Caister Academic Press

Legionnaires Disease: History and Clinical Findings

from Legionella: Molecular Microbiology

Legionnaires Disease: History and Clinical Findings

from Legionella: Molecular Microbiology
See also: Real-Time PCR in Microbiology: From Diagnosis to Characterization

Paul H. Edelstein


The history of Legionnaires' disease began at least 33 years before the 1976 Philadelphia epidemic, when Legionella micdadei was isolated from human blood. Multiple isolations of several different Legionella spp. were made prior to 1976, and it was known by 1968 that tetracycline therapy prevented deaths in L. pneumophila-infected chicken embryos. The 1976 epidemic provided the scientific focus and resources necessary to determine that L. pneumophila caused epidemic pneumonia and to show that epidemics of Legionnaires' disease had occurred worldwide many years before 1976. Despite a surfeit of available resources and expertise, the effort to isolate the aetiological agent succeeded solely on the basis of one person's determination to solve a scientific problem and his willingness to re-examine his assumptions about prior laboratory results.

Pontiac fever, a disease of unknown aetiology, is a self-limiting and short duration febrile illness that has been associated with exposure to L. pneumophila. Because of non-specific clinical findings that overlap with other diseases, accurate diagnosis of Pontiac fever in non-outbreak settings is impossible. Legionnaires' disease can be diagnosed specifically through specialized laboratory tests, but not by clinical findings alone. This is because the clinical findings of Legionnaires' disease overlap with those of other more common causes of community-acquired pneumonia. Antimicrobial therapy of Legionnaires' disease requires the use of drugs that are active against intracellular Legionella spp., such as tetracyclines, macrolides, azalides and antibacterial quinolones. read more ...


Legionnaires' disease is a type of pneumonia caused by Legionella spp., which are environmental Gram-negative bacteria. Pontiac fever is a febrile, non-pneumonic, short incubation period, non-fatal disease that is associated with exposure to water-borne aerosols containing Legionella spp. bacteria, but it is unknown if the disease is caused by the Legionella spp. bacteria. The vast majority of Legionnaires' disease cases are caused by L. pneumophila, and in particular L. pneumophila serogroup 1. A 1976 epidemic of pneumonia in American Legionnaires' attending a Philadelphia, Pennsylvania, congress was the first recognized occurrence of Legionnaires' disease, and it led to the naming of the disease and isolation and characterization of L. pneumophila.

Prompt and specific antibiotic treatment of Legionnaires' disease can markedly reduce the fatality rate of the disease, without which 15-80% of patients die, depending on underlying disease, host immunity, and duration and severity of illness before treatment. Unfortunately there are no clinical features that allow a clinicians to specifically diagnose the disease with sufficient accuracy, which places great emphasis on the need for empiric antibiotic therapy for the disease. read more ...

Historical aspects read more ...

Legionnaires' disease and Pontiac fever epidemics prior to 1976

SPAM City, 1957

The first known epidemic of Legionnaires' disease occurred in Austin, Minnesota, in the summer of 1957 (Osterholm et al., 1983). Seventy-eight people in this small city of approximately 30,000 people developed pneumonia between June and August. A significant risk factor for illness was working in the city's meat packing plant, although 40% of cases did not work at the packing plant. Antibiotic therapy had no apparent benefit, but 97% of affected people recovered in 2-14 days after onset of illness. The source of the epidemic was never determined, and the pneumonia did not recur in subsequent years. The mystery of the cause of the epidemic was solved by investigation of survivors 22 years later. This showed that the survivors had significantly elevated antibodies to L. pneumophila in comparison to matched controls. Because both packing plant employees and non-employees became ill, it is most likely that the epidemic derived from a cooling tower. Of interest, Austin, Minnesota, is the headquarters of the Hormel Foods Corporation, and is known as 'SPAM City' after the name of the canned ham product made there since 1937.

St. Elizabeth's hospital, 1965

Another early Legionnaires' disease epidemic affected residents of a large psychiatric hospital in Washington, DC, St. Elizabeth's hospital (Thacker et al., 1978). The hospital housed about 6000 patients in multiple buildings on a 350 acre (1.4 km2) campus. In July to August 1965, at least 81 hospital patients developed pneumonia, 14 (17%) of whom died. Cephalosporin therapy was ineffective, but tetracycline plus streptomycin treatment appeared to have to some beneficial effect. Sleeping near open windows and having permission to walk on the hospital grounds were found to be the most significant risk factors for acquiring pneumonia. Soil excavation work for a lawn sprinkling system was causing very dusty conditions. A strong wind storm and summer rains occurred concurrent with the outbreak, which ended spontaneously, and contemporaneously with the cessation of the excavation work. A soil or dust borne pathogen was thought to be most likely. Extensive microbiological testing of lungs from dead patients and of the hospital environment, failed to reveal an aetiology for the epidemic. However, analysis of stored serum specimens in 1977 showed that 19 of 26 patients tested seroconverted to L. pneumophila serogroup 1. Subsequent Legionnaires' disease epidemics have rarely implicated soil excavation, although disruptions of potable water systems and contamination of building plumbing systems during construction have been alternative explanations (Mermel et al., 1995). An interesting historical note is that the attempted assassin of US President Ronald Reagan is currently confined to this institution.

