United StatM
Environmental Protection
Agency
Health Effects
Research Laboratory
Research Triangle Park.
North Carolina 27711
EPA/600/1-90/001
Fbo'ruary 1990
Research and Development
Indoor Air - Health
LJ1
Neurotoxic Effects of
Controlled Exposure to a
Complex Mixture of Volatile
Organic Compounds
38101
EJBD
k ARCHIVE
EPA
. 600-
> 1-
90-
s, 001
-------
DISCLAIMER: This manuscript has been reviewed by the Health Effects Laboratory,
U.S. Environmental Protection Agency, and approved for publication as an EPA
document. Approval does not signify that the contents necessarily reflect the
views and policies of the Agency. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
-------
FINAL REPORT (Deliverable 2520)
NEUROTOXIC EFFECTS OF CONTROLLED EXPOSURE TO A COMPLEX
MIXTURE OF VOLATILE ORGANIC COMPOUNDS
Principal Investigator: David A. Otto
2
Co-Investigators: Lars Molhave
3
H.Kenneth Hudnell
George Goldstein
John O'Neil1
4
Statistician: Dennis House
Technical Support: Gregory Rose
Jon Berntsen
Wayne Counts
Shirley Fowler1
Scott Meade
Human Studies Division, Health Effects Research Laboratory,
U.S. Environmental Protection Agency, RTF, NC
2
Institute of Environmental and Occupational Medicine,
University of Aarhus, Aarhus, Denmark
Neurotoxicology Division, Health Effects Research Laboratory,
U.S. Environmental Protection Agency, RTF, NC
Research and Regulatory Support Division, Health Effects Research
Laboratory, U.S. Environmental Protection Agency, RTF, NC
NSI Environmental Services, Inc., Research Triangle Park,
NC 27711
CE Environmental, Inc., Chapel Hill, NC 27514
Sources of Support: EPA Contract 68-02-4179 (Combustion Engineering)
EPA Contract 68-02-3800 (Subject Recruitment)
Repository Material u H „ us EPA
/„ ,, Headquarters and Chemical Libraries
Pern lanent Collection EPA west a* Room 3340
Mailcode 3404T
1301 Constitution Ave NW
Washington DC 20004
202-566-0556
-------
PREFACE
In October 1986 Congress passed the Superfund Amendments and
Reauthorization Act (SARA, PL 99-499) that Includes Title IV - The Radon Gas
and Indoor Air Quality Research Act. The Act directs that EPA undertake a
comprehensive Indoor air research program.
Research program requirements under Superfund Title IV are specific.
They Include Identifying, characterizing, and monitoring (measuring) the
sources and levels of Indoor air pollution; developing Instruments for Indoor
air quality data collection; and studying high-risk building types. The
statute also requires research directed at Identifying effects of Indoor air
pollution on human health. In the area of mitigation and control the
following are required: development of measures to prevent or abate Indoor
air pollution; demonstration of methods to reduce or eliminate Indoor air
pollution; development of methods to assess the potential for contamination
of new construction from soil gas, and examination of design measures for
preventing Indoor air pollution. EPA's indoor air research program is
designed to be responsive In every way to the legislation.
In responding to the requirements of Title IV of the Superfund
Amendments, EPA-ORD has organized the Indoor Air Research Program around the
following categories of research: A) Sources of Indoor Air Pollution; B)
Building Diagnosis and Measurement Methods; C) Health Effects; D) Exposure
and Risk (Health Impact) Assessment; and E) Building Systems and Indoor Air
Quality Control Options.
EPA is directed to undertake this comprehensive research and
development effort not only through in-house work but also in coordination
with other Federal agencies, state and local governments, and private sector
organizations having an interest in indoor air pollution.
The ultimate goal of SARA Title IV is the dissemination of information to the
public. This activity includes the publication of scientific and technical
reports in the areas described above. To support these research activities
and other interests as well, EPA publishes its results in the INDOOR AIR
report series. This series consists of five subject categories: Sources,
Measurement, Health, Assessment, and Control. Each report is printed In a
limited quantity. Copies may be ordered while supplies last from:
U.S. Environmental Protection Agency
Center for Environmental Research Information
26 West Martin Luther King Drive
Cincinnati, OH 45268
When EPA supplies are depleted, copies may be ordered from:
National Technical Information Service
U.S. Department of Commerce
5285 Port Royal Road
Springfield, VA 22161
-------
TABLE OF CONTENTS
A. EXECUTIVE SUMMARY J
B. BACKGROUND AND RATIONALE *
C. METHODS ;
C.I Subject Selection and Characterizaton /
C.2 Physical Facilities 8
C.3 VOC Generation and Monitoring 9
C.4 Equipment Readiness and Pilot Study 9
C.5 Experimental Design 1°
C.6 Behavioral Battery 11
C.7 Subjective Reactions 11
C.8 Confirmatory and Exploratory Hypotheses 12
C.9 Data Analysis Plan 13
C.10 Risks and Safeguards 1*
D RESULTS 15
D.I Comfort Meter (Potentiometer) 15
D.2 Symptom Questionnaire 15
D.3 Auditory Digit Span (ADS) 16
D.4 NES Main Test Results 17
D.4a Finger Tapping 17
D.4b Visual Digit Span (VDS) 17
D.4c Continuous Performance Test (CPT) 17
D.4d Symbol Digit Substitution (SDS) 18
D.4e Serial Digit Learning (SDL) 18
D.4f Pattern Memory (PM) 18
D.4g Switching Attention (SWATT) 18
D.4h Mood Scales 19
D.5 NES Secondary Test Results 19
D.5a Associate Learning/Recall 19
D.5b Pattern Comparison 19
D.5c Simple Reaction Time (SRT) 20
D.5d Grammatical Reasoning 20
D.5e Horizonal Arithmetic 20
D.6 Water Consumption 20
D.7 Learning/Practice Effects 20
E. DISCUSSION 22
E.I Confirmatory Hypotheses ^
E.2 Exploratory Hypotheses 22
E.2a Sensory Irritation 22
E.2b Controlled Climate Variables 23
E.2c Motor Performance 23
E.2d Mood Scale Alterations 24
E.2e Cognitive Performance 24
E.3 Effects of Practice and Learning 25
F. REFERENCES 28
-------
LIST OF TABLES
1. Symptoms of Sick Building Syndrome 30
2. Subject Selection Criteria 30
3. VOC Mixture 31
Experimental Design 31
5. Measurement Variables 32
6. Comfort Ratings (Potentiometer Readings) 34
7. Symptom Questionnaire 35
8. Symptom Response Changes Over Time 37
9. Yes-No Question Responses 38
10. Auditory Digit Span (without 12 subjects) 39
11. Auditory Digit Span (All subjects) 40
12. Finger Tapping 41
13. Visual Digit Span 41
14. Continuous Performance Task 42
15. Symbol Digit Substitution 42
16. Serial Digit Learning 43
17. Pattern Memory 43
18. Switching Attention Test 44
19. Mood Scale 45
20. Associative Learning 46
21. Pattern Comparison 46
22. Simple Reaction Time 47
23. Grammatical Reasoning 47
24. Horizonal Arithmetic 48
25. Serial Order Effects: Auditory Digit Span 49
26. Serial Order Effects: Visual Digit Span 50
27. Serial Order Effects: Pattern Memory 50
28. Summary of Serial Order Effects 51
LIST OF FIGURES
1. Schematic of Test Protocol 53
2. Comfort Rating Scale Results 54
3. Headache, Odor and Air Quality Responses 55
4. Odor, Eye and Throat Irritation 56
5. Control Question Responses 57
6. Yes-No Question Responses 58
7. Auditory Digit Span 59
8. Visual Digit Span 60
9. Continuous Performance Task 61
10. Switching Attention Test 62
11. Mood Scale Difference Scores 63
12. Serial Order Effects: Auditory Digit Span 64
13. Serial Order Effects: Visual Digit Span 65
14. Serial Order Effects: Confusion Scale 66
15. Serial Order Effects: Pattern Memory 67
-------
APPENDICES
A. VOC Generation and Monitoring Methods .68
B. Instructions for Administering and Scoring the
Auditory Digit Span Test 79
C. Description of Neurobehavioral Evaluation System
(NES) Tests 84
D. Symptom Questionnaire 91
E. Linear Analysis Rating Scale (Potentiometer) 93
-------
A. EXECUTIVE SUMMARY
Growing public awareness of the potential health hazards of chemicals found
in the indoor environment has given rise to the definition of a new problem
designated as sick (or tight) building syndrome (SBS). SBS is characterized by
irritation of the eyes, nose and throat. Preliminary evidence from Danish
studies (Molhave et al, 1986; Kjaergaard et al, 1989) also indicates effects on
cognitive functions such as memory. Many of the chemicals found indoors that
have been linked with SBS are volatile organic compounds (VOCs). Congress has
mandated EPA to characterize the sources, real-world exposure levels and health
risks of indoor pollutants. Little information is currently available on the
health effects of exposure to VOC mixtures found in new or renovated buildings.
EPA has established an intramural research program to address this question.
The current report details the results of the initial study in this program.
Objectives of this study were to confirm and extend the pioneering work of
Molhave et al (1986) and to identify sensitive measures for use in subsequent
studies of sick-building syndrome. Molhave et al. demonstrated sensory
irritant and neurobehavioral effects of controlled human exposure to a complex
VOC mixture at concentrations of 5 and 25 mg/m3, representing the usual and
maximal levels found in new homes. The design of the Danish experiment was
problematic, however. A repeated-measures design was used in which subjects
completed clean air and exposure conditions in the morning and afternoon of the
same day. Three types of confounding may have occurred in this design—practice
effects, VOC exposure carry-over effects from morning to afternoon, and
time-of-day differences. The present study was designed to minimize or
eliminate these problems by scheduling a separate training day to minimize
practice effects, testing all subjects at the same time of day (afternoon), and
separating clean air and exposure runs by a minimum interval of one week.
Other important differences between studies included the subject population
-------
and tests used. The Danish study included both males and females aged 18-64,
all with documented histories of chemical sensitivity. The present study was
limited to young adult males aged 18-39 with no history of chemical sensitivity.
Only two neurobehavioral performance tests were used in the Danish study, while
fourteen tests were used in the present study to fully characterize the nature
of the possible neurobehavioral effects of VOC exposure.
Results of this study confirm the adverse subjective reactions of subjects
to a 25 mg/m3 concentration of volatile organic compounds reported by Molhave et
al. Ratings of general discomfort (defined as irritation of the eyes, nose and
throat), symptom questionnaire responses on odor intensity, air quality, eye and
throat irritation, headache and drowsiness, and mood scale measures of fatigue
and confusion all differed in predicted directions between clean air and
exposure conditions. However, no convincing evidence was found of any
neurobehavioral impairment associated with exposure to the VOC mixture. That
is, we were unable to confirm the Molhave et al report that VOCs impair short
term memory as measured by auditory digit span. Nor were other tests of memory,
sensorimotor performance, or more complex cognitive functions affected by
exposure to this concentration of VOCs. In summary, clear adverse subjective
reactions to VOC exposure, but no functional (neurobehavioral) impairments, were
found.
Confirmation of sensory irritation in a normal, healthy adult male
population is quite important. This finding demonstrates that subjective
reactions to VOCs at levels found in new buildings are not limited to
"complainers" or chemically sensitive subgroups in the general population. In
fact, subjects in this study probably represent the subgroup least likely to be
affected by VOC exposure. Sensory irritation, headache and drowsiness are also
of considerable importance in the community and workplace. Perceived
dissatisfaction with air quality in the home, office, market place, or factory
can lead to reduced productivity, decreased quality of life among home-owners
and workers, and could result in mass exodus from buildings with air quality
problems. This is an important factor in the sick-building syndrome. Further
consideration should be given to the relevance of subjective reactions in
-------
evaluating the adverse health effects of indoor pollutants.
Subjects in this study and previous Danish studies found the odor of VOCs
to be very strong and unpleasant. The unmistakable odor of VOCs confounds
efforts to conduct a double-blind experiment. Subjective reactions to the VOC
mixture can be attributed, at least in part, to the unpleasant odor. If this
were the only objectionable characteristic of VOCs, we would probably conclude
that VOCs at concentrations found in the indoor environment do not constitute
any significant health risk. The time course of responses to odor and symptoms
of irritation differ, however, suggesting that odor accounts for only part of
the subjective reaction to VOCs. Further study is needed to differentiate
olfactory and trigeminal components of the VOC response.
Failure to find any functional consequences of exposure to the VOC mixture
however, raises important questions in assessing health risks and planning
subsequent indoor air studies. For instance, we cannot conclude that a 25 mg/m3
concentration of VOCs poses no neurotoxic risk to susceptible subgroups of the
general population, or even to the general population. Further study with a
chemically sensitive group of subjects is needed and the relationship of age and
gender to VOC response needs to be explored. It is also possible that the
neurobehavioral tests used were not sensitive (or difficult) enough to detect
effects. If subjects find that VOCs irritate the eyes and throat and induce
headaches, one would expect performance to be impaired, at least indirectly, by
distraction or impaired attention.
Another possibility to consider is physiological endpoints. If VOCs
irritate the eyes, nose and throat, we should examine secretions of these
tissues for evidence of inflammation. Kjaergaard et al (1989) recently reported
evidence of inflammatory response in eye fluids following VOC exposure.
However, no evidence of inflammatory response in nasal secretions was found,
consistent with symptom response in the present study. Further study of
inflammatory effects of VOC exposure in eye fluid and other mucosal tissues are
needed.
-------
B. BACKGROUND AND RATIONALE.
"Sick Building Syndrome" costs industry and the government considerable
time and dollars in lost vorktime, worker dissatisfaction, and temporary
abandonment of facilities (Melius et al, 1984). Beyond the economic impact is
the question of possible adverse health effects of exposure to indoor air
pollutants found in new and renovated homes and commercial buildings. Most
people spend the majority of their lives indoors (NRC, 1981). Therefore,
agencies and institutions responsible for protecting human health therefore need
to carefully study the potential adverse effects of chemicals commonly found in
indoor environments.
Molhave and his colleagues (1986) have studied the health effects of
controlled exposure to mixtures of volatile organic compounds found in new-
construction Danish homes. Subjects were healthy adults selected by means of a
questionnaire survey to identify individuals who have experienced discomfort
associated with indoor air quality. Subjects were exposed to three
concentrations (0, 5, 25 mg/m3) of a mixture containing 22 VOCs in the climate
chamber at the Institute of Hygiene, Environmental and Occupational Medicine,
University of Aarhus. The composition of this mixture was based on VOCs
commonly found in newly constructed Danish homes (Molhave and Holler, 1979) and
is similar to the levels and types of VOCs found in U.S. homes (White, 1987).
Several measures of performance and sensory irritation were obtained.
Subjective ratings of sensory irritation and air quality varied significantly
with VOC exposure. Results also indicated impaired memory as measured by the
Digit Span Test. The implications of these results are that exposure to low
level mixtures of VOCs commonly found in new or renovated buildings, levels
hitherto considered to be harmless to human health, may impair-the performance,
productivity, comfort and well-being of workers and home-owners.
These findings provide preliminary evidence of neurotoxic effects of
-------
exposure to complex low-level mixtures of VOCs, although further study is needed
to determine the precise nature of the observed effects. For instance, are
these findings limited to a particular subset of the population who dislike the
odor of VOCs? Is the response psychosomatic or does low-level mixed VOC
exposure produce measurable decrements in objective health endpoints? Which
health endpoints are most sensitive to VOC exposure—e.g., sensory, memory
pulmonary, or immunological measures? Are there other vulnerable subsets of the
general population such as young children, elderly, asthmatics or persons with
allergies? Are VOC exposure effects due to single constituents of the mixture
or the combination of several constituents? Systematic study of these questions
under controlled laboratory conditions is necessary to clarify the human health
risks of exposure to indoor air pollutants. Very little evidence is currently
available concerning the neurotoxic effects of low-level human exposure to
complex VOC mixtures. Controlled exposure studies are also necessary before
undertaking related field studies.
