Poor indoor air quality slows down metabolic rate of office workers

Abstract

A re-analysis of two independent laboratory studies was made in which a total of 60 female subjects had been
exposed for several hours to 6 different air quality conditions in groups of 6 people at a time. The subjects
performed typical office tasks at their own pace during exposures. Measured carbon dioxide (CO2) concentrations
and outdoor air supply rates were used to calculate CO2 produced by subjects at each air quality level. The
re-analysis showed that CO2 produced by subjects was affected by air quality (P<0.015). It decreased by ca. 13%
when the percentage dissatisfied with the perceived air quality increased from 8% to 40%, indicating a
dose-response relationship. A change in breathing pattern (shallow breathing) or a slow down of work rate in
polluted air would both reduce metabolic rate and thus the CO2 production rate.

Full text

In: Proceedings of Indoor Air 2005, Beijing, China: in press
POOR INDOOR AIR QUALITY SLOWS DOWN METABOLIC RATE OF
OFFICE WORKERS
Zs Bakó-Biró *, P Wargocki, DP Wyon and PO Fanger
International Centre for Indoor Environment and Energy, Technical University of Denmark,
Nils Koppels Alle, Building 402, DK-2800 Kgs. Lyngby, Denmark (www.ie.dtu.dk)
ABSTRACT
A re-analysis of two independent laboratory studies was made in which a total of 60 female
subjects had been exposed for several hours to 6 different air quality conditions in groups of 6
people at a time. The subjects performed typical office tasks at their own pace during
exposures. Measured carbon dioxide (CO2) concentrations and outdoor air supply rates were
used to calculate CO2 produced by subjects at each air quality level. The re-analysis showed
that CO2 produced by subjects was affected by air quality (P<0.015). It decreased by ca. 13%
when the percentage dissatisfied with the perceived air quality increased from 8% to 40%,
indicating a dose-response relationship. A change in breathing pattern (shallow breathing) or
a slow down of work rate in polluted air would both reduce metabolic rate and thus the CO2
production rate.
INDEX TERMS
CO2 emission rate, perceived air quality, building materials, personal computers, metabolic
rate.
INTRODUCTION
The effects of a poor air quality on human comfort and performance of office work have been
investigated in a number of laboratory experiments (Wargocki et al., 1999, Lagercrantz, 2000;
Wargocki et al., 2002, Bakó-Biró et al., 2004). The air pollution level in these studies was
usually altered by placing typical building products, such as used carpets, linoleum floor
coverings, bookshelves with books and papers, sealant, and personal computers (PCs) with
cathode-ray tube (CRT) monitors, into a ventilated office. These interventions negatively
affected the perceived air quality (PAQ) and caused subjects to report an increased intensity
of such symptoms as headache and difficulty in concentrating, and to perform office tasks less
effectively. The carbon dioxide (CO2) concentration measured in the occupied space was
generally lower when the pollution sources were present compared to the unpolluted
conditions, although the outdoor air supply rate and all other environmental parameters were
kept unchanged, but no detailed analysis of this issue was undertaken. In another study,
Wargocki et al., (2000) reported that the metabolic rates, estimated from the CO2 levels and
actual ventilation rates, significantly increased when the outdoor air supply rate, and
consequently the PAQ, was improved in a polluted office. The higher metabolic rate was
attributed to an increased muscular tonus given by the higher work rate that was seen under
the conditions with improved PAQ. It was mentioned that breathing shallowly when air
quality is poor might be what lowered the metabolic rate. The objective of the current paper
was to analyze the CO2 production rates of subjects participating in two independent
* Corresponding author e-mail: zbb@mek.dtu.dk; from 18.04.2005: zbb@reading.ac.uk, School of Construction
Management and Engineering, The University of Reading, Whiteknights, PO Box 219, Reading RG6 6AW,
United Kingdom (www.cme.rdg.ac.uk)