Pontiac fever, Pontiac, Michigan, 1968

In July 1968 visitors to, and employees of, the Pontiac, Michigan, Health Department developed a relatively mild self-limited febrile illness that was called Pontiac Fever (Glick et al., 1978). Cases became ill over a period of several days, with at least 144 people affected. The attack rate was 95% in building occupants and 29% in building visitors. Being in the building was the sole risk factor for illness, with an increased risk the greater the duration of time spent in the building. After an incubation period of 5-66 hours, fever and other symptoms developed. The illness lasted for 2-5 days and was manifested mainly by fever, headache, myalgia and fatigue. Most cases recovered promptly without sequelae although a small fraction of disease victims required months to recover from assorted neuropsychiatric symptoms. No case was shown to have pneumonia. Re-entering the building after recovery from illness resulted in a lower attack rate, around 10-45%, and was associated with milder symptoms than during the first illness. Investigations were negative for allergic reactions to known environmental allergens, for known toxins and for viral infections. The cause of the illness was traced to a leak in the building's air duct that allowed water from the evaporative condensing system to enter the circulating air. Turning off the building air conditioning eliminated the risk of disease, while turning it back on caused disease.

Guinea pigs that were placed in the building during the outbreak developed pulmonary nodules (Kaufmann et al., 1981). At the CDC, challenge of additional guinea pigs with aerosolized water from the evaporative condenser caused pneumonia in the animals, but not when the water was filtered or autoclaved. Extensive microbiological studies of the guinea pig lungs failed to demonstrate a convincing aetiological agent. The CDC employee performing the aerosol challenge studies became ill with what was later thought to be Legionnaires' disease (US Department of Health and Human Services-Public Health Service-Centers for Disease Control and Prevention and National Institutes of Health, 1993).

Legionella pneumophila serogroup 1was cultured from some of the frozen lungs of the sick guinea pigs after the aetiology of Legionnaires' disease was known (Kaufmann et al., 1981). One of these isolates, Pontiac 1 is the type strain for the Pontiac monoclonal antibody type. In addition, serum from those with Pontiac fever contained antibodies to L. pneumophila.

Hotel outbreaks prior to 1976

A prelude to the 1976 Philadelphia Legionnaires' disease epidemic was an epidemic of pneumonia that occurred in Philadelphia in September 1974 (Terranova et al., 1978). Approximately 1500 members of the Independent Order of Odd Fellows attended a convention at the Bellevue Stratford Hotel, the same hotel involved in the 1976 epidemic. The Independent Order of Odd Fellows are a fraternal order dating from eighteenth-century England. Twenty Odd Fellows developed pneumonia one to nine days after attending a September 16th meeting in main ballroom of the hotel; two of these people died of the pneumonia. The major risk factor for pneumonia was attending the 16 September meeting. Serological studies of survivors and matched controls in 1977 showed that 4 of 11 cases, but no controls, had elevated antibody titres to L. pneumophila serogroup 1. Of historical interest, there is speculation that some members of ancient Odd Fellows organizations were Roman Legionnaires stationed in Britain.

Meanwhile in Spain an unrecognized epidemic of pneumonia had been occurring among British tourists to the Rio Park Hotel in Benidorm, a coastal resort town on the Costa Blanca (Grist et al., 1979; Bartlett et al., 1984). From 1973 to 1980, at least 150 British tourists acquired Legionnaires' disease at the hotel, possibly in several discrete epidemics. In 1973, 3 years before the Philadelphia epidemic, at least 89 hotel guests had febrile respiratory illnesses, three of whom died with pneumonia, one on the plane en route from Spain to the United Kingdom (Grist et al., 1979). Serological testing of saved serum specimens, undertaken after the discovery of the cause of the 1976 Philadelphia outbreak, showed that these illnesses were due to Legionnaires' disease. A 1980 investigation showed that the source of the epidemic was the hotel's potable water system. The hotel had no air conditioning cooling tower and nearby cooling towers could not be implicated based on epidemiological findings. Engineering changes made to the plumbing system, chlorination, and maintenance of hot water at 50-60°C appeared to end the multiyear epidemic. read more ...

The 1976 Philadelphia epidemic

Legionnaires' disease was first recognized as a distinct entity during an epidemic of pneumonia that occurred in Philadelphia, in the summer of 1976. About 4,000 members of the Pennsylvania State American Legion, an organization of former military veterans, met in July for their annual meeting in the city, which lasted from 21 to 24 July (Fig. 1.1). The convention hotel, the Bellevue Stratford, considered the best hotel in the city, was a frequent convention site and the hotel where visiting dignitaries stayed. It also was the site of the 1974 Odd Fellows Legionnaires' disease epidemic.

July, 1976 was a special time in Philadelphia and the nation, as this was the bicentennial anniversary of the United States Declaration of Independence from Great Britain. Philadelphia was where the Declaration of Independence was written and first announced, and the nation's first capital. Celebrations and large parades were held in the city during the time of the American Legion convention. City officials were worried that outside protestors would create violent disruptions of planned bicentennial celebrations. In addition, the hot and humid Philadelphia summer was made even more unbearable by a garbage collector's strike that left high mounds of stinking garbage on city streets.

In the spring of 1976, the US Center for Disease Control warned about the possibility of pandemic swine influenza occurring in the United States after two cases of the disease were diagnosed in army recruits at Fort Dix, New Jersey (Gaydos et al., 2006; Sencer and Millar, 2006). A nationwide influenza vaccination campaign was advised, but this was stalled because of manufacturer's concerns about financial liability.

On Friday 30 July 30 1976, Dr Ernest Campbell, a Bloomsburg, Pennsylvania, physician, realized that he was treating three convention delegates for a febrile illness (Lattimer and Ormsbee, 1981). From Friday until Sunday, other health care providers became aware of clusters of pneumonia (Lattimer and Ormsbee, 1981; Thomas and Morgan-Witts, 1982), and at least one called the Philadelphia Department of Health on Sunday (Thomas and Morgan-Witts, 1982). Since the Pennsylvania State Public Health Department was closed on that day and through the weekend, these reports were made on Monday, 2 August. By that day 18 convention delegates or city visitors had died of acute pneumonia, and about 170 were sick with pneumonia. In retrospect, the first recorded death from the epidemic was on 27 July. Initial recognition of the epidemic was delayed because the convention delegates were widely disbursed after the convention ended, and because many lived in small towns. The Philadelphia epidemic was featured in national television and radio broadcasts on 2 August, with an emphasis on the unknown cause of the epidemic. Convention delegates last became ill on 3 August, with the last known cases of Legionnaires' disease being in non-convention delegates on August 16th. The epidemic was over by the time that the epidemiological investigation began in earnest.

