Indoor Air 2015; 25: 210–219
wileyonlinelibrary.com/journal/ina
Printed in Singapore. All rights reserved

© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
INDOOR AIR
doi:10.1111/ina.12134

The eﬀects of an energy eﬃciency retroﬁt on indoor air quality
Abstract To investigate the impacts of an energy eﬃciency retroﬁt, indoor air
quality and resident health were evaluated at a low-income senior housing
apartment complex in Phoenix, Arizona, before and after a green energy
building renovation. Indoor and outdoor air quality sampling was carried out
simultaneously with a questionnaire to characterize personal habits and general
health of residents. Measured indoor formaldehyde levels before the building
retroﬁt routinely exceeded reference exposure limits, but in the long-term
follow-up sampling, indoor formaldehyde decreased for the entire study
population by a statistically signiﬁcant margin. Indoor PM levels were
dominated by ﬁne particles and showed a statistically signiﬁcant decrease in the
long-term follow-up sampling within certain resident subpopulations (i.e.
residents who report smoking and residents who had lived longer at the
apartment complex).

S. E. Frey1, H. Destaillats2,
S. Cohn2, S. Ahrentzen3,
M. P. Fraser4
1
Department of Chemistry and Biochemistry, Arizona
State University, Tempe, AZ, USA, 2Indoor Environment
Group, Environmental Energy Technologies Division,
Lawrence Berkeley National Laboratory, Berkeley, CA,
USA, 3Rinker School of Construction Management,
University of Florida, Gainesville, FL, USA, 4School of
Sustainable Engineering and Built Environment, Arizona
State University, Tempe, AZ, USA

Key words: Particulate matter; Formaldehyde; Senior housing; Phoenix; Arizona.

H. Destaillats
Indoor Environment Group, Environmental Energy
Technologies Division, Lawrence Berkeley National
Laboratory, Berkeley, CA, USA
Tel.: +1 (510) 486-5897
Fax: +1 (510) 486-7303
e-mail: hdestaillats@lbl.gov
and
M. P. Fraser
School of Sustainable Engineering and Built
Environment, Arizona State University, Tempe, AZ, USA
Tel.: +1 (480) 965-3489
Fax: +1 (480) 965-8087
e-mail: matthew.fraser@asu.edu
Received for review 19 November 2013. Accepted for
publication 26 May 2014.

Practical Implications

The results presented here provide insight into the indoor air quality before, immediately after, and 1 year after an
energy eﬃciency retroﬁt on a federal-subsidized senior apartment complex. With increasing focus on building energy
eﬃciency, it is critical to evaluate possible relationships between resident health and changes in indoor environmental
quality. Initially, formaldehyde exposure was quite high for all study participants, but an overall decrease was measured a year after the construction was completed. Particulate matter, however, was largely impacted by resident
behavior (such as smoking), and a long-term decrease was only observed when combined with particular
subpopulations.

Introduction

As buildings become more energy eﬃcient, there are
concerns about long-term eﬀects of changes in construction practices, materials and building operation.
The need for reduced energy consumption, driven
by rising energy costs and the desire to eliminate
210

dependence on fossil fuels, has become a national
priority. A common approach is to seal the building
envelope, reducing air leakage and unnecessary usage
of heating and cooling units. However, by reducing
ventilation rates, pollutants could become trapped and
increased exposure to toxins becomes a concern (Fisk,
2000; Jones, 1999; Weschler and Shields, 2000).

Indoor air quality and energy eﬃciency retroﬁt
The risk of increased exposure lies in the fact that
Americans spend up to 90% of their time indoors,
whether at work, school, or in their homes (U.S. EPA,
1989; Wallace et al., 2006). Mitigating exposure and
understanding the sources, fate and transport of
indoor pollutants is of utmost importance (Lee et al.,
2002; Wallace et al., 2006). The most vulnerable populations aﬀected are children, the elderly, and people
with existing respiratory diseases. In addition, lowincome populations are less likely to have access to
indoor air pollution intervention information, making
low-income seniors a population most impacted and
least able to respond to the burdens of a toxic indoor
environment (Williams et al., 2000).
Therefore, a combination of increased time spent
indoors and energy eﬃciency building practices has
created the need to assess the impact of renovations on
indoor environmental quality and human health. The
U.S. Department of Housing and Urban Development
promotes energy conservation and healthy home environments and has funded research on the potential
impacts of ‘green’ building methods on both indoor
environments and resident health (US HUD, 2009). It
has been suggested that ‘green’ housing solutions may
actually be detrimental to resident health, by not taking into account low-risk building materials and
neglecting indoor air quality during the design process
(Wargo, 2010). However, a recent study illustrated the
potential for overall improvements in indoor air quality when retroﬁt measures are implemented with the
simultaneous aims of saving energy and improving
indoor environmental quality (Noris et al., 2013).
Here, we evaluate impacts on indoor air quality as it
relates to particulate matter and volatile carbonyl concentrations, speciﬁcally formaldehyde, acetaldehyde,
and acetone.
The US EPA considers particulate matter, or PM,
as a major concern due to the ability of particles with
diameters <10 lm (PM10) to pass through the throat
and nose and into the lungs. This, in turn, has an
impact on both lung and heart health (Dockery
et al., 1993; Pope and Dockery, 2006). There have
also been many studies connecting outdoor PM exposure to increased mortality rates and respiratory diseases (Davidson et al., 2005; Englert, 2004; Fann
et al., 2012; Li et al., 2003). Due to these ﬁndings,
National Ambient Air Quality Standards (NAAQS)
were set to regulate annual ambient concentrations of
PM10 at 150 lg/m³ and PM2.5 at 35 lg/m³. No limits
have been established for indoor PM levels, even
though more evidence of indoor PM exposure being
linked to negative health eﬀects is surfacing (Koenig
et al., 2005).
Carbonyls, especially formaldehyde, are ubiquitous
in the indoor environment and have been associated
with both chronic and acute health eﬀects. The main
sources of indoor formaldehyde include the degradation

