categories were defined by sory perception of the environment, weak inflammatory boiling-point ranges and no sharp limits ex reactions, and environmental stress reactions. Further, ist between them. The VOC category was de the TVOC-indicator concept for exposure is discussed. fined by a boiling-point range with a lower 1'he conclusion is that no experimental or field data contradict the proposed causality. On the contrary, evi limit between 50 °C and 100 °C and an upper dence supports the suggested causality. 1'he biological limit between 240 °C and 260 °C, where the model, however, is not yet based on acceptable meas higher values refer to polar compounds ures of the variables for exposures, co-variables or (WHO, 1989). health effects. A tentative guideline for VOC's in A recent review lists 307 VOC's identified non-industrial indoor environments is suggested. 1'he in indoor air in different countries (Berglund 3 no-effect level seems to be about 0.2 mg/m • A multi et al., 1986). The WHO report (WHO, 1989) factorial exposure range may exist between 0.2 and 3 on VOC's indoor summarized the concentra 3 mg/m • Above 3 mg/m3 discomfort is expected. tions found in four major European studies (Krause et al., 1987; De Bortoli et al., 1986; Le bret et al., 1986; Wallace, 1987). In summary, normally 50 to 300 volatile organic compounds are found in air samples from most non-indus trial environments. Each compound seldom 3 exceeds a concentration of about SO µglm , KEYWORDS: which is 100 to 1000 times lower than relevant Health, Comfort, Perceived indoor air quality, En occupational threshold values (TLV's) (AC vironmental stress, TVOC, Volatile organic com GIH, 1988). An upper extreme average total pounds. concentration of all VOC's in normal homes seems to be 20 mglm3 (M~lhave, 1986). The to Manuscript received: 22 October 1990 3 Accepted for publication: 19 December 1991 tal concentration of all VOC's (mglm ), how 3 ever, is normally well below 1 mglm , which is ' Institute of Occupational and Environmental Medi cine, University of Aarhus, Building 180, Universitet only 0.2% of the occupational threshold limit sparken, DK-8000 Arhus C., Denmark. value (TLV) for toluene. . .

358 M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health

Table 1 Classification of indoor organic pollutants •.

Description Abbreviation Boiling-point range ("C)b

Very Volatile (gas~us) Organic Compounds woe < 0 to 50-100 Volatile Organic Compounds voe 50-100 to 240-260 Semivolatile Organic Compounds svoc 240-260 to 380400 Organic Compounds Associated with Parti culate Matter or Particulate Organic Matter POM > 380

• From WHO 1989. b Polar compounds appear at the higher end of the range.

Health of the special effects are not caused by the air Health has been defined by WHO as "A state pollutants themselves, but by e.g. metabolites of complete physical, mental and social as in the case of the neurotoxicity of n-hex well-being and not merely the absence of dis ane and certain ketones, the ocular toxicity ease or infirmity'' (WHO, 1961). The toxic ef of methanol, or the hepatoxicity of certain fects of volatile organic compounds may be chlorinated hydrocarbons. A list of special defined as any significant change caused by effects of VOC's and commonly used tests VOC's on a person when compared to other for such effects are shown in Table 2, which wise comparable, but unexposed persons. is adapted from Andrews et al. (1986). This Such effects may be classified in effects com table lists effects or tests commonly used in mon to most VOC's and effects specific for human studies of normal industrial organic individual compounds. solvents at or above TLV-levels. Reviews of The special properties of individual com such special effects are found in textbooks. pounds may result in unusually high intens Special effects like genotoxic effects or ef ity or prevalence of common effects (such as fects on the immune system are severe for neurotoxic effects, cancer, or allergy). Many the few unlucky occupants affected by them.

Table 2 Symptomatology and commonly used tests for behavioral effects caused by VOC°.

Symptomatology Test Sensory: Parasthesias, visual or auditory deficts Neurologic, sight, and hearing examinations Cognitive: Memory (both shon-tenn and long-tenn), Wechsler Memory Scale confusion, disorientation Wechsler Adult Intelligence Scale (WAIS) Affective: Nervousness, irritability Eysenck Personality Inventory depression, apathy, compulsive Rorschach Test behavior Digit-symbol Substitution Task Bourdon-Wiersma Vigilance Task Motor: Weakness in hands, incoordination, Neurologic examination fatigue, tremor Santa Ana Dexterity Test Finger-tapping Test Simple or Choice Reaction Time

a Andrews et al., 1980. .

\ ' \ L 1~ :

Mr1.1lhave: Volatile Organic Compounds, Indoor Air Quality and Health 359

They are, compared to what we know, rare ginally were developed for the causality of effects in relation to low-level voe expo diseases described by a spectrum of re :-point range (oC)b sures in normal indoor environments. Such sponses and caused by specific agents. These effects, however, may evade registration if criteria have been revised in this paper for : 0 to 50-100 their prevalence is too small to cause a statis 00 to 240-260 tically significant association in a population Table 3 Henle-Kock-criteria for causation revised for 60 to 380-400 of the size found in most normal non-indus the effects of low-level voe exposure among occu trial buildings. In the future, they should be pants of non-industrial indoor environments". > 380 investigated further. 1) A spectrum of responses should follow exposure to This paper is focused on the causality be VOC's along a predicted biological gradient from hind the more common effects of VOC's ac mild to severe. cording to the above definition. These also 2) A measurable response following exposure to VOC's happen to be the most frequent effects of should regularly appear among occupants lacking low-level voe exposure. For simplicity, the this before exposure or should increase in magni caused by the air tude if present before exposure. r e.g. metabolites effects of low-level exposures to VOC's will be divided into a) disturbances of body func 3) Exposure to VOC's should be present more com oxicity of n-hex monly in those showing the effect than in controls e ocular toxicity tions and b) perception of the environment without the effect when all risk factors are held xicity of certain or of body conditions. Further, to avoid con consomt. A list of special fusion, the term perceived indoor air quality 4) Temporality: The effects should follow exposure to nonly used tests (PIAQ) will be used for the subjectively per VOC's (possibly with a delay time). n Table 2, which ceived air quality in contrast to measured in 5) Elimination ofmodification of the VOC exposure or t al. (1986). This door air quality (MlAQ) which is the chemic prevention or modification of the occupants' sensi ally or physically measured indoor air qual tivity to exposure should decrease or eliminate the mmonly used in effect. idustrial organic ity, e.g. in the form of air temperature, air humidity, pollution concentrations, etc. A 6) Experimental reproduction of the effects should pro :vels. Reviews of duce higher incidence in animals or humans ap l in textbooks. distinction will be made between the occu propriately exposed to VOC's than in those not so >Xie effects or ef pants' adapted perception after exposure for 1 exposed. n are severe for hour or more and the visitors' immediate 7) The etiology must make biological sense. 1ffected by them. and adapted response upon entering the room. Generally, in this paper these two • Evans 1976. types of evaluation are referred to as subacute