The total number of Legionnaires' disease cases was 182, with 29 deaths. Both convention delegates (149 cases) and non-Legionnaires (33 cases) were affected by the disease, including some who never entered the convention hotel. Some members of a marching band and people who watched a parade outside the hotel became ill; this was initially termed Broad Street pneumonia to distinguish the illness from that suffered by the Legionnaires. One dramatic example is that of a marching band bus driver who became ill and died of the disease; he had spent only hours in Philadelphia, and most of the time in his bus. Great public concern and fear arose, such that funerals of those who died with the disease were not attended by the public because of concern of getting the disease from the corpse (Thomas and Morgan-Witts, 1982). Even before the public health investigations had been completed, Bellevue Stratford hotel guests cancelled their reservations and the hotel was left empty. The state governor stayed overnight in the hotel to emphasize its safety, but to no avail; the hotel was forced to close permanently on 18 November 1976. Reopened later under a new name and after major renovations, it is now known as the Park Hyatt Philadelphia at the Bellevue.

Epidemiological studies, led by David Fraser of the US Center for Disease Control (CDC), concluded that the focus of the disease was the Bellevue Stratford Hotel, that both American Legionnaires and non-convention delegates had acquired the disease, and that the disease was most probably spread by the air-borne route (Fraser et al., 1977). However, the cause of the epidemic pneumonia was not determined until December 1976, some 5 months after the outbreak (McDade et al., 1977). Tests for almost all known types of respiratory infections were negative, as were toxicological studies.

The inability of the CDC investigators to quickly determine the cause of the outbreak led to widespread, and often unfounded, speculations about the cause of the epidemic. Some suggested that the epidemic was a US Central Intelligence Agency germ or chemical warfare experiment gone bad. Others suggested that the epidemic was a hoax, designed to gin up support for the CDC and swine influenza vaccination. Some of this speculation was fuelled by the rapid US congressional subcommittee approval of a vaccine manufacturers liability protection act, on August 3rd, 1977, after congressional testimony that swine influenza could not be excluded as a cause of the epidemic (Lattimer and Ormsbee, 1981). Some thought that domestic terrorists had caused the illnesses through chemical or microbiological means. A leading Philadelphia pathologist and toxicologist, Dr William Sunderman, suggested that the epidemic was due to nickel carbonyl poisoning, a disease that he had studied for years. The New York Times excoriated the CDC for its failure to collect adequate specimens for toxicological studies (Anonymous, 1976). The possibility of nickel carbonyl poisoning was supported briefly by autopsy studies until it was realized that the autopsy instruments were the probable sources of the nickel that was detected; repeat testing and detailed analysis showed that nickel carbonyl poisoning was not the cause of the epidemic (Lattimer and Ormsbee, 1981). Regardless, Dr Sunderman and his son, also a toxicologist, maintained for years afterwards that nickel carbonyl poisoning was a strong possibility as the cause of the epidemic (Sunderman, 1977; Sunderman, 1978). Others speculations for an aetiology were toxic fumes from photocopy machines, air conditioning refrigerant decomposition producing phosgene gas, and nefarious foreign or extra-terrestrial forces.

Concern about the inability to identify the aetiology of the epidemic led to US Congressional hearings in November 1976, about a week after the permanent closure of the Bellevue Stratford Hotel. The hearings, representative of the frustration of many that no aetiology had been determined for the epidemic, were primarily political theatre. A month later Joseph McDade had determined that the aetiology of the outbreak was L. pneumophila (McDade et al., 1977; Sharrar et al., 1977). In April, 1977, the term Legionnaires' disease was first published by the Center for Disease Control as the official name of the epidemic disease, although it had been called that in the press as early as September 1976 (Anonymous, 1977; Anonymous, 1976; Lattimer and Ormsbee, 1981). read more ...

Discovery of the aetiological agent of Legionnaires' disease

McDade's achievement of determining that L. pneumophila was the cause of Legionnaires' disease was remarkable when so many others had failed for months to isolate the bacterium (McDade et al., 1977). A rickettsiologist, he was assigned to determine if Legionnaires' disease was caused by Coxiella burnetii, the agent of Q fever (McDade, 2002). Using the technique of guinea pig inoculation followed by embryonated egg passage, he was unable at first to demonstrate an infectious agent affecting the embryos. The guinea pigs became ill after being injected with human lung homogenates, but passage into the embryos resulted in no embryo infections. Initial microscopic examination of guinea pig tissues stained with the Giménez stain, which stains rickettsia as well as many other bacteria, revealed just an occasional rod thought to be the result of contamination. However, when McDade re-examined guinea pig tissue slides he found clusters of rods, and convinced that these were real he repeated the hen's egg inoculations without adding the standard antimicrobials used for rickettsial isolation. By eliminating antimicrobial suppression he successfully infected embryos with L. pneumophila, discovering the aetiology of Legionnaires' disease. Axenic cultivation was achieved by Robert Weaver of CDC, using Mueller-Hinton agar supplemented with IsoVitaleX and haemoglobin, a source of needed l-cysteine and iron (Feeley et al., 1978). Moreover, McDade was able to show that several patients with Legionnaires' disease had antibodies to the bacterial isolate, as did stored sera from the St. Elizabeth hospital and Pontiac epidemics (McDade et al., 1977).