The type of symptoms affected by VOC exposure in the Molhave et al study
has been identified as the Sick Building Syndrome (SBS) by WHO (1982). This
syndrome is characterized by subjective and physiological symptoms of sensory
irritation of mucosal tissue and other symptoms as shown in Table 1.
These symptoms are acute, nonspecific, and cannot be ascribed to any
specific cheraical(s) found in the problem environment. In the Molhave et al
study, individual VOCs in the complex mixture were present at concentrations
well below levels that would be expected to produce irritant effects for any
single chemical in the mixture. Molhave (1986) has hypothecized that the
summation of effects of exposure to the mixture, rather than any single con-
stituent chemical, may be the cause of the effects observed in the study.
Molhave (1986) has further hypothecized that the observed effects are mediated
by trigeminal nerve function which is generally considered to be the neuro-
physiological substrate of the chemical sense.
the
The objective of the present study was to confirm the principal findings of
Aarhus study—i.e., increased perception of sensory irritation and
-------
discomfort associated with exposure to a low-level complex mixture of VOCs.
Molhave et al also reported an effect on digit span performance suggestive of a
VOC-related impairment of short-term memory. The functional significance of the
digit span finding, however, requires clarification. The observed effect could
be a primary neurotoxic effect of VOC exposure or a secondary effect of sensory
irritation (e.g., distraction). Digit span performance also reflects both
attentional and memory processes. The present study included a battery of
neurobehavioral tests designed to clarify the nature and validity of the Danish
findings. Response to low-level VOC exposure was studied in normal healthy
adults.
-------
C. METHODS
C.I Subject Selection and Characterization.
Normal healthy adult volunteers were recruited by the UNC Subject
Recruitment Office (operated under EPA Contract 68-02-3800). Advertisements
posted on campus and published in local newspapers were used to recruit
subjects. Selection criteria based on those used in the Danish study are
specified in Table 2. In brief, healthy, non- smoking Caucasion males aged 18
to 40 with no history of asthma, allergic rhinitis, cigarette use or other
medical conditions were referred for physical examination and allergy testing.
It should be noted that these requirements were more stringent than the Danish
study which included smokers and persons up to 64 years of age. Smokers and
persons over 40 years were excluded from the present study in order to minimize
the variability in performance and sensory irritation that might be associated
with these variables.
Exposure history was assessed by a questionnaire adapted from Molhave et al
(1986). Exposure history was used to exclude individuals with significant
occupational or recreational exposures that could influence neurobehavioral
measures and to exclude individuals with histories of chemical sensitivity.
Volunteers who met selection criteria were referred for physical examination and
allergy skin testing. A standard scratch test for 18 common allergens including
airborne pollens, cat and dog danders, molds and house dust was administered.
Subjects with positive skin allergy tests were not included in the study in
accordance with the selection criteria of Molhave et al (1986).
Payment for participation conformed to guidelines established by the EPA
Clinical Research Branch. A base rate of $7 per hour and bonus for completion
of the study was paid to subjects. $30 was paid to potential subjects who
completed the physical examination. Subjects who qualified for participation in
th* study earned $84 more, depending on the number of hours devoted to training
-------
and running in the control and exposure sessions. Subjects also received a
bonus of $30 for completion of the experiment. Subjects completing all phases
of the study earned a maximum of $144.
Payments were made at the end of the experiment or upon termination of
participation. Although a subject could elect to discontinue the experiment at
any time due to discomfort, concern or other reasons, if the subject chose to
leave for non-medical reasons, only fees earned to that point were paid. On the
other hand, if the investigators terminated an individual on medical or
technical grounds once an exposure session had commenced, the subject(s)
received full payment for that session. Subjects commuting from Durham and
Raleigh were paid an additional $6 and $11, respectively, and all parking fees
were paid.
To conform to the Federal Privacy Act, each subject was assigned a number
at the time of physical examination. Subject privacy was maintained thereafter
by exclusive use of the subject number for subject identification.
Seventy-six subjects completed the VOC study. Data from ten subjects vere
discarded for technical reasons including the addition of the auditory digit
span test after seven subjects had been run and the accidental termination of
exposure during one session with three subjects. Therefore, data from sixty-six
subjects vere used in statistical analyses. The mean age of subjects was 25.0
yrs. (s.d. 5.6).
C.2 Physical Facilities.
The study vas carried out in the exposure facilities of the EPA Human
Studies Division in Chapel Hill (Strong, 1978). The environment including
temperature, humidity, light levels and pollutant concentration vas continuously
controlled and monitored by computer. Personnel operating this facility
included EPA psychologists and technicians and a team of technical and
engineering personnel from CE, Inc. vho operate and maintain the facility under
contract to EPA. These personnel vere highly trained individuals vho routinely
8
-------
administer other similar protocols. A licensed physician was immediately
available to the operating team during all experiments.
C.3 VOC Generation and Monitoring.
The complex VOC mixture described by Molhave et al (1986) was used in the
present study with the substitution of one compound (1,1-dichloroethane for 1,2
dichlorethane). The Molhave mixture of 22 volatile organic compounds is shown
in Table 3. The compounds were selected on the basis of previous Danish surveys
(Molhave and Holler, 1979) including the 10 most commonly found and the 10 found
in greatest concentration in indoor nonindustrial environments, but ex- eluding
known carcinogens. That is, this mixture is characteristic of the chem- icals
and concentrations to which residents in new Danish and American homes are often
exposed. Any compounds classified by the International Agency for Research on
Cancer (IARC monographs) as Group 1 (chemicals for which there is "sufficient
evidence from epidemiological studies to support a causal associa- tion between
the exposure and cancer") or Group 2 (chemicals that are probably carcinogenic
to humans—i.e., "exposures for which there [is] at least limited evidence of
carcinogenic!ty to humans" or "sufficient evidence in animals") were excluded.
The concentration of individual constituents of the mixture shown in Table
3 is specified in ug/m3, each of which is below the threshold limit value (TLV)
for occupational exposure. TLV is the concentration of a chemical to which
workers may be continuously exposed for 8 hours without adverse health effect as
defined by OSHA. A system for generating and monitoring the VOC mixture
equivalent to that used in the Danish study was constructed by CE, Inc. A des-
scription is contained in Appendix A. Subjects were exposed to a 25 mg/m3
concentration of VOCs measured as a toluene equivalent concentration with a
flame ionization detector (FID). The duration of exposure was 2.75 hours.
C.4 Equipment Readiness and Pilot Study.
Prior to initiating the main study, a pilot study was conducted to test
equipment, to train the operating staff, and to check-out the protocol to insure
-------
that the testing schedule ran smoothly. Prior to exposing subjects to the VOC
mixture, extensive testing of the gas generating and monitoring system was
conducted by CE to ensure that the system operated reliably and to ensure that
gas concentrations did not vary by more than +3.0% once the target level (25
mg/m3) was reached. When CE completed testing of the gas generating and
monitoring equipment, two groups of three subjects (total N=6) were run through
the complete protocol in order to work out any scheduling problems and to train
the operating staff. Both groups were run under control (clean air) and
exposure conditions as specified for the main protocol in the following
sections.
C.5 Experimental Design.
Molhave et al (1986) used a repeated measures design in which each sub-
ject completed both clean air control and VOC exposure (either 5 or 25 mg/m3)
conditions on the same day. Half the subjects received clean air in the morn-
ing and half in the afternoon in a counterbalanced design. There are several
problems with this design including possible confounding from carry-over effects
(i.e., carry-over effects on afternoon control performance from morning
exposure), from practice or fatigue effects (session 1 to session 2), and from
time-of-day effects (morning vs. afternoon). Carry-over effects were prevented
by separating control and exposure sessions by a week or more. This separation
also reduced, but did not eliminate practice effects. It also permitted each
subject to complete exposure and control sessions at the same time of day,
eliminating time-of-day confounding. A disadvantage was that subjects were
required to return for an extra day, introducing the possibility of subject
drop-out as well as other possible confounding factors such as different weather
conditions or subject mood and preparedness.
To simplify the design further, only one exposure level—25 mg/m3—was run.
Control and exposure sessions were administered in, a double-blind manner
counter-balanced across subjects similar to the design shown in Table AA.
Although the chamber could accommodate four subjects per run, subject
availability limited testing on many days to less than four as shown on Table
10
-------
4B.
C.6 Behavioral Battery
A battery of behavioral tests was used to assess which behavioral processes
are affected by VOC exposure. The battery included the digit span test reported
by Molhave et al (1986) to be impaired by VOC exposure. This test was
administered in the same manner as Molhave et al—i.e., the paper-and-pencil
form derived from tfechsler (1955) with auditory stimuli presented by tape
recorder. Instructions for the administration and scoring of this test are
contained in Appendix B. Since performance of the digit span test includes
components of attention as well as memory, other behavioral tests were included
to evaluate which processes are affected by VOC exposure. A main set of eight
computerized behavioral tests developed by Baker and Letz (1985) and described
in Appendix C was used. The tests included motor speed, memory, attention and
mood. These tests were administered twice during each testing session—at the
beginning and end. A secondary set of five other sensorimotor, memory and
cognitive tests were administered once midway through each testing session.
These tests are also described in Appendix C.
C.7 Subjective Reactions
A 25-item questionnaire related to sensory irritation, discomfort, and
indoor air quality was used to obtain subjective ratings as described by Molhave
et al (1986). This questionnaire is presented in Appendix D. The feeling of
irritation in the eyes, nose and throat was also measured at 15-30 minute
intervals using a linear potentiometer described in Appendix E. In the
questionnaire subjects were asked to indicate the degree of perceived discomfort
by adjusting the position of a vertical marker on a video screen by means of a
joystick. The range of response varied between two well-defined extremes of
items such as dryness of the nose or skin, irritation of the throat, and room
noise, temperature or humidity. The linear potentiometer is a device that
reduces the questionnaire to a single scale representing the collective
irritation of eyes, nose and throat. The potentiometer consisted of an aluminum
11
-------
box with a sliding lever which electronically and continuously recorded the
comfort rating of a particular subject during the experiment.
C.8 Confirmatory and Exploratory Hypotheses.
Confirmatory and exploratory hypotheses specified a priori are stated
below. Measurement variables from individual tests and questionnaires are
listed in Table 5.
a. Confirmatory Hypotheses.
(1) Exposure to complex VOC mixtures will result in higher ratings
of sensory irritation as measured by the linear analog rating
scale (potentiometer).
(2) Exposure to complex VOC mixtures will result in impaired
cognitive performance relative to clean air as follows:
i. Short-term memory as measured by span length (sum of forward
and backward span lengths) in the auditory digit span test
will be
reduced;
ii. Coding time (msec) for symbol-digit substitution items will
be longer;
iii. Selective attention, as measured by mean time/item in the
switching attention test (incompatible trials) will be
longer.
b. Exploratory Hypotheses.
Exposure to complex VOC mixtures will:
(3) Increase sensory irritation as indexed by questionnaire ratings;
(4) Not influence ratings of controlled climate variables—i.e., air
temperature, humidity, light level and sound level;
(5) Not affect motor performance on the finger tapping test;
(6) Alter selected dimensions of mood—i.e., tension, confusion and
12
-------
fatigue—as indexed by pre-post mood scale difference scores;
(7) Impair other measures of cognitive performance as indexed by
continuous performance, serial digit learning and pattern memory
tests.
C.9 Data Analysis Plan.
The main analyses of the experiment were done to determine if there were
any VOC effects. Secondary analyses of most variables were done to determine if
learning or practice effects were present and, if so, to determine the nature of
those effects. In the main or confirmatory analysis were: (1) auditory digit
span (forward plus backward), (2) potentiometer (comfort meter) measurements,
(3) symbol-digit substitution (mean response time) and (4) switching attention
test (reaction time in alternating condition). All other variables and analyses
are considered as secondary or exploratory in nature.
Variables measured only once during each testing session were analyzed by a
paired t-test on the two measurements for each subject. Variables with pre and
post measurements were analyzed by a paired t-test on the air and VOC change
(post-pre) for each subject. This analysis is equivalent to a two-way analysis
of variance with subjects as one factor and treatment changes as the other
factor. The two response profiles of each of the 25 questions in the comfort
questionnaire and the response profiles of the comfort meter measurements were
tested for parallelism using multivariate analysis of variance of repeated
measures methods. If the parallelism hypothesis for a question was rejected,
the three additional hypotheses were tested in order to explain the nature of
the lack of parallelism. The first hypothesis compared the mid-exposure changes
on air and VOC, the second hypothesis compared the post-exposure changes and the
third hypothesis compared the changes from mid to post-exposure. The first
three questions required either a yes or no response and were analyzed in the
above manner by coding "1" for yes and "0" for no. In other words, the profile
analyzed was the proportion of "yes" responses at each measurement point.
13
-------
Pretest questionnaire responses were analyzed by the nonparametric McNemar
test for the significance of changes. This test was used in order to answer the
question of whether or not potential covariates vere changing from one exposure
to the other.
All variables except questionnaire and comfort meter responses vere
analyzed to determine if learning or practice effects vere present and, if so,
as far as possible, to determine the nature of those effects. For variables
vith pre- and post-measurements during each exposure, the exposure vas ignored
and the actual order of the four measurements vas analyzed by fitting a first,
second and third degree polynomial to the means. For variables vith only one
measurement during each exposure, the exposure vas ignored and the first
measurement vas compared to the second to determine if learning effects vere
present.
C.10 Risks and Safeguards.
The concentrations of all volatile organic compounds used in this study
vere belov the TLVs for occupational exposure by orders of 10 to 1000 (ACGIH,
1982). Although mucosal irritation and discomfort vere reported by subjects in
the study conducted by Molhave et al (1986), these effects vere gone vithin 24
hours as indicated by follow-up questionnaires (Molhave, personal
communication). Reported effects on digit span performance vere marginal and
transient as veil. There is no evidence of any permanent adverse effect of
acute VOC exposure at these levels. There is no known risk involved in the
performance of the computerized behavioral battery or discomfort rating instru-
ments.
In the unlikely event of medical emergency, a licensed physician vas be
present in the EPA facility at all times during the experiment. Comprehensive
emergency procedures used as general practice in the operation of the facility
could have been activated vithin seconds of the occurrence of an emergency.
Emergency equipment vas kept at the site of the exposure chambers at all times.
Furthermore, subjects vere kept under constant surveillance by trained
14
-------
biomedical personnel via closed circuit TV monitoring, direct observation, and
voice (intercom) communication.
D. RESULTS
D.I Comfort Meter (Potentiometer)
Comfort meter response profiles for exposure and clean air were
significantly different (p < .001) as shown in figure 2. That is, subjects
expressed much more discomfort (irritation of mucous membranes) during VOC
exposure than clean air. An abrupt divergence of comfort levels occurred when
VOC exposure commenced at the 75 min. point in testing. The maximum difference
in comfort levels occurred about 75 min. after exposure onset. Mean
potentiometer readings (and standard errors) at each sampling point are shown in
Table 6.
D.2 Symptom Questionnaire
A 25-item symptom questionnaire was administered three times during each
testing session—during pre-exposure baseline, immediately after VOC build-up,
and at the end of the 2.75 hr. exposure period. Mean response levels at the
three measurement points for continuous items 1-22 are shown in Table 7.
Significant main effects of VOC exposure (25 mg/m3) compared to clean air were
found for six items including headache, odor level, eye and throat irritation,
drowsiness and air quality. Figures 3 and 4 illustrate these findings. Five
environmental variables (light and noise levels, temperature, humidity and air
movement) were held constant during testing. Responses to control variables did
not vary with exposure (Figure 5).
Different patterns of change over time in response levels were apparent for
different symptoms. Further analysis of temporal patterns (Table 8) indicated
that the perceived odor level increased dramatically after VOC exposure onset,
but then decreased significantly with continued exposure at a constant level
15
-------
(Figure 3). A similar (but inverse) pattern vas observed for air quality.
Headache, eye and throat irritation also increased significantly after VOC
onset, but remained elevated for the duration of VOC exposure (Figure 4).
Increased drowsiness, relative to clean air, was significant only at the end of
exposure.
