patients [124], but azotemia and even renal failure have been reported in other series (Table 13).

Table 12. - Occurrence of Hypophosphatemia in Le- gionnnires' Disease

Propotion with Numkr of patients Reference

hypophosphaternia (%)

Table 13. - Occurrence of Azotemia in Legionnnires' Disease

Numkr of patients Pmporlion with Reference wotemia (%)

In the Philadelphia outbreak of 1976, presentation to the bospitni with a semm urea nitrogen > 20 mgldl or semm creatine > 2.0 mddl was associated with rela-

v

tive risks of fatd outcome of 4.8 and 4.5, respectively [135].

Liuer Function Tests. - Abnormal lactic dehydrogen- ase (LDH), alanine aminotranferaee (SGOT), alkaline phosphatase, and bilmbin levels are commonly seen [135-l361 (Table 14).

Table 14. - Liver Function Tests in Legionnaires' Dir- eace

Propotion with high:

LDH 88%(36/41) SGOT 49%(21/43) 90%(9/10) Alkdine phmphame 49%(20/41) 45%(10/22) SGPT 24%(4/17) Biiirubin 15%(7/47) 30%(7/23)

Reference 124 138 137

Helms and colleagues found high SGOT values to be more common in patients with legioneilosis than in caaes of pneumococcal or mycoplasmal pneumonia 11371.

Creatine Phosphokinase. - Sharrar and associates re- ported high creatine phosphokinase (CPK) levels in five of aia sporadic caaea of Legionnaires' diaease, but elevat- ed CPK determinations have been very unuaual in larger series[124,135].

Arteria1 Blood Gases. - Arterial hlood gas measure- ments reveal hypoxemia and hypocarbia commenaurate with the radiographic extent of pneumonia 184, 1241. Respiratory failure is the commonest mode of death. In the Legionnaires'outbreak 15((12%)of 123 patienta had an initial PaO, < 50 tom; nine died. Niueteen (15%) of the 123 eventually required mechanical ventilation; the cm-fatality rate in this subgroup was 68%. The relative risks of fatal outcome sesociated with presentation with hypoxemia or eventual development of respiratory fail- ure were 2.8 and 5.2, respectively [135].

Cold Agglutinins nnd Other Serologic Tests. - Kirby and associates reported reciprocal cold agglutinin titers af 16 in 2 of 16 patients and high M. pneumoniae com- plement fixation titern in 2 of 16 [124]. No cold agglu- tinins were found in sera from the Austin pneumonia outbreak [5].

Examination of Respirotory Secretions. - Sputum ie chara~teristicall~ thin and watery or mucoid at the oneet of pneumonia. Polymorphonuclear leukocytes are leea common than erythrocytes and macrophages, and acel- lular materia1 is most common. Polymorphonuclear cene become more frequent several days after institution of antibiotic therapy [l40]. Giménez siains 1351 of speci- mena are lesa sensitive than direct immunofluorescence (DFA), and Gram'e stain is neither aensitive nor specific [140]. The combination of hemoptysis and pleuritic chest pain is not infrequent, and differentiation from pulmonary embolism with infarction may be impossible at the bedside [124,135].

Kirby et al. reported the results of microacopical examination of 29 tramtracheal aspirates; &ere were few polymorphonuclear leukocytes in 10, moderate to many polymorphs in 17, snd mononuclear cells in 2 [124].

Pleuml Fluid. - Kiby et al. reported results on 13 patients who underwent thoracocentesie. Total protein and LDH levels suggested an exudate in eight of ten specheus. The mean leukocyte count was 6,600/~1, with a range of 554-32,50O/@l; polymorphonuclear leukocytes usually predominated [124].

Bronchoscopy. - Thin, watery aecretions without pumlence were found in each of 11 patients under- going brouchoscopy; mucod hyperemia and infiamma- tiou were noted m four [124].

Cerebrospinnl Fluid. - Cerebrospinal fluid i8 almost invariably nomal [124,135].

Radiopphy. - Characteristically, a unilobar patchy alveolar infiltrate extenda locally with or without sub. sequent involvement of other lobes. Progression of the infiltrate may continue for as many as five days after institution of erythromycin therapy, even in the face of l other evidence of clinical improvement. Significant improvement in the radiograph is usudy obaeived with-

in two weeks, although severa1 months may be required €or full radiogaphic remlution. SmaU pleural effusions are common and occasionally precede the appearance of. an infiltrate. Mediastinal adenopathy is not aeen [124, 135,1411.

