United States
Environmental Protection
Agency
Health Effects Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S1-85/019 Sept. 1985
Tl
&EPA Project Summary
Investigation of Legionella
pneumophila in Drinking Water
Linden E. Withered, Robert W. Duncan, Kenneth M. Stone, Lori J. Stratton,
Lillian Orciari, Steven Kappel, and David A. Jillson
An investigation of Legionella pneu-
mophila in drinking water systems and
home plumbing appurtenances was
undertaken in two phases. In Phase 1,
68 water samples for L. pneumophila
analysis were collected from hot and
cold kitchen sink faucets in homes on
17 community water systems.
No L. pneumophila organisms were
isolated from any of the samples in
Phase 1. The relatively small sample
size, the many variables in the sampling
procedure, and potential limitations of
the laboratory detection techniques
employed may have contributed to the
failure of significant recovery.
I n Phase 1 A, the kitchen sink faucets/
aerators and showerheads/supply pipes
were sampled with sterile swabs and a
sample of hot water was collected from
the drains of domestic hot water heaters
in each home. A total of 184 samples
(92 swabs, 92 hot water samples) were
collected from homes on four commun-
ity water systems. In addition, two
samples (one swab, one hot water
sample) were collected from a home
with an individual, shallow, dug well.
Field analyses of pH, temperature,
turbidity, and chlorine residuals and
laboratory analyses for standard plate
count (SPC), total coliform, and iron
content were performed using Standard
Methods, 15th Edition. L. pneumophila
analyses were conducted using a modi-
fication of the Gorman and Feeley
Direct Plating method developed at the
Centers for Disease Control.
During Phase 1A, L. pneumophila
organisms were recovered from two
hot water heater flush samples; one
from a home having a community water
system and the other from a home
having an individual, shallow, dug well.
Both of the positive samples were taken
from electric, non-recirculating, hot
water heaters. This finding together
with the lack of recovery in the home
water distribution plumbing may indi-
cate that L. pneumophila may be trans-
ported from natural aquatic sources in
very low levels in drinking water sys-
tems to hot water heaters in homes
where the organisms may increase to
detectable levels.
The finding that approximately 2 of
the 59 (3.4%) electric, nonrecirculating,
domestic hot water heaters sampled
contain L. pneumophila organisms may
have public health significance. These
sources may serve as a foci of the
pathogen with resultant sporadic en-
demic legionellosis.
This report was submitted in fulfill-
ment of Cooperative Agreement No.
CR-810360 by the Vermont Depart-
ment of Health under the partial spon-
sorship of the U.S. Environmental Pro-
tection Agency. This report covers a
period from November 1982 to March
1984, and work was completed as of
March 1985.
This Project Summary was developed
by EPA's Health Effects Research Lab-
oratory, Research Triangle Park. NC, to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction/Background
The investigation by the Centers for
Disease Control (CDC) of the mysterious
outbreak of pneumonia at the July 1976
annual convention of the Pennsylvania
-------
Department of the American Legion led to
the discovery of the pathogen, Legionella
pneumophila. Investigations of subse-
quent outbreaks, together with review of
stored samples from previous unex-
plained similar outbreaks, have added to
knowledge of the pathogen and its role in
human disease.
Acute human infection caused by L.
pneumophila is now described as legion-
ellosis. Legionellosis is presently recog-
nized with two distinct clinical disease
presentations: Legionnaires' disease and
Pontiac fever.
Both Legionnaires' disease and Pontiac
fever appear to be non-communicable;
person-to-person transmission has not
been documented. The only documented
reservoir for L. pneumophila, and similar
atypical L pnet/mop/i/'/a-like organisms,
is environmental. Epidemiological studies
show that the method of transmission
from environmental sources to suscept-
ibles is airborne with entry into the body
via the respiratory system.
L. pneumophila is a rod-shaped bacillus
measuring 0.3 to 0.4 /urn in width and
normally 2 to 3 /urn long and even 50 /urn in
length. Electron microscopy studies have
found the organisms to be atypical, gram
negative bacilli, with double membranes
and no cell walls. Some of the organisms
have flagella.
