Legionnaires' disease is a form of pneumonia that is caused by Legionella pneumophila, a bacterium that breeds in warm, moist conditions and stagnant water. Legionnaires’ disease can occur in outbreaks. The source of infection is often an air-conditioning system in a large public building; the disease is contracted by the inhalation of droplets of contaminated water.

The first symptoms include headache, muscular and abdominal pain, diarrhoea, and a dry cough. Over the next few days, pneumonia develops, resulting in high fever, shaking chills, coughing up of thick sputum (phlegm), drowsiness, and sometimes delirium. Liver and kidney damage may occur.

Treatment is with antibiotic drugs such as clarithromycin and rifampicin. The majority of people recover but mortality rates are higher among the elderly.

Legionellosis and legionnaires’ disease in more detail

Essentials

In 1976, an outbreak of pneumonia affected 221 and killed 34 members of the American Legion who had attended a convention in a Philadelphia hotel. A newly identified organism, Legionella pneumophila, was discovered and named after the outbreak. Since then many different species of the family Legionellaceae have been discovered. Clinical illness is referred to as legionellosis, and there are two principal syndromes: legionnaires’ disease (pneumonia) and Pontiac fever (a self-limiting influenza-like illness).

Aetiology and pathology

The organism

The Legionellaceae are aerobic nonsporing Gram-negative bacilli whose cell walls contain distinctive branched-chain fatty acids and lipo-oligosaccharide (LOS). Of the 50 formally recognized legionella species, L. pneumophila is the principal cause of human infections. Of the 16 or more serogroups (SG), L. pneumophila SG1 is the most pathogenic and responsible for most cases. L. pneumophila SG1 can be further subdivided by monoclonal antibody and molecular typing, which is useful for outbreak investigations.

Infections with other serogroups or species (e.g. L. micdadei and L. bozemanii) can occur in patients who are highly immunocompromised (e.g. transplant recipients). In some parts of Australia, L. longbeachae is the commonest species causing legionnaires’ disease.

Pathology

Legionellae are intracellular pathogens that are found within protozoa in the environment and in alveolar macrophages in humans. Following inhalation of contaminated aerosol droplets, legionellae reach the alveoli where they are internalized in macrophage endosomes. They block the development of the endosome into a phagolysosome, preventing the normal cellular bacterial killing mechanism through the action of an important virulence factor, the macrophage infectivity potentiator (mip) protein.

The lungs are the principal organ affected and show a severe inflammatory response. The alveoli and terminal bronchioles are distended by fibrin-rich debris, mononuclear inflammatory cells, and neutrophils. Organisms can be demonstrated within alveolar spaces by silver or immunofluorescence stains. In survivors, alveolar and interstitial fibrosis can result.

Epidemiology

The natural worldwide habitats of legionellae are freshwater streams, lakes, thermal springs, moist soil, and mud, where they are found in small numbers. They usually live and multiply within amoebae and other protozoa where they are protected from adverse condition and can survive and disseminate widely.

By contrast, in artificially constructed water systems, legionellae can multiply to extremely high numbers, encouraged by favourable temperatures (20–45° C) and water stagnation. As legionellae are associated with amoebae within the biofilm, complete eradication is difficult once systems are colonized.

The principal source for large outbreaks of legionellosis is wet (or evaporative) cooling systems (cooling towers) used for air conditioning and other cooling equipment. Cooling towers are commonly seen on the outside walls or roofs of buildings such as hotels, office blocks, hospitals, and factories. If poorly maintained, they can become heavily contaminated with legionellae leading to the emission of an infectious aerosol of legionella-containing droplets. Such aerosols can drift 500 m or more, depending on the position of the cooling tower and the climatic conditions.

Within buildings, legionellae commonly multiply in cold-water storage tanks, hot-water calorifiers, and in the hot and cold water distribution pipework, particularly if long and complicated runs of pipework lead to a loss of temperature control, or water stagnation (‘dead-legs’). Contaminated aerosols are most commonly disseminated by showers, but other well-recognized sources include:

• ◆ whirlpool spas and other warm-water baths
• ◆ decorative fountains
• ◆ respiratory therapy equipment rinsed or topped up with contaminated tap water
• ◆ automatic car washes
• ◆ potting compost (for L. longbeachae SG1 in Australia).