experiments who had been exposed to various air quality conditions in a real office space and to
formulate an alternative hypothesis that may explain the alterations in CO2 levels that have
previously been reported.
METHODS
In two independent experiments, the same approach was used to establish different levels of
air quality: in an office with low-polluting floor, walls and furniture (CEN, 1998) common
building-related products were placed to increase the emission of indoor pollutants. With this
method two air quality conditions were created, one with sources present and one with
sources absent in the office, maintaining a given air change rate. This procedure was repeated
at outdoor air change rate of 1 h-1 and 3 h-1 (i.e. 5 and 15 L/s per person) in the first-, and of 2
h-1 air change rate (10 L/s per person) in the second experiment, obtaining thus 6 different air
quality conditions in the office. All other environmental parameters were kept unchanged.
The pollution sources consisted of linoleum, bookshelf with books and papers, and sealant in
the first experiment, and of PCs with CRT monitors in the second study. In each experiment
30 female subjects were recruited, i.e. a total of 60 subjects were exposed in groups of six
people at a time to the air quality conditions created in the office in 30 sessions (one session
per day, 5 sessions per week) using a balanced design for order of presentation. The exposure
period lasted for ca. 3 and ca. 5 hours in the first and second experiment respectively. During
each exposure the subjects performed simulated office work comprising text typing,
proofreading and arithmetical calculations. They remained thermally neutral by adjusting
their clothing. In each experiment the subjects had roughly the same average body size of 1.7
m2 (±0.1 SD). The office where the experiments were carried out was described in detail by
Wargocki et al. (1999). It was divided by a 2-m-high partition into a space for the equipment
used to supply and condition the outdoor air and for the pollution sources to be placed, and a
space where the subjects were exposed. The air was supplied directly from outdoors by an
axial fan mounted in the window and exhausted under the entrance door to an adjacent
corridor. Several desktop fans ensured good mixing of the pollutants released from the
sources in the office. The air was conditioned by oil-filled electric radiators and ultrasonic
humidifiers. If necessary a SPLIT-type air-conditioner was started to cool the air in the office.
The space used for exposure had six workstations, each consisting of a table, a chair, a desk-
lamp and a 6-year-old low-polluting PC with a sensory pollution strength of 0.3 olf/PC as
quantified by sensory evaluations carried out prior to the experiments.
The methodology of subjective and objective evaluation of air quality, indoor climate, sick
building syndrome (SBS) symptoms and performance were previously described in detail
(Wargocki et al., 2002, Bakó Biro et al., 2004). The perceived air quality was assessed on a
continuous acceptability scale upon entering, during exposure and upon re-entering as visitors
after they had left the office for a few minutes. The latter assessment is the most
representative in terms of air quality to which the subjects were exposed, as the office air
contained both pollution from the sources (when present), and human bioeffluents in what
was effectively a steady-state condition. Among the measured parameters the outdoor air
supply rate and indoor/outdoor concentration of CO2 were used in the present analysis. These
were continuously monitored each day before, during and after the exposures with a Inova
Multi-Gas monitor Type 1302 connected with a Inova Multipoint Sampler and Doser Type
1303. The constant concentration method with sulphur hexafluoride (SF6) as a tracer gas was
used to measure outdoor air supply rate. The sampling of CO2 and tracer gas was taken from
the breathing zone of each subject and from a central location in the occupied space at a
height of 1.1 m to ensure that the air was well mixed. The mean CO2 concentration in the
office was calculated as the average of CO2 concentrations measured at each of these 7

sampling points. Using this value, the outdoor CO2 concentration, and the outdoor air supply
rate, the CO2 production rate was calculated using a mass-balance model. As subjects were
exposed to different air quality conditions in groups of 6 subjects at a time on each exposure
day, the CO2 production rate was calculated separately for each group to take account for
small differences in outdoor air supply rate and CO2 concentration occurring between days for
the same exposure conditions. Dividing the group CO2 production rates by the number of
people in the group yielded an estimate of an average CO2 production rate per person. The
CO2 production rates per person calculated in this way were then used to calculate average
CO2 production rate per person in each of the exposure conditions for the whole group of 30
subjects participating in each of the two experiments. The data regarding air acceptability
were analysed using paired t-test. Wilcoxon matched-pairs test and chi-square statistics were
used to analyze the CO2 emission rates as function of the interventions established in the
office. All p-values are 1-tailed of an effect in the expected direction.
RESULTS
Table 1 shows that the acceptability of air quality was significantly lower (higher %
dissatisfied with air quality) in the presence of pollution sources
Table 1. Acceptability of air quality and % dissatisfied with air quality upon re-entering the
office in each condition (assessed as visitors)
Experiment 1 Experiment 2
air change rate = 1 h-1 air change rate = 3 h-1 air change rate = 2 h-1
Sources
present Sources
absent Sources
present Sources
absent Sources
present Sources
absent
Acceptability 0.03 0.15 0.04 0.42 0.11 0.29
p-value (t-test)* <0.016 <0.0003 <0.01
% dissatisfied 42 28 40 8.4 32 15
* difference between condition with sources present and absent
The average CO2 production rates were lower for almost every group of subjects when the
pollution sources were present compared to the conditions with sources absent, regardless
whether the outdoor air change rate was at 1, 2 or 3 h-1 (Figure 1 – left). The mean values for
30 subjects (Figure 1 – right) indicate a ca. 5% lower CO2 production rate in the office with
sources present compared to the condition with sources absent, at each air change rate. It may
be observed that the differences in the CO2 production rate were not always statistically
significant when applying the Wilcoxon test, as the sample size was relatively small – 5,
which is the number of groups exposed at a given air change rate to the polluted/unpolluted
conditions. The p values from each pair of conditions (sources present/absent, Fig. 1 right)
were combined to calculate an overall probability using the chi-square statistic, under the null
hypothesis that the sum of the Natural logarithms of the observed 1-tail probabilities are
distributed as –0.5·χ2. This analysis yielded a probability of p<0.015, indicating that an
overall hypothesis that CO2 production rates are lower in the presence of sources tested in this
analysis may be accepted.