Laboratory studies show that the organ-
isms are extremely fastidious bacteria
with unusual growth requirements. An
interesting characteristic with possible
public health implications is the ability of
the L. pneumophila to persist in distilled
or tap water. The organisms have been
known to survive up to 139 days in
distilled water and for up to 369 days in
tap water. In fact, one report suggests
that the organism may multiply in sterile
water.
Ecological studies suggest that the
organism is part of the natural aquatic
environment and is capable of surviving
wide ranges of environmental conditions.
In one study of 793 water samples
collected from 67 different lakes and
rivers in the United States, virtually all
samples were found positive for L. pneu-
mophila when the direct fluorescent
antibody (DFA) technique was used for
detection. The water samples had the
following characteristics; pH 5.5-8.1,
temperature 5.7-63° C, dissolved oxygen
0.3-9.6 mg/l, and Secchi disk readings
1 -4 m. Chi square analysis of the data
found a significantly greater isolation
efficiency in the 36-60° C range. These
findings suggest that L pneumophila may
favor warm aquatic environments.
Another ecological study reported the
isolation of L. pneumophila from an algal-
bacterial mat community growing at45°C
in association with a blue-green algae
over a pH range of 6.9-7.6. The research-
ers surmised that L. pneumophila was
using algal extracellular products as its
carbon and energy sources. These obser-
vations indicate that in this particular
environment of a man-made thermal
effluent, the temperature, pH, and nutri-
tional requirements were not as stringent
as those observed when the organism is
cultured in the laboratory on complex
media. The suggestion has been made in
the literature that the association be-
tween L pneumophila and certain blue-
green algae could explain the apparent
widespread distribution of the bacterium
in the natural environment.
Another possible explanation of the
widespread distribution of L pneumo-
phila is provided by the finding that the
organism can grow within freshwater
and soil amoebae. Because L. pneumo-
phila thrives within ubiquitous soil and
fresh water amoebae of the genera
Acanthamoeba andNaegleria, it has been
suggested that amoeba-associated L.
pneumophila, rather thanfree/.. pneu/Tio-
phila, could be the infective form for
humans.
Because L. pneumophila species are
neither new nor rare organisms in the
natural environment, it seems probable
that a low level of previously undetected
human infection with the organisms from
natural sources has occurred in the past
and will continue to occur in the future.
Reviews of human blood sera for DFA
positivity have in. some cases (Vermont
studies) revealed relatively high preva-
lence rates of 8-26% (using a titer of
1:128 as a criterion for positivity), indicat-
ing at least a widespread human exposure
to antigens consistent with those of L.
pneumophila. Reviews have also shown
seropositivity to L. pneumophila-Wna or-
ganisms among fatal and non-fatal clin-
ical cases whose probable sources were
aquatic (aspiration in a near-drowning,
contaminated SCUBA equipment). Addi-
tionally, examination of blood sera from a
large (2000) number of animals has
shown evidence of widespread exposure
in horses (>600 samples) to four sero-
groups of L. pneumophila (31.4% titer of
> 1:64 to at least one serogroup antigen).
As of September 30, 1979, 1005
confirmed cases of sporadic legionellosis
in U.S. residents had been reported to
CDC and 19% of these cases were fatal. In
fact, it is estimated that there are approx-
imately 25,000 sporadic cases of legion
ellosis in the U.S. every year. Howevei
outbreaks of legionellosis are relative!
new phenomena. Widespread media cov
erage in 1976 acquainted many with thi
"new" disease, yet careful histories
review revealed that the earliest outbrea
occurred in 1965.
A review of epidemiological investiga
tions of 17 outbreaks of legionellosi
through 1979 finds that most occur ii
buildings: hospitals, hotels, a golf coursi
clubhouse, and a factory were among thi
first sites. Construction and excavatioi
activities have also been implicated ii
some outbreaks. Cooling towers am
evaporative condensers have been as
sociated with nine outbreaks. A history o
recent travel or attendance at a conven
tion were also factors associated witl
legionellosis. Two outbreaks of legion
ellosis in Vermont in 1980 were assoc
iated with a hospital, a medical school
and a cooling tower.