Although uncommon, cavitary lung disease can occur in L. pneurnophilo pneumonia, particularly in immuno- compromised patients [142-1441.

Among those hospitalized in the Austin pneumonia outbreak, radiogaphic ahnormalitiea were most com- mon in those aged 40 years or older [5].

Extrothorocic Infection - Hist~~athologic evidence of extrathoracic infection is uncommon in patients dying o€ Legionnaires' disease [124], but casa of pyelo- nephritis [145], peritonitis [146], perirectal abscess [147], hemodialysis fistula infections [148], and disse- minated infection [l491 with L. pneumophiia have al1 been reported.

Concornitant Infection - The isolation of other po- tential pathogens from clinical specimens does not ex- elude simultaneous infection with L. pneurnophila [124, 1401. Concomitant infections with both L. pneurnophila and Mycobocterium tuberculosu [150], L. micdadei [144], or cytomegalovims [l241 have been documented.

Other Syndrornes. - Despite the association o€ L. pneurnophila with cooling towers and evaporative con- densers, aerologic data do not support an mociation with hypersennit~ty pneumonitis [151].

Pathology

Posmortem examination reveals bronchopneumonia with conmlidation and hepatization. The carinal and hilar nodes may be eularged. There is an intra-alveolar exudate of polymorphonuclear leukocytes, macropha- ges, red blood cells, and fibrin, L. pneumophilo is visible extracellularly as well as within macrophages. Blood veeael walls and large hronchi are spared. Adrenal lipid depletion and adrenal hemorrhage are sometimes oh- served [l24, 152,1551.

L. pneurnophila does not stain with conventional tis- sue Gram's atains but ie readily but nonspecifically vi+ ualized with modification o€ the Dieterle silver impreg- nation stain [154,155].

Pathogeneas

Experirnentai infectioni. - Davis et aL have devel- oped an experimental mode1 of Legionnaires' disease in which guinea pige and rats are exposed to aerosols of L. pneurnophila in a specially constructed chamber. Neh- ulization of approximately 7 x lo9 eolony-forming unita reproducihly yields histologicdiy detectable pneumonia in animala of both species, although illness

io more severe and ktal infection more common in guinea pie. L. pneurnophila begins exponential growth in the lungs of both species without an appreciable lag phase. Infiltration of polymorphonuclear leukocytes and then mononuclear cella into involved areas begins after 24 to 48 hours, but growth of the organism contin- ues for three to six days despite an increme in the in- flammatory response [156].

Guinea pigs but not mice, rabbits, and rhesua mon- keya developed pneumonia on exposure to the Oakland County Health Department at the time of the investiga- tion of the original Pontiac fever outbreak. Naturally and artificially expoaed guinea pige had a nodular bron- chopneumonia that undement centra1 necrosis and eventual Firous encapeulation [lo].

Endotoxin. - L. pneurnophilo cells and cell-free pro. ducta are potent activatora of Lirnulus amoebocyte lysate gelation, a sensitive but not specific test for bacte- riai Lipopolysaccharide [157]. However, the febrile re- sponae of rabbite injected intravenously with theae bac- terial producta is slight [157], andL. pneumophila shows little potency in other hioassays of endotoxin activity [158]. Nonetheleea, Fumarola and his associatee have demonstrated generation o€ procoagulant activity by human monocytes exposed to L. pneurnophila and related organisms [159], as occuw when monocytes are stimulated by E. coli lipopolyaaccharide.

Cytotoxin. - A low-molecular-weight extracellular product that is toxic for Chinese hamster o v q ce& has been identified [160]. Thia suhstance impaira hexose monophosphate shunt activity and oxygen consumption by human polymorphonuclear leukocytes during phagc.. cytosis and reduces iodination and killing of test bacteria in uitro [161E

Hernolysin. - Cultures of L. pneurnophiia ahow he- molytic activity [69], and stede filtrate8 of plaama and urine from experimentally infected rabbita lyse guinea pig red cells [162]. Hemolytic activity is present in the macromolecular fraction of concentrated supemates of L. pneumophiia culture8 [163].