and acute responses. Also, the term "irrita Table 4 Four criteria for acceptable biological expla tion" will be specified as either stimulation nation of the etiology between low-level voe exposure of sensory systems, inflammatory-like skin and observed effects". a.ations reactions, or a psychologic mood condition. 1) Reasonable and documented biological mechan If not specified, sensory irritation is meant. ism, must be used for the explanation of the ob served symptoms or effects. Criteria for Causality 2) The variables (effects, exposures and cofactors) in Tables 3 and 4 show a suggested set of criter volved in the suggested etiology must be measur ia for causality and for biologically reason able: alternatively acceptable indicators must be available. able etiologies in relation to voe exposures at low levels and health effects as they may 3) The effects or symptoms and expected mechan isms should preferably be known from similar or occur in non-industrial indoor environ higher exposure levels to VOC's. ments. In epidemiology, the criteria in Tables 4) The etiology must explain any delay (latency) in 3 and 4 are known as the Henle-Kock-criter responses following exposure to VOC's. ia (Evans, 1976). The Henle-Kock criteria ori- • Evans 1976. 360 Mriilhave: Volatile Organic Compounds, Indoor Air Quality and Health the causality relation between VOC's and in The first class contains the primary per door air quality (IAQ). ceived stimulation of the sensory system caused by the air pollution. Especially olfac Biological Model for Health tion and the trigeminal nerve seem to be in Effects caused by VOC volved in these reactions. The second class includes effects related to changes of the skin Exposures at Low Levels or other exposed tissue. These effects are ei The Background for a Biological ther directly caused by the exposure or sec Model of Human Readions to ondarily caused by perception of changes in Low-level Exposure to VOC's the tissues. Such changes may be mediated One of the Henle-Kock criteria in Table 3 for by nervous or biochemical reflexes. The a causality between VOC's and health effects third class includes observable changes in (e.g. complaints about reduced indoor air human behavior (Cain, 1989). quality) was that a reasonable biological The sub-acute effects are observed (M~l model exists which can explain the known have, 1986) as headache and other weak sub features of the suggested causality. In the fol jective nervous-related effects or weak in lowing, such a model will be suggested with flammatory-like reactions. Chronic effects the aim of explaining as many as possible of like systemic, genotoxic or immune-system the indications found in the literature and effects caused by absorbed or metabolized leaving a minimum of unexplained evidence VOC's are not frequent consequences of of effects. The suggested model is still a post low-level VOC-exposures (M~lhave, 1986). ulate and should be challenged in future ex periments and investigations. A Biological Model of Human Little is known about the effects of the Readions to Low-Level VOC Exposures low-level voe exposures, which are charac From the evidence summarized above, it ap teristic of non-industrial environments. Evi pears that the most frequent acute or sub dence from experiments and investigations acute effects of voe exposure at a low level indicates that, generally, the responding tis fall into three main classes: a) perception of sues are mucosa! membranes in eyes, nose the environmental exposure caused by acute and throat, skin on the face, neck and hands, stimulations of senses, b) perception or ob and the upper and lower airways (M~lhave, servation of weak acute or subacute inflam 1986). The most frequent effects seem to be matory-like reactions in the exposed tissues, consequences of reactions close to the surface and finally c) a number of effects which may of the tissue exposed to air. Generally, the ef be described as a group of subacute environ fects are reversible and disappear shortly mental stress reactions caused by the percep after the exposure is ended (M~lhave, 1986; tions (M~lhave, 1990b ). M~lhave, 1990a; M~lhave, 199la). Perceived air quality: Our present knowl The most frequent effects seem to be edge about the senses indicates that VOC's acute. They may, like perception of odors, are sensed by either the odorous sense in the show adaptation (Clausen et al., 1985). Some top of the nasal cavity, the gustatory senses effects may be sub-acute in which case they, on the tongue or the chemical sense (Cain, like headache, are expected to increase in fre 1989). The three sensory systems - odor, taste quency and intensity with increasing expo and chemical sense - respond to airborne sure time (Otto et al., 1990; M~lhave, 1990a). chemicals, but to different qualities of the ex The acute effects appear to fall into three posure. Stimulation of one, two or three of classes (Cain, 1989; Cain et al., 1986; Comet these sensory systems seems to result in a to-Muniz et al., 1984; Medina et al., 1982): combined perception of something which ,.. .< [ 1~ ;

M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health 361

be primary per may be called the perceived indoor air qual the low-level voe exposures in non-indus sensory system ity (PIAQ). This perceived air quality may trial environments. Generally, the first sign Especially olfac even include additional nerve signals from of acute inflammation is indications of peri ve seem to be in other senses like vision (e.g. haze), or the pheral dilatation of vessels causing color and f he second class thermal environment. temperature changes of tissue. Subsequently 1anges of the skin The chemical sense includes both the tri granulocytes and other cell types are activa ese effects are ei geminal nerve in facial skin and in mucosa! ted (Thaysen et al., 1980). exposure or sec membranes of eyes, nose and mouth, as well Most acute inflammatory reactions are on of changes in as other similar non-myelinated nerves in supposed to be activated by chemical medi nay be mediated other skin areas (Cain, 1989). These nerves ators released in the exposed tissues. More al reflexes. The have polymodal receptors. They, therefore, than ten different chemical classes of medi 'able changes in can respond to many different types of stim ators have been identified (like histamine ). uli. The receptors are supposed to respond to and kinines). These mediators are produced ! observed (Mi;;;l environmental chemicals following a chemi after an external exposure to irritants. Most other weak sub- cal reaction or a physical adsorption of the of these mediators are themselves known to cts or weak in compounds to the receptor proteins. be sensory irritants and are believed to sti Chronic effects Activation of the senses leads to two ef mulate sensors in the tissue and cause a sec immune-system fects. First, a sensation of, for example, an ir ondary perception of the exposure (Thaysen or metabolized ritation, a burning, smarting or stinging feel et al., 1980). consequences of ing and, secondly, protective reflexes. These If the exposure increases in intensity or li;;;lhave, 1986). may, for example, be tearing, changed respir duration beyond the point of comfort or safe atory frequency, cough or sneezing (Brink et ty, the body may react by initiating protec ;1man al., 1948; Nielsen et al., 198Sa; Nielsen et al., tive reflexes or mechanisms. These reflexes roe Exposures 1988). may be activated either by chemical medi fzed above, it ap At low-level exposures the common che ators or through sensory perception and ner nt acute or sub mical sense exhibits some spatial summa vous signals. Examples are running eyes or Ire at a low level tions i.e. the more tissue and sensors exposed nose, cough, changes in respiratory pattern, a) perception of the stronger is the perceived irritation (Cain, increased mucosa! secretion, increased blood caused by acute 1989). Time summation also seems to appear flow to exposed skin areas, etc. ;erception or ob at least over short-time exposure of a dur Environmental stress: The constant effort subacute inflam- ation of seconds (Cometto-Muniz et al., 1984) needed to identify the wanted and to over

1 exposed tissues, or minutes (Cain et al., 1986). Adaptation ride the unwanted sensory information, as 'ffects which may may appear at high exposure levels, but time well as the efforts needed to maintain protec I ubacute environ- summation for exposures of hours' duration tive reflexes, is a strain to humans and may ,_ d by the percep- has not been investigated in detail. A sum by itself cause secondary effects. If such mary of the chemical sense is found in Cain stress situations are continued for an exten present knowl (1989) which discusses some of the possible ded period of time, stress-like complaints Ftes that VOC's mechanisms behind summation and adapta will arise, of which headache seems to be the rous sense in the tion. most important. [gustatory senses Inflammation: In medicine, inflammatory Symptoms of weak environmental stress lical sense (Cain, reactions are related to microbiologic, meta are well known from both the indoor and the rems - odor, taste bolic or immune system reactions and are outdoor environment. Many different physi ond to airborne generally considered to be a protective reac cal or chemical exposures have been shown l1atities of the ex- tion to a potential cell damage. Inflammation to cause these typical stress symptoms which two or three of is known both as acute and subacute reac have been reviewed by Evans et al. (1989). to result in a tions (Thaysen et al., 1980). Only the acute Some typical symptoms are shown in Table 1 mething which reversible reactions seem to be relevant to s. 362 M"lhove: Volatile Organic Compounds, Indoor Air Quality and Health

Table 5 Some typical symptoms of environmental stress•. mary processes are stimulation of sensory nerve endings or initiation of weak inflam Increase in stress hormone levels. Blood pressure increase. matory tissue reactions. The secondary acute Fatigue. effects in the exposed tissue are perceptions Irritability and reduced tolerance. of the tissue reactions, initiated reflexes due Reduced productivity, errors. Psychological symptoms. to the primary perception of exposures, or Attempts to change stressors . . changed sensitivity of the senses due to tis Feeling of helplessness, lack of control of stressors. sue changes. Subacute effects may also occur. Changes in feeling of job satisfaction, or life quality. They are environmental stress reactions or more severe skin reactions. The three types • Evansecal.1989. of effect expected to follow from low-level ex posure to voe according to this model are Figure 1 shows a biological model, which summarized in Table 6. may explain the combined reactions of hu The intensity of each of the symptoms mans to multifactorial exposures in the in may be modified by additional factors such door environment (MIAQ) including VOC's. as age, smoking history, or gender. Further, Most environmental exposures or stressors the number of symptoms observed and their cause overlapping spectra of health effects. intensity may cause a feedback on the indivi According to the model, the combined re dual's behavior, thereby causing them, for versible effects of a MIAQ exposure which example, to modify their environment, or to include a major voe component fall into focus on certain symptoms and thus suppress the three classes described above. They are: others. Consequently, each subject may react perception, inflammatory-like and stress-like differently to the mixed exposure and exhibit reactions. only a few of the symptoms from the spec In the exposed tissue both primary and trum of symptoms observed in the exposed secondary acute processes occur. The pri- population as a group.