To date, no known outbreak of legion
ellosis has been directly associated with <
natural environment such as a lake, pond
or stream. Apparently, in outbreak situa
tions, human activities or man-made
facilities which concentrate and/or dis
seminate L. pneumophila organisms ir
sufficient numbers and in a respirabk
form, are responsible for large numbers
of disease cases. A review of the environ-
mental aspects associated with pas:
legionellosis outbreaks may suggest ways
to prevent future occurrences.
The role of buildings in legionellosis
outbreaks is not clear. Certainly, buildings
are necessary for bringing together large
numbers of people, a critical requiremem
for an outbreak. There is close association
between travel and/or conventions anc
buildings. Both of these activities usually
require the use of buildings such as
hotels and convention centers. In hospi-
tals, of course, there are concentrations
of people who are ill, and consequently,
more at risk of contracting diseases such
as legionellosis. However, not all build-
ings have been involved in outbreaks nor
are all victims of Legionnaires' disease
hospitalized. In addition to the recognition
of buildings as areas of concentrations of
people, consideration must be given to
activities and/or facilities in and near
buildings such as construction, cooling
towers, and water distribution systems,
which have been associated with legion-
ellosis outbreaks.
Another environmental source of L.
pneumophila organisms associated with
nosocomial legionellosis is water distri-
-------
bution systems in hospitals. The literature
notes that the two largest sustained
outbreaks of nosocomial Legionnaires'
disease have been at hospitals in Los
Angeles and Pittsburgh where/., pneumo-
phila organisms have been isolated from
the potable water supplies of each facility.
Where L pneumophila organisms have
been isolated from hospital potable water
distribution systems, the investigations
were prompted by occurrence of noso-
comial Legionnaires' disease cases.
The recovery of Legionella organisms
in hospital water systems may have
important public health significance be-
cause of the possible generation of con-
taminated aerosols from showerheads in
buildings containing large numbers of
immuno-compromised hosts. A possible
explanation for the occurrence of L.
pneumophila organisms in hospital water
systems is the normal low operating
temperature (43 to 49°C) of the hot water
used in such systems. This temperature,
used for patient safety, falls within the
environmental temperature range at
which optimal isolation is achieved.
Worthy of note is the fact that many other
hot water systems are now being oper-
ated at similar temperatures for energy
conservation reasons.
L. pneumophila organisms have also
been recovered from hot water systems
in other institutions and homes. While
the existence of these bacteria in other
plumbing systems is not necessarily
associated with an outbreak of disease,
they could be a reservoir and means of
transmission of legionellosis.
Legionella in drinking water systems
has also been investigated. In a study of
three municipal drinking water reservoirs,
it was concluded that Legionella is either
absent or present in undetectable num-
bers in these waters. In another study, 5
of 856 samples taken from chlorinated
public water supplies yielded Legionella
isolates.
Water is an interesting and perhaps
very important element common to many
of the ecological findings regarding L.
pneumophila organisms and environ-
mental health investigations involved
with legionellosis outbreaks. L. pneumo-
phila organisms are found in water in the
natural environment and in water in
cooling towers, evaporative condensers,
hospital shower heads, and whirlpools
associated with legionellosis disease
outbreaks. I n fact, except for one report of
the isolation of L pneumophila organisms
from riparian mud of a thermally polluted
stream, there are no known isolations
from natural environmental sources other
than water.
Water systems may play a role in the
transmission of L. pneumophila organ-
isms which cause legionellosis. L. pneumo-
phila organisms common to the natural
aquatic environment may be taken into
water systems and delivered to devices
such as cooling towers/evaporative con-
densers, whirlpools, showerheads, etc. In
these water-using devices, L. pneumo-
phila organisms may increase in numbers
due to mechanical concentration or
growth resulting from favorable environ-
mental conditions such as elevated tem-
perature, existence of algae, amoeba,
and/or other factors not yet clearly
understood. Such water-using devices
generate aerosols which may contain L.
pneumophila organisms possibly trans-
mitted by airborne hosts entering the
respiratory system. This scenario may
explain the method of transmission of L.
pneumophila organisms from the natural
environment to the human host not only
for outbreak situations but also for the
estimated large number of sporadic cases
which occur each year.