Three other questions solicited binary (yes-no) responses concerning air
quality as follows:
(1) Do you feel now that staying in the chamber is comfortable?
(2) Would you be satisfied with this air quality in your own home?
(3) Would you ventilate your home more?
VOC exposure in each case elicited significant change in perceived comfort as
shown in Figure 6 and Table 9. Sixty-four percent more of the subjects, for
instance, indicated that they would not be satisfied with the air quality
perceived after VOC exposure build-up (mid) than at the same point during clean
air. A small (14%) but significant reduction in dissatisfaction occurred by the
end of exposure.
D.3 Auditory Digit Span (ADS)
Six scores were derived from the ADS test including the standard measures
of the longest spans recalled in forward and backward conditions and the total
(forward + backward) span. Three variant measures, designated as "weighted"
scores, were also obtained as described in the methods section.
A problem was encountered in the administration of ADS to some subjects who
failed to reverse numbers in the backward condition. That is, numbers were
written down from right-to-left on the answer sheets, enabling these subjects to
complete the backward condition without mentally reversing numbers. The problem
was observed in 12 subjects, despite explicit instructions and training to the
contrary. Therefore, ADS results were analyzed both with and without the 12
16
-------
aberrant subjects. Hean pre- and post-exposure scores are shown in Tables 10
(without 12 subjects) and 11 (all subjects). VOC exposure did not affect
auditory digit span in the predicted manner. One measure (standard backward
span length) varied significantly with VOC exposure, but in the wrong
direction—i.e., that is, auditory digit span (backwards) was slightly longer
during exposure than clean air. ADS results are shown in Figure 7.
D.4 NES Main Test Results
D.4a Finger Tapping
As predicted, finger tapping speed was not altered by VOC exposure. The
mean numbers of finger taps obtained with preferred, non-preferred and
alternating hands are shown in Table 12.
D.4b Visual Digit Span (VDS)
Three scores were derived from the VDS test including the longest spans
recalled in forward and backward directions and the total (forward + backward)
span. These scores are comparable to the standard scores obtained form the
auditory digit span test. Mean pre- and post-exposure scores are shown in Table
13 and Figure 8. Forward span length increased slightly but significantly more
under clean air than VOC exposure. Backward and total span lengths did not vary
with exposure.
D.4c Continuous Performance Test (CPT)
Three CPT measures were analyzed—mean reaction time (RT), errors of
omission, and false positives. The primary measure (RT) did not vary with VOC
exposure, nor did omission errors. A marginal trend (p = .061) toward increased
false positives was observed during VOC exposure. Mean scores are provided in
Table 14 and illustrated in Figure 9.
17
-------
D.4d Symbol Digit Substitution (SDS)
Two scores were obtained from the SDS test—the mean number of incorrect
responses per trial, and the mean time to complete a trial (requiring nine digit
substitutions). Hean scores and results are shown in Table 15. Neither score
was affected by VOC exposure.
D.4e Serial Digit Learning (SDL)
Two measures were analyzed from the SDL test, a performance score described
in the methods section, and a variant of the performance score in which the
total number of trials to reach criterion was added to the performance score.
Unlike related auditory and visual digit span tests, smaller scores indicate
better performance in the SDL test. Mean scores and results are detailed in
Table 16.
SDL performance was not affected by VOC exposure.
D.4f Pattern Memory (PM)
Three measures were analyzed including the mean response latency of all
trials, mean response latency for correct trials only, and the mean number of
correct trials. Results and means are shown in Table 17. VOC exposure did not
affect pattern memory.
D.4g Switching Attention (SWATT)
Seven response-time measures were derived from the SWATT test including
mean response times for three basic conditions—side, direction and the
alternation of side and direction (switching). The switching condition was then
broken down into four subsets of trials—side compatible, side incompatible,
18
-------
direction compatible, and direction incompatible. Mean response times for each
of these measures are presented in Table 18. Response times improved over the
testing session (pre - post difference) for all measures during clean air. The
pre - post difference was less or in the opposite direction for six of seven
measures during VOC exposure, although none of the differences were
statistically significant. Response times also increased markedly relative to
the complexity of the type of trial—the side direction is easiest, the
direction condition is more difficult, and the alternating (switching) direction
condition is most difficult as shown in Figure 10.
D.4h Mood Scales
Mean scores obtained for the five mood scales during clean air and VOC
exposure are shown in Table 19. Post - pre change scores differed significantly
during VOC exposure on two scales (confusion and fatigue) as predicted. That
is, increases in fatigue and confusion over the 4-hour testing session were
greater during VOC exposure than during clean air as shown in Figure 11.
D.5 NES Secondary Test Results
D.5a Associate Learning/Recall
Two measures were obtained for this test—the total number of correct
pairings over 3 trials and the number of correct pairings made during the
subsequent recall trial. Means and results are shown in Table 20. Associate
learning and recall were not affected by VOC exposure.
D.5b Pattern Comparison
Three measures were analyzed including the number of correct trials, mean
latency for all trials, and mean latency for correct trials only. Mean scores
and results are presented in Table 21. VOC exposure did not affect any pattern
19
-------
comparison measures.
D.5c Simple Reaction Time (SRT)
Mean reaction-times were examined for preferred and non-preferred hands.
Means and results are shown in Table 22. SRT was not altered by VOC exposure.
D.5d Grammatical Reasoning
Three measures were obtained from this test including the mean response
time for individual trials, the total number of errors, and a composite score of
the total number of correct responses minus the total number of errors. None of
these measures vas affected by VOC exposure as shown in Table 23.
D.5e Horizontal Arithmetic
Three measures were obtained including the number correct, mean response
time for all trials and the mean response time for correct trials only.
Slightly more problems were answered correctly during VOC exposure (p = .031)
than during clean air as shown in Table 24.
D.6 Water Consumption
Subjects consumed slightly more water during VOC exposure (2.60 cm) than
during clean air (2.36 cm), but this difference was not significant.
D.7 Learning/Practice Effects
First, Second and third degree polynomials were fitted to the four serial
order means of measures from each of the NES main tests to determine whether
learning or practice effects occurred beyond testing. A significant linear
contrast implies that there are learning or practice effects across the two
training sessions (or four serial test administrations). A significant
20
-------
quadratic effect implies that the learning or practice effect is approaching or
has reached asymptote. A significant cubic effect within the present design
suggests that learning or practice effects occur within each testing session
(pre-post comparison). The absence of linear, quadratic or cubic effects
implies no learning or practice effects. Most measures showed significant
serial order effects within or across testing sessions. The auditory digit span
test illustrates significant linear trends for all measures (Figure 12, Table
25). The total span length measure on the Visual Digit Span Test (Figure 13,
Table 26) demonstrates a clear quadratic pattern. An example of a cubic effect
in the absence of linear or quadratic trends occurred on the Confusion Mood
Scale (Figure 14). Response latencies on the pattern memory test exhibited a
complex serial order pattern in which all three contrasts were significant
(Table 27, Figure 15). Table 28 summarizes the results of polynomial contrasts
for all computerized NES tests used in this study.
21
-------
E. DISCUSSION
E.I Confirmatory Hypotheses.
Four confirmatory hypothese were specified, only one of which was
confirmed. As hypothesized, subjects expressed considerably more general
discomfort on the linear analog rating scale of eyes, nose and throat irritation
during VOC exposure than clean air. This finding is consistent with previous
Danish reports (Molhave et al., 1986; Kjaergaard et al, 1989).
Other hypotheses that VOC exposure would impair cognitive performance of
short-term memory on the auditory digit span test, coding time on the symbol
digit substitution test, and selective attention on the switching attention test
were not confirmed. Contrary to Molhave et al's (1986) report, short-term
memory was not affected by VOC exposure.
E.2 Exploratory Hypotheses
Five exploratory hypotheses were specified, four of which were confirmed.
E.2a Sensory Irritation.
Symptoms of sick building syndrome, listed in Table 1, include irritation
of the eyes, nose and throat; mental fatigue, and headache. Subjects expressed
more irritation of eyes and throat (but not nose) during exposure to VOCs than
clean air. Subjects also reported more headache and drowsiness during VOC
exposure, consistent with Danish studies (Molhave et al, 1986; Kjaergaard et
al, 1989). Symptom questionnaire results provide a more detailed picture of the
irritant effects of VOC exposure than the general measure of discomfort provided
by the linear analog rating scale. Although the potentiometer rating was
defined as an index of "eye, nose and throat irritation", responses to specific
symptom questions suggest that VOCs irritated the nose less than eyes and
throat. This finding has implications for future VOC studies—i.e., that
measures of inflammatory process in eye fluids and throat secretions may yield
more useful information than in nose secretions. Failure to find a significant
22
-------
increase in nose irritation is not in accord with the SBS profile, although it
is consistent with Kjaergaard et al's (1989) report that VOC exposure did not
produce any signs of inflammation in nasal secretions.
The most dramatic reactions to VOC exposure in the present study and in
previous Danish studies were obtained on ratings of odor intensity and air
quality. Subjects exposed to the 25 mg/m3 VOC mixture found the odor very
strong and unpleasant. Subjects similarly reported that VOCs degrade air
quality very much. Molhave et al (1988), moreover, found a systematic
dose-dependent relationship of perceived odor intensity and air quality at VOC
exposure levels of 3, 8, and 25mg/m3.
The strong and unmistakable odor of the VOC mixture poses a methodologic
challenge—i.e., VOCs and clean air cannot be administered to subjects in a
double-blind manner. Subjects know unequivocally when exposure begins. This
problem was reduced by using a counterbalanced design. The rapid habituation of
olfactory sensation also helps to mitigate this problem. Results of the present
study, for instance, indicate that perceived odor intensity decreased at the end
of the testing period, while eye, throat, headache, fatigue and confusion scores
either increased or remained elevated during the 2.75 hour exposure period. The
irritant effects of VOC exposure thought to be mediated by the trigeminal nerve
(Molhave, 1986), can thus be distinguished from olfactory response over time.
Further study is needed to characterize in detail the time course of olfactory
and trigeminal response to VOCs.
E.2b Controlled Climate Variables
It was further hypothesized that ratings of five climate variables (air
temperature, humidity, light and noise levels, and air flow), which were held
constant throughout testing, would not vary with VOC exposure. This hypothesis
was confirmed, providing construct validity to the symptom questionnaire.
E.2c Motor Performance.
The finger tapping test provides a simple, straightforward measure of motor
23
-------
performance. Since motor dysfunction has not been associated with VOC exposure,
we predicted no effects on finger tapping. This hypothesis was confirmed.
E.2d Mood Scale Alterations.
Pre-post differences on three mood scales—tension, confusion and fatigue—
associated with VOC exposure were hypothesized. This hypothesis was confirmed
for confusion and fatigue, but not tension. These findings are consistent with
SBS syndrome as veil as the drowsiness item on the symptom questionnaire. The
lack of effect on the tension and anger scales is of some interest, particularly
in view of the perceived increase in discomfort and sensory irritation. Many
constituents of the mixture are organic solvents with known anesthetic or CNS
depressant properties—e.g., n-butylacetate, 1-1-dichloroethane, n-hexane, and
p-xylene (Anger and Johnson, 1985). The anesthetic or narcotic effects of the
solvents may account for the absence of tension or anger in subjects during VOC
exposure.
E.2e Cognitive Performance.
The final exploratory hypothesis concerning the impairment of performance
on three different tests of cognitive performance (continuous performance test,
serial digit learning and pattern memory) was not confirmed. The results of all
memory, sensoriraotor and other types of cognitive tests were very consistent.
None of these tests showed any convincing evidence that VOC exposure impaired
performance, despite marked subjective reactions on the symptom questionnaire
and comfort rating scale. Nor is this finding particularly surprising in view
of the low aggregate exposure level (7 ppm toluene equivalent) of the VOC
mixture. Neurobehavioral effects have seldom been observed below 300 ppm
toluene (Johnson et al, 1987). Concentrations of individual chemicals in the
mixture were below TLVs by factors of 10 to 100 or more.
Molhave et al (1986) reported a small reduction in digit span during VOC
exposure. The consistent negative results on fourteen neurobehavioral tests
argue rather persuasively against any functional impairment in healthy young
adult males exposed to this concentration (25 mg/m3) of volatile organic
24
-------
compounds. Exposure and clean air sessions were run in the morning and and
afternoon of the same day in the original Danish study, giving rise to possible
confounding from exposure carry-over effects, practice or fatigue effects, and
time-of-day effects. Modifications in the experimental design, to mitigate
these confounding factors render the present results more reliable than the
original study.
On the other hand, several important differences in the subject samples
used in the two studies could possibly account for the discrepant digit span
findings. (1) Molhave et al carefully selected chemically sensitive subjects
with documented histories of indoor climate problems. Chemically sensitive
subjects were excluded from the present study. (2) The range of ages in the
Danish study (18-60 years) was much wider than the present study (19-39 years).
Older subjects could be more sensitive than younger subjects to VOC exposure.
(3) Finally, the Danish study included equal numbers of males and females.
Maizlich et al (1987) have shown that females perform more poorly than males on
many NES tests. Whether cognitive performance is more susceptible to impairment
by VOC exposure in women then men, in older than younger subjects, and in
chemically sensitive compared to normal healthy subjects are questions that
require further study.
E.3 Effects £f Practice and Learning
The question of practice or learning effects is an important concern in
repeated-measures studies employing neurobehavioral tests. Such effects could
exaggerate or obscure effects of chemical exposure if not controlled properly in
the experimental design. It is also of considerable interest to the research
community to characterize the effects of practice and learning in a widely used
neurotoxicity testing system. For the present discussion the terms "practice"
and "learning" will be used synonymously. Although the study was not
specifically designed to examine practice effects, the study provided a
convenient and extensive dataset to evaluate this question.
Nine tests were administered twice on each of two test days, yielding four
25
-------
replications. Test days were separated by at least one week. Practice effects
were evaluated by testing for linear, quadratic and cubic trends in the data
arrayed in the serial order of presentation. As noted previously, a significant
linear trend implies extended practice effects across the four test
administrations, quadratic effects imply that practice effects approach or reach
asymptote levels, and cubic effects imply transient practice effects (within a
testing session).
Results of polynomial tests for serial order effects are summarized in
Table 28 including data on nine tests repeated four times. Evidence of extended
practice effects (significant linear trends) were found on five tests—auditory
and visual digit span, continuous performance, symbol digit substitution, and
switching attention tests. Results of three secondary NES tests (associate
learning, pattern comparison, and grammatical reasoning tests), administered
once during each testing session, also suggest extended practice effects. Only
one of five mood scales (tension) exhibited a significant linear trend. Tests
which do not appear to have extended practice effects include finger tapping,
serial digit learning, other mood scales, simple reaction time, and horizontal
addition. Finger tapping and two mood scales (fatigue and confusion), however,
showed significant cubic trends which imply within-session effects, but no
between-session effect. Mood scales, in particular, are designed to assess
short-term subjective reactions to experimental variables. Fatigue and
confusion scales, if valid, should vary over a demanding task and then return to
previous values after resting or at the beginning of another testing session.
Significant quadratic effects were observed in two tests—visual digit span and
pattern memory—and marginal quadratic trends were found for some variables in
serial digit learning, switching attention, and the tension scale. All three
trends occurred in the pattern memory test.
The implication of the serial order analysis is that practice or learning
effects occur in the performance of most neurobehavioral tests commonly used in
neurotoxicity testing of humans. It should be noted that these effects were
observed despite an initial training session. Data from the training session
were not included in the present analysis because the length and complexity of
26
-------
some tests varied on training and testing days. However, a single training
session does not appear to be adequate to eliminate practice/learning effects on
most NES tests. Further parametric studies need to be done to systematically
evaluate the practice/learning characteristics of each NES test. Task
complexity, length of task and interest trial interval must be carefully
assessed in order to determine optimal parameters for the use of NES tests in
repeated-measures studies in the laboratory as well as in the field.