Environment The human body Effects

Skin and other tissue Perception of: - The environment Processes ••• - nssue or skin in condition Sensory.._...... _, peripheral stlmu· .- or latlon central nervous system Environmental stress reactions: - Immediate neuro logical effects - Delayed or accu mulated neuro logical effects - Changed percept ual function

Observed changes in skin or tissue

Fig. 1 A biological model for human responses to exposures to low levels of mixtures of volatile organic compounds (VOC's) as air pollution in non-industrial indoor environments. \ '"

M01have: Volatile Orgonic Compounds, Indoor Air Quality and Health 363

~ation of sensory In this model, the type of effects associated The Exposure Measures of the Model J. of weak inflam with voe exposures are unspecific and may At present, few methods are available for ob fe secondary acute be caused by environmental exposures other jective measurements (MIAQ) of voe expo ;1e are perceptions than chemicals. For example, physical expo sures and co-factors. Measuring techniques tiated reflexes due sures, such as temperature or inert dust, may exist for the concentrations of most VOC's, ~ of exposures, or cause a similar spectrum of symptoms. Any but two problems often prevent the use of lsenses due to tis discussion of a causality between VOC's and these methods at low exposure levels. The ':ts may also occur. the type of symptoms appearing in Table 6, first is the extreme sensitivity of humans to ;tress reactions or therefore, must discuss not only the voe ex many types of exposure, e.g. to odorants. l. The three types posure levels but also the level of other con This sensitivity is difficult to match with .from low-level ex tributing exposure factors. existing measuring techniques. Another ;to this model are The discussions in occupational hygiene problem is the many compounds present at normally focus on one health risk factor at a the same time which make any detailed mea of the symptoms time and in such ranges of it that the con suring program time-consuming and expen ional factors such sidered exposure factor can be assumed to be sive. r gender. Further, the only, or at least the primary, cause of Three different short-cuts are used to >bserved and their health effects. Such extreme exposures are ty overcome these two difficulties related to lack on the indivi pically not found in non-industrial build measurements of exposure. The first is to use :ausing them, for ings. In consequence, the dose-response rela human subjects as detectors, e.g. as panels. nvironment, or to tionship may be multifactorial and include a An example of this is the olf7decipol theory. and thus suppress number of exposure factors (temperature, This indicator measure is described else subject may react concentration of pollutants, etc.) in the etio where (Fanger et al., 1988). The idea is to ?Osure and exhibit logy. quantify pollution sources in buildings by ns from the spec Under field conditions, three exposure comparing them to a well-known pollution ·~d in the exposed ranges are of interest. They are defined by source. The unit for the emission rate of the relative contribution ofVOe exposure to bio-effluents (air pollutants) is taken to be the prevalence of effects or symptoms. Below the emission from a standard person in ther Effects a lower threshold (the no-effect level) no ef mal comfort. The unit is called an olf. One Perception of: fects are expected to follow from the expo olf is the perceived air pollution caused by - The environment sure to VOC's despite any other simultan one standard person ventilated by 10 l/s of - Tissue or skin condition eously occurring exposure. Above an upper unpolluted air. Other sources, therefore, are threshold (the effect level), an effect of quantified by the number of standard per VOC's is expected even when all other expo sons (olfs) required to cause the same dissa Environmental sure factors are controlled and acceptable. tisfaction (which is expressed in decipol) as stress reactions: - Immediate neuro Between the two thresholds a correlation the actual pollution source. Given a certain logical effects may or may not occur between VOC's expo percentage of dissatisfied among a panel of - Delayed or accu mulated neuro sure and the prevalence of effects, depending judges, the number of standard persons (olfs) logical effects - Changed percept on the interactions from other exposure fac that would cause the same dissatisfaction can ual function tors or the composition of the exposure. This be calculated. This number is the olf value of range is called the multifactorial exposure the pollution source. The olf value is not range. In this exposure range a consistent or supposed to provide information on possible Observed changes in skin or tissue monotonous dose-response relationship may health risks of the pollution source other not exist and complaints may not necessarily than the perceived air quality. The value re disappear if one of the relevant exposure fac fers to the visitors' situation and not to that tors is removed from the environment. of the occupants. organic compounds The second short-cut focuses on tracer compounds which are used as indicators for 364 Mtiilhave: Volatile Organic Compounds, Indoor Air Quality and Health

Table 6 Three classes of human responses to VOC' s in normal indoor air. Primary reactions are observed at acute low-level exposures. secondary effects ore observed ofter prolonged or more intense exposures. Their intensity de- pends on the properties of the dominating compounds in the exposures and on the sensitivity of the subjects. The toble further shows examples of effects found in controlled experiments with humans.

A: Acutely perceived deterioration of the quality of the environment.

Concencration (mg/ml) Type of Examples mechanism of effect 3 5 8 15 25 40 Primary Recognition Odor +2 +l 0 0 +4 +3 of exposures perception +s Stinging, 0 (-1) +2 0 +4 +3 itching etc. +5 Reduced air +l quality, 0 0 +2 0 +4 +3 Need more +5 ventilation Secondary Reflexes in eyes, nose 0 0 0 0 0 0 0 and airways Changed Changed mucosa! tearfilm 0 0 0 0 0 +3 secretion stability Changed cell counts 0 0 0 0 +4 +3 in eye liquids Difficulties 0 0 0 0 0 0 0 in breathing Activities to Need more 0 0 +2 0 0 0 change the ventilation environment

B: Acute or subacute reactions in skin or mucous membranes similar to beginning inflammatory reactions.

Concenttation (mg/ml) Type of Examples mechanism of effect 3 5 8 15 25 40 Primary Dilation of 0 0 0 0 0 0 0 peripheral vessels Stinging, Perceived 0 (-1) +2 0 +s +3 itching or irritation tingling feeling Secondary Pain 0 0 0 0 0 0 0 Changed skin Perceived 0 0 0 0 ( + 1) 0 temperature skin temperature \ \' r -·

M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health 365

! observed at acute C: Subacute and weak stress-like reoctions ("Environment stress"). ';. Their intensity de 3 . the subjects. The Concentration (mg/m ) Type of Examples mechanism of effect 3 s 8 15 25 40 Primary Discomfort Headache 0 0 0 -2 -1 0 and complaints 4 ~) +S 25 40 Drowsiness 0 0 0 0 +S 0 +4 +3 Secondary Complications Changed 0 0 0 0 ( + 1) 0 +5 in body composition functions and of eye and +4 +3 physiological nose liquids +5 effects Changed 0 0 0 0 ( + l) 0 + l odor +4 + 3 threshold +5 Changed 0 0 0 0 ( + l) 0 performance ( +4) -S 0 0 Changed 0 0 0 0 ( + 1) 0 mood +5

0 +3 Changed 0 0 0 0 ( +4) 0 lung function

+4 +3 + Significant effect. References: - No effect seen. 1. M!!llhave et al., 1986 4. Kjzrgaard et al., 1991 0 No information. 2. M!!llhave et al., 1990 5. Otto et al., 1990. 0 0 () Indication of effecr/no effect. 3. Kjzrgaard et al., 1989