The Giménez stain used by McDade to visualize L. pneumophila in guinea pig and chicken embryos contains carbol fuchsin, which intensely stains Gram-negative rods, as well as rickettsia. Use of this stain was a key factor in McDade's discovery of L. pneumophila in guinea pig tissue, as was staining of fresh rather then fixed tissues. A number of conventional stains used to detect bacteria in tissues failed to detect L. pneumophila in the lungs of patients with Legionnaires' disease, despite the tremendous number of the bacteria in the lungs of fatal cases. In particular the usual Gram stain formula uses safranin as the counter stain, which barely stains the bacterium when taken from fresh growth on culture medium, let alone from infected tissue or sputum. In contrast, the Giménez stain is quite good at detecting the bacterium in culture, in infected macrophages, and in infected fresh and Formalin-fixed lung. However, Giménez stain does not stain the bacteria in paraffin-embedded lung, or in deparaffinized tissue, making it unlikely that use of the stain by pathologists investigating the disease would have been helpful, unless they had studied non-embedded tissues, something not ordinarily done by pathologists (Chandler et al., 1977). The poor staining of L. pneumophila by conventional stains undoubtedly contributed to earlier failures to recognize the bacterium in the lungs of fatal cases.

The methods used by McDade to isolate and identify L. pneumophila, as well as to determine the link between the bacterium and human disease, were worked out by Walter Reed Army Institute of Research scientists over 30 years previously to detect non-conventional microbes causing human diseases. In fact, these prior efforts led to the first identification of several different Legionella spp., although at the time they were thought to be adventitial rickettsia-like agents. read more ...

First isolations of Legionella spp. bacteria

The first known isolation of a Legionella spp. bacterium was in 1943. Captain Hugh Tatlock, a military physician, was assigned by the Army to investigate an epidemic of a febrile rash illness at Fort Bragg, North Carolina, termed Fort Bragg fever (Daniels and Grennan, 1943). Many affected soldiers had bathed in a local creek. Tatlock injected the blood of ill soldiers into guinea pigs, and from one soldier's blood isolated what was thought to be a rickettsia-like agent organism from three of five inoculated guinea pigs (Tatlock, 1944). However, Tatlock could not demonstrate that serum from patients with Fort Bragg fever reacted with the bacterium, nor could the bacterium infect other animal species, including mice, monkeys, rats, or rabbits. He concluded that this bacterium was not the cause of Fort Bragg fever, and later expressed concerns that the bacterium was enzootic in the guinea pig colony (Tatlock, 1947). This rickettsia-like agent, which became known as the Tatlock strain, was identified in 1980 as L. micdadei (Hébert et al., 1980a; Hébert et al., 1980b). In 1944, Tatlock isolated a second microbial agent from the blood of soldiers with Fort Bragg fever, which when injected into human volunteers caused Fort Bragg fever (Tatlock, 1947). This second agent was later shown to be Leptospira autumnalis (Gochenour, Jr. et al., 1952). In 1981 Tatlock reiterated that the Tatlock strain was not the cause of Fort Bragg fever, although he suggested that perhaps that isolate caused a disease that was not Fort Bragg fever (Tatlock, 1981; Tatlock, 1982). Since guinea pig isolation of Legionella spp. bacteria is insensitive for environmental samples, and presumably for clinical sources, it is difficult to understand how the L. micdadei isolate could have been a contaminant (Edelstein et al., 1982). Whether it was enzootic is difficult to know, but Tatlock was unable to isolate a Tatlock-like bacterium from other guinea pigs in the same colony as those that yielded this bacterium (Tatlock, 1944). Whether some soldiers had both leptospirosis and Legionnaires' disease, from the same creek source, is impossible to know, but is an intriguing possibility. Apart from his investigative career, Dr Tatlock served as General Dwight Eisenhower's physician at Walter Reed Army Hospital, and later as a respected New Hampshire physician. He died at age 92 in 2005.

Guinea pig inoculation was used by the Rickettsial Laboratory at the Walter Reed Army Institute of Research in the 1940s and 1950s to diagnose human rickettsial infections. This led to the discovery of three rickettsia-like bacteria which were later shown to be Legionella spp. bacteria. The first isolate, Olda, shown in 1979 to be L. pneumophila, was isolated in 1947 from the blood of a patient with a fever of unknown origin (McDade et al., 1979). In 1959 Bozeman and colleagues at the same laboratory isolated a rickettsia-like agent from the lung of a Navy skin diver with fatal pneumonia (Bozeman et al., 1968; Cordes et al., 1979). The diver became ill in 1958, two days after undergoing scuba training in a Florida freshwater swimming pool, and died six days later of fatal pneumonia that failed to respond to penicillin and streptomycin therapy. This isolate, known as Wiga, was identified as L. bozemanae in 1979 (Cordes et al., 1979; Brenner et al., 1980). Bozeman and colleagues also reported the 1959 isolation of strain Heba, later shown to be L. micdadei (Hébert et al., 1980b). Heba was isolated from the blood of a patient with pityriasis rosea; whether the patient had another disease is unknown, but likely, as L. micdadei is not a cause of pityriasis rosea (Chuh et al., 2004). Tetracycline was shown by Bozeman and colleagues to be effective in preventing hen's egg embryo deaths when they were infected with Wiga, Tatlock, Heba or Olda (Bozeman et al., 1968), the first evidence that Legionella infections could be potentially cured with tetracycline therapy. read more ...