Dopo. - L. pneurnophila appem to possess the enzyme tyrosine-3- monoxygenase, which catalyzes the convewion of tyrosine to dihydroxyphenylalanine (dopa) [67,68,164].Thus one component of pulmonary infection with this organism may he liberation of dopa into the syatemic circulation, with poeaible effects on the centrai and autonomie nervous systems.

Host Defenae Mechaniws

Antibody. - Infection with L. pneurnophila stimu- lates an antibody reaponse. The immunoglobulin classes involved vary between individuala (1651.

L. pneurnophilo maintained by paasage in ouo is

resiatant to the bactericidal effect of human semm even in presence of specific antibody, which ia required t o fix complement to the audace of these bactena [166]. How- ever, mice and guinea pigs paasively immunized with spe- cificIgG raised by immunization with the antigen charac- terized by Wong e t al. were protected against experi- menta1 challenge with the organism [50].

Phagocytosis. - Optimal phagocytosis oof L. pneumo- phila by human polymorphonuclear leukocytes repuires both complement and specific antiboby. These phago- cytes kill L. pneumophila only in the presence of complement and specific antibody, but bacterial killing ia incomplete; however polymorphonuclear c e b do not eupport the growth of L. pneumophila [166].

In contraet, L. pneumophila multipliee within mem- brane-hound cytoplasmic vacuoles of human monocytes in vitro under conditions that do not permit extraceUu- lar multiplication\!167]. Specific antibody promotee the binding of L. pneumophilo t o monocytes. Monocytes require both specific antibody and complement t o kill L.pneumophila, but as in the case of polymorphonuclear leukocytes, killing is incomplete. In contrast to poly- morphs, monocytes permit the growth of L. pneumo- phila, even in the presence of antihody and complement, under conditions that do not permit extracellular multi- plication [167, 1681.

Cell-Mediated Immunity. - Human monocytes in cell culture inhibit multiplication of L. pneumophila if incubated with both human lymphocytes and conca- navalin A or with ceU-free supemates of concanavalin A-sensitized mononuclear cells. The activated mono- cytee inhibit multiplication by ingesting fewer bactena than d o nonactivated monocytes and slowing the rate of intracellular replication of L. pneumoph170 [169].

Primary Isolotion. - Primary isolation of L. pneumo- phila from contaminated environmental and clinical specimens was initially performed after paseage in guinea pigs and embryonated eggs [lo, 201.

Edelstein e t al. reported positive cultures on CYE agar from endotracheal and transtracheal aapirates, lung tieaue, and blood but not from pleural fluid. Cultures from respiratory tract aecretions required three to seven days to became positive [140]. L. pneurnophila may still be ieolated from respiratory aecretions or lung days after initiation of erythromycin therapy [140,170].

Recovery of \L. pneumophiia bom lung tieaue may sometimes be enhanced by dilution of the specimen be- fore culture on agar media [ l701

Antigen Detection. - Antigenuria h a been detected in patienta with L. pneumophila infection by radioim- mnnoaeaay [171], enzyme-linked immunosorhent aeaay [172,173], and k tex agglutination [174].

Direct Immunofluorescence. - DFA (Direct Fluores cent Antibody) may be uaed t o detect the organism in environmental and clinical specimens. including respi- ratory secretions and tissue from biopsies and autopsies [175, 1761. By defmitiou, DFA staining is serogroup- specific. Organisme belonging to serogroups not encom- passed by &e reagente used by a particular laboratory will stain poorly or not a t ail. Seven serogroups of L. pneumophiia have been described [177], and recogni- tion of additional serogroups can be expected in the future. Staining of the specimen with fluoresceinated pre-immune serum from the same animal as that from which the DFA antibody was obtained is required as a contro1 to minimize the risk of falsely positive examina- tions. Unrelated bactena that stain specifically with DFA reagente for L. pneumophih are infrequently en- countered [175,176].

Edelatein e t al. reported that eputum specimens gave a higher yield than did transtracheal aspirates. Two bronchial washinga and one transbronchial biopey were even leea satisfactory. In some instances the morphology of the DFA-staining bactena became atypical after the patients began receiving antibiotics [140]. Prior erythro- mycin therapy apparently reduces the eensitivity of DFA inpost mortem diagnosis [12S].

