This article was downloaded by: [Monash University Library]

On: 01 February 2015, At: 11:59

Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Archives of Environmental Health: An International

Journal
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/vzeh20

Self-Reported Short- and Long-Term Respiratory Effects

among PVC-Exposed Firefighters
a b
Dr. Jeffrey S. Markowitz P.H.
a
Clinical and Genetics Epidemiology Unit , New York State Psychiatric Institute , USA
b
Public Health , School of Public Health Columbia University , New York, USA
Published online: 03 Aug 2010.

To cite this article: Dr. Jeffrey S. Markowitz P.H. (1989) Self-Reported Short- and Long-Term Respiratory Effects
among PVC-Exposed Firefighters, Archives of Environmental Health: An International Journal, 44:1, 30-33, DOI:
10.1080/00039896.1989.9935869

To link to this article: http://dx.doi.org/10.1080/00039896.1989.9935869

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Self-Reported Short- and Long-Term Respiratory Effects
among PVC-Exposed Firefighters
Downloaded by [Monash University Library] at 11:59 01 February 2015

JEFFREY S. MARKOWITZ, Dr.P.H.

Epidemiologist
Clinical and Genetics Epidemiology Unit
New York State Psychiatric Institute
and
Adjunct Assistant Professor
of Public Health
School of Public Health
Columbia University
New York

ABSTRACT. Firefighters exposed to burning polyvinyl chloride (PVC) were studied to assess
respiratory effects at 5-6 wk post-incident and again 22 mo following the fire. Exposed
subjects reported significantly more frequent and severe respiratory symptoms at both
time points than did firefighter controls. In longitudinal analyses, a number of symptoms
persisted over time, and acute symptom scores were significantly correlated with chronic
scores. At Time 2, approximately 18% of exposed firefighters, compared with none of the
controls, reported that since the time of the PVC exposure, a physician had told them that
they had either asthma andfor bronchitis.

WIDESPREAD USE of plastics has resulted in a new
set of health risks to firefighters and other exposed indi-
viduals.' For example, when polyvinyl chloride (PVC)
burns, large quantities of hydrogen chloride (HCI) are
Hydrogen chloride is a strong acid that can
cause mucosal burns of the respiratory tract.4 In addi-
tion, HCI is capable of adsorbing onto tiny carbon par-
ticles, which can be inhaled and carried to the alve01i.~
Inhalation injury associated with PVC exposure is,
therefore, likely to be respiratory in nature. Respiratory
symptoms associated with PVC inhalation injury may
include tachypnea, cough, hoarseness, dyspnea, chest
painhightness, and ~ h e e z e . ~
Evidence linking PVC exposure to adverse respiratory
effects i s primarily derived from case and observational
studies, laboratory data, studies of medically treated in-
dividuals, and litigants. Few studies have used a control
group, and long-term respiratory effects data are lim-
ited to one cross-sectional study. Longitudinal study re-
sults that have assessed changes in respiratory effects
over time have never been reported.
A large warehouse fire began in Plainfield, New Jer-
sey, on March 20, 1985, and continued through the
next day. Initially, firefighters had no idea that the un-
occupied, burning warehouse was being used for the il-
legal storage of pure, bulk polyvinyl chloride (PVC)
scrap. Consequently, protective measures, like the use
of masks, were only ordered once the chemical nature
of the fire became known.' This resulted in additional
chemical exposure among the firefighters.
Approximitaley 1 000 community residents were
evacuated from their homes, and 28 firefighters re-
ported to a local hospital with symptoms associated
with the fire. Two-thirds of these firefighters had abnor-
mal pulmonary function tests.6 Chemical analvses of
debris specimens taken directly from the scene of the
fire showed that HCI was a significant combustion
product of the fire.7