Development of an artificial growth medium

Robert Weaver of CDC was able to grow L. pneumophila on an artificial medium from the infected chicken embryo yolk sacs that Joseph McDade gave him. Weaver did this by plating the yolk sacs to 18 different bacteriological media under different atmospheric conditions (McDade et al., 1977), finding that only Mueller-Hinton agar supplemented with 1% haemoglobin and IsoVitaleX supported growth. Growth was optimal in 2.5% CO2. At approximately the same time, Morris Dumoff, a clinical microbiologist in Flint, Michigan, had independently isolated L. pneumophila by plating autopsy lung and pleural fluid specimens onto a commercially made chocolate agar supplemented with haemoglobin and l-cysteine (Dumoff, 1979). Dumoff was the first person known to have isolated L. pneumophila from an artificial medium; he was honoured for this achievement by the naming of L. dumoffii (Brenner et al., 1980). Growth of L. pneumophila on supplemented Mueller-Hinton and chocolate agars was very poor, and the lack of a good selective medium made continued use of guinea pig and embryonated hen's eggs inoculations important in some cases. Jim Feeley and George Gorman of CDC devised an improved non-selective medium, F-G medium(Feeley et al., 1978). The formulation of F-G medium was based on an analysis of the composition of prior successful media, and of IsoVitaleX, a nutritional supplement devised to enhance the growth of Haemophilus influenzae. This analysis established that l-cysteine and iron were required for optimal bacterial growth. However, F-G medium was difficult to prepare and not very supportive of bacterial growth. The next iteration, buffered yeast extract charcoal medium, was superior to F-G medium as it was more supportive of L. pneumophila growth and much easier to prepare (Feeley et al., 1979). Addition of Ó-ketoglutarate to the medium formulation enhanced the L. pneumophila growth rate, plating efficiency, and colony size; this compound is now part of most commercial formulations of buffered yeast extract medium (Edelstein and Finegold, 1979). However, there was still a problem in isolating the bacterium from contaminated clinical and environmental specimens, which required guinea pig inoculation with or without chicken embryo passage. The development of several different selective media resulted in the modern formulations of selective media for the clinical cultivation of Legionella spp (Edelstein and Finegold, 1979; Bopp et al., 1981; Edelstein, 1981; Wadowsky and Yee, 1981), and eliminated the need for guinea pig inoculations (Edelstein et al., 1982). read more ...

Discovery of the mode of spread of nosocomial Legionnaires' disease

During the period from 1976 to 1982 there were several multiyear epidemics of Legionnaires' disease in the US, the best example of which was at the Los Angeles Wadsworth Veterans Administration Hospital (Bock et al., 1978; Haley et al., 1979; Kirby et al., 1980). A new hospital building was occupied in March 1977, and it soon became apparent that renal transplant patients were dying of nosocomial pneumonia. These pneumonias were confirmed to be Legionnaires' disease through a lengthy process involving sending serum and tissue specimens to the CDC. Once diagnostic testing became available on site, many cases of Legionnaires' disease were diagnosed in patients, hospital visitors, and some hospital staff workers. Because the environmental reservoir, and mode of spread of Legionnaires' disease, was unknown at the time, it was unclear what could be done to prevent further cases of the disease. The hospital cooling towers were chlorinated in November 1978 because the cooling tower contained L. pneumophila (Shands et al., 1985). Cooling tower chlorination resulted in a temporal decline in cases from an average of 10 cases per month to five cases per month but no further. Air lock doors were installed in the hospital to prevent infiltration of contaminated air, and a HEPA-filtered nursing unit was bought, but never used, to protect immunocompromised patients. Isolation of the bacterium from environmental sources required guinea pig inoculation and was insensitive. It was not known at the time that L. pneumophila was common in building water systems, although the ubiquity of the bacterium in natural water sources had been shown by Morris and colleagues (Morris et al., 1979). Several attempts were made to isolate L. pneumophila from nearby environmental sources that might be reservoirs and disseminators. For example, I was instructed to culture captured gophers to determine if they were a reservoir of the epidemic; a gopher hunter flushed the animals out of their tunnels, and their tissues cultured without avail. I was also assigned to culture a water culvert about 500 meters from the hospital, again without yield; these studies give a flavour of the desperation and ignorance of the time. A breakthrough was made in 1979 by John Tobin in Oxford (Tobin et al., 1980). Using guinea pig inoculation to detect L. pneumophila in shower heads, Tobin reported that the source of a Legionnaires' disease outbreak in a renal transplant unit was caused by bacterial colonization of the hospital showers and potable water system. The big break at Wadsworth VA came in March 1979, when a huge rise in the numbers of nosocomial cases at Wadsworth VA occurred, up from an average of 4.5 cases per month to 25 cases in 1 month. It was discovered that a test of an emergency water pump failed immediately prior to the upswing in cases, resulting in a pressure shock and brown-black water throughout the plumbing system. This pressure shock was replicated a month later in an unoccupied hospital ward; a comparison of pre-shock and post-shock samples demonstrated a 35-fold increase in L. pneumophila concentration in the water, a convincing experiment which took several months to complete (Shands et al., 1985). The epidemic was ended by hyperchlorination of the potable water system in September 1980, three and a half years after the start of the epidemic. By the end of the epidemic at least 250 inpatients, outpatients, visitors and employees had acquired Legionnaires' disease from the hospital water in the plumbing system.

Multiyear Legionnaires' disease nosocomial outbreaks still occur, although these are now due either to inadequate case surveillance or inability to disinfect plumbing systems, not ignorance about the reservoir and disseminators of the bacterium (Anonymous, 1997; Kool et al., 1998; Lepine et al., 1998). read more ...

Discovery of the role of amoebae in the environmental ecology of Legionella spp.