27
-------
14. REFERENCES
American Conference of Government Industrial Hygienists. Threshold Limit Values
for Chemical Substances and Physical Agents in the Workroom Environ- ment with
Intended Changes for 1982. ACGIH Publ. Office, Cincinnati, 1982.
Bach B, Molhave L, Pedersen OF. Human reactions during controlled exposures to
low concentrations of organic gases and vapors known as normal indoor air
pollutants: Performance tests. In Berglund et al (eds) Indoor Air, Vol 3,
Sensory and Hyperreactivity Reactions to Sick Buildings. Swedish Council for
Building Research, Stockholm, 1984, pp397-401.
Hathaway SR, McKinley JC. The Minnesota Multiphasic Personality Inventory. The
Psychological Corporation, 304 East 45th St, New York, NY 10017, 1966.
Hollingshead AB. Two-factor index of social position. Unpublished ms., Yale
University, 1957 (Available from the author, 1965 Yale Station, New Haven, CT
06520).
IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to
Humans, Volumes 1-41, Lyon, 1971-1986.
Johnson BL (ed). Prevention of Neurotoxic Illness in Working Populations.
Wiley, New York, 1987.
Kjaergaard S, Molhave L, Pedersen O.F. Human reactions to a mixture of indoor
air volatile organic compounds. Environment Int., 1989 (in press).
Letz R, Baker EL. Computer-administered neurobehavioral testing: On defining
its limitatins. In: Environmental Health, Doc 3, World Health Organization,
Copenhagen, 1985:158-162.
Maizlich N, Schenker M, Weisskopf C, Seiber J, Samuels S. A behavioral
evaluation of pest control workers with short-term, low-level exposure to the
organophosphate Diazinon. Am J Indus Med, 1987, 12:153-172.
Melius J, Wallingford K, Carpenter J, Keenlyside R. Indoor air quality: the
NIOSH experience. Am Conf Indus Hyg, Rep 10, 1984: 3-7.
Molhave L. Indoor air quality in relation to sensory irritation due to volatile
organic compounds. ASHRAE Transactions, Vol 92(1), paper 2954, Dec 1986.
Molhave L, Bach B, Pedersen OF. Human reactions during controlled exposures to
low concentrations of organic gases and vapors known as normal indoor air
pollutants. In Berglund et al (eds) Indoor Air, Vol 3, Sensory and Hyperreac-
tivity Reactions to Sick Buildings. Swedish Council for Building Research,
Stockholm, 1984, pp431-436.
28
-------
Molhave L, Bach B, Pedersen OF. Human reactions to low concentrations of
volatile organic compounds. Environment Int., 1986, 12: in press.
Molhave L, Jensen JG, Larsen S. Acute afnd subacute subjective reactions to
volatile organic air pollutants. Unpublished Report, Inst. Environ. Occup.
Med., Univ. of Aarhus, Denmark, Dec. 1988.
Molhave L, Holler J. The atmospheric environment in modern Danish dwell-
ings-Measurements in 39m flats. in Indoor Climate, P Fanger and 0 Valbjorn
(eds), SBI, Horsholm, Denmark, 1979, pp!71-186.
National Research Council, Committee on Indoor Pollutants. Indoor Pollutants.
National Academy Press, Washington, DC, 1981.
Strong AA. Description of the CLEANS Human Exposure System. EPA-600/1-78-064,
Health Effects Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC, Nov 1978.
Wechsler D. Wechsler Adult Intelligence Scale Manual. Psychological Corpora-
tion, New York, 1955.
White JR. A Review of Indoor Air Contaminants and Recommendations for a
"Typical" Indoor Air Mixture. Internal Report, Air and Energy Engineering
Laboratory, U.S. Environmental Protection Agency, Research Triangle, 1987.
World Health Organization. Indoor Air Quality Research. EURO Reports 103,
Stockholm, 1986.
29
-------
TABLE 1. SYMPTOMS OF SICK BUILDING SYNDROME (WHO, 1982)
Irritation of eye, nose and throat
Dry mucous membrane and skin
Erythema
Mental fatigue, headache
Airvay infections, cough
Hoarseness of voice, wheezing
Nonspecific hyperreactivity reactions
Nausea, dizziness
TABLE 2. SUBJECT SELECTION CRITERIA
Age: 18-40 years
Race: Caucasion
Gender: Male
Non smokers
No pre-existing diseases or
abnormal conditions such as:
*chronic airway disease
*cardiac disease
*Asthma
*Chronic bronchitis
*tuberculosis
*Psychiatric disorder
*Neurological disorder
*Skin disease
*Chronic eye disease
*0besity
*Cancer or other severe disease
*Alcohol or Drug Addiction
*IgE Atopy (negative skin test
to common agents)
Native language must be English
Subjects must be able to perform
tasks without difficulty
Normal psychological profile
as determined by MMPI
30
-------
TABLE 3. MOLHAVE et al (1986) VOC MIXTURE
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Compound
n-Hexane
n-Nonane
n-Decane
n-Undecane
1-Octene
1-Decene
Cyclohexane
3-Xylene
Ethylbenzene
1,2, 4-Trimethylbenzene
n-Propylbenzene
a-Pinene
n-Pentanal
n-Hexanal
Iso-propanol
n-Butanol
2-Butanone (MEK)
3-Methyl-3-Butanone
4-Methyl-2-Pentanone
n-Butylacetate
Ethoxyethylacetate
1,1-Dichloroe thane
Ratio
1
1
1
0.1
0.01
1
0.1
10
1
0.1
0.1
1
0.1
1
0.1
1
0.1
0.1
0.1
10
1
1
cone. (ug/m3)
825
825
825
75
8
825
75
8250
825
75
75
825
75
825
75
825
75
75
75
8250
825
825
*Based on a mixture with a total concentration of 25 mg/m3
TABLE 4. EXPERIMENTAL DESIGN
A. TREATMENT SESSION 1 SESSION 2
A 0 25
B 25 0
B.
No. of Subjects
in Chamber
1
2
3
4
Frequency of Occurrence
during Testing
20
23
15
10
31
-------
TABLE 5. MEASUREMENT VARIABLES
A. MAIN NES TESTS
TEST
MEASURE
C/E*
Finger Tapping
Visual Digit Span
Continuous
Performance
Test
Symbol Digit
Substitution
Serial Digit
Learning
Pattern Memory
Switching Attention
Mood Scales
flaps for preferred, non preferred,
and alternating condition
Total span length (forward+backward)
Forward span length
Backward span length
Mean Reaction Time
tMisses
#False Alarms
Mean msec/digit
terrors
^trials attempted + score
Mean msec/trial
#correct
Mean msec/side condition
Mean msec/direction condition
Mean msec/switching condition
Tension
Depression
Anger
Fatigue
Confusion
E
E
E
E
E
E
C
E
E
E
E
E
C
E
E
E
E
E
32
-------
TABLE 5 (cont.)
B. SECONDARY NES TESTS
TEST
MEASURE
C/E
Associate Learning
Associate Recall
Pattern Comparison
Simple Reaction Time
Grammatical Reasoning
Horizontal Arithmetic
# Correct (trials 1-3)
# pairs correctly recalled
mean latency of trials 2-25,
excluding incorrect trials
# trials correct
mean RT for preferred
hand (P2-P5) and
non preferred hand (N6-N9),
excluding block 1
mean response time
# errors
# correct
mean response time
E
E
E
E
E
E
E
E
C. OTHER TESTS
Auditory Digit Span
Comfort Scale
(Potentiometer)
Symptom Questionaire
total span length (Fwd + Bwd)
forward span length
backward span length
difference profile
between VOC exposure
and clean air
difference between VOC
exposure and clean air
for individual items
C
E
E
E
*C=Conformity, E=Exploratory Measures
33
-------
TABLE 6
MEAN POTENTIOMETER READING
BY EXPOSURE AND TIME
EXPOSURE
AIR
VOC
TIME
(MINUTES)
0
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
0
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
MEAN*
.048
.059
.234
.235
.235
.310
.321
.361
.362
.438
.461
.531
.467
.476
.586
.572
.570
.103
.106
.373
.382
.380
.570
.921
1.084
1.136
1.277
1.303
1.131
1.166
1.178
1.223
1.222
1.223
S.E.
.017
.020
.045
.045
.045
.056
.055
.056
.056
.069
.070
.081
.077
.078
.097
.093
.094
.033
.033
.055
.054
.054
.074
.096
.100
.106
.117
.123
.115
.116
.118
.123
.131
.131
* N=63
Ho: F(16,47) = 3.98 (P<.001)
-------
TABLE 7
QUESTIONNAIRE
QUESTION
1. BODY/SKIN
IRRITATION
2. NOISE LEVEL
3. EYES DRY / WATERY
4 . HEADACHE
5. FACIAL SKIN
TEMPERATURE
6. SLUGGISHNESS
7. GDOR LEVEL
8. EYE IRRITATION
9. BODY TEMPERATURE
10. -I5HT INTENSITY
11. ZOUGHING
TIME
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
: POST
ft
MEAN(tt)
1.77
3.35
5. 18
11.48
12.91
13.80
12.71
12.23
11.30
2.00
2.12
3.41
14.00
12.30
11.33
6.33
8.27
9.83
8.73
6.53
0.68
2.56
5.21
7.59
14.03
11.05
10.05
17.11
17.35
18.08
1.38
3.94
1.05
tIR
S.E.
0.51
0.73
0.80
0.76
0.82
0.89
0.36
0.45
0.56
0.61
0.53
0.76
0.37
0.55
0.63
0.85
0.95
1.09
0.95
0.89
0.91
0.57
0.87
0.97
0.46
0.56
0.66.
0.44
0.48
3. 55
3.41
0.32
0.37
V
MEAN ( * )
1.73
5.26
6.39
10.88
13.45
14.09
13.30
12.39
11.24
1.64
5.05
6.97
14.53
12.82
13.21
5.70
8.70
11.71
B. 96
21.68
17.77
3.70
9.68
10.71
13.86
11.58
11.55
17.44
17.68
18.42
1.1*
2.06
1.82
OC
S.E.
0.48
0.90
0.92
0.75
0.81
0.88
0.34
0.59
0.60
0.48
0.82
1.04
0.39
0.63
0.60
0.84
0.94
1.10
0.99
0.80
1.09
0.76
1.04
1.13
0.49
0.62
0.64
0.50
a. 52
: 0.53
: 13.41
: 3.57
: 0.55
F«
(P)
1.71
( .190)
1.37
( .261)
0.47
( .626)
9.52
( < . 001 )
2.05
(.138)
2.42
C . 098 )
77.90
«.001)
4.13
(.021)
1.84
( .167)
0.00
( . 999 )
: 2.31
( . 108 ) ;
* N
«
•66
& 64 Degre<
es of Freedom
35
-------
TABLE 7
(Can t)
QUESTIONNAIRE
QUESTION
12. FACE IRRITATION
13. HUMIDITY
14. AIR MOVEMENT
15. NOSE IRRITATION
16. TEMPERATURE
17. TIRED SLEEPY
IB. SKIN. BODY
no IST /DRY
19. THROAT IRRITATION
20. AIR QUALITY
21. ^ACIAL SKIN
MOIST/DRY
22. CONCENTRATION
TIME
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
f.
MEAN(*») !
1.70
3.35
4.55
11.62
11.38
11.44
17.86
19.79
20.09
4.33
5.26
6.85
10.82
8.61
7.70
6.97
8.79
10.45
12. 88
12.03
11.76
2.48
3.92
4.45
11.00
11.59
11.52
12.65
12.23
11.86
2.42
2.55
3.20
IIR
S.E.
0.46
0.76
0.84
0.56
0.61
0.65
0.67
0.59
0.62
0.80
0.89
1.03
0.42
0.47
0.53
0.89
1.02
1.02
0.37
13.60
13.56
0.60
0.79
0.92
0.74
0.80
0.87
0.51
0.56
0.59
0.59
0.55
0.66
V,
MEAN ( * )
2.36
4.33
5.17
12.24
12.50
12.24
17.88
19.41
20.23
5.12
7.77
9.41
10.47
8.36
7.95
5.98
8.08
12.15
13.02
12.00
12.24
1.71
5.65
6.18
10.74
19.98
18.36
12.45
12.17
11.53
2.26
3.42
4.03
'DC
S.E.
0.56
0.79
0.86
0.54
0.62
0.60
0.65
0.62
0.57
0.85
0.97
1.11
0.35
0.46
0.44
0.90
0.91
1.05
0.42
0.42
0.47
0.46
0.94
1.00
0.77
0.70
0.80
0.49
0.52
(3.56
0.56
0.69
0.75
F*
(P)
0.08
( . 924 )
0.24
( .784)
0.50
( .612)
1.16
( .319)
0.60
( .550)
4.38
( .017)
3.36
i .696)
4.41
( .016)
32.86
( <.001)
0.08
( . 922 5
1.75
( .183)
« 2 4 e>4 Degrees of
36
-------
TABLE 8
ANALYSIS OF QUESTIONS
4, 7, 8, 17, 19, AND 20
QUESTION
4 . HEADACHE
7. ODOR LEVEL
8. EYE IRRITATION
17. TIRED / SLEEPY
19. THROAT
IRRITATION
20. AIR QUALITY
CONTRAST
(*)
1
2
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
CONTRAST
MEAN
3.29
3.92
0.63
15.02
10.96
-4.06
3.33
1.98
-1.35
0.28
2.69
2.41
2.50
2.50
0.00
8.65
7.10
-1.55
F(*«)
16.04
15.99
0.70
153.16
79.04
12.79
8.24
2.34
1.01
0.05
4.65
7.80
6.73
6.84
0.00
66.47
39.92
5.30
P
<.001
<.001
0.406
<.001
<.001
<.001
0.006
0.131
0.318
0.815
0.035
0.007
0.012
0.011
0.999
<.001
<.001
0.024
*1: Change to mid on VOC - Change to mid on AIR
2: Change to post on VOC - Change to post on AIR
3: Contrast 2 - Contrast 1
** 1 and 65 Degrees of Freedom
37
-------
TABLE 9
QUESTIONNAIRE
(YES-NO QUESTIONS)
QUESTION
23
24
25
TIME
PRE
MID
POST
PRE
MID
POST
PRE
MID
POST
(•
PROP.
"YES"
0.95
0.92
0.80
0.79
0.73
0.74
0.35
0.3B
0.36
>IR
5.E.
0.03
0.03
0.05
0.05
0.06
0.05
0.06
0.06
0.06
V
PROP.
"YES"
0.97
0.62
0.59
0.79
0.09
0.23
0.33
0.83
0.73
/DC
S.E.
0.02
0.06
0.06
0.05
0.04
0.05
0.06
0.05
0.06
F*
(P)
15.19
«.001)
44.68
«.001)
22.35
( <.001 )
# N=66
t 2 & 64 Degrees of Freedom
TABLE 9A
QUESTIONNAIRE
(YES-NO QUESTIONS)
: QUEST I ON
i
: 23
1
: 24
1
25
CONTRAST
(*)
1
2
3
1
2
3
1
2
3
CONTRAST
MEAN
-0.32
-0.23
0.09
-0.64
-0.51
0.13
0.47
0.39
-0.08
F(«>
26.62
12.60
1.51
90.14
50.63
4.84
42.18
28.59
2.03
P
<.001
<.001
0.223
<.001
<.001
0.031
<.001
< .001
0.159
*1: Change to mid on VOC - Change to mid on AIR
2: Change to post on VOC - Change to cost on AIR
3: Contrast 2 - Contrast 1
** 1 and 65 Degrees of Freedom
38
-------
TABLE 10
AUDITORY DIGIT SPAN
(WITHOUT 12 SUBJECTS)
TEST
FORWARD
(STANDARD)
BACKWARD
(STANDARD)
TOTAL
(STANDARD)
FORWARD
(WEIGHTED)
BACKWARD
(WEIGHTED)
TOTAL
(WEIGHTED)
TRT
AIR
VOC
AIR
VOC
5 S5SS^SS
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
7.85
8.09
7.89
8.09
7.43
7.24
7.22
7.54
15.28
15.33
15.11
15.63
60.61
62.74
60.96
61.91
53.43
54.43
53.81
58.59
114.04
117.17
114.80
120.50
S.E.