30 Archives of Environmental Health

 Using a longitudinal study design and a control group
of nonexposed firefighters, this study was designed to
assess respiratory effects among a sample of firefighters
exposed to burning PVC at the Plainfield warehouse
fire. Exposed subjects were not limited to medically
treated individuals. Both short- and long-term respira-
tory effects were evaluated, as was the relationship be-
tween symptoms reported at the two time points.
Methods
Plainfield and North Plainfield firefighters who fought
the fire on the first day, Wednesday March 20, 1985,
and/or the following day, were designated as “exposed
subjects.” Firefighters from the same two companies
who did not fight the warehouse fire were designated
as “nonexposed controls.”
All 125 firefighters of the Plainfield and North Plain-
field fire departments were asked to volunteer to par-
ticipate in the Time 1 study. Ninety-five questionnaires
Downloaded by [Monash University Library] at 11:59 01 February 2015
(76%) were completed (75% or 78/104 in Plainfield and
80% or 17/21 in North Plainfield) between April 15,
1985, and May 22, 1985.
Approximately 22 months after the PVC warehouse
fire, all 101 members of the Plainfield Fire Division only
were asked to participate in the Time 2 component of
the study. Attrition associated with a 22-mo follow-up
due to retirement or reassignment was expected to be
minimal.
A total of 86 men completed questionnaires at Time
2, including 64 exposed firefighters and 22 nonexposed
controls. Thus, the Time 2 response rate was 85% Ke.,
86/101 = 85,1%). Administrative approval to survey
North Plainfield firefighters a second time could not be
obtained. Among Plainfield firefighters, attrition be-
tween time points was random rather than systematic.
A total of 55 exposed firefighters from Plainfield were
studied at both time points. Analyses of matched pairs
of firefighter controls were not conducted because they
numbered only 8.
A total of four items pertaining to cough, wheeze,
shortness of breath, and chest pains were included as
part of the self-report assessment battery at both time
points. The assessment time interval for both subjects
and controls at Time 1 was ”since the Plainfield ware-
house fire”; at Time 2, “during the past month” was
used to preface each of the items. Each of the four
items was rated on a four-point scale from absent or
zero to present and severe or three. Mildly severe
symptoms were scored as one, and moderately severe
symptoms were scored as two. In addition, two scales
were formed (one for each time point), which were the
sum of the four item values. The two “total” scales had
good internal consistency reliability, as was reflected
by Cronbach alpha coefficients8 of 0,83 at Time 1 and
0,73 at Time 2.
Only at Time 2 were the firefighters asked whether a
physician had informed them since the time of the
warehouse fire that they had asthma. A second item
about physician-diagnosed bronchitis was similarly in-
cluded. Also at Time 2, four questions pertaining to
phlegm and sputum production were asked and formed
a phlegm scale which was internally consistent (alpha =
JanuarylFebruary1989 (Vol. 44 (No. l ) ]
0,911. This scale had an assessment time interval of
“usually” and was adapted from the American Thoracic
Society and the Epidemiology Standardization P r ~ j e c t . ~
The items tapped into bringing up phlegm from the
chest, number of times, and whether it occurred upon
getting up in the morning and/or during the rest of the
day. The total phlegm scale score was an unweighted
count of the items endorsed as positive and could range
from zero to four.
Ideally, complete physical examinations and labora-
tory and other tests should be collected. However, this
significantly increases the costs of any study and may or
may not be necessary. Self-reports can be used as a
”barometer” to obtain quickly and inexpensively an
overall picture of a group of exposed individuals. Self-
reports can provide insights regarding whether more
extensive, clinical examinations are required.
It was predicted that exposed firefighter subjects