of additives used in wood-based building materials,
furniture, and sealants as well as combustion and chemical reactions common to the indoor environment
(Destaillats et al., 2006, 2011; Hodgson et al., 2002;
Sidheswaran et al., 2013; Singer et al., 2006). Potential
health concerns include irritation to the eyes, nose,
throat, and lungs. Chronic exposure to formaldehyde
has also been shown to lead to asthma symptoms, allergic sensitization, and overall reduction of lung function
(LBNL, 2008; Salthammer et al., 2010; California EPA,
2007). Formaldehyde has been listed among the top ﬁve
indoor pollutants causing chronic health eﬀects in US
residences (Logue et al., 2012), has been identiﬁed as a
potential carcinogen by the US EPA (group B1, US
EPA, 1999a), and classiﬁed as carcinogenic to humans
(Group 1) by the International Agency for Research on
Cancer (IARC, 2012). Acceptable exposure levels for
formaldehyde determined by various countries have
been summarized by Salthammer et al. (2010). In 2007,
the California EPA established an 8-h Reference Exposure Level (REL) of 7 ppb and an acute REL of 44 ppb.
Health Canada, however, has 8-h REL of 40 ppb, over
ﬁve times higher than the CA EPA requirements (2006).
In 2010, the World Health Organization released a less
stringent guideline of 80 ppb for both short-term and
long-term risks (WHO, 2010).
This research is a subset of a larger study evaluating
the overall impact of an energy eﬃciency retroﬁt on a
vulnerable population, including cost eﬀectiveness and
health beneﬁts (Ahrentzen et al., 2013). Here, particulate matter and volatile carbonyl concentrations are
the benchmark for measuring how the retroﬁt aﬀects
the indoor air quality both immediate post-renovation
and 1 year following renovation. We seek to understand what sources and behaviors may impact
increased PM and aldehyde exposure. We also examined whether indoor air quality improvements resulted
in changed health conditions or health-related behaviors (such as improved sleep).

Materials and methods
Sampling campaign and health survey

A study was conducted at a local apartment complex,
operated by the City of Phoenix Housing Department,
for seniors who qualify for subsidized rent. Originally
built in the early 1970s, this three-story, 116-unit structure underwent unit renovations and energy eﬃciency
improvements in spring and early summer of 2011. The
HVAC system for each apartment included a throughwall package terminal air conditioner (PTAC) unit, a
bathroom exhaust fan, a range hood exhaust fan, and
doors and windows. The retroﬁt of each apartment
included upgrades in PTAC units, both exhaust fans,
and installation of energy eﬃcient, double pane exterior windows and sliding glass balcony doors. Other
211

Frey et al.
improvements included installation of low VOC ﬂooring, new cabinetry (natural oak product), paint (zero
VOC), low VOC carpet and carpet pad (Green Label
Plus Certiﬁed), Energy Star kitchen appliances (refrigerator, electric range, microwave, and garbage disposal), and the addition of a bedroom ceiling fan.
Researchers tested air quality in the self-contained
apartments a total of three times: in the summers of
2010, 2011, and 2012. Panel 1, before the renovation,
was conducted during the summer of 2010. Panel 2
was conducted immediately after construction was
completed from late April through September 2011.
One year later, Panel 3 was conducted from June
through early August 2012. For air quality sampling,
a total of 72 apartments were studied in Panel 1,
which have been reported separately (Frey et al.,
2014). A total of 55 and 53 units were studied in
Panels 2 and 3, respectively. However, only 47 units
participated in all three panels, corresponding to an
attrition rate of 35%. As the research design was a
longitudinal panel study examining the changes in
each resident’s apartment over time, data from residents in the ﬁrst panel who did not participate in
later panels were eliminated from the analyses. This
type of research design, however—in which each
apartment is its own control group—allowed us to
examine improvements or changes from the ﬁrst to
subsequent panels using paired t-tests and ﬁxedeﬀects regression models, which worked eﬀectively
with this smaller sample size.
Simultaneous measurements of indoor air pollutants, particulate matter and volatile carbonyls were
collected in the living room and kitchen, and outdoor
pollutant concentrations on the balcony of each unit.
All units have 619 ft2 of livable space and are identical
in interior layout and are all-electric homes (i.e. no ﬁreplaces, gas stoves, etc.) with individual through-wall
package terminal air conditioner (PTAC) units. Onehour samples were collected in each unit between the
hours of 9 am and 5 pm. Repeated testing in a subset
of units (7% replicate) ensured that no time-of-day
bias was present. The summer season was selected for
sampling as local hot weather would result in the
apartment units being sealed (i.e. windows closed) with
air conditioning running. During instrument setup and
sampling, residents did not cook, smoke, or clean.
At the same time as air quality testing, a health survey of over 100 questions was given to the residents to
solicit information about personal habits and selfreported health conditions. Performing the air quality
measurements and questionnaire simultaneously
ensured that resident activities (i.e. cooking, smoking)
did not bias the data and would not be present during
sampling.
The indoor air quality testing, presented here, is a
subset of a larger-scale study in which cost eﬃciency
and health beneﬁts of the renovations were also
212

analyzed (Ahrentzen et al., 2013). Additional information about temperature, humidity, air leakage, and
the health questionnaire can be found in that report.
While 1-h sampling periods are relatively short and
may be more susceptible to short-term sources, the
sampling plan was carefully designed to minimize the
impact of resident activity during sampling, thus
reducing that risk. In addition, the sampling plan
and the panel survey research design maximized the
number of participants to account for an expectedly
high attrition rate among a low-income elderly population over a 2-year period. Short-duration samples
allowed coordinating for a large number of units
over the month-long sampling period obtaining a
representative data set for the building. Although
cross-contamination across diﬀerent units could
impact the results of this study (e.g. due to all units
sharing common vertical exhaust ducts), it was not
quantiﬁed. In addition, because ventilation rates were
not tracked during sampling, changes in concentrations could be due to either changes in sources or
changes in ventilation rates.
PM measurements