Indirect Immunofluorescence. - The most widely used aseay of antibody in convalescent serum specimens is the IFA (Indirect Fluorescent Antibody) test [20, 178, 1791.

Fluoreweinated anti-human immunoglobulin ia nsed as the conjugate, so the test detecte lgG, IgM, and IgA [180]. Since the antigen uaed in the test is whole heat- killed L. pneumophila cells, which presumably expose multiple determinante to the test aerum, there is oppor- tunity for cross-reaction with antibody to heterologous serogroups of L. pneumophila. other species of Legionel- la, and even completely unrelated bactena.

Th iq -one (84% ) of 37 Pontiac fever patienta de- monstrated 8eroconversion t o L. pneumophilo serogroup 1 [lo] as did 90.9% of 111 patients from the 1976 Le- gionnaires' outbreak [20].

Nine (41% ) of 22 patiente with seroconveraion by iFA reported by Edelstein e t al. did not show a rise in titer to serogroup 1 antigen [140].

Most patiente in whom seroconversion to L. pneumo- phiia occurs have diagnoatic titer elwations within three weeks of onset of illness. The optimal time to draw the convalescent-phase serum specimen is 1 4 to 60 days after onset (Table 15). Although a single IFA titer of 256 is sometimes taken as aufficient to eatablisb that a recent illness representa infection with L. pneumophilo [20], eerodiagnosis from standing titem ia hazardous, particularly if data are not available on the background prevaience of antibody in the refenal population from which the patient cornea.

Administration of phamacologic doses of glucocorti-

'I'able 15. - Time Required for Seroconuersion in Sero- l ~ ~ i c o l l y Prouep Cases

Froportion (%) exhibiiing IFA titer af at least 128

coids does not appear to retard the antihody response in those who show seroconversion [ l 2 4 1

Microagglutination. - A microagglutination test has been described [181]. I t chiefly detecta IgM antibody [182E

Complement Fixation. - L. pneumophiin appears to fix complement poorly 11661, a phenomenon that may have contnbuted t o the failure to appreciate the signif- icance of Jackson's original isolate [l1

Other Serologic Tests. - Immune adherence hemag- glutination, indirect hemagglutination [183], ELISA [183], and immunodiffusion [l841 for antibody have all heen described hut are not widely used at present. Some serogroup crospreactivity is seen with the indirect he- maggiutination test. Differentiation between species and groupspecific reactivity can be achieved by adsorption of sera with gronpspecific antigen [ l855

Combined Modalities. - The prolonged outbreak of nosocomial Legionnaires' disease at the Veterana Admin- istration Wadsworth Medical Center in Los Angeles per- mitted an evalnation of available methods for the diag- nosis of legionellosis in a uniform setting. The diagnosi8 was con6rmed in 32 patients studied from October 1978 t o May 1979. Direct immunofluorescence exmination of respiratory tract secretions and indiiect immunofluc- rescence aseayof serum antihody detected 62%and 75%, respectively, of 1 3 culturepositive cases. When a posi- tive result with any of the three techniques was taken as the definition of a case, IFA serologic testa (serogroups 1 - 4) were 81% sensitive (22/27), culture on charcoal yeast-extract agar was 62%sensitive (13/21), and direct immunofluorescence of respiratory tract secretions was 47% sensitive (15132). The use of al1 three techniques in conjunction was recommended for optimal results [140].

No controlled trials of tberapy in humans have been reported. The available data come from serie8 of patients treated with various combinations of antimicrobials a t the diicretion of the responaible physicians.

In the 1976 Legionnaires' outbreak the case-fatality rates were worst for patienta treated with cephalosporins (41 40 ), intermediate for those receiving aminoglycoaidic aminocyclitols (36% ), chloramphenicol (30%), ampicil- lin (24%), or penicillin G (20%), and best for those receinng erythromycin (11%) or tetracycline (10%) [135].

Beaty e t al. reported case-fatality rates of 50% for untreated patients. The case-fatality rates were 32% in those receiving gentamicin, 26%with penicillin G, 25% with a cephalosporin, and 6% with erythromycin [86].