would report more respiratory symptoms at both time
points compared with controls. The author expected a
reduction in respiratory symptomatology between time
points and a significant positive correlation between
symptom reports over time.
The response categories for the four respiratory
symptom items used at both time points were ordered.
In addition, the frequency distributions of these items
were not distributed normally. Consequently, nonpara-
metric tests of significance were used. Mann-Whitney
tests were employed to compare subjects and controls
on individual symptom reports and total respiratory
scores. To assess changes in symptom reports over time
among matched pairs of subjects, Wilcoxon matched-
pairs signed-ranked tests were used. The correlation
between Time 1 and Time 2 symptom scores were ana-
lyzed using Spearman’s rank correlation coefficient
(rho). These three procedures have been detailed by
Siegel.” Chi-square was used to compare rates of
physician-diagnosed asthma and bronchitis between
Time 2 subjects and controls. An independent t test
was used to assess differences between the subject and
control group on the phlegm scale. All tests of signifi-
cance were two-tailed, and an alpha level of 0,05 was
used.
Results
At Time 1, the firefighters ranged in age from 21 to 57
y (mean = 37,2 y; standard deviation [SD] = 8,8). The
firefighters had worked as members of their respective
fire departments an average of 12,3 y (SD = 8,O); ap-
proximately 28% had < 5 y of firefighting experience.
Seventy percent were white, and approximately 27%
were black. The firefighters had completed an average
of 12,6 y of school. More than 68% were currently mar-
ried. All of the firefighters were male, and 40% currently
smoked cigarettes.
At Time 1, exposed firefighter subjects scored signifi-
cantly higher than nonexposed controls on cough,
wheeze, shortness of breath, chest pains, and the total
respiratory symptom score. With the exception of
wheezing, all Time 2 symptom scores plus the total
score were again highly significant. At Time 2, cough-
31
 ing was reported more frequently by exposed subjects
as was shortness of breath and chest pains. These re-
sults are detailed in Table 1.
Among exposed subjects, the incidence of chest
pains dropped significantly between time points, as did
the total scores. Also, there was a trend for cough
scores to drop between time points; however, wheeze
and shortness of breath did not drop significantly be-
tween time points.
Among exposed firefighter subjects, all four Time 1
respiratory symptoms were significantly correlated with
their respective Time 2 item counterparts, with coeffi-
cients ranging from 0,27 on cough to 0,53 for chest
pains. Spearman's rho for wheeze was 0,33 and 0,46
for shortness of breath. The total scores at the two time
points were also significantly correlated (rho = 0,6).
Thus, high respiratory symptom scores at Time 1 meant
higher scores at Time 2 and vice versa.
On average, nonexposed controls at Time 2 endorsed
Downloaded by [Monash University Library] at 11:59 01 February 2015
jects (14,3%) were positive for bronchitis. In all, 12 of
64 firefighter subjects (18%)had been told by a doctor
that they had either asthma and/or bronchitis since the
time of the Plainfield warehouse fire. None of the 22
controls had been told the same. The difference in
these proportions was nearly statistically significant
(corrected chi-square = 3/36; p = 0,067).
In a final set of analyses, exposed firefighter subjects
were grouped into one of three categories as follows:
(1) never smoked cigarettes, (2) ever smoked cigarettes
but not currently, and (3) currently smokes cigarettes.
At both Time 1 and Time 2, there was no significant dif-
ference among these groups with respect to the total
respiratory score. Furthermore, neither the interaction
terms of smoking status by age (under vs. over 40 y) nor
smoking status by years of experience (under or over
10 y) was significantly related to total respiratory
scores. Finally, among those who currently smoked,
there was no relationship between number of cigarettes