Indoor air quality sampling included real-time measurement of PM using TSI DustTrak DRX aerosol
monitors (model 8533; TSI, Inc., Shoreview, MN,
USA) sampler. This instrument contains a light-scattering laser photometer to detect various particle sizes,
including PM1, PM2.5, PM4, PM10, and PM-total. The
maximum size measured for the PM-total is approximately 15 lm based on manufacturer speciﬁcations.
Three samplers were deployed to the apartment
kitchen, living room, and balcony to simultaneously
collect particle data over a 1-h period during which the
resident was given the health survey. By sampling both
indoor and outdoor air, we are able to calculate
indoor/outdoor ratios with the goal of quantifying the
impact of inﬁltration of outdoor particles vs. indoor
sources on indoor air quality. Dusttraks were labeled
and used in a consistent manner among units, were calibrated prior to the study, and tested for reproducibility by collocated sampling. While Dusttraks have been
shown to overestimate PM compared with gravimetric
measurements, the use of a consistent sampling platform was designed to minimize bias due to sampling
technique (Jenkins et al., 2004 and Wallace et al.,
2011).
Carbonyl measurements

Samples of indoor and outdoor formaldehyde, acetaldehyde, and acetone were collected using commercial
samplers
containing
dinitrophenyl
hydrazine
(DNPH)-coated silica gel (Sep-Pak XPoSure Aldehyde Sampler, # WAT047205; Waters Corp., Milford,

Indoor air quality and energy eﬃciency retroﬁt
MA, USA). The cartridges were preceded by an ozone
scrubber (Sep-Pak Ozone Scrubber, # WAT054420;
Waters Corp.) to eliminate ozone from the incoming
air. Air was drawn through the samplers by means of
pumps operating at  2 L/min (determined with a
precision better than  3%). Samples were collected
over 1-h periods using portable gas pumps (Universal
XR Pump, Model PCXR4; SKC Inc., Eighty Four,
PA, USA). The sampling ﬂow of each pump was calibrated in the laboratory before and after the sampling period using a bubble ﬂow meter and a
primary airﬂow calibrator (Gilibrator-2 Sensidyne,
St. Petersburg, FL, USA). Three samples were collected simultaneously with and in close proximity to
the PM samplers in the living room, kitchen, and
balcony of each unit.
After collection, each DNPH cartridge was capped,
labeled, and stored at 4°C until it was extracted and
analyzed. Acetonitrile extracts were analyzed by
high-performance liquid chromatography (HPLC)
with UV detection at 360 nm following a US EPA
method (US EPA, 1999b). The concentration value
reported in each case corresponded to a time-integrated average over the sampling period. Calibration
curves for quantiﬁcation were determined with
authentic standards of the dinitrophenylhydrazones
of formaldehyde, acetaldehyde, and acetone (SigmaAldrich). The detection limit for each volatile carbonyl was typically 10 ng or lower, corresponding to
air concentrations < 0.1 ppb. Laboratory and ﬁeld
blank samples (at least three laboratory and six ﬁeld
blanks) were also analyzed, showing non-detectable
values of the three analytes.
Reported health measures

The resident survey created and used in this study
contained over one hundred ﬁxed-response and openended questions pertaining to health conditions, resident assessments of the environmental quality of their
apartments, and household activities and behaviors
relevant to environmental quality. The health-related
questions were derived from standardized instruments
developed by the Centers for Disease Control: the
National Health Interview Survey (NHIS) and the
Behavioral Risk Factor Surveillance System (BRFSS).
The same questions were asked of residents at each
panel. Pertinent to the analyses presented here were
questions regarding smoking behavior; use of cleaning
and odor-masking products; an index of respiratory
conditions (derived from single questionnaire items
pertaining to snoring, asthma, emphysema, hay fever,
bronchitis, sinusitis); index of quality of health/life
(derived from three questionnaire items); index of emotional distress (derived from six standardized questionnaire items; see Pilkonis et al., 2011); and sleep
(number of hours).

Results and discussion

While the eﬀectiveness of the retroﬁt is beyond the
scope of this particular manuscript, energy and
water savings have been quantiﬁed. To summarize,
the retroﬁt resulted in a reduction of 12.6% in water
consumption and 19.4% in electricity consumption
based on analysis of 39 months of metered electrical
and water use between July 2009 and September
2012 (Ahrentzen et al., 2013).
Data analyses procedures

As mentioned previously, the particulate matter (PM)
and aldehyde concentrations of each resident’s kitchen,
living room, and balcony were recorded. However,
because linear correlations between an apartment’s
kitchen and living room PM data were 0.90 or higher,
measurements were combined from these rooms into
one composite measure (by averaging room-level data)
to represent the unit.
Trends are evaluated as the change of conditions
between Panels 1 and 2, labeled the ‘Short Term’ and
between Panels 1 and 3, labeled the ‘Long Term’
where only units that participated in both Panels are
included in the statistical analysis. Given the panel
research design, we used ﬁxed-eﬀects models when
comparing diﬀerences in an apartment’s conditions
between panels (all statistics presented will be ﬁxedeﬀects regressions, unless otherwise noted). As we did
not have a control group but did have a longitudinal
panel research design, these models were quite appropriate to the panel nature of our study, where each
individual’s apartment acts as his or her own control.
There are two basic data requirements for using ﬁxedeﬀects methods (Allison, 2005), both of which were
addressed in our study: (i) the dependent variable
must be measured for each unit on at least two occasions and those measurements must have the same
metric and (ii) the predictor variable must change in
value across those two occasions for some substantial
portion of the sample.
Sample characteristics