0.14
0.13
0.15
0.14
0.11
0.15
0.13
0.11
0.22
0.22
0.24
0.21
1.99
1.71
1.95
2.04
1.61
1.73
1.74
1.69
3.31
2.89
3.32
3.14
CHANGE
(POST
-PRE)
0.24
0.20
-0.19
0.31
0.06
0.52
2.13
0.94
1.00
4.78
3.13
5.70
P-value
**
0.869
0.028
0.132
0.638
3.096
0.462
* N=54 for all means
** Ho: Change on VOC -
Change on AIR = 0
39
-------
TABLE 11
AUDITORY DIBIT SPAN
(ALL SUBJECTS)
TEST
FORWARD
( STANDARD )
BACKWARD
(STANDARD)
TOTAL
( STANDARD )
FORWARD
(WEIGHTED)
BACKWARD
(WEIGHTED)
TOTAL
(WEIGHTED)
TRT
AIR
VOC
AIR
voc
AIR
VOC
AIR
VOC
AIR
VOC
AIR
voc
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
7.80
8.11
7.86
8.03
7.17
7.17
7.14
7.38
14.97
15.27
15.00
15.41
59.95
62.47
60.62
61.58
50.95
50.20
52.52
54.47
110.91
115.67
113.15
119.05
S.E.
0.13
0.12
0.13
0.13
0.15
0.13
0.13
0.14
0.24
0.20
0.22
0.22
1.72
1.54
1.76
1.90
1.82
1.68
1.58
1.65
3.20
2.69
3.02
3.00
CHANGE
(POST
-PRE)
0.30
0.17
0.00
0.24
0.30
0.41
2.52
0.95
2.24
4.95
4.76
5.89
P-value
**
0.503
0.259
0.721
0.481
0.208
0.715
* N=66 for all means
*» Ho: Change on VOC -
Change on AIR = 0
40
-------
TABLE 12
'INGER TAPPING
CHANGE
NUMBER
TAPS
'REFERRE
HAND)
; NUMBER
ITAPS (NON-:
I PREFERRED
HAND)
NUMBER
: TAPE
: (ALTER-
: NAT ING)
TRT
, AIR
1
: voc
;
: AIR
!
: VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POET
MEAN*
116.6
113.3
114. 1
113.5
106.3
102.7
104.7
102.1
5 . E •
1.9
1 .9
1.8
1.9
1.7
1.4
1.5
1 e
(POST
-PRE )
-3.3
-0.6
-3.6
-2.6
P— vai ue
x*
0.213
0.396
AIR
VOC
PRE
POET
PRE
POST
101.0
99.9
102.4
98. 9
* . c
1.9
-1 .1
-3.5
* N=66 for all means
*» Ho: Change on VOC - Change on AIR = 0
TABLE 13
VISUAL DIGIT SPAN
TEST
LONGEST
FORWARD
SPAN
LENGTH
g=^^»^sg?~ •— -•
LONGEST
BACKWARD
SPAN
: LENGTH
TOTAL
TRT
AIR
VOC
B*^^B«v^e^*e
^V^^^E^Bfi
AIR
voc
AIR
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
7.68
8.20
7.97
8.06
7.53
7.62
7.59
7.83
15.21
15.62
S.E.
0.17
0.16
0.17
0.16
0.21
0.20
0.21
0.21
0.34
0.32
CHANGE
(POST
-PRE)
0.52
0.09
0.0"
• 0.24
0 . 61
P-value
XX
0.030
0.512
VOC
PRE
POS"
15.56
15.8*
E.34
0.41e
« N=do *or all means
** no: Change on VOC - Change on .'SIR = C
-------
TABLE 14
CONTINUOUS PERFORMANCE TASK
TEST
MEAN
REACTION
TIME
TOTAL
OMISSION
ERRORS
============
TOTAL
FALSE
POSITIVES
TRT
AIR
VOC
AIR
VOC
=======
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
EtEssrsass
PRE
POST
PRE
POST
MEAN*
376.3
380.4
375.5
383.7
0.26
0.48
0.26
0.35
==========
0.91
0.88
0.74
1.20
5 . E .
4.23
4.08
4.14
4.37
0.11
0.14
0.08
0.09
========
0.14
0.19
0.14
0.15
CHANGE
(POST
-PRE)
4.1
8.2
0.22
0.09
==========
-0.03
0.46
P- value
**
0.156
0.406
=========
0.061
* N=66 for all means
»> Ho: Change on VOC - Change on AIR = 0
TABLE 15
SYMBOL-DIGIT SUBSTITUTION
TEST
MEAN
NUMBER
CORRECT
as • SB • ' ' I sssssssa
MEAN
TIME
;
TRT
AIR
VOC
AIR
VOC
TIME
ESSSSSSS&
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
.028
.027
.034
.029
15.05
14.61
14.94
14.50
S.E.
'35^5 XS SS ^E ^5 S
.008
.009
.007
.009
0.27
0.27
0.24
0.25
CHANGE
(POST
-PRE)
-0.001
-0.005
-0.44
"
-0.44
P-value
**
0.812
0.954
* N=66 for all means
** Ho: Change on VOC - Change on AIR = 0
42
-------
10
TABLE 16
SERIAL DIGIT LEARNING
TEST
SCORE
SCORE
-»• NUMBER
OF
TRIALS
TRT
AIR
VOC
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
2.55
2.48
2.70
2.89
6.45
6.30
6.80
7.06
S.E.
0.2B
0.34
0.27
0.34
0.44
0.53
0.45
0.52
CHANGE
(POST
-PRE)
-0.07
0.19
-0.15
0.26
==========
P-value
X*
0.54B
0.539
* N=66 for all means
** Ho: Change on VOC - Change on AIR = 0
TABLE 17
PATTERN MEMORY
TEST
LATENCY
(ALL
TRIALS)
LATENCY
( CORRECT
TRIALS
ONLY)
NUMBER
CORRECT
TRT
AIR
VOC
AIR
VOC
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
5.21
4.29
5.08
4.31
4. 98
4.02
4. 82
4.08
23.3
22.7
23.2
22.7
S.E.
0.1B
0.14
0.16
0.14
0.17
0.13
0.15
0.14
0.20
0.23
0.17
0.24
CHANGE
(POST
-PRE)
-0.92
-0.77
-0.96
-0.74
-0.6
-0. 5
P-value
**
0.359
0.157
0.B56
* N=66 for all means
** Ho: Change on VOC
Change on AIR = 0
43
-------
11
TABLE 18
SWITCHING ATTENTION TEST
TEST
SIDE
CONDITION
DIRECTION
CONDITION
SWITCHING
CONDITION
SWITCHING-
SIDE
COMPAT .
SWITCHING-
SIDE
INCOMPAT.
SWITCHING
DIRECTION
COMPAT .
SWITCHING
DIRECTION
INCOMPAT.
TRT
.AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
319.5
315.7
320.3
321.1
432.2
428. 0
430.9
419.9
489.3
474.9
485.9
483.4
426.4
416.4
410.6
414.6
444.1
427.5
446.0
438.7
550.4
537.0
551.5
553.7
536.1
518.8
535.6
526.8
S.E.
6.2
7.1
6.6
6.7
6.9
8.4
7.7
7.7
13. B
13.8
16.6
15.4
14.7
14.6
14.5
15.0
15.4
17.0
22.1
18.8
14.0
13.6
17.6
16.9
13.8
13.2
15.6
14.4
CHANGE
(POST
-PRE)
-3.B
0.8
-4.2
-11.0
-14.4
-2.5
-10.0
4.0
-16.6
-7.3
-13.4
2.2
-17.3
-8.8
P-value
0.611
0.473
0.341
0.337
0.561
0.321
0.532
* N=66 for all means
** Ho: Change on VOC - Change on AIR = 0
-------
12
TABLE 19
MOOD SCALE
============
TEST
TENSION
SCORE
DEPRESSION
SCORE
ANGER
SCORE
FATIGUE
SCORE
CONFUSION
SCORE
TRT
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
AIR
VOC
TIME
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
PRE
POST
MEAN*
1.97
1.89
1.94
1.98
1.57
1.58
1.50
1.57
1.18
1.20
1.23
1.22
2.68
2.75
2.58
2.82
1.94
1.97
1.93
2.10
5 .E .
0.07
0.06
0.08
0.08
0.05
0.06
0.05
0.05
0.04
0.05
0.06
0.04
0.09
0.09
0.09
0.08
0.06
0.06
0.07
0.08
CHANGE
(POST
-PRE)
-0.08
0.04
0.01
0.07
0.02
-0.01
0.07
0.24
0.03
0.17
P-vai ue
0 . 1 53
0.197
0.722
0.045
0.026
* N=66 for all means
*» Ho: Change on VOC - Change on
AIR
0
-------
13
TABLE 20
ASSOCIATIVE LEARNING
TEST
TOTAL
CORRECT
RECALL
TRT
AIR
VOC
AIR
VOC
MEAN*
21.15
22.06
7.94
8.02
S.E.
0.62
0.55
0.17
0.21
DIFF-
ERENCE:
VOC-AIR
0.91
0.0B
V ^HB ^ ^ ^^ •«
P-value
X*
0.132
0.747
* N=66 for all means
«» Ho: VOC - AIR = 0
TABLE 21
PATTERN COMPARISON
TEST
NUMBER
TRIALS
CORRECT
LATENCY
(ALL
TRIALS)
LATENCY
( CORRECT
: ONLY )
TRT
AIR
VOC
AIR
VOC
AIR
VOC
MEAN*
24.80
24.85
3.07
3.07
3.07
3.07
S.E.
0.06
0.05
0.09
0.08
0.09
0.08
DIFF- ;
ERENCE: IP-value
VOC-AIR I »*
0.05 ! 0.581
t
0*00 : 0.968
0.00 : 0.969
* N=fa6 for all means
** Ho: VOC - AIR = 0
46
-------
14
TABLE 22
SIMPLE REACTION TIME
TEST
REACTION
TIME
(PREFERRED
HAND)
REACTION
TIME (NON-
PREFERRED
HAND)
TRT
AIR
VOC
AIR
VDC
MEANX
244.3
246.4
253.7
254.3
S . E .
3.38
3.41
3.91
3.54
: DIFF-
ERENCE:
VOC-AIR
2.1
p-value
x»
0.455
0.6 0.862
* N=66 for all means
*« Ho: VOC - AIR = 0
TABLE 23
GRAMMATICAL REASONING
TEST
MEAN
RESPONSE
TIME
SUM
OF
ERRORS
SCORE
TRT
AIR
VOC
AIR
VOC
AIR
VOC
MEAN*
3396
3375
6.7
6.2
: 54.5
: 54 . o
: DIFF-
S.E. 1ERENCE:
! VOC-AIR
96 I
101 : . -21
0.9 :
0.9 : -0.5
1 t ;
: l.o. 0.4
._______.
P-value
IX
0.737
0.4B2
: 0.694
* N=66 for all means
xx HO: VOC - AIR = 0
47
-------
15
TABLE 24
HORIZONTAL ADDITION
TEST
NUMBER
CORRECT
RESPONSE
TIME (ALL
TRIALS)
=======:===
RESPONSE
TIME (COR-
RECT ONLY)
TRT
AIR
voc
AIR
VOC
AIR
voc
EAN*
58.5
56.9
3.26
3.25
3.25
3.24
S.E.
0.18
0.11
0.09
0.0B
0.09
0.08
DIFF-
ERENCE:
VOC-AIR
0.4
-0.01
-0.01
P— va lue
**
0.031
0.766
0.806
* N=66 for all means
** Ho: VOC - AIR = 0
-------
16
TABLE 25
AUDITORY DIGIT SPAN
(ALL SUBJECTS)
TEST
FORWARD
( STANDARD )
BACKWARD
( STANDARD )
TOTAL
( STANDARD )
FORWARD
(WEIGHTED)
BACKWARD
(WEIGHTED)
TOTAL
(WEIGHTED)
ORDER
1
2
3
4
1
1
^
3
4
i
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
MEAN*
7.74
8.06
7.92
8.08
7.03
7.11
7.27
7.44
14.77
15.17
15.20
15.52
58.76
61.64
61.82
62.41
50.73
54.32
52.74
56.35
109.50
115.95
114.56
118.76
S.E. i CONTRAST
i
0.13 !
0.13 ILINEAR
0.13 ! QUADRAT 1C
0.12 JCUBIC
0.15 !
0.14 iLINEAR
0.13 ! QUADRAT 1C
0.13 ICUBIC
0.24 !
0.22 ILINEAR
0.21 ! QUADRAT 1C
0.19 ICUBIC
1.68 !
1.82 ILINEAR
1.77 I QUADRAT 1C
1.64 1 CUBIC
1.80 !
1.64 ILINEAR
1.60 1 QUADRAT 1C
1.72 ICUBIC
3.22 ;
2.94 ILINEAR
2.96 [QUADRATIC
2.75 ICUBIC
P-value
0.042
0.379
0.081
0.002
0.644
0.836
<.001
0.788
0.301
0.022
0.291
0.521
0.002
0.994
0.036
<.001
0.465
0.053
N=66 for all means
49
-------
17
TABLE 26
VISUAL DIGIT SPAN
TEST
LONGEST
FORWARD
SPAN
LENGTH
LONGEST
BACKWARD
SPAN
LENGTH
TOTAL
1 ORDER
: 1
> 4U
• 3
: 4
: 1
: 2.
• 3
: 4
1 <^
: 3
: 4
MEAN*
7.55
8.00
B.li
6.26
7.26
7.71
7.B6
7.74
i4.ee
15.71
15.97
16.00
5 .£ .
0.17
0.15
0.17
0.17
0.21
0.20
0.20
0.21
0.34
0.31
0.32
0.33
; CONTRAST
! LINEAR
! QUADRAT I C
! CUBIC
J
I LI NEAR
! QUADRAT 1C
i CUBIC
I LINEAR
! QUADRAT 1C
.'CUBIC
P-vaiue
<.001
0.129
0.377
0.005
0.024
0.957
<.001
0.007
0.556
N=66 for all means
TABLE 27
PATTERN MEMORY
TEST
LATENCY
(ALL
TRIALS)
LATENCY
( CORRECT
TRIALS
ONLY)
NUMBER
CORRECT
ORDER
1
2
3
4
1
2
3
4
i
_j"
T
', **
MEAN*
5.73
4. SB
4.56
4.02
5.47
4.25
4.33
3.84
23.1
22.2
23.4
23.1
S.E.
0.17
0.16
0.14
0.12
0.16
0.14
0.14
0.12
0.20
0.23
0.17
: 0.23
CONTRAST
LINEAR
QUADRATIC
CUBIC
LINEAR
QUADRATIC
CUBIC
LINEAR
QUADRAT I C
CUBIC
P— value :
<.00l :
<.001 :
<.00l :
<.001 :
<.001 :
<.00i ;
0.10B •
0.053
X .001
for ail means
50
-------
TABLE 28. SERIAL ORDER/PRACTICE EFFECTS (p-Values)
A. NES MAIN TESTS
1.
2.
3.
4.
5.
7.
8.
TEST
Finger
Tapping
Visual
Digit
Span
Continuous
Performance
Test
Symbol Digit
Substitution
Serial Digit
Learning
Pattern
Memory
Switching
Attention
Mood
Scales
MEASURE
Preferred Hand
Nonpreferred
Alternating
Fvd Span
Bvd Span
Total
React ion- time
Omissions
False Positives
Correct
Response Time
Score
Score + tt trials
Latency (all trls)
Latency (corr. trls)
Side
Direction
Switching
Sw. Side Compat.
Sw. Side Incomp.
Sw. Dir. Compat.
Sw. Dir. Incomp.