about one phlegm item (of four), whereas the exposed
subjects averaged more than 1,8. An independent t test
determined that the difference in the mean total phlegm
score between subjects and controls approached statisti-
cal significance (p < 0,07). None of the individual items
within the phlegm scale were significant, but p values for
two of the four items were in the 0,l range.
Six of 64 Time-2 firefighter subjects (9,4%) reported
that since the time of the warehouse fire, a doctor had
told them they had asthma, and 9 of 63 firefighter sub-
smoked per day and the total respiratory scores at Time
1 or Time 2 (r = 0,14 and r = -O,ll, respectively).
Discussion
In this study, random assignment could not be con-
sidered in order to allocate subjects and controls to the
exposed and nonexposed groups, Consequently, there
could have been inequivalencies for both measured
and unmeasured baseline variables. Any variables
where between-group baseline differences existed

Table 1.-Comparison of Subjects and Controls on Short- and long-Term Respiratory Symptoms

Time 1 Time 2
Subjects Controls Mann- Subjects Controls Mann-
N % N % Whitney N 010 N % Whitney

Cough
Absent 28 36 11 79 33 52 18 82
Mild 22 28 3 21 12 19 1 4
Moderate 21 27 0 0 3,2t 13 20 2 9 2,3$
Severe 7 9 0 0 6 9 1 4
Totals' 78 100 14 100 64 100 22 100
Wheeze
Absent 47 61 14 100 46 72 20 91
Mild 15 20 0 0 8 13 0 0
Moderate 15 20 0 0 2.85 7 11 1 4 1,7
Severe 0 0 0 0 3 5 1 4
Totals' 77 100 14 100 64 100 22 100
Shortness of breath
Absent 46 58 13 100 41 64 21 96
Mild 19 24 0 0 12 19 0 0
Moderate 11 14 0 0 235 9 14 1 4 2,7s
Severe 3 4 0 0 2 3 0 0
Totals' 79 100 13 100 64 100 22 100
Chest pains
Absent 46 58 13 93 41 64 21 96
Mild 16 20 1 7 12 19 0 0
Moderate 13 17 0 0 2.55 10 16 0 0 2,6§
Severe 4 5 0 0 1 2 1 4
Totals' 79 100 14 100 64 100 22 100

*Ns fluctuate slightly due t o missing data. Percentages are rounded

tp < 0,001.
Sp < 0,05.
fjp < 0,Ol.