The questionnaire given to residents during testing was
essential to characterize the demographics of the apartment units sampled. Most units (88%) were occupied
by a single individual. Average age of residents at the
beginning of the study was 73 years, and 65% reported
at least one respiratory health problem at the ﬁrst
panel. The average length of stay of living in the apartment was 5.5 years.
Behavioral questions most relevant to the data
reported here are summarized in Table 1. In addition,
this resident behavior information aids in the interpretation of the indoor air quality data collected, and a
213

Frey et al.
The ranges of PM10 concentration were 8–783,
13–1375, and 11–600 lg/m3 for Panels 1, 2, and 3,
respectively. The best way to visualize these concentrations and changes between the short and long
term is through a cumulative frequency plot, as seen
in Figures 1 and 2.
While mean PM counts did show changes over time,
the variance was so large that statistical signiﬁcance
was not achieved. Overall, there was no statistically signiﬁcant change in PM levels before the renovation and
afterward (either in the short or long term). However,
if the top 25th percentile of Panel 1 (who also participated in Panel 3) is isolated, there is a statistically signiﬁcant decrease in both PM2.5 and PM10 in the long
term (paired t-test, N = 13 of 53: t = 2.167, P < 0.05
and t = 2.219, P < 0.05, respectively).
One of the largest factors connected to increased
indoor PM concentrations was smoking. This is shown
speciﬁcally for each panel in Table 4. Mean and median PM2.5 concentrations and indoor/outdoor ratios
are given.
Statistical analysis using various covariates was used
to see if resident demographics or habits had an
impact. In the short term, the resident’s length of stay,
whether the resident smoked, and use of odor-masking
products were covariates that had statistical impact.
In the short term (between Panels 1 and 2), both
PM2.5 and PM10 concentrations increased as the length
of time residents lived at the apartment complex
increased (PM2.5 t = 3.063, P = 0.003; PM10 t = 3.041,
P < 0.003). However, the indoor/outdoor ratios
decreased with length of time living there (I/O PM2.5
t = 3.721, P < 0.001; I/O PM10 t = 3.732, P < 0.001);
no coherent or consistent explanation could be found
for this association. The units of those residents who
used odor-masking products showed increased levels
of PM2.5 and PM10 in the short term (PM2.5 t = 1.963,
P = 0.052; PM10 t = 1.972, P = 0.051), but there was
no similar change of indoor/outdoor PM ratios. Not
surprisingly, PM concentrations and I/O ratios were
signiﬁcantly higher in homes of those residents who
smoked than in the units of non-smokers (PM2.5
t = 3.717, P < 0.001, PM10 t = 3.960, P < 0.001; I/O

Table 1 Examples of relevant questions asked to residents during air sampling. This is
a subset of a questionnaire of over 100 questions
Relevant questionnaire inquiries
Do you smoke?
Do you use bug sprays?
Do you use anything to change the smell of the air in your home more than once per
week?
If yes, does that include candles? Incense? Air freshener? Or Other?

Table 2 Questionnaire responses for each panel, number of responses (percentage)

Panel 1
Panel 2
Panel 3

Total
units

Smoker (%)

Use insecticide
(%)

Change smell
of air (%)

72
55
53

16 (22)
9 (17)
11 (21)

17 (24)
19 (36)
1 (2)

46 (64)
34 (64)
33 (62)

summary of participant behavioral data is presented in
Table 2 below.
Particulate matter

Mean and median indoor concentrations, as well as
indoor outdoor ratios, can be found in Table 3. Based
on all indoor/outdoor ratios being greater than 1, measured levels of indoor PM often far exceeded measured
outdoor concentrations, an indication of the importance of indoor PM sources for the units participating
in the study. Although indoor particle concentration
averages are higher, they are also widely variable compared with outdoor PM concentrations, although part
of the diﬀerence may be due to varying particle morphology between indoor and outdoor PM, which alters
instrument response. When comparing Panel 1 indoor
PM10 to outdoor PM10, there is a mean diﬀerence of 42
(lg/m3) (paired t-test: t = 2.665, P < 0.01) and a low
correlation of 0.29 (P < 0.05). This trend is also found
in Panels 2 and 3. However, as the bias of PM mass
concentrations has been reported with the use of lightscattering instruments, the focus will be on the relative
change of PM concentrations between panels.

Table 3 Means and medians for particulate matter concentrations and indoor/outdoor ratios for all three panels. Panel 1 (N = 72) from 2010, Panel 2 (N = 55) from 2011, and Panel 3
(N = 53) from 2012
PM2.5

N = 72
N = 55
N = 53

214

Panel 1 Mean
Panel 1 Median
Panel 2 Mean
Panel 2 Median
Panel 3 Mean
Panel 3 Median

PM10

Indoor concentrations
(lg/m3)

Outdoor concentrations
(lg/m3)

Indoor/Outdoor
ratio

Indoor concentrations
(lg/m3)

Outdoor concentrations
(lg/m3)

Indoor/Outdoor
ratio

58  125
13
67  145
20
37  87
19

20  21
13
17  13
13
10  5
10

3.0
1.1
2.8
1.6
2.9
1.9

62  125
18
74  146
25
41  87
22

24  21
17
26  18
19
16  7
15

2.5
1.0
2.9
1.5
2.2
1.5

Indoor air quality and energy eﬃciency retroﬁt
Cumulative frequency (%)

100
90
80
70
60

Panel 1

50

Panel 2

40
Panel 3

30
20
10
0
1

Fig. 1 Cumulative frequency plot of
PM_10 from Panel 1 (solid), Panel 2
(dot), and Panel 3 (dash)