The ubiquitous presence of Legionella spp. bacteria in the environment was known by 1980, and by 1981 it was known that the bacteria were found in man-made aqueous environments (Morris et al., 1979; Tobin et al., 1981). However, how the Legionella spp. bacteria survived and grew in low-nutrient waters was unclear, in view of the fastidious nutritional requirements of the bacterium when cultivated on artificial media. Tim Rowbotham, a public health microbiologist in Leeds, made seminal observations on the growth of L. pneumophila in free-living amoebae (Rowbotham, 1980). These studies were performed almost exclusively based on the microscopic observations of the bacteria interacting with the amoebae, using the most basic equipment and facilities. Rowbotham hypothesized that inhalation of Legionella-infected amoebal cysts is the means of transmission of Legionnaires' disease. Based on these studies, several investigators determined that environmental Legionella spp. are dependent on the presence of amoebae for their growth and persistence, serving as an explanation for the growth of the bacteria in otherwise nutrient-poor water (Fields et al., 1984; Wadowsky et al., 1988; Fields et al., 1989; Fields et al., 2002). The formation of a consortium of microorganisms in a biofilm also helped to explain the erratic phenomena of some Legionnaires' disease epidemics, based on sloughing of the biofilm in response to physical or chemical changes. read more ...

Clinical features of Pontiac fever and Legionnaires' disease read more ...

Pontiac fever

Pontiac fever is a self-limited short duration febrile illness of unknown aetiology. It is unclear whether the illness results from non-Legionella spp. bacteria endotoxin inhalation, from the inhalation of live or dead Legionella spp. bacteria, or from inhalation of a toxic mix of Legionella and non-Legionella microbes and their toxins (Edelstein, 2007). There is neither an agreed-upon standard definition of Pontiac fever nor any specific laboratory test that can be used to diagnose the disease, although there is at least one proposal to standardize the definition (Tossa et al., 2006; Edelstein, 2007). Most experts agree that a reasonable definition of epidemic Pontiac fever includes the following: (a) an epidemic febrile short duration illness without pneumonia or severe illness; (b) evidence of exposure to a microbe-containing aerosol, one component of which must be a Legionella spp. bacterium; (c) demonstration of an immune response to the Legionella spp. bacterium found in the aerosol source in some of the affected people AND evidence that a matched unexposed control population has a significantly lower antibody response to the same bacterium; d) lack of other credible explanations for the illness. Some would also add as a diagnostic test the presence of L. pneumophila serogroup 1 antigenuria in a number of the affected people, in the absence of documented pneumonia or prolonged illness, and with spontaneous recovery without antibiotic therapy (Burnsed et al., 2007). It is unknown how to specifically diagnose single cases of Pontiac fever, and case reports of sporadic cases should be viewed with scepticism. The reason for this scepticism is that Legionella spp. are commonly found in environmental water without causing disease, and because up to 25-40% of healthy people may have detectable Legionella spp. antibody - especially against non-L. pneumophila Legionella spp. (Edelstein, 2006b). Thus, a high proportion of healthy people might meet laboratory diagnostic criteria for Pontiac fever. The requirement for an epidemic, and use of a matched control population, increases the positive predictive value of laboratory testing.

Whether Pontiac fever signs and symptoms can occur in the absence of exposure to Legionella spp. bacteria is unstudied; however based on the overlap of its clinical findings with a number of infectious disease, such as influenza and bacterial sepsis, this seems likely. Also where Pontiac fever ends and Legionnaires' disease begins is not defined, so in some cases people suspected of having Pontiac fever might really have mild Legionnaires' disease, and vice versa. Because of uncertainties about the pathogenesis and definition of Pontiac fever, the full clinical spectrum of the disease is unclear.

The incubation period after exposure to a bacteria-contaminated aerosol has a wide range, 4-120 hours, with variable median intervals ranging averaging around 1-2 days (Table 1.1). The sites of Pontiac fever outbreaks have included workplaces, hotels, recreational spas and restaurants, but disease can occur almost anywhere there is the possibility of encountering a bacterial aerosol. In some cases, exposed people may develop either Pontiac fever or Legionnaires' disease. Attack rates are very high, with more than 80-90% of such exposed people becoming ill. Fever, myalgia, headache and fatigue are the dominant symptoms (Table 1.2). Cough, dyspnoea, anorexia, arthralgia and abdominal pain occur less frequently. There is little information about physical examination findings in the first day of illness; examination 2-5 days after onset may show fever and tachypnoea, but little else. Pneumonia does not occur. In one outbreak pneumonia and hypoxaemia were documented, contrary to reported findings for all other outbreaks (Castor et al., 2005). The authors reporting this outbreak speculated that the workers inhaled a very high concentration of endotoxin. Fatigue and non-focal neurological complaints have been reported to persist for up to several months in the minority of affected patients. Most people with Pontiac fever recover within 2-4 days, although some are sick for up to a week (Table 1.1). Illness severe enough to result in hospitalization is exceptional, so much so that a requirement for this should bring into question the diagnosis and the possibility of other diseases including Legionnaires'
disease. read more ...

Legionnaires' disease


Legionnaires' disease is a pneumonia caused by Legionella spp. that may or may not be associated with extrapulmonary infection. There are very rare case reports of Legionella spp. extrapulmonary infection in the absence of pneumonia, for which an appropriate term is either legionellosis or a description of the site of infection and infecting species, e.g. 'L. longbeachae pyarthrosis'. read more ...