32 Archives of Environmental Health

 could potentially result in confounding between ex-
posure status and outcome scores. At Time 2, the
groups were not found to be significantly different on
any baseline characteristics. At Time 1, there were
several baseline differences between subjects and con-
trols, e.g., marital status (more controls were married),
cigarette smoking (more controls smoked), and years of
firefighting experience (more experience among con-
trols). Nevertheless, in post hoc regression analyses,
controlling for these Time 1 baseline equivalencies did
not reduce or diminish the effects presented in this
paper.
In future studies, self-report symptom data could be
used in conjunction with medical examination, x-ray,
and laboratory data, if possible. This is required to help
validate the self-reports, which could be subject to a
number of sources of error variance.
The fact that Time 1 and Time 2 respiratory symptom
scores tended to correlate has potential intervention
Downloaded by [Monash University Library] at 11:59 01 February 2015
implications. Firefighters who scored high on the vari-
ous Time 1 respiratory symptoms were the ones most
likely to have persisting, long-term problems. It is this
group of acutely distressed individuals who would,
therefore, be prime targets for medical treatment and
monitoring programs directed at reducing levels of
respiratory distress and dysfunction.
Since the time of the PVC warehouse fire, nearly one
of every five Time-2 firefighter subjects had been told
by a doctor that they had either asthma or bronchitis.
The wording of the asthma and bronchitis items (e.g.,
“Since the time of the Plainfield warehouse fire has a
doctor told you that you have asthma?”) may or may
not argue for these being new cases. Therefore, causal
claims that attribute these asthma and bronchitis cases
to the occurrence of the PVC Plainfield fire should be
avoided. In any event, these results were clinically
meaningful, biologically plausible, and consistent with
previous research results reported by Wallace in her
study of firefighters who fought the New York City PVC
telephone company fire in 1975.” In that study, 8,3%
had repetitive bronchitis and 6,9% had asthma. It is
recommended that all individuals exposed to PVC be
evaluated carefully for asthma and bronchitis. Further-
more, future studies in this area should pay careful at-
tention to the distinction between prevalence and
event-related incidence of these potentially serious
respiratory problems.
Many previous studies have shown a relationship be-
tween cigarette smoking and pulmonary abnormalities
among firefighters.I2-l4One study by Miller et aI.l5
demonstrated significant dimunition in air flow among
200 of 348 workers in a vinyl chloride (VC)/PVC poly-
merization plant. These pulmonary abnormalities were
associated with age and level of exposure. These prior
studies, however, did not focus on a specific exposure
event involving burning PVC. Only one small study of
21 Los Angeles firefighters exposed to burning PVC
reported that all 4 men with abnormal values of
FEV 75% 1 mo following the fire were cigarette smok-
ers.?‘The results of the present study, however, show
January/February1989 [Vol. 44 (No.l)]
no relationship between cigarette smoking and self-
reported respiratory symptoms-both short- and long-
term. Additional studies, with follow-up periods ex-
ceeding 2 y, are indicated to further examine this issue.
This study represents an additional piece of evi-
dence, which supports previous findings of invidious
health consequences of PVC fire exposure because it
uses a larger sample size, a nontreated subject group,
and a control group. This study builds on previous
work in this area because it used a longitudinal study
design and documented evidence of long-term respira-
tory distress among PVC-exposed firefighters. The ef-
fects of exposure to the combustion products of this
chemical warrant further attention.
**********
The computer costs of this study were funded through a grant to the
Computer Center of the New York State Psychiatric Institute from the
National Institute of Mental Health (MH-3090608).
I also wish to acknowledge the support and encouragement of Dr.
Elane Gutterman, Dr. Sheila Gorman, Dr. Bruce Link, and Talia Car-
son Markowitz of New York City. A special note of thanks to the
members of the Plainfield and North Plainfield fire departments, who
participated in this study, and the administrative supervision provided
by Captain Chet Puri, Firefighter Jan Brothers, and Mark Ciarlariello.
Submitted for publication February 10, 1988; revised; accepted for
publication June 22, 1988.
Requests for reprints should be sent to: Jeffrey S . Markowitz,
Dr.P.H., Clinical & Genetic Epidemiology, NYS Psychiatric Institute,
722 W. 168th St., New York, NY 10032.
**********
References
1. Eckardt RE, Hindin R. The health hazards of plastics. J o f c u p
Med 1973; 15:808.
2. Stone JP, Hanzlett RN, Johnson JE, Carhart HW. The transport of
hydrogen chloride by soot from burning polyvinyl chloride. J Fire
Flammability 1973;442.
3. Dyer RF, Exh VH. Polyvinyl chloride toxicity in fires: Hydrogen
chloride toxicity in fire fighters. JAMA 1976235:393.
4. Cohen MA. Inhalation of Droducts of combustion. Annals
Emergency Med 1983;12:628.
5. Puri Lt Chet. Unpublished report on the Plainfield polyvinyl
chloride warehouse fire, June 1985; pp. 8-9.
6. Kale S. Muhlenberg Hospital report on March 20, 1985 Plainfield
New Jersey PVC warehouse fire, p. 1.
7. Gollob Analytical Service Report No. 577356, June 14, 1985,
Berkeley Heights, New Jersey.
8. Cronbach LJ. Coefficient alpha and the internal structure of tests.
Psychometrika 1951;16:297.
9. Ferris BG. Epidemiology standardization project: American
Thoracic Society. Am Rev Respir Dis 197O(Suppl.):12.
10. Siege1 55. Non-parametric statistics. New York: McGraw-Hill,
1956.
11. Wallace D. Dangers of polyvinyl chloride wire insulation decom-
position. I. Long-term health impairments: studies of firefighters
of the 1975 New York Telephone Fire and the survivors of the
1977 Beverly Hills Supper Club Fire. Combustion Toxicol 1980;8:
205.
12. Sidor R, Peters JM. Prevalence rates of chronic non-specific
respiratory disease i n firefighters. Am Rev Respir Dis
1974;109:255.
13. Sidor R, Peters JM. Firefighters and pulmonary function. Am Rev
Respir Di5 1974;109:249.
14. Peters JM, Theriault GP, Fine LJ, et al. Chronic effect of fire
fighting on pulmonary function. N Engl J Med 1974;291:1320.
15. Miller A, Teirstein AS, Chuang M, Selikoff IJ. Changes in
pulmonary function in workers exposed to vinyl chloride and
polyvinyl chloride. Ann NY Acad Sci 1975;246:42.
16. Tashkin DP, Genovesi MG, Chopra S, Coulson A, Simmons M.
Respiratory status of Los Angeles firemen. Chest 1977;71:445.
33