0 0 the level of pollution, such as carbon dioxide 1986, 1990b ). This measure is easily obtained (C02) and water (H20), or compounds which through the chemical analysis (e.g. using an themselves are potent air pollutants, such as integrating FID detector). From a biological 3 formaldehyde (CH20). This short-cut is de point of view, number as molecules per m , reactions. scribed in most textbooks on indoor climate. (the molar concentrations in PPM or PPB) The third and last short-cut focuses on indi may be more relevant. Mathematical func i) cator measures of the total measurable level tions based on combinations of other vari I 25 40 of air pollution. One of these indicators is ables - like type of radicals, vapor pressure or 0 0 called Total Volatile Organic Compound polarity of the compounds - have also been (TVOC), and is discussed in Mjijlhave et al. suggested as indicators. (1992). According to our present knowledge, the + 5 +3 response of the nerves may be described in The Concept of Total Volatile Organic the following way: the receptors in the thin Compounds (TVOC) mucosa! membrane are positioned within the 0 0 At present, no proper measuring unit has mucosa close to the air-mucosa interface. ( + l) 0 been established for the combined effects of This causes a rapid establishment of equi the many different compounds in the atmo librium between concentrations of chemicals sphere. Addition in mg/m3 of masses of pol in the air, the concentration of absorbed mo luting molecules has been suggested and in lecules in mucosa and the number of activa relation to volatile organic compounds it is ted receptors. often called the TVOC indicator (Mjijlhave, A simplified chemical and physical model, 366 M121lhave: Volatile Organic Compounds, Indoor Air Quality and Health therefore, includes three steps which de based on experimental evidence. However, it scribe the absorption of airborne pollutants must be emphasized that the TVOC concept into the liquids of the mucosal membrane, has not yet been thoroughly tested in prac and the subsequent binding to the receptor, tice and therefore is still a postulate. which is followed by reactivation of the re ceptor. Such a model has previously been de scribed in the literature (Nielsen et al., Some Investigations and 1985b; Nielsen et al., 1988; Kristiansen et al., Controlled Experiments with 1988) and among other things assumes ideal Low-level VOC Exposure and gas conditions of the air phase and a lipophi lic receptor compartment. In M0lhave et al. Health Effects (1992) this model has been used to identify Field Investigations the short-cuts made in the derivation of the The dominating effects or symptoms caused TVOC indicator. It appears that the TVOC by low-level exposures to VOC's are similar indicator under specified assumptions is an to those associated with the sick building estimation of the lower level of perceived un syndrome as defined by WHO (WHO, 1982). specific stimulation of nerves in a population These acute or subacute effects appear after exposure to VOC's. The response may among a major part of the occupants in be proportional to the sum of mass-concen buildings. Generally, no excessive environ 3 trations (mg'm ) of the compounds in the air mental exposure can be identified and the if the following assumptions are made: occupants seem to be without any unusual sensitivity. The WHO group (WHO, 1982) stated that • The response is caused by unspecific sti more than 30% of all new buildings seem to mulation of lipophilic sensors which re be affected by these indoor climate problems spond additively to a multicomponent ex which further seem to have no evident cause. posure. The symptoms included in the syndrome • The compounds can be assumed to react may be observed in any group of persons and equally strongly with the unspecific sen the "sick buildings" are characterized by a sors, i.e., the equilibrium constants of the large fraction of the occupants having the compounds can be assumed to be within symptoms. The syndrome, therefore, seems limited ranges. to be a normal reaction of the normal popu • The molecular weights and the vapor lation to unfavorable indoor climates. pressures of the compounds are assumed No investigation of this postulated SBS to be within limited ranges. syndrome has been reported in which a com • The exposures do not include compounds plete and well-defined spectrum of symp that react chemically with the receptor, e.g. toms was used. Generally, the descriptions of formaldehyde or acrolein. Such reactions the symptoms in the literature are anecdotal may cause an additional reaction to be ad and unsystematic. Criteria for causality such ded to the estimated lower limit for unspe as those shown in Tables 3 and 4 are not yet cific response. fulfilled and, therefore, SBS still appears to • The indicator cannot be used to predict be a postulate. other types of effect, for example effects on A review of investigations dealing with CNS, tissue changes or cancer. low-level exposures to VOC's and their influ ence on health should preferably deal with The simplifications used in M0lhave et al. both acute and subacute effects. In most in (1992) to develop this TVOC concept are vestigations this distinction is reflected in a 1~ .

M121lhave: Volatile Organic Compounds, Indoor Air Quality and Health 367

dence. However, it distinction between job- and building-related may be as low as 0.2 mg/m3 depending on :he TVOC concept effects or symptoms. Such a time selectivity other simultaneous exposures. hly tested in prac is inadequate in relation to most sensory ef postulate. fects, which may show strong intensity varia Controlled Experiments with Low-level tions within seconds or minutes. In the in VOC Exposure and Health Effects vestigations, the symptoms further should Investigations and experiments with con 1s and show a correlation to VOC exposures. Due to trolled low levels of volatile organic com 1ents with the many VOC's normally found in indoor pounds (VOC's) are, for several reasons, osureand environments, the results of concentration more difficult to interpret than traditional measurements in practice are often imposs clinical investigations and experiments with ible to interpret as measures of exposure for toxic compounds (M0lhave, 199lb ). At pres occupants. Such measurements, further, are ent, few acceptable objective measures exist symptoms caused expensive. Only a few investigations, there for effects. These low-level experiments often VOC's are similar fore, report adequate information on voe. use unspecific subjective reactions as the the sick building exposure levels. In some reports, indications only available measures of effects. This has a 1IO (WHO, 1982). of possible voe influence on indoor air number of consequences for the experiments. te effects appear quality may be deduced if the type or intens These limitations reduce their usefulness for the occupants in ity of ventilation is correlated to the preva conclusions about causality at low exposure !Xcessive environ lence of symptoms. Such indications, how level. dentified and the ever, are not used in this summary as causali hout any unusual ties other than VOC's may be active in these • Group or population responses are often cases (e.g. humidifier fever). used instead of the objective measures of ), 1982) stated that In a previously reported summary of in effects. buildings seem to door air investigations Gohansson, 1982) • More co-factors must be controlled than · climate problems eight of the U investigations described hu in traditional experiments both in relation ! no evident cause. man reactions. These eight and a few more to exposure conditions, to the subjects and in the syndrome investigations were summarized in a pre to their sensitivity. Exposure duration, time mp of persons and vious publication (M0lhave, 1986). The aver of day, self-reported sensitivity, and occupa '.haracterized by a age voe concentrations in houses with in tion seem to be important co-variables. lpants having the door climate problems were 1.3 mg/m3 with a • A spectrum of several symptoms or signs 3 , therefore, seems range from 0.09 to 13 mg/m , while the con must be included in the tests due to the the normal popu centration in houses where no problems were unspecific nature of the effects. r climates. reported showed an average of 0.36 mg/m3 • The number of possible interactions at 3 s postulated SBS ranging from 0.02 to 1.7 mg/m • Complaints low-level exposure is so large that at pres d in which a com seem to appear in all field investigations re ent there seems little chance of investiga 3 ectrum of symp porting concentrations exceeding 1. 7 mg/m • ting their etiology in detail. ihe descriptions of The "no-effect level" could not be estima • Intersubject variation is so large that the rure are anecdotal ted from these field investigations. In a more subjects are used as their own controls. !for causality such recent investigation (The Danish "Town Repeated measurements may, however, and 4 are not yet Hall" study (Zweers et al., 1990)) a significant lead to effects from learning or from dif lS still appears to correlation was found between both TVOC ferent times of the day, week or year. and olf or decipol exposure indicators in the Proper baseline measurements, therefore, ns dealing with rooms and mucosa! irritation or work-related are difficult to arrange. 's and their influ symptoms. The threshold for effects of • New undocumented methods or old me ferably deal with VOC's on air quality was estimated to be be thods outside their documented range are ' ects. In most in- tween 0.19 mg/m3 and 0.66 mg/m3 (TVOC). often used and often without proper qual 1 is reflected in a The lower threshold for effects, therefore, ity assurance programs.

f - . --- ·· -~ 368 M11.1lhave: Volatile Organic Compounds, Indoor Air Quality and Health

• Double-blind designs are impossible to ar Few occupational experiments, therefore, range because of the perceived exposure. are relevant for extrapolation to the low Special considerations are needed in low range of concentrations found in non-indus level experiments to establish causality. trial indoor environments. Some experi • Registration of motivation and attitudes ments have been performed in which hu are especially important for experiments mans have been exposed to low levels of using registration of perceived exposure VOC. Four controlled exposure experiments and symptoms as measures of effects. were established in the climate chamber at our institute to test if low exposure levels of Experiments at higher exposure level, e.g. ex VOC's may cause reduced well-being or dis periments around or above occupational comfort (Mi.;lhave et al., 1986; Mf

Table 7 Summary of 5 exposure experiments.

