|dc.description.abstract||Several Legionella species have been recognized as human
pathogens for more than a decade. The manifestations of
legionellosis vary from asymptomatic seroconversion or mild
febrile illness to severe, fatal pneumonic disease with
multiple system failure. Because of this wide clinical
spectrum legionellosis probably goes unrecognized especially
if it occurs where it is not suspected. The distribution of
Legionella appears to be world wide, and retrospective
serological investigations have revealed the presence of
Legionella infections long before the first isolation of
Adequate information is not available on the
distribution of legionellosis on the African continent.
While readily recognized as a human infection, there is
hardly any mention of legionellosis in domestic or wild
animals in spite of the apparent wide distribution in some
regions of these organisms in natural or man-made water
sources. Consequently, more information is needed on the
geographical distribution of these infections, the
susceptibility of various animal species and the reservoirs.
potential cases of legionellosis as well as of recognized
sources may provide this information.
Although natural Legionella infections have not been
reported in animals, experimentally induced infection has
succeeded in a variety of animal species. The presence of
naturally occurring Legionella antibodies in animals would
suggest that these animals are susceptible to Legionella
With the assumptions that the main reservoirs and
sources of infection are natural or man-made water sources,
and that a variety of animals are susceptible to Legionella
infection, it would appear consequential that animals might
serve as seniinels -for the presence of these organisms in a
given environment. Demonstration of significant antibody
titres to Legionella would only be indicative of their
presence in the environment, and confirmation would have to
be obtained by isolation and identification.
A vast literature on Legionella and legionellosis has
been published in the last few years. In this study an
attempt has been made to provide a fairly comprehensive
review of the literature. Investigations were carried out to
establish the presence or absence of Legionella by testing
animal and human sera for significant levels of antibody
titres, and also by attempting isolation of these organisms
from environmental sources.
Serum samples from cattle, sheep, goats, pigs, horses,
camels, dogs, wild animals, healthy humans and human
pneumonia cases were titrated using the microagglutination
test for antibodies to~. pneumophila serogroups 1-6, ~.
longbeachae serogroups 1 and 2, ~. bozemanii, L. dumoffii, L.
gormanii, ~. micadadei and~. jordanis.
Although the serum samples were not randomly selected,
they represent nevertheless populations from several
geographic areas of the country.
The majority of the serum samples from all species
showed agglutination ,reactions up to a dilution of 1/16 to
all Legionella antigens. It is unlikely that these reactions
are indicative of exposure to all the Legionella strains.
However, antibody titres in the range 64-512 to at least one
Legionella species or serotype were found in 25/168 (14.9%)
cattle, 5/148 (3.4%) sheep, 18/204 (8.8%) goats, 2/107 (1.9%)
pigs, 2/19 (10.5%) horses, 2/65 (3.1%) camels, 6/51 (11.8%)
dogs, 1 eland out of 10 elands from a total of 196 wild
animals, and, 3/224 (1.3%) human serum samples. Most
investigators would consider titres in this range indicative
of infection. In the present study, however, titres above a
value of four times the estimated means of the titre
distributions were considered indicative of infection, or
possibly due to exposure to organisms sharing antigens with
Legionella. The lowest titre which was considered suggestive
of infection would therefore vary not only between animal
species but also with the antigens.
While in most animals and humans the titres to most
Legionella antigens had a monophasic distribution, cattle
sera gave biphasic distributions to~. pneumophila serogroup
3 and 6, and to L. jordanis. The second peak appears to
constitute a distinct population of animals which have
suffered past infection. Although some cattle sera gave
titres to~. pneumophila serogroups 1 and 5, and to L.
bozemanii, which fe~l within the range of the first peak, the
titres were sufficiently high to warrant their inclusion
among positive reactors.
Sheep sera gave elevated titres to L. pneumophila
serogroup 2, ~. bozemanii, ~. dumoffii and~. jordanis.
Goat sera showed elevated titres to L. pneumophila
serogroups 2,4, and 5, L. bozemanii, L. dumoffii and
A few of the sera from dogs, horses and camels gave
titres to L. pneumophila serogroup 2, or L. bozemanii or L.
jordanis, which may be significant.
All sera from pigs and wild animals gave only low titres
which were not considered significant.
Seven of 224 sera from healthy humans gave titres in a
range which may indicate past exposure to ~. pneumophila
serogroups 1,3,5 and 6. Only one of 29 human pneumonia
patients showed a titre of 32 to~. pneumophila serogroup 3.
Although of possible significance, this could not be
ascertained since a second sample could not be obtained for
In spite of the serological evidence for possible
expsoure to several Legionella species in a variety of
animals, it cannot be excluded that these observations may be
due to exposure to cross-reacting organisms, such as certain
strains of Pseudomonas fluorescens, Pseudomonas alcaligenes,
Pseudomonas maltophila, Bacteroides fragilis, Flavobacterium
sp.and Xanthomonas sp. Cross-reactions have also been
observed with sera from humans suffering from plague,
tularemia and leptospirosis. Leptospirosis may be
particularly relevant in Kenya since the infection has been
documented in several geographic areas of the country.
To conclusively demonstrate the presence of Legionella,
attempts were made to isolate the organisms. A total of 41
water samples from evaporative condensers, cooling towers,
rivers, streams and ponds in the Nairobi area and from Lake
Victoria, were tested. These water samples were selected
from water sources similar to those which had been found to
yield Legionella in other parts of the world. Although the
procedures used for the examination of water samples were
identical to those recomended by previous investigators, none
of the 636 bacterial isolates initially suspected to be
Legionella could be confirmed as such upon further
Although the results of the serological investigations
are indicative of the presence of ~. pneumophila serogroup I
to 6, ~. bozemanii, ~. dumoffii and L. jordanis, it has not
been possible to demonstrate conclusively the presence of
these organisms in.Kenya, neither as causative agents of
infections in man or animals, nor as free-living organisms in