10

100

1000

PM10 concentration (μg/m³)
[log scale]

Cumulative frequency (%)

100
90
80
70
Panel 1

60
50

Panel 2

40

Panel 3

30
20
10
0

Fig. 2 Cumulative frequency plot of
PM_2.5 from Panel 1 (solid), Panel 2
(dot), and Panel 3 (dash)

1

10

100

1000

PM2.5 concentration (μg/m³)
[log scale]

PM2.5 t = 6.592, P < 0.001, I/O PM10 t = 6.957,
P < 0.001). However, there was no signiﬁcant change
in the short term when smoking was added as a covariate.
In the long term (between Panels 1 and 3), the resident’s length of stay, resident’s age, and whether the
resident smoked were statistically signiﬁcant covariates. Contrasting to the short-term changes, as a resident’s length of stay increased, a decrease in long-term
PM2.5 and PM10 concentrations was identiﬁed (PM2.5
t = 1.865, P = 0.065; PM10 t = 1.897, P = 0.061).
This change was not reﬂected in the indoor/outdoor

ratios. When compared to the increasing age of a
resident, both long-term PM concentrations and
indoor/outdoor ratios decreased (PM2.5 t = 2.214,
P = 0.029; PM10 t = 2.151, P = 0.034 and I/O PM2.5
t = 2.151, P = 0.034; I/O PM10 t = 1.929,
P = 0.057, respectively). Finally, units occupied by residents who smoke had higher PM levels than units with
non-smoking residents (PM2.5 t = 6.186, P < 0.001;
PM10 t = 6.161, P < 0.001). The higher PM concentrations measured in units with smokers have a signiﬁcant
decrease compared with non-smokers (PM2.5
t = 3.078, P < 0.001; PM10 t = 3.059, P < 0.003).

Table 4 Means and medians for PM2.5 concentrations and indoor/outdoor ratios for all three panels, separated by smoking units and non-smoking units: Panel 1 (N = 16, N = 56) from
2010, Panel 2 (N = 9, N = 44) from 2011, and Panel 3 (N = 11, N = 42) from 2012
Smoking PM2.5

Panel 1 Mean
Panel 1 Median
Panel 2 Mean
Panel 2 Median
Panel 3 Mean
Panel 3 Median

Non-smoking PM2.5

Indoor concentrations
(lg/m3)

Outdoor concentrations
(lg/m3)

Indoor/Outdoor
ratio

Indoor concentrations
(lg/m3)

Outdoor concentrations
(lg/m3)

Indoor/Outdoor
ratio

209  232
99
361  430
257
82  173
25

24  18
20
28  20
51
13  6
10

8.5
4.6
12
5
4.3
2.5

20  38
12
22  14
19
25  41
16

19  22
12
15  11
12
10  5
10

1.4
1
2.2
1.6
2.5
1.9

215

Frey et al.
Table 5 Mean and median concentrations of acetone, acetaldehyde, and formaldehyde for each Panel
Acetone

N = 72
N = 54
N = 55

Panel 1 Mean
Panel 1 Median
Panel 2 Mean
Panel 2 Median
Panel 3 Mean
Panel 3 Median

Acetaldehyde

Formaldehyde

Indoor concentrations 
standard deviation (ppb)

Indoor Outdoor
Ratio

Indoor concentrations 
standard deviation (ppb)

Indoor/Outdoor
Ratio

Indoor concentrations 
standard deviation (ppb)

Indoor/Outdoor
Ratio

41  41
28
91  45
90
52  42
39

8.7
8.1
14
11
11
9.7

20  9
18
34  17
33
20  7
20

11
9.8
13
10
10
9.1

39  11
38
42  13
43
27  7
26

8.7
7.7
9.5
7.1
7.1
6.8

Carbonyl measurements

Table 5 summarizes acetone, acetaldehyde, and formaldehyde concentrations and indoor/outdoor ratios for
each panel. We also illustrate the cumulative frequencies of the formaldehyde and acetaldehyde data in
Figures 3 and 4. Figure 3 also includes reference lines
for the most recent California 8-h Reference Exposure
Level (REL), the Health Canada 8-h REL, and the
California acute REL.
As seen in Figure 3, 100% of samples in all three
panels, with levels ranging from 17 to 69 ppb, exceeded
the California EPA 8-h REL of 7 ppb. When compared to the CA acute REL standard of 44 ppb, 32%
of Panel 1 and 43% of Panel 2 units were above the
standard. Additionally, 40% of Panel 1 samples, 56%
of Panel 2 samples and only 4% of Panel 3 samples
exceeded the Health Canada REL of 40 ppb. However,
no unit exceeded the WHO guideline of 80 ppb. These
formaldehyde levels are comparable or higher than
those reported for US buildings and homes. Hun et al.,
(2010) reported a mean and median formaldehyde concentration of 17 ppb in 179 US residences. Oﬀermann
(2009) determined a median of 29 ppb formaldehyde in
new homes, and Hodgson and Levin (2003) reported a
median formaldehyde level of 17 ppb, with a 95
percentile of 61 ppb. Similar residential formaldehyde
levels have been reported in other countries, with a
mean of 18 ppb in the UK (n = 833, 1997–1999),