Kirby e t ol.reported a 13%case-fataiity rate in 46 pa- tienta who received erythromycin, including 7%fatalities in 29 patients without immunosnppression and 24% fa- taìities in 17 immunoauppressed patients. In contrast, there wae a 55 %case-fataiity rate in 18 patients who re- ceived no erythromycin (25% of eight patients without immunosnppression and 80% of ten immunosuppressed patienta). In nine of the ten fatalities in patients not re- ceiving erythromycin, death occurred within four t o 1 3 days (mean 6 days) of onset. Subjective improvement wae noted from 12 to 96 hours after initiation of oral or intraveuous erythromycin with defervescence complete in a mean of 4.2 days (range 1 - 15). Kirhy e t al. recom- mended a dose of 2 - 4 giday for a total of three weeks t o avoid relapse or protracted convalescente [124].

Seven patients developed Legionnaires' disease while receiving ampicillin (two patients), peniciilin, cefazolin, cephalexin, cephapirin, or tetracycline. Legionnaires' disease progressed in most patients given ampicillin, carbenicillin, oxacillin, penicillin, cefazolin, cephalexin, cephalothin, cefoxitin, amikacin, gentamicin, vancomy- cin, or clindamycin instead of erythromycin. A possible response was seen in two of four patients receiving tetracycline and in two patients given trimethoprim- sulfamethoxazole [124].

Honvitz and Silventein have found that growth of L. pneumophila in human monocytes in uitro is rapidly but reversibly inhibited by concentratione o€ erythromy- cin or rifampin comparable t o those that inhibit multi- plication of the organism in BYE broth. However neither antimicrohial in concentrations equal to or greater than attninable peak serum levels effectively kills L. pneumo- phila that are multiplying within monocytes or suepend- ed extracellularly in medium that does not support multiplication of these organisms [186].

Detection of Legionella in the Inanimate Enuiron- rnent. - Environmental specimens can be screened for L. pneumophila using direct immunofluorescence [187]. Paesage of environmental specimens through guinea pigs

and embryonated egga has been used successfully to iso- late L. pneumophila in the prewnce of large numbera of other micmorganisme [188]. Enrichment steps using low pH [l891 or heat [l901 have also heen employed as an alternative to theae biologic filtem. Recent imprwe- menta in selective media for L. pneumophila permit iso- M o n from potable water specimens after direct plating onto agar medium [33].

Decontamination. - Decontamination of cooling towera has been attempted using chlorination at neuhd pH with the goal of achieving a free chlorine residua of 100 . 250 ppm [119]. Periodic flushing of the potable water system of a building with superheated water may be the beat way of dealing with contamimtion of storage tanks, pipea, and outiets 1951.

The dmdopment of CYE agar and ita modifications and the search f a L, pneumophila in environmental ape. cimens has led to the discovery of the existence of many other pm-negative aerobic nonfermentative rode that do not grow on conventional bacteriologic media and that arecharacterized by a high proportion of branched- chain fatty acida. Most are known to have hh-water aquatic habitats, and at l w t many are proven eti- ologic agente of pneumonia in man. These organisms have been claaeified together ae memhers of the f a d y Legioneliaceoe on the basis of their phenotypic eimilar- ities [191]. Since these organime share little to no homologous sequencee as detemined by hybridization of intact chromosomal DNA, they represent distinct spe. cies. Some workem hwe chosen to emphasize this lack of genetic relatedness by pmposing new genera. Fluori- bocter and TatlockiB, for mme of these species, but despite a d o n e to the conhary [l921 them designa- tions do not have priority. Since their intmduction of. fem no real advmtage while complicating the terminol- ogy of workera in microbiology, clinical infectious dieease, and public health, these alternate generic terms should he avoided [193].

Legionello micdadei

In the early 1940's outbreaks of an acute febrile iil. ne@ characterized by an exanthem over the anterior aspect of the le@ were reported from Wrene, Georgia [194], and Fort Bragg, North Carolina [195]. The etiol- ogic agent of Fort Bragg fever was ultimately isolated by Tatlock 1196, 1971 and submquentiy ehown to be a strain of the Autumnalis serogoup of Leptospim interrogans [198].

Early in the coume of hie investigations in North Car- olina, Tatlock imlated a different microorganism from guinea piga inoculated with blood from a case of Fort

Bragg fever [199]. This agent, a poorly staining gram- negative pleomorphic bacillus, was pathogenic for guinea piga and could be maintained by paaeage in yolk sacs of fertile hens' eggs. No complernent-fixing or ag. glutinating antibody to the agent wan found in conva- lescent sera of patienta who had had Fort Bragg fever and the organism was preaumed to be enzootic in the guinea piga used in the laboratory [197, 1991. Jackmn reported antigenic differences between the Tatlock agent and the organism that she ieolated in 1947, which was ultimately shown to be L. pneumophiia [l].