Tension
Depression
Anger
Fatigue
Confusion
LINEAR
NS
NS
NS
<.001
0.005
<.001
0.024
0.087
0.012
0.090
<.001
NS
NS
<.001
<.001
(0.085)
NS
<.001
<.001
0.004
<-001
<.001
0.002
NS
NS
NS
NS
QUADRATIC
NS
NS
NS
NS
0.024
0.007
NS
NS
NS
NS
NS
(0.102)
NS
<.001
<.001
NS
NS
NS
(0.065)
NS
NS
(0.072)
(0.089)
NS
NS
NS
NS
CUBIC
0.014
<.001
0.009
NS
NS
NS
0.003
NS
NS
NS
NS
NS
NS
<.001
<-001
NS
NS
NS
(0.097)
NS
NS
NS
NS
(0.062)
NS
0.004
0.02
51
-------
Table 28 (cont.)
B. AUDITORY DIGIT SPAN TEST
Forward (Std)
Backward (Std)
Total (Std)
Forward (Wtd)
Backward (Wtd)
Total (Wtd)
0.042
0.002
<.001
0.022
0.002
<.001
NS
NS
NS
NS
NS
NS
(0.081)
NS
NS
NS
0.036
(0.053)
C. NES SECONDARY TESTS DIFFERENCE (ORDER 2 - ORDER 1)
1. Associate # Correct 0.022
Learning Recall NS
2. Pattern # Correct (0.095)
Comparison Latency (All) <.001
Latency (Corr. trls) <.001
3. Simple Preferred Hand NS
Reaction-Time Nonpreferred NS
4. Syntactic Mean Response Time <.001
Reasoning f Errors <.001
Score 0.001
5. Horizontal # Correct NS
Arithmetic Response Time (All) NS
Response Time (Corr.) NS
52
-------
FIGURE 1
SCHEMATIC OF TEST PROTOCOL
COMFORT RATING SCALE
SYMPTOM QUESTIONNAIRE
NES MAIN BATTERY
AUDITORY DIGIT SPAN
NES FILLER TESTS
PRE-EXPOSURE
BUILDUP
CONSTANT EXPOSURE
I I I I I I I I I I
I
I I
0
60 120 180
Time Expired (Minutes)
240
-------
FIGURE 2
COMFORT RATING SCALE RESULTS
1.4 r
Ln
AIR
VOC
60
90 120 150
TIME (minutes)
180
210
240
-------
FIGURE 3
HKADACHE, ODOR & AIR QUALITY
Ln
24
20
^D
ID
II 16
12
8
O
o
CO
<
LJ
0 '—
AIR VOC
HEADACHE
AIR VOC
ODOR
AIR VOC
AIR QUALITY
-------
FIGURE 4
ODOR, EYE & THROAT IRRITATION
cr»
AIR VOC
THROAT
VOC
EYE
AIR VOC
ODOR
-------
FIGURE 5
CONTROL QUESTION RESPONSES
Ln
0
AIR VOC
NOISE
AIR VOC
LIGHT
AIR VOC
TEMPERATURE
-------
FIGURE 6
YES-NO QUESTION RESPONSES
"DO YOU FEEL NOW THAT STAYING
to IN THE CHAMBER IS COMFORTABLE?"
| ,i . . y
I ,f !"^~
&J t
£UJ-
O Mr-
O r
O 014.
§ '
t &3»-
O
g -
fe oL
K PRE MID POST
EXPOSURE PERIOD
"WOULD YOU BE SATISFIED WITH THIS
« AIR QUALITY IN YOUR OWN HOME?"
«f -^ _ . JS.
f -^— . • ,
a. : ~ _
i PRE MID POST
EXPOSURE PERIOD
"WOULD YOU VENTILATE YOUR HOME MORE?"
to
W 1, . «
1 !
< OJf
r
*.__ I VDC
PRE MID POST
EXPOSURE^gERIOD
-------
FIGURE 7
AUDITORY DIGIT SPAN
Ln
vo
AIR
VOC
PRE POST
FORWARD
PRE POST
BACKWARD
PRE POST
TOTAL
-------
FIGURE 8
VISUAL DIGIT SPAN
PRE POST
FORWARD
PRE POST
BACKWARD
PRE POST
TOTAL
-------
FIGURE 9
CONTINUOUS PERFORMANCE TASK
400
MEAN REACTION TIME
voc
I o.u
TOTAL OMISSION ERRORS
VOC
TOTAL FALSE POSITIVES
-------
FIGURE 10
SWITCHING ATTENTION TEST
UJ
600
550
500
450
UJ
en
z
CL 400
cn
UJ
CE
350
300
-G
-A
PRE
POST
L_
PRE
-•
. J_-_
POST
600
550
500
u
Q)
UJ
450 £
UJ
en
400 a.
in
UJ
tr
350
300
SIDE
DIRECTION
SNITCHING
SHITCHING-SIDE
COMPATIBLE
SHITCHING-SIDE
INCOMPATIBLE
SWITCHING-DIRECTION
COMPATIBLE
SWITCHING-DIRECTION
INCOMPATIBLE
AIR
VOC
-------
FIGURE 11
MOOD SCALE DIFFERENCE SCORES
ON
U>
LiJ
O
^
X
O
bJ
OH
Q-
Z>
H-
CO
O
Q.
0.30
0.25
0.20
0.15
0.10
0.05
0.00
-0.05
r\ 1 n
-
-
-
-
.
AIR
VOC
TENSION DEPRESSION ANGER
(p=.153) (p=-197) (p=.722)
EATIGUE CONEUSION
(p=.045) (p=.026)
-------
12
16.5
16
15.5
TOTAL
FORWARD
BACKWARD
15
(O
CO
FIGURE 12
SERIAL ORDER EFFECTS
AUDITORY DIGIT SPAN
b:
o
Q
7.5
6.5
-j-
_i_
2 3
ORDER
-------
13
16.5
16 -
15.5
15
CO
co
2 H.5X
'*^
fc
O
o
7.5
6,5
FIGURE 13
SERIAL ORDER EFFECTS
VISUAL DIGIT SPAN
ORDER
65
-------
UJ
UJ
2.06
2.04
2.02
2.00
CL 1.98
(/I
LU
^ 1.96
1.94
1.92
1.90
CONFUSION
FIGURE 14
SERIAL ORDER EFFECTS
CONFUSION SCALE
CONTRAST
p-VALUE
LINEAR
QUADRATIC
CUBIC
0.639
0.761
0.027
L-
3
4
SERIAL ORDER
-------
FIGURE 15
SERIAL ORDER EFFECTS
PATTERN MEMORY
6.00
Ld
O
CL
CO
5.50 -
5.00
4.50
4.00
3.50
CONTRAST
LINEAR
QUADRATIC
CUBIC
p-VALUE
< 0.001
< 0.001
< 0.001
SERIAL ORDER
-------
APPENDIX A
DESIGN AND PERFORMANCE OF A SYSTEM TO CONTROL VOC CONCENTRATIONS
WITHIN AN ENVIRONMENTAL LABORATORY USED FOR HUMAN EXPOSURE
Jon Berntsen
C-E Environmental, Inc.
Chapel Hill, NC 27514
INTRODUCTION
The Environmental Protection Agency's Clinical Environmental Laboratory
(CEL) on the campus of the University of North Carolina at Chapel Hill, is used
to provide human exposure data which can withstand vigorous technical and legal
3 3
challenge. This facility consists of of two 84 m (3000 ft ) environmental
laboratories which are used to expose human volunteer subjects to constant or
varying concentrations of 03> N02, S02, CO and water soluble aerosols. The
pollutant delivery system and the aerosol delivery system are under computer
control. The computers are used to monitor the concentration of the pollutants,
control the pollutant delivery systems, validate the data as it is taken, alarm
the operator of unsafe conditions and record the data for later use. These
chambers have been in continuous use since 1977 exposing subjects to pollutants
and measuring responses with a computerized Physiological Data Acquisition
System.
The EPA's interest in indoor air pollution and the effects of these
pollutants on humans has expanded the requirements of the facility. To enable
the exposure of humans to carefully controlled concentrations of Volatile
Organic Compounds (VOCs), the pollutant delivery system was modified to allow
injection and monitoring of the VOC compounds in one of the chambers. The
following sections discuss the design and operation of the VOC generation and
monitoring system.
68
-------
SYSTEM DESIGN
Existing System Constraints
The VOC monitoring and delivery system was designed to be compatible with
the existing pollutant control system and chambers. In contrast to other
chambers, these chambers recirculate air with a flow of 3.8 mVsec (8000
ftVmin) through the main loop. Fresh (make-up) air is provided at a rate of
0.56 m3/sec (1200 ftVmin) to replace the same amount of recirculating air
exhausted through a purge damper, leaks or the chamber door when opened. The
total volume of the system, chamber and associated duct work, has been
calculated to be 175 m3 (6180 ft3) with a time constant of about 5 minutes for
exchanging 67K of the chamber air. For the purpose of controlling a pollutant
gas concentration, the chambers behave as a one pass chamber with a total volume
of 175 m3 and a flow rate of 0.56 m3/sec except for the amount of pollutant
needed to bring the concentration to the desired value.
Due to space constraints, the VOC generator was connected to the aerosol
generator. The aerosol generator uses .38 mVsec (800 ftVmin) for sweep air to
dry and remove the aerosol particles from the generator. This air also moves
aerosols from the generator into the main air duct where the aerosols are mixed
with chamber air. This sweep air and the aerosol delivery ducts are used by the
VOC generator.
The chamber design is shown in Figure 1. This figure shows recirculated
air through the main chamber air loop, the fan and bypass to force air through
the VOC generator, the location of the makeup air injection point and the
location of the exhaust air outlet.
VOC Generator Design Considerations
The VOC generator was designed to meet a variety of stringent requirements.
69
-------
The generator had to deliver precisely controlled amounts of a vaporized mixture
of highly flammable VOC's to the chamber. The amount of vapor had to be
automatically controlled by an existing computer and associated system. Since
the mixture contains compounds with boiling points between 69 °C and 181 °C, the
generator must vaporize all constituents without changing the relative
concentrations of any components. The generator must also vaporize the
components with a minimum of chemical reactions so that only the components in
the mixture are in the chamber without generating impurities. Since VOCs are
flammable, the generator must minimize the danger of explosion or fire.
The other major factor is the chamber design. Since the chamber is
multipass with large internal volume and a relatively low air exchange rate, the
generator uses an on-off type of control instead of proportional control. Since
the design of the existing system has a 30 second control cycle, the VOC
generator also incorporates this 30 second controlling period. Thus the VOC
concentration can be controlled at a steady concentration by turning the
solenoid valve on at the start of the 30 second period and off sometime in the
middle of the period. Since the computer reports 2 minute averages, the effects
of this on and off control are further minimized.
VOC Generator Design
The VOC generator (Figure 2) is a modification of a design by Molhave,
which was scaled up and changed to meet increased flow requirements. The
mixture of hydrocarbons is placed in an Erlemeyer flask. The top of the flask
has been modified to accept a screw type fitting. This fitting allows the
connection of a N» pressure line and a mixture delivery line. A Fairchild Model
81 precision pressure regulator is used to control the pressure over the liquid
mixture. This pressure was held at 3 psig for these tests but can be varied to
control the liquid feed rate. A dip tube which almost touches the bottom of the
flask removes liquid from the flask. The flask is placed on a magnetic stirrer
and agitated throughout the run to keep the mixture homogeneous.
The mixture delivery tube was connected through a T-connection to a Nupro
70
-------
SS-SS2-KZ fine metering valve. The metering valve vas fitted with Kalrez and
Teflon seals. The other end of the T-connection was connected to a tube several
feet long. This tube was uncapped after the mixture was pressurized to allow
the mixture to fill the delivery tube. This tube was securely capped after the
delivery tube was filled. The outlet of the metering valve was connected to the
inlet of a Sierra Model 840VO-020 solenoid valve. The valve has a 0.02 inch
orifice with Kalrez seals. The outlet of the solenoid valve was connected to
the (evaporator) through a sprayer. This sprayer was constructed of two
concentric tubes. The inner capillary tube allowed the mixture to enter the
evaporation container. About 0.05 liters/sec (0.1 CFM) of nitrogen was passed
through the outer tube. The flow of nitrogen helped spray the mixture on the
bottom of the evaporation container and also forced the vapor out of the
container.
The evaporator is made of glass and during operation is placed in a large
brass block (15 cm in diameter and 15 cm high). The evaporator is about 8 cm in
diameter and 10 cm high. A fritted glass disk separates the evaporator into two.
sections. The fritted glass keeps any liquid from moving from the bottom
section to the top section. An arm on one side of the bottom section allows
access for the sprayer. The angle of this arm allows the VOC mixture to be
sprayed on the bottom of the container. Another arm, sealed at the bottom and
almost touching the bottom of the container, contains a thermocouple (Omega
Model C03-E) probe. The large brass block contains six 250 V cartridge heaters
(McMaster-Carr P/N 3618K615) which were controlled by an Omega Model CN5001E1
proportional controller. For this mixture of compounds, the block was heated to
250 °C. After the VOC mixture has evaporated, the nitrogen forces it through
the fritted glass into the top section. Here the vapor is diluted with another
0.5 liters/sec (1 CFM) of nitrogen. The flow of nitrogen was measured and
controlled with an Omega Model FM064-63ST flowmeter.
The concentrated VOC vapor is then diluted with about .02 m3/s (40 CFM) of
air from the chamber. This air cools the mixture and dilutes the VOC
concentration below the lower explosive limit of the compounds. A vertical 4
inch glass pipe 5 feet long allows any liquid to condense and run down the glass
71
-------
into a trap. After passing through this condenser, the VOC mixture is then
diluted with about .36 m3/s (760 CFM) of the chamber air. This air is then
center injected into the main air duct where it is further diluted and mixed
with main recirculated air and passed through the chamber.
The use of nitrogen to dilute the VOC vapor and to force it out of the
evaporator reduces the explosion and fire risk, prevents the oxidation of VOCs
and minimizes the time for reactions caused by the heat required to evaporate
the compounds.
VOC Monitoring and Control System
The VOC concentration in the chambers is monitored by two GowMac 23-500R
total hydrocarbon analyzers. These analyzers use a flame ionization detector to
monitor the hydrocarbon concentration in the chamber.
The analyzers sample from a 3 inch glass sample manifold used by all of the
gaseous pollutant analyzers in the facility. This manifold draws a sample
through an inverted glass funnel. The opening of this funnel is 3 meters from
the floor and in the center of the chamber ceiling. The manifold exits the
chamber near the ceiling of the chamber and passes over the back top of the
instrument racks containing the gas analyzers. Chamber air is pushed through
the sample manifold by the 0.1 inch of water positive pressure in the chamber
and two sample manifold exhaust blowers. The flow though the sample manifold is
0.2 m3/s.
Each analyzer has a tap consisting of a I/A inch teflon line inserted into
the glass pipe so the opening is in the middle of the glass pipe. For the
hydrocarbon analyzers a Mace Model 800-2334-2-0 solenoid valve is used to allow
the analyzers to sample from the sample manifold or a calibration manifold. The
sample is pushed into the analyzer with a Thomas Model 107CA14TFEL pump with
teflon coated piston, valves and cylinder.
A PDF 11/35 computer measures the output of each analyzer once per second.
These samples are used to generate 2 minute averages which are reported to
72
-------
investigators as the chamber concentration. For control purposes, the computer
samples the concentration in the chamber every 30 seconds and determines the
percent on time for the VOC valve accordingly. This value is converted to a
voltage and sent to the VOC control card. Depending on the state of the system,
this card uses a voltage from a potentiometer in the manual mode or the voltage
from the computer in the automatic mode, to control the duty cycle of the valve.
This allows either automatic or manual control of the generator. In the
automatic mode, another signal from the computer forces the card to synchronize
with the computer.
System Calibration The VOC concentration in the Molhave study was 25 mg/m3
toluene equivalent. Molhave calibrated FID analyzers with toluene and
controlled the VOC concentration with instrument output at the same level
obtained with the toluene calibration standard. The 25 mg/m3 was converted to 7
ppm toluene equivalent for these studies. The VOC analyzers were calibrated
with toluene and the concentrations reported in terms of toluene. A standard of
toluene diluted with nitrogen was diluted to the various concentrations with
ultra high purity zero air with the dilution calibration system which has been
used for 12 years. This system is automatically controlled by the computer
which after the calibration has been specified by the operator, automatically
calibrates the analyzers. The standard calibration consisted of three span
points (5, 7 and 9 ppm) and a zero point. The zero was taken first, then the
highest point was taken twice with the first one thrown out.