10
12
14
16
18
20
10 12 14 16 18 20
CO
2
production rate (L/h per person)
Source absent (PD%=15±8)
CO
2
production rate (L/h per person)
Source present (PD%=39±6)
14
15
16
17
18
CO
2
production rate (L/h per person
)
Sources absent Sources present
p<0.06
p<0.03
p<0.15
p<0.015
1 h
-1
buiding mat. 2 h
-1
PCs 3 h
-1
buiding mat.
Figure 1. Left: Average CO2 production rate per person in conditions with sources absent
compared with average CO2 production rate per person in conditions when sources were
present. Each point represent data for the same group of 6 subjects exposed in the office with
sources present and absent ventilated with air change rates of 1, 2 or 3 h-1 (i.e 5, 10 and 15
L/s per person); Right: Average CO2 production rate per person at 6 different air quality
levels created in the office; each bar represents average CO2 production rate per person for
30 subjects exposed calculated from data for 5 groups of 6 subjects each
The average CO2 production rates per person (for 30 subjects exposed in each of the 6
experimental conditions, Fig. 1-right) were regressed against the assessments of acceptability
of air quality in these conditions made immediately upon re-entering the office after exposure
(Table 1) and yielded an approximately linear relationship (R2=0.6; p<0.07). This relationship
is plotted in Figure 2, after acceptability ratings were expressed in % dissatisfied with air
quality (Wargocki, 2004), and suggests that subjects start to produce less carbon dioxide as
the air quality decreases (increased % dissatisfied): a change by 13% when the percentage
dissatisfied with the quality of air changes from 8% to 40%.
14
15
16
17
18
0 102030405060
% Dissatisfied
CO
2
production rate (L/h per person *)
Figure 2: Carbon dioxide production rate per person as a function of the % dissatisfied with
air quality; * each point represents average CO2 production rate per person with an average
body size of 1.7 m2 (±0.1 SD) for 30 subjects; the bars represent standard errors.