Clinical findings

Patients with Legionnaires' disease have pneumonia and in addition may have clinical findings suggestive of a systemic disease (Edelstein and Cianciotto, 2005). Symptoms and signs of the disease are often quite variable. The majority of patients have fever, which is usually one of the earliest signs of the illness. Accompanying the fever may be anorexia, myalgia, rigors, and headache. Chest pain, shortness of breath and cough may or may not be prominent findings. The cough may or may not be productive, and when it is productive the sputum can be bloody, purulent, or scant and mucoid. In some patients the absence of purulent sputum production, chest pain and cough may fool clinicians into discarding pneumonia as a possibility. When chest pain and haemoptysis are prominent the patient may be suspected of having a pulmonary infarction. Abdominal pain, diarrhoea, nausea and vomiting may occur as well, symptoms that have led to consideration of intra-abdominal infections and inflammatory conditions such as appendicitis, peritonitis, abscesses, inflammatory bowel disease and diverticulitis. Elderly and immunocompromised patients may not have fever or findings that localize to the lung. Confusion and memory loss are common presenting findings. Much less common are frank encephalopathy, focal neurological findings, seizures and

As the untreated disease progresses, the major findings include fever or hypothermia, dense consolidation of the lung, and often respiratory failure. The majority (~70%) of non-immunocompromised patients without significant underlying diseases recover without specific therapy after being quite ill for 5-7 days. Death is usually the consequence of respiratory failure, often combined with sepsis

Hyponatraemia, leucopenia or leucocytosis, thrombocytosis or thrombocytopenia, elevated serum creatinine kinase and elevated liver-associated tests are all non-specific laboratory findings that can be commonly observed. In the presence of severe respiratory failure, thrombocytopenia and evidence of disseminated intravascular coagulation is common. Chemical evidence of pancreatitis is an occasional finding, as is evidence of myocarditis.

Chest radiography almost always reveals alveolar filling infiltrates, often with consolidation (Kirby et al., 1979; Tan et al., 2000). Purely interstitial lung infiltrates are very uncommon although such infiltrates may be the only finding very early in the disease process. Cavitation of prior areas of consolidation occurs in up to 10% of immunocompromised patients. Pleural effusions are seen in about 40% of patients, often in patients with other causes for this such as congestive heart failure and renal failure. However, a small fraction of people may have a small pleural effusion without apparent chest infiltrates as the only radiographic manifestation of the disease.

Physical examination findings that are most common include pulmonary râles and other evidence of lung consolidation, tachypnoea and tachycardia. Confusion and memory loss are also relatively common findings.

Patients treated with specific antibiotic therapy usually improve promptly, with the systemic signs of infection such as confusion and sepsis clearing most rapidly. Up to a week may be required for patients to become completely afebrile, although the fever starts to decrease within 12 hours of the initiation of therapy. Immunocompromised patients or patients with very advanced pneumonia may have either no response or a very slow response to specific therapy, with prolonged fever and respiratory failure. Fatality rates are the highest in patients treated late in the disease, with the fatality rate approaching 70% in those with severe respiratory failure on presentation, even with specific therapy.

Extrapulmonary manifestations of Legionnaires' disease can be of two types: organ or tissue disease without obvious infection, and metastatic or contiguous infection. Examples of the former type include sepsis syndrome, hepatitis, pancreatitis, myositis, myopericarditis, and encephalitis; the most common of which are sepsis syndrome, myositis and pancreatitis (Edelstein and Cianciotto, 2005; Edelstein, 2006a). Examples of the latter type include pleural empyema, purulent pericarditis, vascular graft infections, renal and brain abscesses, purulent peritonitis, bowel abscesses, pyarthrosis, and others; of these only pleural empyema has been reported in more than a handful of cases (McClelland et al., 2004).

Not all of the clinical manifestations of Legionnaires' disease may be due to infection with Legionella spp., as up to 10% of patients with this disease have coinfections with other microbes. Documented coinfecting bacteria have included Streptococcus pneumoniae, Haemophilus influenzae, Staphylococcus aureus, Escherichia coli, Aspergillus, Rhizopus, cytomegalovirus, Pneumocystis jirovecii and many other organisms (Meyer et al., 1980; Ruutu et al., 1987; Marrie et al., 1992; Edelstein and Cianciotto,

Legionella spp. infection in the absence of Legionnaires' disease is very uncommon (McClelland et al., 2004). Most such cases involve direct inoculation of injured tissues with water containing Legionella spp. One example includes L. pneumophila and L. dumoffii sternal wound infections after bathing post-cardiac surgery patients with water colonized with L. pneumophila and L. dumoffii (Lowry et al., 1991). Another example includes L. pneumophila wound infection after bathing a wound in L. pneumophila-colonized water (Brabender et al., 1983). read more ...

Distinguishing Legionnaires' disease from other causes of pneumonia

Rapid diagnosis of Legionnaires' disease allows the use of therapy specific for the disease, and prompt notification of public health authorities that may curtail epidemics. Therefore, major efforts have been made to discover clinical findings specific for the disease. However, the presenting signs and symptoms of Legionnaires' disease are indistinguishable from those found in people with other common causes of community-acquired pneumonia, such as that due to Streptococcus pneumoniae. Although Legionella-specific laboratory testing is required to accurately diagnose the disease, these tests may be expensive, insensitive, unavailable or require many days to perform; details of these tests are found in Chapter 2. The similarity of the clinical findings of Legionnaires' disease to those of other pneumonias, and the deficits of specific diagnostic testing, led to interest in defining specific clinical findings that could be used to diagnose the disease quickly without resort to specific laboratory testing.