19 ppb in Germany (n = 586, 2003–2006), 33 ppb in
Finland, 20 ppb in Austria (n = 160), and 25 ppb
in Japan (n = 1181, 2005), and a median of 16 ppb in
France (n = 554, 2003–2005) and 24 ppb in Canada
(n = 96, 2005) (WHO, 2010). By contrast, very high
levels have been reported in recently remodeled Chinese homes, with a mean of 190 ppb (n = 6000, 1999–
2006) (WHO, 2010), and in trailers supplied by the US
Federal Emergency Management Agency (FEMA) to
shelter evacuees from Hurricanes Katrina and Rita in
2005, with a mean formaldehyde concentration of
77 ppb (n = 519) (Murphy et al., 2013).
All measured acetaldehyde levels, reported in Figure 4, were below the health-based exposure levels
recommended by the California EPA (8-h
REL = 160 ppb and 1-h REL = 260 ppb). Acetone
levels measured in this study do not pose any health
hazards, as the National Institute for Occupational
Safety and Health (NIOSH) recommended exposure
limit (REL) is 250 ppm as a time-weighted average for
up to 10-h work shift over a 40-h work week (NIOSH,
1994). Acetone levels are included in this report, even
though levels are far below health guidelines, to illustrate the behavior of a common indoor VOC generated
by sources predominantly related to human activities.
As can be seen in Table 5, long-term changes were
notably diﬀerent from those of the short term for formaldehyde. While there was no change in the short term,
there is a signiﬁcant decrease in the long term

Fig. 3 Cumulative frequency plot of
formaldehyde concentrations from
Panel 1 (solid), Panel 2 (dot), and
Panel 3 (dash)

216

Indoor air quality and energy eﬃciency retroﬁt

Fig. 4 Cumulative frequency plot of
acetaldehyde concentrations from
Panel 1 (solid), Panel 2 (dot), and
Panel 3 (dash)

(t = 6.376, P < 0.001). This decrease held after
controlling for most of the mediating building characteristics (orientation, wing, ﬂoor) and other covariates.
While older residents had higher formaldehyde
concentrations in their apartments, there was no
change between Panels 1 and 3.
Interestingly, acetaldehyde and acetone did not
follow the same trends as formaldehyde. In contrast,
both acetone and acetaldehyde had a statistically
signiﬁcant increase in the short term (t = 5.928,
P < 0.001 and t = 4.924, P < 0.001, respectively), even
after controlling for mediating factors. Additionally,
residents who have lived longer at the residence or
began using odor-masking products had a higher
increase in acetaldehyde concentrations in their homes
(t = 2.180, P = 0.031 and t = 1.934, P = 0.056, respectively) while those who stopped using indoor insecticide saw a decrease (t = 2.483, P = 0.015). In the
long term, neither chemical experienced a change from
panel 1, although acetaldehyde concentrations were
higher in units where residents indicated they smoked
(t = 5.290, P < 0.001).
Correspondence between improvements in air quality and reported
health

Given the relatively brief scope of this study and lack
of a control group, we did not expect to ﬁnd deﬁnite
changes in health conditions after the retroﬁt, particularly among the more serious medical and health
diagnoses. Nonetheless, we did examine whether the
signiﬁcant decrease in formaldehyde levels in the

long-term measurements also resulted in improved
health conditions. For example, an index of respiratory
conditions, an index of quality of life/health, emotional
distress, and number of hours sleeping during the night
are conditions that may be responsive to improved
indoor air quality.
As shown in Table 6, diﬀerences between individual
unit formaldehyde concentrations are associated with
self-reported health conditions, particularly in the
short term. Using ﬁxed-eﬀects regression of data
between Panels 1 and 2, changes in formaldehyde concentrations were correlated with residents’ reported
quality of life/health and reduction in emotional distress. That is, as the formaldehyde levels in one’s apartment improved (i.e. declined in concentration levels),
residents expressed greater satisfaction with their quality of life/health and less emotional distress. Between
Panels 1 and 3, formaldehyde change correlated with
reduced emotional distress scores, but only at marginal
statistical signiﬁcance possibly because all formaldehyde concentrations showed a signiﬁcant decrease in
Panel 3. These ﬁndings between formaldehyde reductions and emotional distress improvements are suggestive only. The larger study (Ahrentzen et al., 2013)
noted correspondence between emotional improvements and other environmental improvements (e.g.
temperature) and physiological factors (e.g. functional
limitations). Multivariate modeling to examine interrelationships between these variables was not conducted
because of the small sample size. However, given the
prominent improvement in emotional distress in the
larger study, future research with larger sample sizes

Table 6 Fixed-effects regression of formaldehyde change in unit and resident reported emotional distress and life/health quality, both short term and long term
Short term

Formaldehyde level in unit

Long term

Quality of health/life

Emotional distress

t

P-value

t

2.624

<0.01

3.912

Quality of health/life

Emotional distress

P-value

t

P-value

t

<0.001

1.257

n.s.

P-value
1.781

<0.08

217

Frey et al.
should examine the role of multiple environmental factors in improving mental health of older adults. There
were no signiﬁcant correlations between formaldehyde
changes and sleep or respiratory conditions.
Conclusions

The research presented here reports key indoor air
quality parameters, including PM levels and carbonyl
concentrations, for a low-income senior apartment
complex before and after an energy eﬃciency retroﬁt.
The air quality sampling was combined with a detailed
health questionnaire and educational material on
‘green’ and healthy homes. The questionnaire was used
to garner information on the personal habits and general health of residents. The educational booklet was
designed speciﬁcally for the residents at the apartment
complex and was distributed prior to the Panel 3 data
collection (Ahrentzen et al., 2013).
Although it was expected to have a short-term
increase in PM concentrations after the retroﬁt, this
was only statistically apparent with two covariates:
length of stay for the occupant and the use of odormasking products. In general, smokers had higher PM
concentrations in all three panels, but no short-term
change and a slight decrease in the long term. In the
long term, a decrease in PM concentrations occurred in
units with residents who had lived longer at the apartment complex.
The Panel 1 and Panel 2 formaldehyde levels measured in this study were comparable or higher than
other US buildings and homes, as described in the

literature. Panel 3 concentrations, however, showed a
signiﬁcant decrease with only 4% of units exceeding
the Health Canada 8-h REL of 40 ppb and virtually
none exceeding the California acute REL of 44 ppb.
The units tested here are much smaller than the
reported literature studies, so increased surface-to-volume ratios could be a factor in the increased concentrations. The signiﬁcant decrease in formaldehyde levels
in Panel 3 is most likely a result of the replacement of
building materials and furnishings during the retroﬁt.
This is supported by the fact that only formaldehyde,
but not acetaldehyde and acetone (which are measured
simultaneously with the same method), showed a signiﬁcant reduction in concentration. Changes in ventilation would have aﬀected all three carbonyls similarly.
Other factors, such as variations in the use of insecticide, were also not correlated with long-term changes
in carbonyl levels.
While both acetone and acetaldehyde concentrations
experienced an increase in the short term, long-term
concentrations were unchanged and well below any
deﬁned risk levels.
Acknowledgements