Ref. Exposure Population Measures of effects type mg/m3 type number M22 0, 5, 25 Randomly 64 Subjective sensory responses, selected but indications of neurologic SBS-sensitive effects and changes in eye and nose liquids 2 M22 0, 1, 3, 8, 25 Randomly selec- 25 Sensory symptoms, headache ted and healthy and general well-being 3 n-decane O, 40, 140, 400 Random, healthy 63 Sensory symptoms. Tear film stability. Leukocytes in eye li quids Healthy 21 Sensory symptoms 4 M22 0,25 Lung function SBS subjects 14 Leukocytes in eye liquids and nasal secretions. Performance 5 M22 0,25 Healthy males 76 Sensory symptoms Neurobehavioral tests

M22 = standard mixture of 22 indoor air pollutants.

1 M0lhave et al., 1986. 4 Kjzrgaard et al., 1991 . 2 M01have et al., l 990a. S Otto et al., 1990. 3 Kjzrgaard et al., 1989. 1~ . [

Meilhave: Volatile Organic Compounds, Indoor Air Quality and Health 369

,icperiments, therefore, ly reduced at TVOC concentrations of 25 probably have been significant at lower expo 3 3 tpolation to the low mglm and that odor appears at 5 mg/m • sure levels if more subjects had been exam s found in non-indus The effects are acute and occur within min ined, if a longer exposure time had been nents. Some experi utes after the start of exposure and no statis used, or if other indoor climate factors had ;ormed in which hu tically significant adaptation is seen except been slightly uncomfortable. .1sed to low levels of for odor intensity (Mf6lhave et al., 1986). The third of the Danish experiments was a exposure experiments These findings have been confirmed in all dose-response study of human reactions to e climate chamber at subsequent experiments. the indoor air pollutant n-decane (Kjrergaard ow exposure levels of In the first Danish exposure experiment, et al., 1989). 63 healthy subjects, randomly )::ed well-being or dis some indications were found of objective ef selected from the normal population, were ' 1986; M!6lhave et al., fects related to odor threshold, to chemical exposed to n-decane concentrations of either ;989; Kjrergaard et al., changes in eye and nose liquids, and to per 0, 10, 35 or 100 µ1/1. Of these exposures, only 3 periment in the USA formance and mood (Mf6lhave et al., 1986). 10 µ1/1 (or about 40 mglm ) is relevant for these Danish experi- Irritation of eye and nose, therefore, may not new buildings. The statistically most signifi 1). These experiments be the only effect to consider as a result of cant findings were dose-dependent changes ;:s 6 and 7. exposure to VOC's. in perceived irritation of mucous mem experiment, humans The major aim of the second Danish expo branes, increased sensation of odor intensity concentrations of a sure experiment was to measure dose-re and reduced air quality. Adaptation was seen :· occurring air pollu- sponse relationships between human sensory at the highest exposure levels, but not at the 1986). These com reactions and exposure to the same mixture levels relevant for a non-industrial environ to be emitted from of VOC's as used in the first experiment. ment. The physiological measurements ;ollow.s from the ex This second experiment focused on the showed decreased tear film stability at all ex y may be significant- dose-response relationships of sensory irrita posure concentrations (Kjrergaard et al., tive symptoms and headache (Mf6lhave et al., 1989). 199la). Further, the importance of these two The number of leukocytes in eye liquids symptoms for the feeling of general well increased in a dose-related manner. Predic ~es of effects being was examined. tors of the sensitivity to exposure, e.g. thres In the experiment, the subjective reactions hold for mucous membrane irritation and indicate that a statistically significant odor skin irritation (Stinging skin test), were cor ;ive sensory responses, 3 ions of neurologic was registered at 3 mglm • This indicates related to subjective ratings of odor intensity md changes in eye and that the lower limit for complaints resulting and irritation of mucous membranes (Kjrer 1uids from this type of air pollution in general liv gaard et al., 1989). 3 ( symptoms, headache ing spaces is at or below 3 mglm • The fourth Danish experiment was a con ;1eral well-being The air quality was rated unpleasant only trolled experimental study of human reac ~ symptoms. Tear film at concentrations at or above 8 mglm3 where tions to 25 mglm3 of the mixture of the 22 r. Leukocytes in eye Ii- the need for additional ventilation or remo volatile organic compounds used in previous val of sources became evident. Also, the irri experiments (Kjrergaard et al., 1991). 21 heal •symptoms mction tation of the mucous membranes was statis thy subjects were compared with a group of es in eye liquids and tically significant only at concentrations at or 14 subjects suffering from the sick building ~ccio n s . Performance higher than 8 mglm3 for SO-min exposures. syndrome (SBS subjects), i.e. having symp Isymptoms In the first study, no statistically significant toms such as irritated mucous membranes ;:havioral tests irritation was found at S mglm3 after 2.75 and headache related to buildings. Both hours' exposure (Mf6lhave et al., 1986). groups reacted subjectively to the indoor air In the experiment the same general trend pollutants and reported worse odor in a ques was found at concentrations lower than the tionnaire, worse indoor air quality and more lowest exposures causing statistically signifi irritated mucous membranes in eye, throat cant effects. The effects, therefore, would and nose than in the clean environment. 370 M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health ·. A tendency to a stronger response was seen among the SBS subjects. Objective Discussion measures indicated exposure-related reduc Henle-Kock Criteria tion of lung function among the SBS sub In order to be accepted, any proposed causal jects. Both groups had an increased number ity between voe and health must fulfill cri of polymorphonuclear leukocytes in tear teria such as the Henle-Kock criteria shown fluid as a result of exposure. This was not in Tables 3 and 4. It must be emphasized seen in nasal secretions. Psychological per that, although fulfillment of these criteria is formance was diminished by exposure strong evidence for a proposed causality, this (Kjrergaard et al., 1991). may not be a sufficient or a final proof. The American experiment (Otto et al., 1989) is the most recent and was aimed at Are the Criteria for an Acceptable confirming and extending the first Danish Causality between VOC Exposure experiment. Only two neurobehavioral per and Health Fulfilled? formance tests were used in the Danish Table 6 summarizes the conclusions of the study, while 14 tests were used in the Ameri five previously mentioned experiments deal can study to fully characterize the possible ing with low-level VOC exposures. The neurobehavioral effects of voe exposure on Table shows the positive or negative findings young healthy white males. at different TVOC levels in relation to the Results of the study confirm the adverse main types of effects expected according to subjective reactions of subjects to a 25 mg/m3 the suggested biological model (M~lhave, concentration of volatile organic compounds. 1990b). Ratings of general discomfort (defined as ir The experiments indicate that, in con ritation of the eyes, nose and throat), symp trolled exposures, effects follow exposure to tom questionnaire responses on odor intens VOC's as required according to criterion 3 in ity, air quality, eye and throat irritation, the Henle-Kock criteria shown in Table 3. headache and drowsiness, and mood scale Perceptive effects are observed at TVOC con 3 measures of fatigue and confusion all dif centrations higher than 3 mg/m • Other sub fered in predicted directions between clean jective effects follow higher exposures. air and exposure conditions. No convincing Few inflammatory reactions were reported evidence was found of any neurobehavioral although some of the irritative effects may be impairment associated with exposure to the of inflammatory origin or may be caused by voe mixture. chemical mediators. Changed tearfilm stab 3 However, it could not be concluded that a ility and cell counts were seen at 25 mg/m • 25 mg/m3 concentration of VOC's poses no Subjects also reported changed temperature neurotoxic risk to susceptible subgroups or sensation in the exposed skin areas at 25 3 even to the general population. In summary, mg/m • These observations may indicate clear adverse subjective reactions to voe ex weak inflammatory reactions. posure, but no functional (neurobehavioral) A 25 mg/m3 level seems to cause weak en impairment, were found (Otto et al., 1989). vironmental stress symptoms like headache The American experiment further con and drowsiness. Associated psychological ef firmed that subjective reactions to VOC's at fects such as changed performance, confu 3 levels found in new buildings are not limited sion and fatigue are also found at 25 mg/m • to ''complainers" or chemically sensitive sub These effects are known from higher expo groups in the general population. sure levels, but have not been consistently found in low-level exposure experiments. At present, 25 mg/m3 seem to be the lowest con- \ \ , . ~·