System Operation The following procedure was used to run a session of the
protocol.
1. The instruments were calibrated the morning of the protocol using the
method given above.
2. A new solution was mixed up. Several safeguards were followed to
assure that the mixture was made correctly. The operator used a check
sheet to remove a bottle of each of the chemicals from a storage
cabinet. Each chemical was pipetted using its own premarked pipette
73
-------
(both with chemical name and the volumne) from the bottle to the flask.
As this was done, the chemical was checked off on the check sheet. The
bottle was then placed back in storage.
3. After the mixture was prepared, it vas connected to the system.
A. The generation system vas turned on and heated to the proper
temperature during instrument calibration and chemical mixing.
5. The results of the calibration were then checked and the new
calibration constants loaded into the computer.
6. The computer vas instructed to monitor the VOC concentration in the
chamber.
7. Just before VOC generation was initiated, the ambient chamber VOC
concentration was measured and the exposure level (7 ppm) was
incremented by the ambient VOC level. For example, if the ambient VOC
level was 0.37 ppm, the exposure level vas set at 7.37 ppm.
8. Seventy-five minutes after commencement of testing, the operator
initiated exposure by activating the automatic control of the VOC
concentration in the chamber.
SYSTEM PERFORMANCE
This system has been used in two different studies for a total of 110 hours
of VOC control during a protocol. The two studies used a mixture very similar
to the tventy-tvo compounds Molhave used. Figure 3 shovs a plot of the VOC
concentration in the chamber over the period of the protocol. For this study,
the subjects were in the chamber while the VOC control was established and the
system equilibrated to the required valve.
-------
MAKEUP AIR
COMPUTER
SVI
CTRL
FID
ANALYZER
PURGE
r
MIXER
A
CHAMBER
t
voc
GENERATOR
FAN
Figure 1
Overall Chamber Design
-------
Spil
NITROGEN
o\
VOC
MIXTURE
Figure 2
VOC Generator Design
METERING SOLENOID
VALVE VALVE
NITROGEN 2 l/m
SPRAYER
GLASS
EVAPORATOR
BRASS BLOCK
DAMPER
HEATERS
-------
H« IN OHMM* ft
I'll/il IHUlll
I .
4 .
3 -
i .
I .
I . —
lit HI 9M 911 IN (41
I I I I I I I
t»M Itl* 1IM »•<• UN lilt I1M
1141 II
»CtU»l
MIIWD
TIM
Figure 3
Typical Plot of VOC Control
-------
REFERENCES
Molhave L, Bach B and Pedersen 0. Klimakammerundersogelse af geneforekomster
hos personer, der udsaettes for organiske gasser og dampe fra byggematerialer
[Human reactions to volatile organis gases]. Project Report ROO-52(22411),
Institute of Hygiene, Universitetsparken, Aarhus, Denmark, 1985.
Strong AA; "Description of the CLEANS Human Exposure System". U.S. EPA.
EPA-600/1-78-064, 1978.
Crider VL, Landis DA, Weaver FH. In "Generation of Aerosolos and Facilities for
Exposure Experiments". Villeke K, (ed.) Ann Arbor, Michigan, 1980:493-515.
Glover DE, Berntsen JH, Crider, VL, Strong AA. "Design and Performance of a
System to Control Concentrations of Common Gaseous Air Pollutants within
Environmental Laboratories Used for Human Exposure". J. Environ. Sci.
Health, A16(5), 1981, pp501-522.
78
-------
APPENDIX B
AUDITORY DIGIT SPAN TEST DESCRIPTION AND INSTRUCTIONS
H. Kenneth Hudnell
Neurotoxicology Division, Health Effects Research Laboratory
U.S. Environmental Protection Agency, RTF, NC
The following tape-recorded instructions were given to subjects.
"Attention Subjects: you will now be given another version of the digit span
test. Unlike the Digit Span Test you took earlier on the computer, this test
will be administered by a cassette recording played over your speakers. Please
write your responses on the answer sheet attached to the clipboard. Please pick
up the clipboard at this time and assume a comfortable writing position.
The answer sheet should say "auditory digit span - pretest" at the top.
The test will have two parts, Digit Forward and Digit Backward. Please find the
section labeled Digit Forward. As you can see on the answer sheet, blank spaces
are provided for you to write down each of the digits in the string. The first
string will contain 3 digits. Please listen carefully as the digits are called
out. Be sure not to repeat the digits verbally, but you may rehearse them in
your mind. After the string has been called out, there will be a pause during
which you are to write down the digits as you remember them and in the same
order as they were presented. Please do not begin writing until all digits in
the string have been called out. After the pause, a different string of 3
digits will be presented and then followed by a pause during which you can write
your answers. This process will be repeated until 2 trials with strings of up
to 9 digits have been presented.
After the digit forward test has been completed there will be a 1 min.
pause. Then, the digit backwards test will begin. This test is exactly like
the digit forward test except that you are to write down the digits in the
reverse order of their presentation. That is, the last digit you hear should be
79
-------
the first digit you write on the answer line. The digit backwards test will
begin with a 2 digit string and end with an 8 digit string. Please remember, do
not repeat the digits out loud and do not begin to write your answers until the
pause begins after all the digits in the string have been called out. Please
prepare to begin the digit forward test. Write the digits in the same order as
they are presented. The test will begin now.
Trial 1 of 3 digits (say 1-4-7 and pause 10 sec.)
Trial 2 of 3 digits (say 3-7-2 and pause 10 sec.)
" 1 " 4 " ( » 2-9-5-6 and " 12 sec.)
" 2 " 4 " < " 2-6-5-4 and " 12 " )
" 1 " 5 " ( " 2-8-7-4-5 and " 14 " )
" 2 " 5 " ( " 2-9-6-7-5 and " 14 ")
" 1 " 6 " ( n 1-6-9-2-4-1 and M 16 " )
" 2 " 6 " ( " 4-7-6-1-2-8 and " 16 " )
M 1 " 7 " ( " 5-7-1-6-2-4-9 and M 18 " )
" 2 " 7 " ( " 1-5-2-8-9-4-6 and " 18 " )
" 1 " 8 " ( " 7-3-9-8-4-5-2-7 and " 20 " )
" 2 " 8 " < " 8-7-4-2-5-4-3-1 and " 20 " )
•" 1 " 9 " ( " 3-7-4-9-7-3-8-4-2 and " 22 " )
" 2 " 9 " ( " 1-3-2-5-2-7-8-4-8 and " 22 ")
The digit forward test is now completed. We will pause for 1 min. before
beginning the digit backwards test, (pause 1 min.) Please prepare to begin the
digit backwards test. Please write the digits in the order opposite that in
which they are presented. The test will begin now".
Trial 1 of 2 digits (say 3-5 and pause 10 sec.)
Trial 2 of 2 digits (say 4-2 and pause 10 sec.)
" 1 " 3 " ( " 2-9-2 and " 12 " )
" 2 " 3 " ( " 3-2-5 and " 12 " )
" 1 " 4 " ( " 2-7-9-4 and " 14 " )
" 2 " 4 " ( " 2-8-9-4 and " 14 " )
" 1 " 5 " ( " 3-6-7-5-1 and " 16 " )
" 2 " 5 " ( " 2-6-9-7-3 and " 16 " )
M 1 " 6 " ( " 1-4-8-3-9-3 and " 18 " )
" 2 " 6 " ( " 7-3-6-2-4-1 and " 18 " )
" 1 " 7 " . ( " 4-1-6-7-9-5-1 and " 20 " )
" 2 " 7 " ( " 2-9-6-7-2-1-4 and " 20 " )
" 1 " 8 " ( " 1-4-2-6-7-5-6-9 and " 22 " )
M 2 " 8 " ( " 1-6-1-3-9-8-3-7 and " 22 " )
Similar instructions were given for the post test. Digits presented in the post
test were:
80
-------
FORWARD
2-8-6
3-2-6
1-5-7-3
9-3-5-1
6-8-1-9-2
8-2-5-7-3
2-8-6-1-3-7
2-4-8-7-5-1
9-5-1-3-1-7-2
6-3-5-7-2-8-5
1_5_9_5_7_3_9_2
1-6-9-3-4-8-6-2
2-8-1-5-2-6-1-7-3
4-1-4-8-2-6-9-4-1
Digit Sequences used on day 2 vere as follows:
FORWARD PRE-TEST
6-1-3
3-5-1-7
8-9-4-2
5-9-8-1-6
2-8-9-4-6
3-1-4-2-4-7
3-7-9-2-6-1
2-9-3-7-9-4-6
8-5-1-7-2-4-3
7-1-4-6-8-3-2-9
9-4-7-3-2-9-5-6
1-6-2-5-1-8-2-7-3
7-3-9-3-8-4-5-7-2
BACKWARD
TT~
2-8
5-2-7
1-3-8
3-7-9-5
6-1-4-7
2-8-9-5-6
2-9-5-2-1
1-3-4-8-2-1
5-9-3-8-7-2
4-5-7-4-1-6-7
3_9_5_9_1_5_7
1-7-4-3-2-9-1-6
9-4-8-2-7-1-3-2
BACKWARD PRE-TEST
T-3
2-6
1-4-2
1-8-7
5-2-4-1
2-4-6-5
3-1-5-4-6
6-3-5-1-7
2-9-1-3-8-4
3-6-7-5-2-8
7-3-7-1-8-1-4
6-1-3-7-9-8-2
2-9-4-7-5-1-4-9
4-2-4-1-7-3-6-8
POST TEST-FORWARD
7-4-1
2-8-9
1-6-8-7
5-2-7-1
2-7-3-1-4
l_5_8-3-9
9-3-4-7-6-1
4-6-2-8-1-3
3-7-8-9-1-3-5
3-2-4-5-7-9-6
2-7-8-6-9-3-4-8
9-4-3-5-8-2-1-5
8-3-7-4-6-9-2-3-1
5-2-3-1-4-9-2-6-8
POST TEST-BACKWARD
539
3-1
2-9-5
2-6-9
1-4-5-8
3-2-8-6
7-3-2-4-1
5-6-1-4-3
7-3-5-1-2-6
8-5-2-7-3-1
3-2-3-5-4-2-1
4-3-6-5-7-8-2
1-5-3-9-4-8-2-6
1-6-9-3-8-2-7-1
81
-------
Digit Span Answer Sheet
Subject ID Number :
Date :
Session # (Circle one): 147
Day (Circle one) : 0 1 2
A. Digit Forward
3. (1)
(2)
4. (1)
(2)
5. (1)
(2)
6. (1)
(2)
7. (1)
(2)
8. (1)
(2)
9- (1)
(2)
B. Digit Backward
2. (1)
(2)
3. (1)
(2)
4. (1)
(2)
5. (1)
(2)
6. (1)
(2)
7. (1)
(2)
8- (1)
(2)
82
-------
INSTRUCTIONS FOR SCORING AUDITORY DIGIT SPAN DATA
ADS data will be scored in two ways, the standard method (std) and the weigh
ted
method (wtd).
Procedure.
Both Methods. Start with the 3 digit forward string, trial 1. Use the
answer key to see if all 3 digits are correctly recorded on the answer sheet.
If so, put a checkmark at the right end of the string. If no, put an X at the
right end. Repeat this process for each string in the digit forward test.
Then score the digit backward test.
Std Method. At the bottom of the answer sheet under the forward responses,
write the number of the string that corresponds to the longest string in which
at least 1 of the 2 trials was answered correctly for that string and all
previous strings. For example, if at least 1 trial is correctly answered on
strings 3, 4, 5 and 6 but neither trial is answered correctly on the string of
7 digit, the score is "std = 6". Repeat this procedure for the digit backwards
test. At the center, record the total score. For example, if the forward
score is 6 and the backwards score is 5, write "Tstd = 11".
Wtd Method. For the digit forward test, multiply each string number (3-9)
by the number of trials answered correctly (0,1 or 2). Add all products
together. For example, if the scores are:
string # trials correct the products are
3 2 3 x 2 = 6
4 2 4x2 = 8
5 2 5 x 2 = 10
6 1 6x1 = 6
7 0 7x0 = 0
8 1 8x1=8
9 0 9x0 = 0
The product sum = 38. Write "Wtd = 31" under the forward responses. Repeat
this procedure for the backwards test. Record the total Wtd score at the
bottom of the answer sheet in the center. For example, if the forward score is
38 and the backwards score is 26, write "Twtd = 64".
Note. For the backwards test, remember that the answer must be in the
reverse order of presentation to be correct.
83
-------
APPENDIX C
NEUROTOXICITY TESTING OF VOC MIXTURES IN HUMANS
In order to characterize the neurobehavioral effects of VOC exposure, a
microprocessor-based test battery was used. The neurobehavioral evaluation
system (NES) developed by Baker et al (1985) for neurotoxicity field testing
contains a menu-driven battery of tests selected on the basis of proven
sensitivity to environmental toxicity and WHO recommendations (Johnson et al.,
1987). A variety of sensorimotor and cognitive functions including reaction
time, motor coordination, attention, memory and spatial integration were
assessed. The NES battery has been used previously in the study of nitrous
dioxide (Greenberg et al, 1985), styrene, alcohol, is currently being used by
CDC in the AGENT ORANGE project, and in NHANES III (National Health and
Nutrition Survey). Specific tests to be used in the VOC study included finger
tapping, continuous performance, digit span, symbol-digit substitution, visual
memory and switching attention. Individual tests are described below.
NES TESTS USED IN VOC STUDY
(1) Finger tapping (mean no.) 3 min.
(2) Digit span (no. correct fwd, bvd) 7 min.
(3) Continuous Performance Test (msec, errors) 7 min.
(A) Symbol-digit substitution (sec/digit, errors) 5 min.
(5) Serial digit learning (no. correct, no. trials) 7 min.
(6) Pattern memory (no. correct, msec) 5 min.
(7) Switching attention ((msec/no, correct) 7 min.
(8) Mood scales (5 scale avs.) 5 min.
84
-------
Total time: 46 min.
Descriptions of Individual Tests
(1) Finger-Tapping. This test evaluates motor response speed. The
subject is instructed to press a button as many times as possible within a
1-second interval. In the first trial, the subject presses one button with the
left hand, The second trial tests the right hand, and in the third trial, the
subject alternately taps two buttons with the preferred hand. This test
provided a straightforward measure of VOC effect on motor response. It was a
negative control measure since VOCs were not expected to affect motor response
per se.
(2) Digit Span. The auditory version of this widely used clinical test
has been adapted from the Vechsler Adult Intelligence Scale (Wechsler, 1955).
Subjects must enter into the computer progressively longer series of digits
which are presented visually at a rate of one per sec by the computer. New
digit sequences are created at each span length. After incorrectly responding
to two trials at a span length, the task changes such that the digits must be
entered in reverse order. This test has been used as a measure of short-term
memory and attention in studies of solvent and lead toxicity (Hanninen, 1979).
Molhave et al (1986) reported that digit span performance was impaired by
exposure to the complex VOC mixture, a finding that the present study attempted
to replicate.
(3) Continuous Performance Test (CPT). This test measures sustained
visual attention by having the subject press a button upon seeing a large
letter "S" when it is projected on a video display. Letters flash briefly (for
about 50 msec) on the screen at a rate of one per second for five minutes.
Recording and storage of individual response latencies permits the computation
mean (and standard deviation) of reaction times. Omission and commission
errors are also recorded. Molhave et al (1986) failed to observe any effect of
VOC exposure on "graphic" CPT performance, but the present test is very
85
-------
different and has proven to be sensitive to toxicant exposure in previous
studies (Baker et al, 1988; Erstin et al, 1988; Mahoney et al, 1988).