It was originally thought that Legionnaires' disease was a distinct and easily distinguishable clinical syndrome, characterized by rigors, absence of a productive cough, high fever with a disproportionately low pulse rate, headache, myalgia, anorexia, nausea and diarrhoea. In addition it was thought that several non-specific laboratory abnormalities were characteristic of the disease, including hyponatraemia, elevated liver-associated enzyme levels, hypophosphataemia, proteinuria and myoglobinuria, and elevated levels of muscle-associated enzymes (Tsai et al., 1979; Kirby et al., 1980; Lattimer and Ormsbee, 1981). Some thought that the radiographic presentation of the disease was also characteristic, with rapidly progressive multifocal consolidating infiltrates (Kirby et al., 1979). However, a seminal study by Yu and colleagues showed that it was impossible to distinguish the presenting clinical findings of Legionnaires' disease from other common causes of community-acquired pneumonia (Yu et al., 1982). This study was validated by several others, including studies of laboratory findings, radiography and nosocomial pneumonia (Sopena et al., 1998; Mulazimoglu and Yu, 2001).

In each of the prospective studies comparing the clinical presentation of Legionnaires' disease with that of other causes of pneumonia some clinical or laboratory finding was found to be more common in patients with Legionnaires' disease. The most common ones were a greater frequency of hyponatraemia and of diarrhoea in patients with Legionnaires' disease. For example, in one study hyponatraemia was found in 44% of patients with Legionnaires' disease and in 14% of patients with other pneumonias (Yu et al., 1982). However, there was enough overlap in these findings to make them unsuitable to distinguish unequivocally between Legionnaires' disease and other pneumonias (Edelstein, 2006a). In addition, some significant differences in laboratory abnormalities differed by relatively small amounts. In one study patients with Legionnaires' disease had mean serum sodium concentrations of 132.6 meq./L versus 135.7 meq./L for the patients with other pneumonias; both groups had hyponatraemia with a difference that can be observed from day to day in the same patient (Fernandez et al., 2002).

Attempts at combining typical signs and symptoms of Legionnaires' disease into a diagnostic scoring system have failed. Two scoring systems, the Winthrop University criteria and the CBPIS criteria, were unable to accurately classify patients with and without Legionnaires' disease in prospective studies (Gupta et al., 2001; Fernandez-Sable et al., 2003). Even still, it appears as if there may be a clinically recognizable and distinct presentation of Legionnaires' disease which has yet to be captured by a clinical scoring system. In the study of the CBPIS, emergency department physicians were able to correctly suspect Legionnaires' disease in 64% of patients diagnosed with the disease, whereas they suspected the disease in only 4% of patients with pneumococcal pneumonia (Fernandez-Sabe et al., 2003). In addition, specific therapy for Legionnaires' disease was given to 89% of those who turned out to have the disease, but only to 18% with pneumococcal pneumonia. While imperfect, experienced clinicians can accurately suspect Legionnaires' disease. Unfortunately, depending exclusively on clinical impressions would mean that at least 10% of Legionnaires' disease would not be treated for the disease, an unacceptable figure. Clinical diagnostic inaccuracy would likely be much higher if the physician assessors had less experience with Legionnaires' disease. So it is fair to say that many people with Legionnaires' disease have a distinct clinical syndrome, but that so far it is impossible to detect all patients with Legionnaires' disease using solely clinical and non-specific laboratory findings.

Analysis of inflammatory marker levels has been another approach to aid in the diagnosis of Legionnaires' disease. Despite the theoretical appeal of using markers that are most elevated in intracellular (neopterin) versus extracellular (procalcitonin) infections, measurement of relative levels of these two biomarkers is not useful for making a specific diagnosis of Legionnaires' disease (Prat et al., 2006). This should not be surprising as procalcitonin is driven by inflammatory cytokines that L. pneumophila apparently induces to downregulate immune control of the bacterium by interferon gamma (Yoshizawa et al., 2005). Depending on the stage of the infection, variable ratios of neopterin (interferon gamma driven) to procalcitonin would be expected. Whether there is a specific inflammatory signature of Legionnaires' disease has not been examined, but studies in macrophages show promise for this type of approach (Losick and Isberg, 2006; Stetson and Medzhitov, 2006). Acquiring microarray data on Legionnaires' disease patients during different stages of infection would help answer this question once daunting technical issues are addressed (Liu et al., 2006). In the meanwhile preparation and storage of cDNA from the blood of such patients should be considered. read more ...

Treatment of Legionnaires' disease

Effective treatment of Legionnaires' disease requires the use of antimicrobial agents that are active against intracellular Legionella spp. Many different types of antimicrobial agents are active against the bacteria in vitro, but not necessarily in vivo, because of this need to inhibit or kill the intracellular bacterium. For example, both cefoxitin and gentamicin are very active against L. pneumophila in vitro, but are ineffective for treating Legionnaires' disease. Neither of these two drugs inhibits or kills a sufficient number of intracellular L. pneumophila to prevent intracellular growth of the bacterium. Examples of drug classes that are active for the treatment of Legionnaires' disease include the macrolides, azalides, ketolides, tetracyclines and antibacterial quinolones. Erythromycin, dirithromycin, clarithromycin, azithromycin, telithromycin, tetracycline, levofloxacin, and ciprofloxacin are all drugs that can be used to treat the disease. The most active drugs, used for severely ill or immunocompromised patients, are azithromycin and levofloxacin. Treatment duration is dependent on the drug given, the degree of immunosuppression, the presence of metastatic or contiguous infection, and the clinical course of the patient. Details about antimicrobial agent therapy of this disease can be found in Chapter 2 of this book and elsewhere (Edelstein, 1995; Roig and Rello, 2003; Sabria and Campins, 2003; Edelstein and Cianciotto, 2005). Recovery from the acute infection may be slow and plagued by fatigue, memory loss, and post-traumatic stress disorder, complications common to many types of community-acquired pneumonia (Marrie et al., 2000; Lettinga et al., 2002; Marrie et al., 2004).

from Legionella: Molecular Microbiology