Funding was provided by the US Department of Housing and Urban Development. We gratefully acknowledge the assistance of the staﬀ of the City of Phoenix
Housing Department, Patrick Montgomery, Drew
Bryck, Ernesto Fonseca, John Ball, Kim Shea, William
Johnson, Mookesh Patel, and Amandine Montalbano.

References
Ahrentzen, S., Ball, J., Destaillats, H., Dwyer,
S., Erickson, J., Fonseca, E., Fraser, M.P.,
Frey, S.E., Johnson, W., Patel, M. and
Shea, K. (2013) The green apple research
project: health outcomes of a green housing retroﬁt for older adults in phoenix,
Arizona, Final report for HUD Oﬃce of
Healthy Homes and Lead Hazard Control.
Grant Number: AZLHH0200-09.
Allison, P. (2005) Fixed Eﬀects Regression
Methods for Longitudinal Data: Using
SAS, Cary, NC, SAS Institute.
California EPA (2007) Formaldehyde reference exposure levels, Public review draft.
http://www.oehha.ca.gov/air/hot_spots/
pdf/FormaldehydePR.pdf
Davidson, C.I., Phalen, R.F. and Solomon,
P.A. (2005) Airborne particulate matter
and human health: a review, Aerosol Sci.
Tech., 39, 737–749.
Destaillats, H., Lunden, M.M., Singer, B.C.,
Coleman, B.K., Hodgson, A.T., Weschler, C.J. and Nazaroﬀ, W.W. (2006)
Indoor secondary pollutants from household product emissions in the presence of

218

ozone. A bench scale study, Environ. Sci.
Technol., 40, 4421–4428.
Destaillats, H., Chen, W., Apte, M.G., Li,
N., Spears, M., Almosni, J., Brunner, G.,
Zhang, J. and Fisk, W.J. (2011) Secondary pollutants from ozone reactions with
ventilation ﬁlters and degradation of ﬁlter
media additives, Atmos. Environ., 45,
3561–3568.
Dockery, D.W., Pope, C.A. III, Xu, X.,
Spengler, J.D., Ware, J.H., Fay, M.E.,
Ferris, B.G. and Speizer, F.E. (1993) An
association between air pollution and
mortality in six US cities, N. Engl. J.
Med., 329, 1753–1759.
Englert, N. (2004) Fine particles and
human health – a review of epidemiological studies, Toxicol. Lett., 149,
235–242.
Fann, N., Lamson, A.D., Anenberg, S.C.,
Wesson, K., Risley, D. and Hubbell, B.J.
(2012) Estimating the national public
health burden associated with exposure
to ambient PM2. 5 and ozone, Risk Anal.,
32, 81–95.

Fisk, W.J. (2000) Health and productivity
gains from better indoor environments
and their relationship with building
energy eﬃciency, Annu. Rev. Energ. Env.,
25, 537–566.
Frey, S.E., Destaillats, H., Cohn, S., Ahrentzen, S. and Fraser, M.P. (2014)
Characterization of indoor air quality
and resident health in an Arizona
senior housing apartment building,
J. Air Waste Manage. Assoc., 64, 1251–
1259.
Health Canada (2006) Residential indoor air
quality guideline. http://www.hc-sc.gc.ca/
ewh-semt/alt_formats/hecs-sesc/pdf/
pubs/air/formaldehyde_e.pdf
Hodgson, A.T. and Levin, H. (2003) Classiﬁcation of measured indoor volatile
organic compounds based on noncancer
health and comfort considerations,
LBNL report # 53308. http://www.inive.
org/members area/medias/pdf/Inive/
LBL/LBNL-53308.pdf
Hodgson, A.T., Beal, D. and McIlyaine,
J.E.R. (2002) Sources of formaldehyde,

Indoor air quality and energy eﬃciency retroﬁt
other aldehydes and terpenes in a new
manufactured house, Indoor Air, 12,
235–242.
Hun, D.E., Corsi, R.L., Morandi, M.T. and
Siegel, J.A. (2010) Formaldehyde in residences: long-term indoor concentrations
and inﬂuencing factors, Indoor Air, 20,
196–203.
IARC. (2012) A review of human carcinogens: chemical agents and related occupations, IARC Monogr. Eval. Carcinog.
Risks Hum., 100F, 401–435.
Jenkins, R.A., Ilgner, R.H. and Tomkins,
B.A. (2004) Development and application
of protocols for the determination of
response of real-time particle monitors to
common indoor aerosols, J. Air Waste
Manage., 54, 229–241.
Jones, A.P. (1999) Indoor air quality
and health, Atmos. Environ., 33,
4535–4564.
Koenig, J.Q., Mar, T.F., Allen, R.W., Jansen, K., Lumley, T., Sullivan, J.H.,
Trenga, C.A., Larson, T.V. and Liu,
L.J.S. (2005) Pulmonary eﬀects of
indoor- and outdoor-generated particles
in children with asthma, Environ. Health
Persp., 113, 499–503.
Lawrence Berkeley National Laboratory
(LBNL). (2008) Indoor air quality scientiﬁc ﬁndings resource bank. http://www.
iaqscience.lbl.gov/voc-sensory.html
Lee, S.C., Li, W.-M. and Ao, C.-H. (2002)
Investigation of indoor air quality at residential homes in Hong Kong—case
study, Atmos. Environ., 36, 225–237.
Li, N., Hao, M.Q., Phalen, R.F., Hinds,
W.C. and Nel, A.E. (2003) Particulate air
pollutants and asthma—a paradigm for
the role of oxidative stress in PM-induced
adverse health eﬀects, Clin. Immunol.,
109, 250–265.
Logue, J.M., Price, P.N., Sherman, M.H.
and Singer, B.C. (2012) A method to estimate the chronic health impact of air pollutants in US residences, Environ. Health
Persp., 120, 216–222.
Murphy, M.W., Lando, J.F., Kieszak, S.M.,
Sutter, M.E., Noonan, G.P., Brunkard,
J.M. and McGeehin, M.A. (2013) Formaldehyde levels in FEMA-supplied travel
trailers, park models, and mobile homes
in Louisiana and Mississippi, Indoor Air,
23, 134–141.
NIOSH (1994) Documentation for
immediately dangerous to life or health