M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health 371

trolled exposure which has indicated such 0.05 to 1.38 mg/m3 (Nordbiick, 1990). The ex psychological effects. posure range from 0.19 mg/m3 to 0.66 mg/m3 These experimental findings show that was estimated in the Danish Town Hall any proposed causal- measurable group responses are found in Study for the lower threshold of no-effects 1ealth must fulfill cri controlled exposure experiments as required (Zweers et al., 1990). This range corresponds -Kock criteria shown in criterion 2. The responses follow a gra to the range of the lower limit of concentra 3 must be emphasized dient from sensory effects (odor 3 mg/m ) tions in buildings with complaints and is at nt of these criteria is and indications of subacute inflammatory re present the best estimate of the lower expo ·oposed causality, this actions (changed leukocytes in liquids, per sure limit for no-effects ofVOC's. 3 >r a final proof. ceived skin temperature at 25 mg/m ) and in Reactions do not seem to be related to a dications of subacute stress-reactions at 25 hypersensitive group of subjects. The same 3 n Acceptable mg/m • The criterion 1 of progression of the reactions were found among normal healthy OCExposure effects, therefore, also seems fulfilled. subjects in the Danish experiments (M(l.llhave ~ The total environmental exposure in most et al., 1986, 1991; Kjrergaard et al., 1989, 1991) e conclusions of the field investigations is multifactorial as factors and in the American experiment (Otto et al., ed experiments deal other than voe exposure may exceed their 1989). Subjects responding most strongly did, )C exposures. The no-effect levels. Most of the effects reported however, seem to have special characteristics or negative findings in field investigations, therefore, may have such as stronger skin-response to irritating ls in relation to the more than one cause. Consequently, it is not compounds (Kjrergaard et al., 1991) and, :pected according to surprising that effects of voe exposures in among persons claiming often to suffer from tl model (Mf21lhave, field investigations seem to occur at lower Sick Building Syndrome, a significant, but exposure levels than in controlled experi only slightly stronger response was found licate that, in con ments where exposure factors other than (Kjrergaard et al., 1991). These SBS subjects ; follow exposure to VOC's are supposed to be below their no-ef also showed indications of lung-functional ling to criterion 3 in fect levels. Further, in the clinical experi changes (Kjrergaard et al., 1991). However, shown in Table 3. ments the exposure times were less than 3 the five controlled experiments contain few erved at TVOC con- hours which, from field experience, seems or no consistent investigations of lung func 3 3 mg/m • Other sub too short a period to cause severe subacute tion, allergic or systemic effects. :er exposures. effects at low exposure levels. Much more re No definitive conclusion can be drawn :tions were reported search is needed to finally establish whether with respect to the influence of other co-fac tative effects may be subacute effects may occur after prolonged tors. In those few investigations dealing with ,r may be caused by exposures. such factors, both positive and negative indi mged tearfilm stab In a review of field investigations, it was cations were found. The information, how e seen at 25 mg/m3• found that complaints seem to arise when ever, indicates that future research with more ianged temperature the concentrations exceed 1.7 mg/m3• Below sensitive experimental designs and analytical 3 d skin areas at 25 1. 7 mg/m , complaints may arise if other methods may show such effects. tons may indicate types of simultaneous exposures are present In the published field investigations, the ons. (Mf21lhave, 1986). The concentrations repor observed symptoms are not systematically s to cause weak en ted from field investigations are improperly described. However, they appear to be more :oms like headache documented and they may be biased. The frequent among exposed than among non !d psychological ef published investigations do, however, indi exposed persons. The field investigations, erformance, confu cate that the concentrations of volatile orga therefore, also indicate that criterion 3 may found at 25 mg/m3• nic compounds are generally higher in prob be fulfilled. Future research, however, will from higher expo lem houses than in the houses without prob have to prove this. : been consistently lems (M(l.llhave, 1986). The laboratory experiments indicate that u-e experiments. At A recent field investigation finds an effect the effects which are expected according to i be the lowest con- ofVOC's at concentrations in the range from the biological model can be experimentally 372 M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health reproduced and acutely follow the exposure. The model does not include acceptable The criteria 4 and 6, therefore, are also ful measures of a combined exposure to the filled. No field investigations have been: repor many compounds simultaneously found in ted of tests of the effects of elimination or mod doors. Some indicator measures have been ifications ofVOC exposure. Post-exposure meas suggested. These indicator measures are all urements during the controlled experiments, based on simplifications of which only some howevei; indicate that the effects are reversible have been justified and fewer thoroughly in and disappear shortly after exposure. The 5th vestigated. At present, they cannot be used criterion, therefore, may also be fulfilled. for risk assessment, mitigation, etc. Further, In conclusion, no evidence contradicts the co-variables, e.g. in relation to non-chemical proposed causality between the effects tentat exposure and subject sensitivity, have not ively related to low-level exposure to VOC's. been investigated in detail, and few accept On the contrary, evidence from both field in able measures or indicators exist for possible vestigations and controlled exposure experi co-variables related to the model. Criterion 2 ments does support the causality. The field dealing with acceptable measures or indica investigations and controlled experiments tors of an exposure, therefore, is not yet ful are, however, as yet too few to allow a final filled. conclusion to be drawn. Tentative Guidelines for VOC's Does an Acceptable Biological Model in Non-industrial Environments exist for the Readions to Low-level The observations summarized here have ma Exposures to VOC? jor limitations. They do, however, indicate In Table 4 a set of four criteria was estab that VOC's may be important for indoor air lished which any acceptable biological model quality, especially in the form of discomfort for the causality of a proposed etiology must due to odors and irritative symptoms in eyes, fulfill. The first criterion was that the model nose, and throat, and headache. The list may must explain the observed effects. The list of include other effects, for example related to effects expected - according to the model in productivity and performance. Such effects Figure 1 - to follow from low-level exposure have not yet been positively identified. to VOC's coincides very well with the effects The tentative conclusion from the avail observed in epidemiological field investiga able epidemiological studies (Ml<)lhave, 1986; tions and controlled experiments. It appears Ml<)lhave, 1990a) and the exposure experi that sensory irritation, olfaction, irritation, ments is summarized in Table 8. It indicates and weak neurological effects seem to be the that no effects are expected as a result of ex 3 dominating effects of low-level voe expo posure to VOC's below about 0.2 mg'm • sure. The first criterion for an acceptable At concentrations higher than about 3 model, therefore, seems to be fulfilled. mg'm3, complaints seem to occur in all in The effects are all known from similar or vestigated buildings with occupants having higher exposure levels. The third criterion, symptoms. In controlled exposure experi 3 therefore, is fulfilled. Any delay in effects ments, odors are significant at 3 mg'm • 3 caused by low-level exposure has yet to be At 5 mg'm , objective effects were indica identified. Some indications of a sub-acute ted besides the subjective irritation. Expo latency exist. According to the model, such a sures for SO minutes to 8 mg'm3 led to signif latency of the effects is explained by the sub icant irritation of mucous membranes in acute skin and stress reactions following pro eyes, nose and throat. longed exposures. Therefore, the fourth cri In the reviewed literature, few acceptable terion also seems to be fulfilled. indications of exposure levels are given ; .