In 1959 Marilyn Bozeman and her colleagues recov- ered an organism, designated Heòa, from the spleen of a guinea pig inoculated with blood from a patient with euspected pityriams rosea. The organism was pathogenic for guinea pigs and could be propagated in yolk sscs. It did not grow on any of a variety of bacteriologic media. Heba was found to be antigenicdly related to the Tatlock agent. Complement-fixing antibody t o these agenta was not detected in sera from humans and from thiieen genera of small mammals. Heba was pre- sumed to have been latent in the guinea pig from which it had heen isolated [101].

In 1979 Panculle and his colleagues desmibed an ap- parently newly recognized bacterium responsihle for pneumonia in patients receiving glucocorticoids or other immunosuppreseive medications. A peripheral, pleurally baeed nodular or spherical infiltrate was felt to be radio- graphically typical [ZOO]. Thie organiam and one de- scribed in immuno~om~romised patienta in Charlottes. ville[201] have been shown to represent the same species a6 the Tatlock and Heba organisma. These have been given the designation Legionella micdodei (Pittnburgh pneumonia agent, L. pittsburgensis, Tatlockin micdodeil [202-2041. Tatlock himself has recently emphasized that the evidence fora leptospiral etiology of Fort Bragg fever is subatantial and that the agent that he isolated in 1943 waa in aii likelihood a contaminant in his cultures [197]. Like L. pneumophilo, L. micdadei has an aquatic habitat and has been isolated from reepiratory therapy equipment [205] and potable water systems [206].

When tested on BCYE agar, L. micdadei ia highly suaceptible to erythromycin, rossramycin, rifampin, clorarnphenicol, and the penicillins, susceptible to cephalosporins. cefoxitin, moxalactam, and aminogly co- sidic aminocyclitols, and resistant t o clindamycin and vancomycin. No P-lactamase b produced [107l.L. mic- dodei is also susceptible in vitro to other macrolide anti- biotica [l06].

Deapite apparent differences in their effieacy in c h i - cal infections, erythromycin, rifampin, peniciuin G, cephalothin, and gentamicin are bactericidal against L. micdadei in vitro. Mutants resiatant to rifampin can be selected by exposure to low dosa of thia agent [207].

Clinical experience indicate6 that 0-lactam mtib- iotics are inefficacioua and that erythromycin is the dnig of choice in infections with L. micdodei [208]. Respi- ratory secretions may be positive by culture or DFA for days after ht i tut ion of erythromycin therapy [144].

Legionella bozemanii

In 1958 a 38-year-old Navy SCUBA diver àied after developing fulminant hronchopneumonia whiie in train- ing [209]. In January 1959 Madyn Bozeman and her colleagues a t the Walter Reed Army Institute of Re- search recovered an organiem, designated Wiga, from the brain of a guinea pig inoculated with lung &sue from the dead man. The organinm was pathogenic for guinea p i e and could he propagated in yolk sacs. I t did not ROW on any of a variety of bacteriologic media. Wigo was found t o share an antigenic determinant common to Olda, Tatlock, and Heba. It was resiatant to penicii- lin G in ovo. Complement-fixing antiboàies were not detected in sera from humans and 1 3 genera of s m d mammals. Ae with the other "rickettsia-like agent", it was conaidered to have heen Iatent in the guinea pig from which i t had heen inolated [101].

An identical organism waa isolated from the lung o€ a patient with chronic lymphocytic leukemia who developed a fata1 case of pneumonia after a hoating accident in a hrackieh swamp [209]. Thii organiam has heen given the designation L. bozemanii (Fluoribacter bozemanne) [210].

Although L. bozemanii does produce 8-lactamase, its antimicrohial susceptihility pattern in uitro on BCYE agai resembles that of L. micdndei [107]. L. bozemanii is susceptihle to a variety of macrolides [106].