concentrations (IDLHs): acetone.
http://www.cdc.gov/niosh/idlh/67641.
HTML
Noris, F., Adamkiewicz, G., Delp, W.W.,
Hotchi, T., Russell, M., Singer, B.C.,
Spears, M., Vermeer, K. and Fisk, W.J.
(2013) Indoor environmental quality beneﬁts of apartment energy retroﬁts, Build.
Environ., 68, 170–178.
Oﬀermann, F.J. (2009) Ventilation and
Indoor Air Quality in New Homes. California Air Resources Board and California Energy Commission, PIER Energy
Related Environmental Research Program, Collaborative Report. CEC 500
2009 085. http://www.arb.ca.gov/
research/apr/past/04-310.pdf
Pilkonis, P.A., Choi, S.W., Reise, S.P.,
Stover, A.M., Riley, W.T. and Cella,
D. (2011) Item banks for measuring
emotional distress from the PatientReported Outcomes Measurement
Information System: depression, anxiety, and anger, Assessment, 18,
263–283.
Pope, C.A. III and Dockery, D.W. (2006)
Health eﬀects of ﬁne particulate air pollution: lines that connect, J. Air Waste
Manage. Assoc., 56, 709–742.
Salthammer, T., Mentese, S. and Marutzky,
R. (2010) Formaldehyde in the indoor
environment, Chem. Rev., 110,
2536–2572.
Sidheswaran, M., Chen, W., Chang, A.,
Miller, R., Cohn, S., Sullivan, D., Fisk,
W.J., Kumagai, K. and Destaillats, H.
(2013) Formaldehyde emissions from
ventilation ﬁlters under diﬀerent relative
humidity conditions, Environ. Sci. Technol., 47, 5336–5343.
Singer, B.C., Destaillats, H., Hodgson, A.T.
and Nazaroﬀ, W.W. (2006) Cleaning
products and air fresheners: emissions
and resulting concentrations of glycol
ethers and terpenoids, Indoor Air, 16,
179–191.
U.S. Department of Housing and Urban
Development (USDHUD) (2009) Leading Our Nation to Healthier Homes,
Washington, D.C., U.S. Dept. of Housing and Urban Development, Oﬃce of
Healthy Homes and Lead Hazard
Control.
U.S. EPA (1989) Report to Congress on
Indoor Air Quality, Volume II: Assessment and Control of Indoor Air Pollution,

Washington DC, Oﬃce of Air and Radiation. EPA 400-1-89-001C.
US EPA (1999a) Integrated Risk Information
System (IRIS) on Formaldehyde.
National Center for Environmental Assessment, Washington, DC, Oﬃce of
Research and Development.
US EPA (1999b) Compendium Method
TO-11A – Determination of Formaldehyde in Ambient Air Using Adsorbent
Cartridge Followed by HPLC [Active
Sampling Methodology], Cincinnati,
OH, Oﬃce of Research and Development – US Environmental Protection
Agency.
Wallace, L., Williams, R., Rea, A. and Croghan, C. (2006) Continuous weeklong
measurements of personal exposures and
indoor concentrations of ﬁne particles for
37 health-impaired North Carolina residents for up to four seasons, Atmos. Environ., 40, 399–414.
Wallace, L.A., Wheeler, A.J., Kearney, J.,
Ryswyk, K.V., You, H., Kulka, R.H.,
Rasmussen, P.E., Brook, J.R. and Xu, X.
(2011) Validation of continuous particle
monitors for personal, indoor, and outdoor exposures, J. Expo. Sci. Environ.
Epidemiol., 21, 49–64.
Wargo, J. (2010) LEED Certiﬁcation: Where
Energy Eﬃciency Collides with Human
Health, New Haven, CT, Environment+Human Health, Inc.
Weschler, C.J. and Shields, H.C. (2000) The
inﬂuence of ventilation on reactions
among indoor pollutants: modeling and
experimental observations, Indoor Air,
10, 92–100.
WHO (2010). WHO Guidelines for Indoor
Air Quality: Selected Pollutants,
Copenhagen, Denmark, World Health
Organization Regional Oﬃce for Europe, ISBN 978 92 890 0213 4. http://
www.euro.who.int/__data/assets/pdf_ﬁle/
0009/128169/e94535.pdf
Williams, R., Suggs, J., Creason, J., Rodes,
R., Lawless, P., Kwok, R., Zweidinger,
R. and Sheldon, L. (2000) The 1998 Baltimore particulate matter epidemiologyexposure study: part 2-personal exposure
assessment associated with an elderly
study population, J. Expo. Anal. Env.
Epid., 10, 533–543.

219