. r ~

M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health 373

ot include acceptable which allow an estimate of the threshold for Table 8 Tentative dose-response relationship for dis ined exposure to the 3 comfort resulting from exposure to solvent-like volatile headache. Concentrations below 3 mg/m in organic compounds as air pollutants in non-industrial .ultaneously found in field investigations were found to show a sig indoor environments. ~ measures have been nificant difference in frequencies of headache ::ator measures are all between problem buildings and ·control Total con Irritation and Exposure is of which only some centration• discomfon range buildings. On the other hand, significant mg/m3 i fewer thoroughly in- headache was found in only one of the expo < 0.20 No irritation or Thecomfon 3 they cannot be used sure experiments and then at 25 mg/m • discomfon range itigation, etc. Further, The reason for the lower threshold in field 0.20-3.0 Irritation and dis- The range of ation to non-chemical investigations may be either the interaction comfon possible if multifactorial sensitivity, have not of other exposures, or the effect of longer ex other exposures exposures etail, and few accept posure durations. Therefore, based on the interact nors exist for possible present information, the threshold for head 3.0-25 Exposure effect and The range of he model. Criterion 2 probable headache discomfon ache and other weak neurotoxic . effects possible if other ex- :: measures or indica caused by exposure of less than a few hours' posures interact ::refore, is not yet ful- duration is expected to be between 3 and 25 Headache. The toxic mg/m3. Additional neurotox- exposure range These conclusions refer to the more preva ic effects other than ;forVOC's lent of the effects caused by voe exposure headache may occur 1vironments among normal subjects. Risk groups may ex • Measured as TVOC according to M91lhave et al. (1992). iarized here have ma ist that respond more than the normal popu b This range is only panly discussed in this paper. lo, however, indicate lation. Indications of lung function reactions >ortant for indoor air were found in a pilot study on allergic per associated indirect effects on productivity, e form of discomfort sons exposed to 25 mg/m3 VOC (Harving et etc., will continue if development of objec ve symptoms in eyes, al., 1989). Further, future investigations deal tive measurements of exposures and effects ~adache. The list may ing with larger groups of persons may reveal are neglected and discomfort continues to be >r example related to special effects such as allergy or carcinogeni considered irrelevant to health effects in the :mance. Such effects city from low-level exposures to VOC's. evaluation of indoor air quality. vely identified. These special effects, however, have not been An important consequence of the lack of :sion from the avail demonstrated to follow from exposure to the objective measures of exposure and effect in tdies (M!lllhave, 1986; type of compounds and concentrations of relation to low-level exposures to VOC's is he exposure experi- VOC's found in indoor air. that dose-response relationships have not yet Table 8. It indicates been established for discomfort. A dose-re red as a result of ex Consequences of the Proposed Model sponse relationship is the relation between bout 0.2 mg/m3• for Building Construdion exposure (concentration and duration) and gber than about 3 A sense of well-being is an essential part of effect and may be used for risk assessment to occur in all in general health. This is not always reflected in and regulation, etc. For most diseases or ad occupants having discussions of indoor air quality, which are verse health effects such relations exist. Toxi :d exposure experi often focused only on absence of objectively cological risk assessment, therefore, has been 3 t at 3 mg/m • defined diseases. One reason for this one-si developed and is operational for mitigation effects were indica ded discussion is that annoyance and IAQ of these diseases. Quantitative evaluations of ve irritation. Expo have not yet been defined and at present can the risks of adverse irreversible health effects mg/m3 led to signif not be objectively measured. Further, no rel (e.g. related to CO or Radon) or to the risks ous membranes in evant measures exist of combined multifac of reversible or irreversible changes of the torial exposures. However, the unacceptably body's physiological functions (e.g. CNS or ture, few acceptable high prevalence in buildings of complaints PNS effects) are used in traditional occupa levels are given about reduced comfort or well-being and the tional or environmental evaluation of health 374 MG'Jlhave: Volatile Organic Compounds, Indoor Air Quality and Health risks related to airborne agents. These assess to allow the occupants to use their senses to ments are made according to standard toxi pick up wanted environmental signals undis cological principles and the results are, e.g., turbed by exposures carrying unwanted sig TLV's for air pollutants, occupational stan nals or information. This means that unwan dards for light levels (LUX), and sound le ted environmental information, like the vels (dB(A)). sound of typing from a colleague in the same The comfort dimension of health refers to room, or the neighbor's radio in a home, qualitative evaluations of the environment should be mufiled. On the other hand, con and, in many respects, it is a new concept for versation among the occupants or their per regulation. The air-deodorants, paints, wall ception of their own activities should be papers or music that are liked by some per eased. In short, this second principle calls for sons are disliked by others and generally ac an optimization of the signal-to-noise ratio cepted principles for regulation of the quality for the senses by allowing the wanted signals of the indoor environment with respect to to propagate to the occupants and the un odors, sounds, colors etc, may seem imposs wanted sensory signals to be dampened. ible to establish, if desired at all. Some gen This second principle, if true, explains eral conclusions for the principles for opti why different occupants have different opti mizing of the indoor climate, however, may mal environments. The signals which bear be drawn if references are made to regula information to one person about his or her tions, e.g., in building codes for the acoustic own activity and environment create sensory or the lighting environment which contain "noise" for another person. Therefore, the additional qualitative concerns besides those signals relevant for one person differ from used for the setting of the TLV type of those relevant for another. guidelines. In the special case of voe pollutants and The first basic principle implicit for these occupants' comfort, the relevant senses are guidelines is that the building must support the combined set of senses, including olfac a specified range of human activities, habits tion and the common chemical sense. A way and preferences. Complaints and decreased of explaining occupant responses to VOC's, performance will automatically follow if the therefore, could be that the combined sen occupants try to perform activities outside sory signals resulting from chemical expo this range, e.g. reading in too dark a room sures inform the occupants about the pres with disturbing intermittent noise peaks, or ence of dangerous or beneficial sources of vo working with a video screen with many light latile organic compounds. These senses are reflections. This range of activity may be dif probably very old in terms of evolution and ferent for homes and offices etc., as the activ they help us to avoid dangers such as fires or ity patterns in homes include recreation, rest wild animals, to find edible food, and to and sleep, and other activities not normally identify other humans. In offices, as in found in offices. Further, the occupants of homes, the sensory signals indicate the pres homes may be more sensitive than the work ence of sources such as special activities, pro ing population as they include the sick, cesses or utensils. This may be meaningful young and old fraction of the population. information to the occupants and, as such, The second basic principle originates from desired information. the assumption that humans do not feel well Not all VOC's bring positive information if they are deprived of the optimal use of to each individual. An office worker may not their senses to perceive their environment feel relaxed and comfortable if he or she con and the activities they are performing. The stantly feel a weak eye or nose irritation indi ideal indoor environment, therefore, seems cating the presence of an unknown health \ r