Legionelin dumoffii

During investigations of an outhreak of infection with L. pneumophiin in the garment district of New York [117], a similar hut dietinct organinm was ieolated from a cooling tower [209]. Thin inolate in the type strain of the species deeignated L. dumoffii [ZIO].

An isolate of L. dumoffii has heen recovered from the lung of a man with small-cell carcinoma of the lung who died of pneumonia in Houston [211]. This organiem has aleo heen recognized by DFA in a specimen of lung from a woman in San Antonio who developed fatal pneumonia while receiving glucocorticoid therapy for systemic lupus erythematoaus [212].

L. dumoffii haa 8-lactamase activity. Itn antibiotic susceptibility pattern in vitro on BCYE agar resembles that of L. pneumophih [107]. L. dumoffii is susceptible in vitro to a variety of macrolide antihiotics [106].

Legionella gormanii

The fifth species of Legionelln to be described was L. gormanii [213]. The type strain was recovered from soil collected hom a creek bank a t a country club in Atlanta that was the site of an outhreak of disease caused hy L. pneumophila [84]. The organiem has been recognized hy DFA in the lung of a man in Con- necticut who died of pneumonia [213].

Although L. gormanii does produce 8-lactamase, it has an antimicrohial susceptihility pattern in uitro on BCYE agar similar t o that of L. micdadei [107].

Legionella longbenchae

Four caees of pneumonia in hoth previously healthy and immunosuppreeaed patients that were associated with recovery of a new species of Legionelin from respi- ratory secretions were reported in 1981. The organiem h a heen designated L. longbeachae. Evidence of clinical henefit of treatment with erythromycin with or without rifampin was seen in three of the four [214]. A fatal case of pneumonia with a strain of L. longbeachae helonging t o a distinct second serogcoup of this species hae heen re- ported [215].

L. longbeachae is susceptihle in vitro t o various macrolide antihiotics [106].

Legionella jordanii

Two isolates from specimena of fresh water have heen clmified as helonging t o a species designated L. jordania [216], named after the Jordan River on the campus of Indiana Univemity, where the fimt of these waa recovered during the investigation of an outhreak of infection with L. pneumophila [83]. Serologic evidence suggests a role for thie organism in case6 of pneumonia in man [179,216].

Legionelin wndsworthii

L. wadsworthii ie represented hy an ieolate recovered from the sputum of a patient with pneumonia and un- derlying chronic lymphocytic leukemia who was ad. mitted to the Veterans Adminietration Wadaworth Med- icai Center in Novemher 1981. The patient reaponded slowly t o antimicrohial therapy that included erythro- mycin and rifampin [217].

Legionelln oakridgensii

The most recently named species of Legionelin ie L. oakridgensii, which is represented hy ten straim iso- lated from industrial cooling towem. These organiems produce h e a s in experimentally infected guinea pigs hut have yet t o he associated with human disease [218].

A Legionelin-like organism represented hy the isolate W 0 - 44 has heen shown to he responsible for an out- break of illness resemhling Pontiac fever that affected 395 engine assemhly plant workem at a factory in Wind- sor, Ontario, in August 1981. Epidemiologic studies sug- gested airborne trammiesion from the coolant system in the piston department [219].

Ricevuto il 4 aprile 1983. Accettato il 26 aprile 1983

I. BRENNER. DJ., STEIGERWALT. A.G. & McDADE, J.E. 1979. Classifieation of the Legionnaires' diaeaae bacterium: Lsgionalb pneumophib, genua novum. specie8 novi, of the family LegioneUaceae. familia novs. Ann. Intern. Mad. 90 : 656 - 658.

l. FEELEY, J.C. & GORMAN, G.W. 1980. Legioneh. In: Monuai of clinica1 microbiology. 3. ed., E.H; Lennette (Ed.) Wnshington.

D.C. American Society for Microbiology. pp. 318 - 324.

I. FEELEY. J.C., GORMAN, G.W., WEAVER, R.E.. MACKEL, D.C. & SMITH. H.W. 1978. Primary isolation media for Legion.

naires diaeaae bacterium.J. Clin. Microbiol. 8 : 320 - 325.