DISCUSSION
This meta-analysis of the results of two independent experiments indicates that the pollution
level of inhaled air may significantly affect the CO2 production rate of occupants. This must
be assumed to be due to changes in metabolic rate. It may have a psychological origin,
reflecting people’s unwillingness to work in poor air quality conditions, or it may be due to
intensified SBS symptoms resulting from exposure to indoor pollutants (e.g., headache) that
reduce work rate. The question of how these symptoms are caused then arises. Changes in
breathing pattern may affect the CO2 content of exhaled air: if breathing does not correspond
to metabolic demand, i.e. if the rate at which CO2 is produced at the cellular level is greater
than the rate at which it is exhaled, due to shallow breathing or other dysfunction in the
respiratory system, CO2 concentration in the blood will increase, inducing physiological
effects such as symptoms similar to SBS (Martin, 1987; Paulev, 2000; Resta, 2000). If this
leads to a decrease in work rate, metabolic rate will be reduced and less CO2 will be exhaled.
This hypothesis is a feasible, simple and thus tempting explanation for the high frequency of
occurrence of complaints in so-called “sick buildings” that is commonly reported. Alteration
in the breathing pattern as a result of exposure to environmental chemicals has been reported
not only in mouse bioassays (Larsen et al., 2000) but also in human subjects (Danuser, 2001),
supporting the present hypothesis. Danuser (2001) reviewed a number of studies reporting
perceptual and breathing responses induced by exposure to environmental chemicals. She
pointed out that a decline in tidal volume (the amount of air inhaled per breath) may occur
even in response to stimuli that are not detected by anosmics and which elicit very little
olfactory response in normal individuals. Crucial experiments to investigate these
hypothesized mechanisms are required. CO2 levels in blood will differ depending on which
mechanism, i.e. psychological effects or changes in breathing pattern, is active. Measurements
of End-Tidal CO2 would reveal this difference non-invasively.
CONCLUSIONS AND IMPLICATIONS
• The rate of CO2 production by occupants decreased significantly, by about 5%, when they
were exposed to emissions from typical indoor pollution sources, compared to conditions
in which these sources were not present. This effect can be caused by change in the
breathing pattern (shallow breathing) or by a slower work rate in polluted air. Both
changes would cause a reduction in metabolic rate, which may be either the cause or the
effect of reduced performance.
• The effects of perceived air quality on the CO2 production rate are expected to be higher
than 5%, following a dose-response relationship. A change in the perceived air quality
level from 8% to 40% may decrease the CO2 production rate and consequently the
metabolic rate by ca. 13%.
• The results of the present investigation imply that an adequate ventilation rate in buildings
is not only necessary to comply with human comfort requirements, but also to prevent a
direct negative effect of a mediocre indoor air quality manifested in an alteration of the
breathing pattern that may induce further physiological effects in humans, including
symptoms similar to SBS.
ACKNOWLEDGEMENTS
This work has been supported by the Danish Technical Research Council (STVF) as part of
the research programme of the International Centre for Indoor Environment and Energy
established at the Technical University of Denmark for the period 1998-2007.

REFERENCES
Bakó-Biró, Z., Wargocki, P., Weschler, C.J. and Fanger, P.O. (2004) “Effects of pollution
from personal computers on perceived air quality, SBS symptoms and productivity in
offices”, Indoor Air, 14(3), 178-187.
CEN (1998) Technical Report CR 1752, Ventilation for Buildings: Design Criteria for the
Indoor Environment, Brussels, European Committee for Standardization.
Danuser, B. (2001) “Candidate physiological measures of annoyance from airborne
chemicals”, Chemical Senses, 26, 333-337.
Lagercrantz, L., Wistrand, M., Willén, U., Wargocki, P., Witterseh, T and Sundell, J (2000)
“Negative impact of air pollution on productivity: previous Danish findings repeated in
new Swedish test room” In: Proceedings of Healthy Buildings 2000, Espoo, Finland, Vol.
1, 653-658.
Larsen, S. T., Hougaard, K. S., Hammer, M., Alarie, Y., Wolkoff, P., Clausen, P. A., Wilkins,
C. K. and Nielsen, G. D. (2000) “Effects of R-(+)- and S-(-)-limonene on the respiratory
tract in mice”, Human and Experimental Toxicology, 19(8), 457-466.
Martin, L (1987) “Pulmonary physiology in clinical practice: the essentials for patient care
and evaluation”, St. Louis; http://www.mtsinai.org/pulmonary/books/physiology.
Resta, O. Foschino Barbaro, M. P., Bonfitto, P., Talamo, S., Mastrosimone, V. Stefano, A.
and Giliberti, T (2000) “Hypercapnia in obstructive sleep apnoea syndrome”, The
Netherlands Journal of Medicine, 56(6), 215-222.
Paulev, P. E. (2000) “Textbook in Medical Physiology and Pathophysiology” Copenhagen,
ISBN 87-984078-0-5; http://www.mfi.ku.dk/ppaulev/content.htm.
Wargocki P. 2004. “Sensory pollution sources in buildings,” Indoor Air 2004; 14 (Suppl 7): 82-
91.
Wargocki, P., Wyon, D.P., Baik, Y.K., Clausen, G. and Fanger, P.O. (1999) “Perceived air
quality, Sick Building Syndrome (SBS) symptoms and productivity in an office with two
different pollution loads”, Indoor Air, 9(3), 165-179.
Wargocki, P., Wyon, D.P., Sundell, J., Clausen, G. and Fanger, P.O. (2000) “The effects of
outdoor air supply rate in an office on perceived air quality, Sick Building Syndrome
(SBS) symptoms and productivity”, Indoor Air, 10(4), 222-236.
Wargocki, P., Bakó-Biró, Zs., Clausen, G. and Fanger, P.O. (2002) “Air quality in a simulated
office environment as a result of reducing pollution sources and increasing ventilation”,
Energy and Buildings, 34(8), 775-783.