Obviously, more information concern-
ing L. pneumophila in drinking water
systems is required, especially the meth-
od of determining the presence of L.
pneumophila organisms in drinking wat-
er. The methods commonly used for
determining microbiological quality of
drinking water are not suitable for deter-
mining the presence of L. pneumophila
organisms. For example, a , study by the
Vermont Department of Health hasfound
no correlation between coliform and/or
standard plate count results and the
presence of L. pneumophila in water in
cooling towers.
The literature identifies studies which
determine the presence of L. pneumo-
phila in drinking water using the DFA
analysis technique. This method is lim-
ited, however, in that there is cross
reaction between L. pneumophila and at
least two Pseudomonas strains with the
DFA procedure. Recent attempts to over-
come another problem with DFA (staining
of non-viable cells) by measuring dye
uptake of the electron transport system
(DFA-INT) did not overcome the problems
of cross reactivity between L. pneumo-
phila and Pseudomonas.
In view of the limitations of the DFA
method, culture methods of analysis have
been developed. The Gorman-Feeley cul-
ture method using synthetic media was
developed at CDC. A recent study compar-
ing the GPP method with direct plating on
synthetic media have shown significantly
greater sensitivity of the plating method
(p<0.01) for environmental samples.
The mere detection of L. pneumophila
in a drinking water system, in itself,
provides little information concerning the
public health significance. Epidemiolog-
ical studies are required to determine the
incidence of legionellosis on systems with
and without L. pneumophila and other
Legionella organisms. The present study
was undertaken in an attempt to deter-
mine the prevalence of Legionella organ-
isms in household plumbing fixtures and
distribution systems. If a high prevalence
rate was found, then an epidemiological
study would be conducted to determine
the significance of the organisms in home
distribution systems.
Conclusions and
Recommendations
In Phase 1, 68 water samples for L.
pneumophila analysis were collected
from hot and cold kitchen sink faucets in
two homes on each of 17 community
water systems in Vermont. No/., pneumo-
phila organisms were isolated from any of
these samples.
During Phase 1A, the kitchen sink
faucets/aerators and showerheads/sup-
ply pipes were sampled with sterile swabs
and a sample of hot water flushed from
the drains of domestic hot water heaters
in each home. A total of 184 samples (92
swabs, 92 hot water samples) were
collected from homes on four community
water systems. In addition, two samples
(one swab, one hot water sample) were
collected from a home with an individual,
shallow, dug weN. L. pneumophila organ-
isms were recovered from two samples;
one from a home on a community water
system and the other from the home with
the individual, shallow, dug well. Both of
the positive samples were taken from
electric, non-recirculating, hot water
heaters.
Of the 93 samples collected from
domestic hot water heater tanks, one had
fungal overgrowth which prevented com-
pletion of L. pneumophila analysis. L.
pneumophila organisms were recovered
from two of the remaining 92 samples
(2.17%). The recovery of L. pneumophila
from the two domestic hot water heaters
in Phase 1A is of interest. This finding
together with the lack of recovery in the
home water distribution plumbing may
indicate that L. pneumophila may be
transported from natural aquatic sources
in very low levels in drinking water
systems to hot water heaters in homes
-------
where the organisms may increase to
detectable levels. However, using the
methods employed in this study, the
occurrence of Legionella in home water
systems was rare.
Continued study of this matter is rec-
ommended. Also, further information is
needed regarding the sensitivity and
specificity of the various laboratory meth-
ods for L. pneumophila analysis. An
accurate method is needed for the isola-
tion of L, pneumophila from environ-
mental samples. In addition, further infor-
mation regarding the epidemiology of
legionellosis is required to determine the
population size that must be sampled to
obtain statistically significant results.
LindenE. Witherell. Robert W. Duncan, KennethM. Stone, LoriJ. Stratton, Lillian
Orciari, StevenKappel, andDavidA. Jillsonare with Vermont State Department
of Health, Burlington. VT 05402-00070.
Walter Jakubowski is the EPA Project Officer (see below).
The complete report, entitled "Investigation of Legionella pneumophila in
Drinking Water," (Order No. PB 85-237 733/AS; Cost: $10.00, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Health Effects Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S1-85/019
0000*29 PS
U 3 ENVIR PROTECTION AGENCY
REGION 5 LIBRARY
230 S OfARSORN STREET
CHICAGO IL 60604
------- |