I. THORNSBERRY, C,, GAVAN, T. L. & GERLACH, E. H. 1977. New developmenta in antimicrobid agent auanptibility testing. h: Curnitech 6. J.C., Shcmia (Ed.). Wash'ington. American Soeieiy for Microbiology. pp. 1 - 13.

i. FEELEY,J.C.,GIBSON, R.J., GORMAN, G.W., LANGFORD, N.C. RASHEED. J.K.. MACKEL. D.C. & BAINE. W.B. 1979. Charcoal-yeast extrait agar: primary isolstion mcdium for Legionelb pneumophih. J. Clin. Microbiol. 10 : 437 - 441.

26. PASCULLE. A.W., FEELEY.J.C., GIBSON, RJ., CORDES. L.G.. MYEROWITZ. R.L., PATTON. C.M.. GORMAN, G.W.. CARMACK, C.L.. EZZELL. J.W. & DOWLING, J.N. 1980. Pitteburgh pneumonia agent: direct iaolation from h u m n lung tisauc. J. Infect. D". 141 : 727 - 732.

27. EDELSTEIN, P.H. 1981. Improved semiseleetive medium for isalstion of Legionelb pneumophib from contaminated dinical and environmentd apecimens. J. Clin. Mioricrobiol. 14 : 298 - 303.

28. MUtLER, D,, EDWARDS. M.L. & SMITH, D.W. 1983. Changes in iran and transferrin levels and body tempera* in experimental airborne legionelloeis. J. Infect. Dis. 147 : 302 - 307.

29. YICKERS. R.M., BROWN, A. & GARRITY, G.M. 1981. Dye-eontaining buffered eharcoai-yeaat extrset medium for differentia- tion of membcrs af the family Le@onelbceae. J. Clin. Microbiol. 13 : 380 - 382.

30. HOLMES, R.L. 1982. Aniline blue-containing buffered charcod-yeast cxtract medium for presumptive identification af L&onsl. 1. apecica. J. Clln. Mimobiol. 15 : 723 - 724.

31. EDELSTEW, P.H.. & FINEGOLD, S.M. 1979. Uae of a semiselestive medium t0 culture Le6ionelbpnaumophib from cantaminat- ed lungspccimens. J. Clin. Microbiol. 10 : 141 - 143.

32. WADOWSKY, R.M. & YEE, R.B. 1981. Glycinc-containing aelective mcdium for ieolstion of L e ~ o n e l b e m e fram enviranmental specimens.App1. Enuiron Microbid. 42 : 768 - 772.

33. EDELSTEIN, P.H. 1982. Comparative study of seleetive media for isolation of Legionelbpncurnophib from potable water. J. CIim Microbiol. 16 : 697 - 699.

34. WEAVER, R E . & FEELEY, J.C. 1979. Culturd and biochemieal eharaeterization of the Le@onnairea' diseaae bacterium. In : "Lqionnaires': The d imie , ths bmteriurn m d methodology. G. L. Jones. & G. A. Hébert (Eds). Atlmta. Centa for Dweiae Control. pp. 20 - 25.

35. GIMÉNEZ, D.F. 1964. Staining rickcttsiae in yolk-sac culNres. Stoin. Technd. 39 : 135 - 140.

36. THOMASON, B.M.. CHANDLER E. W. & HOLLIS, D.G. 1979. ElageUa on Legionnaires' diaeaae baeteria: an interim report.Anm Intein. Med. 91 : 224 - 226.

37. WEST, M., BURDASH. N.M. & FREIMUTH. F. 1977. Simplified silvcr-plating atsin for flagells. J. Clin. Miembiol. 6 : 414 .419.

38. CLARK. W.A. 1976. A simplified Leifson flagella stain. I. Clin. Microbiol. 3 : 632 - 634.

39. CHANDLER, F.W., THOMASON, B.M. & HÉBERT, C.A. 1980. FlageUa a n Legionnaires'disense bacterinin the huminlung. Ann Intern. Med. 93 : 715 - 716.

40. FLESHER, A.R., ITO, S., MANSHEIM, B.J. & KASPER, D.L. 1979. The celi envelopc of the Legionnaires' diseaae baeterium. Ann. Intem. Med. 90 : 628 - 630.

I l . CHANDLER. F.W., COLE, R.M.. HICKLIN, M.D., BLACKMON, J.A. & ~ L L A W A Y , C.S. 1979. Ultraabueture af the Legion.

naires' diaeaae baeterium. A sNdy using tranwiwian elechon microscopy. Ann. htm. Med. 90 : 642.647.