M0lhave: Volatile Organic Compounds, Indoor Air Quality and Health 375 s to use their senses to risk. Such a risk may be caused, e.g., by an in the indoor and outdoor air in northern Italy", En vironment InternatUmal, 12, 343-356. 1 nmental signals undis unidentified chemical exposure brought to Evans, A.S. (1976) "Causation and disease. The Henle :arrying unwanted sig the occupants through the ventilation system Kock postulates revisited", Yale Journal of Biology and his means that unwan from unknown sources somewhere in the Medicine, 49, 175-195. Evans, G.W, Carrere, S. and Johansson, G. (1989) "A nformation, like the building. The office worker, of course, will multivariate perspective on environmental stress",Ar t colleague in the same cry to remove any suspected source, open chives of Complex Environmental Studies, 8(1), 1-5. )r's radio in a home, windows, etc., in order to mitigate the uni Fanger, P.O., Lauridsen, J., Bluyssen, P. and Clausen, G. (1988) "Air pollution sources in offices and assembly 1 the other hand, con- dentified "enemy" and thus be constantly halls, quantified by the olf unit'', Energy and Build 1ccupants or their per- alert. ings, 12, 7-19. activities should be The present practice of setting standards Ferguson, J. (1939) "The use of chemical potentials as indices of toxicity'', Proceedings of the Royal Society of :ond principle calls for or guidelines for perceived air quality London, Series B U7, 387-404. e signal-to-noise ratio (PIAQ) takes the form of concentrations Harving, H., Dahl, R. and M0lhave, L. (1989) "Lung ing the wanted signals which are expected to result in a certain pro function and bronchial reactivity in asthmatics dur ing exposure to volatile organic compounds", Ameri ::cupants and the un portion of a population detecting or recog can Review of Respiratory Diseases, 143, 751-754. to be dampened. nizing the exposure (WHO, 1989). To the ex Johansson, I. (1982) Kemiska Luftfdroreningar Inomhus. ple, if true, explains tent that these guidelines do not distinguish En Lilteraturs Sammenstdllning. (Chemical Indoor Air Pollution, a Swedish Review). Repon 6/1982. Stock ts have different opti between wanted and unwanted exposures, holm, Sweden, National Institute of Environmental .e signals which bear they contradict the principle of signal-to Medicine. :-son about his or her noise optimization. Kja:rgaard, S.K., M0lhave, L. and Pedersen, O.F. (1989) "Human reactions to indoor air pollutants: n-De mment create sensory cane'', Environment Intemalional, 15, 473-482. !rson. Therefore, the Kja:rgaard, S.K., M0lhave, L. and Pedersen, O.F. (1991) e person differ from References "Human reactions to a mixture of indoor air volatile organic compounds", Atmospheric Environment, 2SA, er. American Conference of Governmental Industrial Hy 1417-1426. : voe pollutants and gienists (ACGIH) (1988) TLV's, Threshold Limit Values Krause, C., Mailahn, W, Nagel, R, Schulz, C., Seifen, e relevant senses are und Biological Exposure Indices for 1988-1989, Cincinna B. and Ulrich, D. (1987) "Occurrence of volatile orga ti, USA, American Conference of Governmental In nic compounds in the air of 500 homes in the Federal nses, including olfac dustrial Hygienists (ACGIH). Republic of Germany". In: Seifert, B., Esdom, H., hemical sense. A way Andrews, L.S. and Snyder, R. (1986) "Toxic effects of Fischer, M., Ruden, H. and Wegner, J. (eds) Proceed responses to VOC's, solvents and vapors". In: Klaassen, C., Amadur, M. ings ofIndoor Air '87, Berlin (West), Institute of Water, and Doull J. (eds.) Casarett and Doull's Toxicology; the Soil and Air Hygiene, Vol. 1, pp. 102-106. t the combined sen Basic Sciences of Poisons, London, 3rd edition, Chap. Kristiansen, U. and Nielsen, G.D. (1988) "Activation of rom chemical expo- 20, p. 636. sensory irritants receptors by C7-Cll alkanes", Ar 1ants about the pres- Brink, F. and Postemak, I. (1948) "Thermodynamic ana· chives of Toxicology, 61, 419-425. lyses of the relative effectiveness of narcotics'',]. Cell Lebret, E., Van de Wiel, Bos, H.P., Noij, D and Boleij, 1eficial sources of vo ular. Comp. Physiology, 32, 2ll-233. J.S.M. (1986) "Volatile organic compounds in Dutch ds. These senses are Cain, WS. (1989) Perceptual Characteristics of Nasal Irrita homes", Environment International, 12, 323-332. :ms of evolution and tion. Paper presented at: Course on Sick Building Medina, G. and Cain, WS. (1982) "Bilateral integration Syndrome, October 1989, National Danish Institute in the common chemical sense", Physiology and Beha ngers such as fires or on Occupational Health, Copenhagen, Denmark. vior; 29, 349-353. edible food, and to Cain, WS., See, L.C. and Tusun, T. (1986) "Irritation M0lhave, L. (1986) "Indoor air quality in relation to sen . In offices, as in and odor from formaldehyde; chamber studies". In: sory irritation due to volatile organic compounds'', IAQ'86 Managing the Indoor Air for Health and Energy ASHRAE Transactions, 92, Pt 1 (Paper 2954). als indicate the pres Conservation, Atlanta, USA, American Society of Hea M0lhave, L. (1990a) "The Sick Building Syndrome pecial activities, pro ting Refrigeration and Air conditioning Engineers (SBS) caused by exposures to volatile organic com may be meaningful (ASHRAE), pp.126-137. pounds (VOCs)". In: Weekes, D.M. and Gammage, Clausen G.H., Fanger, P.O., Cain, WS and Leaderer, B.P. R.B. (eds) The Practitioner's Approach to Indoor Air 1pants and, as such, (1985) "The influence of aging panicle filtration and Quality Investigations, Akron, Ohio, USA, American humidity on tobacco smoke odor''. In: Fanger, P.O. Industrial Hygiene Association. positive information (ed.) Clima 2000. Copenhagen, Denmark, National M0lhave, L. (1990b) "Volatile organic compounds, in Institute of Building Research. door air quality and health". In: Walkinshaw, D.S. lice worker may not Cometto-Muniz, J.E. and Cain, WS. (1984) "Temporal (ed.) Proceedings of Indoor Air '90, Ottawa, Inter tble if he or she con integration of pungency", Chemical. Senses, 8, 315-327. national Conference on Indoor Air Quality and Cli nose irritation indi- De Bonoli, M., Knoppel, H., Pecchio, E. and Peil, A. mate, Vol. 5, pp. 15-23. m unknown health (1986) "Concentrations of selected organic pollutants M0lhave, L. (199la) "Indoor climate, air pollution and

---!__ _ 376 M01have: Volatile Organic Compounds, Indoor Air Qualify and Health

comfort", Journal of Exposure Analysis and Environ Otto, D.A., M111lhave, L., Hudnell, H.K., Goldstein, G. mental Epidemiology, I, 63-81. and O'Neil, J. (1990) Neurotoxic Effects of Controlled M111lhave, L. (199lb) "Design considerations for exposure Exposure to a Complex Mixture of Volatile Organic Com experiments at exposure levels below TLV". Accepted pounds, USEPA, HEARL, Research Triangle Park, for European Journal ofApplied Psychology. NC 277ll, USA. EPN600/l-90/001. M111lhave, L., Bach, B. and Pedersen, 0.E (1986) "Hu Thaysen, J.H., Christensen, L.K. and Lorenzen, I. man reactions to low concentrations of volatile orga (1980) Medicinsk Kompendium, Copenhagen, Den nic compounds", Environment International, 12, 167-175. mark, Vol. 1. M111lhave. L., Gr111nkja:r, J. and Larsen, S. (1991) "Subjec Wallace, L. (1987) The Total Exposure Assessment Method tive reactions to volatile organic compounds as air ology (TEAM) study. Summary and Analyses, Washing pollutants",Atmospheric Environment, 25A, U83-U93. ton. USA, US Environmental Protection Agency, Vol. M111lhave, L. and Nielsen, G.D. (1992) "The TVOC indi 1. cator of human response to low level exposures to vol World Health Organization (WHO) (1961) "Constitution atile organic compounds (VOC)". Submitted for In of the World Health Organization" ,Basic Documents, dour Air 1991. 15th edition, Geneva, Switzerland. Nielsen, G.D. and Bakbo, J.C. (1985a) "Sensory irritating World Health Organization (WHO) (1982) lndour Air effects of allyl halides and a role for hydrogen bond Pollutants: Exposure and Health Effects. Copenhagen, ing as a likely feature at the receptor site". '~eta Phar WHO Regional Office for Europe (EURO Reports macologica et Toxicologica, 57, 106-ll6. and Studies No. 78). Nielsen, G.D. and Bakbo, J.C. (1985b) "Exposure limits World Health Organization (WHO) (1989) lndour Air for irritants". Annals· of tM American Conference of Quality: Organic Pollutants, Copenhagen, WHO Re Governmental Industrial Hygienists, U, ll9-133. gional Office for Europe (EURO Reports and Studies Nielsen, G.D. and Vinggaard, A.M. (1988) "Sensory irri No. Ill). tation and pulmonary irritation of C3-C7 n-Alkyla Zweers, T., Skov, P., Valbj111m, 0., M111lhave, L. and the mines; mechanisms of receptor activation", Pharma Danish Indoor Climate Study Group (DISG) (1990) cology and Toxicology, 63, 293-304. "The effect of ventilation and air pollution on per Nordbiick, D. (1990) Environmental Exposures and Per ceived indoor air quality in five town halls", Energy sonal Factors related to Sick Building Syndrome. Thesis and Buildings, 14, 175-181. from Uppsala University, Sweden, Acta Universitatis Upsaliensis.