(4) Symbol-Digit (SD) Substitution Test. This test is similar to the
Digit-Symbol Substitution Test from the WAIS. SD evaluates speed and coding
ability and has been found useful in previous studies of lead, carbon disulfide
and solvent mixtures (Anger and Johnson, 1985). In this adaptation, nine
symbols and digits are paired at the top of the screen and the subject has to
press the digit keys corresponding to a re-ordered test set of the nine
symbols. The time required to complete each SD set and the number of digits
incorrectly matched are recorded. Five sets of nine SD pairs are presented in
succession. The pairing of symbols vith digits is varied between sets to avoid
learning.
(5) Serial Digit Learning (SDL). This is a test of short-term memory and
attention in which subjects are presented a series of 10 digits on the display
screen at a rate of one per second. Upon completion they are asked to enter
that series into the computer. The series is presented until two successive
trials are performed correctly, or at most 8 times. The number of correct
digits entered on each trial is recorded. The series is changed for each test
session to avoid learning effects. This test is a variant of the digit-span
test found to be sensitive to solvent and lead exposure (Anger and Johnson,
1985) and is similar to Supraspan in the neurotoxicity test battery developed
by Eckerman et al (1985). Based on the digit span findings of Molhave et al
(1986), we predicted that SDL performance would be impaired by VOC exposure.
(6) Pattern Memory (PM) Test. This test of short term visual memory
involves the brief presentation of a block-like pattern followed by three
similar figures, one of which is identical with the original pattern. The
degree of correspondence of the two incorrect patterns to the correct one
varies. The number of correct responses and the response time are recorded.
Vhile digit span and serial digit learning assessed verbal memory, PH examined
if VOC exposure impaired visual memory.
86
-------
(7) Switching Attention (SVAT). This test evaluated the effect of VOC
exposure on selective attention with little memory demands. During two initial
sets the subject is trained to respond differentially to target stimuli on the
left or right side of the screen, and then to stimuli in the center of the
screen pointing to the left or right. The target stimuli are large rectangles
with inset arrows pointing left or right which may appear on the right or left
side of the screen. In the third set of trials, these two response conditions
are mixed. When the word on the screen is "SIDE", the subject must press the
left button if the target stimulus is on the left and vice versa. When the
word on the screen is "DIRECTION", then the subject must press the right button
if the arrow points to the right and vice versa. The word on the screen is
randomly assigned each trial, the numbers of incorrect responses and mean
react ion-times are recorded. This test was chosen to clarify whether the
Molhave et al digit span finding was due to memory or attentional impairment.
(8) Mood Scales. This test consists of a series of adjectives or phrases
with which subjects rate themselves on their feelings. Similar instruments
have been used to evaluate the efficacy of psychotherapeutic drugs and to
classify individuals with various neurobehavioral disorders (McNair et al,
1971) The present test contains 25 items and yields a five-dimensional mood
profile (tension, depression, anger, fatigue, confusion) by combining ratings
on individual items. Prior studies have shown that this approach is sensitive
in the evaluation of central nervous system effects of occupational lead
exposure (Baker et al, 1984). Molhave et al (1986) reported that subjective
ratings of sensory irritation (comfort) and air quality were altered by VOC
exposure. The mood scale was chosen to clarify these findings by assessing
which dimension(s) of mood were affected by VOC exposure.
(9) Associate Learning/Recall. This test is similar to the paired
associate learning test in the Wechsler Memory Scale (Wechsler, 1945). Three
letter names (e.g. Jan, Pat, Tom) are paired with occupations (e.g. florist,
lawyer, plumber). Nine pairs were presented on the video monitor for two sees.
each with one sec. intervals between pairs. After all pairs were presented,
each name was displayed and subjects were required to select the appropriate
87
-------
alternative from a list of occupations. Three trials were presented with the
same pairings, but in different order. The score consisted of the total number
of correct pairings in trials 1-3. A single recall trial vith the same
pairings vas presented about 20 mins. later. This test provided an alternative
type of memory challenge for testing VOC exposure effects.
(10) Pattern Comparison. This test is similar to the Pattern Memory Test
described above and utilizes the same 10x10 character arrays. Three patterns
are presented simultaneously on the video screen. Two patterns are identical
and the third is slightly different. The task is to identify the odd array. 25
trials were presented. Two measures were recorded—the mean latency of correct
trials 2-25 and the number of correct trials. Schmedtje et al (1988) found
that performance on this test improved after administration of
dextroamphetamine and worsened after scopalomine.
(11) Simple Reaction Time (SRT). SRT is the oldest tool in the
psychologist's armamentarium. This visual task requires subjects to
button-press as fast as possible when a large square is displayed on the
screen. Nine blocks (five with the preferred hand and four with the
nonpreferred hand) containing 12 trials each and variable intertrial intervals
of 2-4 sees were used. Mean reaction times, excluding block one, were
calculated for each hand. SRT tests have been used extensively in occupational
field studies including workers exposed to solvents (Anger and Johnson, 1985).
(12) Grammatical Reasoning. This test was adapted from Baddeley (1968).
Each trial consists of a logical sequence such as "A FOLLOWS B: AB" to which
the subject must respond "true" or "false". Sequences vary along four
dimensions: active-passive, positive-negative, true-false, and whether the
first letter is A or B. Two blocks of 32 trials each were presented. Mean
response time for all correct trials and the number of correct trials were
recorded. This was the most difficult NES test used. Little evidence of the
utility of this test in neurotoxicity testing is available.
(13) Horizontal Addition Test. Subjects were presented with 60 simple
88
-------
addition problems in which three single digits were presented in a row.
Subjects were instructed to sum the digits as rapidly and accurately as
possible. Responses had to be entered on the keyboard. Mean response times
and number of correct trials were recorded. Subjects made very few errors.
This test served primarily as a filler.
REFERENCES
Anger WK, Johnson BL. Chemicals affecting behavior. In: O'Donoghue JL (ed)
Neurotoxicity of Industrial and Commercial Chemicals. CRC Press, Baton
Rouge, 1985.
Baddeley, AD. A 3 minute reasoning test based on grammatical transformation.
Psychonom Sci 1968, 10: 341-342.
Baker E, Letz R, Fidler A, etal. A computer-based neurobehavioral evaluation
system for occupational and environmental epidemiology: Methodology and
validation studies. Neurobehav Toxicol Teratol 1985, 7:369-377.
Baker E, Letz R, Eisen EA et al. Neurobehavioral effects of solvents in
construction painters. J Occup Med, 1988, 30:116-123.
Eckerman D, Carroll J, Force D, et al. An approach to brief field testing
for neurotoxicity. ibid:387-393.
Estrin WJ, Moore P, Letz R, Wasch HH. The P300 event-related potential in
potential in experimental nitrous oxide exposure. Clin Pharmacol Ther,
1988, 43:86-90.
Greenberg B, Moore P, Letz R, Baker E. Computerized assessment of human
neurotoxicity: Senitivity to nitrous oxide exposure. Clin Pharmacol Ther-
apeutics, 1985, 38:656-660.
Hanninen H. Psychological test methods: Sensitivity to long term chemical
exposure at work. Neurobehav Toxicol Teratol, 1979, l(supp 1):157-161.
Mahoney P, Moore PA, Baker EL, Letz R. Experimental nitrous oxide exposure as
a model system for evaluating neurobehavioral tests. Toxicol., 1988,
49:449-457.
McNair DM, Lorr M, Drpopleman LF. BITS Manual—Profile of Mood States. San
Diego: Educational and Testing Service, 1975.
Molhave L, Bach B, Pedersen OF. Human reactions to low concentrations of
volatile organic compounds. Environ Internal, 1986, 12:167-175.
89
-------
Schmedtje, J.F., Oman C., Letz, R., Baker, E.L. Effects on scopalomine and
dextroamphetamine on human performance. Aviat Space Environ Med, 1988,
59:407-410.
Vechsler D. A standardized memory scale for clinical use. J Psychol 1945,
19:87-95.
Vechsler D. Vechsler Adult Intelligence Scale Manual. Psychological Corp-
oration, Nev York, 1955.
90
-------
APPENDIX D
SYMPTOM QUESTIONNAIRE
SUBJECT ID:
DO YOU FEEL NOW THAT STAYING IN THE CHAMBER IS COMFORTABLE? (Y/N)
WOULD YOU BE SATISFIED WITH THIS AIR QUALITY IN YOUR OWN HOME? (Y/N)
WOULD YOU VENTILATE YOUR HOME MORE? (Y/N)
(1) DOES THE SKIN OF YOUR BODY FEEL IRRITATED OR ITCHY?
NOT AT ALL | 1 1 VERY MUCH
(2) HOW IS THE NOISE LEVEL IN THE ROOM?
VERY QUIET | 1 1 VERY NOISY
(3) DO YOUR EYES FEEL DRY OR WATERY?
VERY DRY | 1 1 VERY WATERY
(4) DO YOU HAVE A HEADACHE RIGHT NOW?
NOT AT ALL | 1 1 VERY STRONG
(5) DOES THE SKIN ON YOUR FACE FEEL WARM OR COLD?
VERY COLD | 1 1 VERY WARM
(6) DO YOU FEEL SLUGGISH?
NOT AT ALL | 1 1 VERY MUCH
(7) HOW IS THE ODOR LEVEL IN THE ROOM?
NO ODOR | 1 1 VERY STRONG
(8) DO YOU FEEL ANY IRRITATION IN YOUR EYES?
NOT AT ALL | 1 1 VERY STRONG
(9) DOES THE SKIN OF YOUR BODY FEEL WARM OR COLD?
VERY COLD | 1 1 VERY WARM
(10) HOW IS THE LIGHT LEVEL IN THE ROOM?
VERY DARK | 1 1 VERY BRIGHT
(11) HAVE YOU BEEN COUGHING?
NOT AT ALL | | 1 A LOT
(12) DOES YOUR FACE FEEL IRRITATED OR ITCHY?
NOT AT ALL | 1 1 VERY MUCH
91
-------
(13) HOW IS THE AIR HUMIDITY IN THE ROOM?
VERY DRY | 1 1 VERY HUMID
(14) HOW IS THE MOVEMENT OF AIR IN THE ROOM?
VERY STILL | 1 1 VERY DRAFTY
(15) DOES YOUR NOSE FEEL DRY OR IRRITATED?
NOT AT ALL | 1 1 VERY MUCH
(16) HOW IS THE AIR TEMPERATURE IN THE ROOM?
VERY COLD | 1 1 VERY HOT
(17) ARE YOU TIRED OR SLEEPY?
NOT AT ALL | 1 1 VERY MUCH
(18) DOES THE SKIN OF YOUR BODY FEEL MOIST OR DRY?
NOT AT ALL | 1 1 VERY MUCH
(19) DOES YOUR THROAT FEEL IRRITATED?
NOT AT ALL | 1 1 VERY MUCH
(20) HOW IS THE QUALITY OF AIR IN THE ROOM?
VERY GOOD | 1 1 VERY BAD
(21) DOES THE SKIN ON YOUR FACE FEEL DRY OR MOIST?
VERY DRY | 1 1 VERY MOIST
(22) HOW HARD DID YOU HAVE TO CONCENTRATE ON THESE QUESTIONS?
VERY LITTLE | 1 1 VERY MUCH
92
-------
APPENDIX E
THE USE OF LINEAR POTENTIOMETERS FOR SUBJECTIVE
ASSESSMENTS OF INDOOR AIR QUALITY
12 3
Lars Molhave , David Otto , Gregory Rose
Institute of Environmental and Occupational Medicine
University of Aarhus
Aarhus, Denmark
2
Human Effects Research Laboratory
U.S. Environmenatal Protection Agency
Chapel Hill, NC 27599-7315
3
NSI Environmental, Inc.
Research Triangle Park, NC 27711
93
-------
Introduction
The principle of potentiometer assessment is called modified magnitude
estimation and is based on a linear analog rating scale (Bond and Lader, 1974).
Subjects may continuously indicate individual or collective ratings of any
specified indoor climate factor presented in the form of questions with defined
continuous answers between two well defined extremes.
Measuring Principles
Each subject has his own rating box (Fig. 1) with a sliding pointer that can
be
positioned along a 60 mm track. The device can be used in two modes. Ratings
can be signalled by external prompt (e.g., a visual prompt on the subject's
video screen or verbal instruction from the investigator) or ratings can be
initated spontaneously by the subject whenever a change in indoor air quality
is perceived. In this study comfort ratings were signalled every 15-30 minutes
by video prompts or verbal instruction. The extremes were defined as "no
irritation" and "unacceptably strong irritation". Prior to commencement of
testing, the potentiometer pointer was positioned at the "no irritation"
extreme.
Instructions to the Subject should include:
a) Definition of the endpoints and rating scale.
b) The nature of the indoor climate factor (s) to be rated.
c) That the measurement scale must be established by the subject
himself.
d) That only the pointer should move and that nobody else should see
the setting of the pointer.
94
-------
e) That the scale includes a neutral range (e.g., at the "no
irritation" end).
f) That the initial position of the pointer has no particular
relationship to the present IAQ.
g) Operation of the box and pointer must be demonstrated and explained
to subjects prior to its use.
h) The time intervals at which ratings are to be made should be
specified.
i) If external timing is used, the subjects should be instructed to go
to the box and make a simulated rating even if they don't feel any
change. They may delay their rating if they are occupied with
other tests when the sound signal appears.
j) The pointer must be set in a neutral position (i.e., no irritation)
at the start of the tests.
k) Subjects should not move the pointer back to neutral and then to
a new position when a rating is required. That is, if a change in
rating is desired, the pointer should be moved from its previous
position directly to the new position.
The Equipment
The rating box is shown in Figure 1. The box contained a linear
potentiometer with sliding pointer which could be positioned along a
one-dimensional 60 mm axis. A 5 volt power source in the breakout box supplied
the voltage across the resistive arm to give a O to 5 volt output range. Due to
a non-linear response in the 4 to 5 volt range, the slide was limited to a
maximum output of 4 volts. Thus 0 volts corresponded to "no irritation" and 4
volts corresponded to "unacceptabley strong irritation". Each of the four
potentiometer outputs was directed to an input channel of an analog-to-digital
95
-------
realtime module mounted in the backplane of the Compaq Plus microcomputer
control station. A BASIC algorithm automatically sampled each of the four
channels simultanelusly every 15 minutes throughout the test day. Sampling was
initiated manually two minutes after testing commenced to permit subjects to
complete the initial comfort rating. Subjects vere prompted on video monitors
while testing and by operators over loudspeakers during break periods to set
their comfort level prior to the 15 minute auto sampling. The algorithm
converted the sampled digital units back to voltage units to store and report
irritation in a range of 0.0 to 4.0.
96
-------
4 VOLTS 0
RATING BOX 1
RATING BOX 2
RATING BOX 3
RATING BOX 4
i
BREAKOUT BOX
(WITH 5 VOLT
POWER SUPPLY)
FIGURE 1
COMPAQ PLUS
WITH
A/D
CONVERTER
-------
The calibration of the potentiometers with the analog-to-digital converter
was checked prior to each test day and stored by the computer in 0, 25%, 50%,
75%, 100% increments of the slide of the potentiometer which corresponded to
measures of 0.0, 1.0, 2.0, 3.0, and 4.0.
Instructions to the Subjects
"An aluminum Comfort Meter is located to the right of your computer. The
lid should be open to show a sliding brass knob with a pointer that can be moved
from "No irritation" to "Strong irritation". You will use this comfort meter to
rate the degree of comfort or irritation that you feel in your eyes, nose or
throat during the experiment. There are no scale markings on the box, so you
will have to establish your own scale when you start".
"Remember, we want to monitor the level of comfort or irritation that you
feel in your eyes, nose and throat during the experiment. You will be
instructed at regular intervals, both by computer prompts and over the intercom,
to indicate your comfort level. Do not interrupt an ongoing test to reset the
comfort meter unless instructed to do so. During rest periods between tests,
however, you are encouraged to indicate any change in comfort level that you
feel. Please close the lid after each rating in order to prevent others from
seeing your comfort rating".
REFERENCE
Bond A., and Lader H. The use of analoque scales in rating subjective feelings.
Br. J. Med. Psycholo. 1974, 47:211-218.
98
------- |