In temperate countries, legionellosis is seasonal with most cases occurring in the summer and autumn. The same number are related to travel (either within the same country or more commonly abroad) as are acquired locally. A history of recent travel can be an important pointer to legionella infection. Locally acquired legionellosis is increasingly recognized as being domestically acquired.

Hospital-acquired legionellosis is uncommon, but may involve less pathogenic legionella strains affecting a highly susceptible or immunosuppressed patient population in small clusters.

Prevention

Several primary preventive measures can be taken to minimize the risks of acquiring legionellosis from water systems. Cooling towers must be registered with local authorities and regularly maintained, using biocide treatment to inhibit legionella growth; sampling for the presence of legionellae within the recirculating water must be carried out regularly. Hot and cold water systems must be adequately designed and maintained to minimize legionella growth, either through temperature control or use of chemicals, ozonation, or point-of-use filtration. In health care facilities, a balance needs to be struck between adequate water temperature at outlets and risks of scalding. Outlets that are not in regular use should be regularly flushed through to avoid water stagnation.

Although legionellosis is not a formally notifiable disease, even single cases should be reported to public health authorities for investigation of possible environmental sources. Continuing surveillance of legionellosis is important at local, national, and international levels. By collating data, coordinated European surveillance systems have been able to pinpoint outbreaks of legionnaires’ disease associated with a particular holiday resort or hotel.

Clinical features

Legionella pneumonia

Legionella infection tends to lead to moderate or severe illness usually requiring hospital admission within 5 to 7 days. It causes 2 to 5% of cases of community-acquired pneumonia admitted to hospital (but with wide geographical and seasonal variation) and is the second commonest community-acquired pneumonia requiring intensive care.

The incubation period is usually 2 to 10 days. Men are 2 to 3 times more frequently affected than women. Infection in children and elderly people is unusual and the highest incidence is in 40- to 70-year-old people, with a mean age of 53 years. People particularly at risk include cigarette smokers, alcoholics, diabetics, and those with chronic illness or who are receiving corticosteroids or immunosuppressive therapy.

Clinical features

Typically, the illness starts fairly abruptly and progresses quickly with high fever, shivers, bad headache, and muscle pains. Respiratory symptoms such as cough and breathlessness can sometimes be minimal, with confusion and diarrhoea dominating the clinical picture, masking the true diagnosis of pneumonia. The patient commonly looks ill, with a high fever over 39° C, signs of pneumonia, and confusion (in one-half of patients).

Differential diagnosis

Table 1 below compares features of legionella pneumonia with other types of community-acquired pneumonia. No unique pattern allows the early clinical differentiation of legionella infection from other, more common, causes of pneumonia. Important clues include epidemiological pointers (e.g. recent foreign travel or a local epidemic), high fever, confusion, multisystem involvement, absence of a predominant bacterial pathogen on sputum examination, and lack of response to β-lactam antibiotics.

Clinical investigation

The total white cell count is usually only moderately raised (to 15 × 10⁹/litre), often with a lymphopenia. Hyponatraemia, hypoalbuminaemia, and abnormal liver function tests are detected in more than one-half of the cases (Table 1). Other nonspecific features may include raised blood urea and muscle enzymes, very high C-reactive protein, hypoxaemia, haematuria, and proteinuria. Gram’s staining of sputum typically shows few pus cells and no predominant pathogen. Initial blood and sputum cultures are negative.

Radiographic features

Radiographic shadowing is usually homogeneous. Characteristically, radiographic deterioration occurs within the same or opposite lung. Radiographic improvement is particularly slow. Only two-thirds of radiographs clear within 3 months and some take more than 6 months.

Prognosis and complications

A wide variety of complications have been reported, the most important being acute respiratory failure requiring assisted ventilation which occurs in up to 20% of patients. In addition to confusion, various neurological complications have been reported, leading to the suggestion of a neurotoxin. Acute but usually reversible renal failure may be seen in severe disease. Clinical recovery appears to be very slow in some patients, particularly of symptoms such as tiredness, weakness, breathlessness, memory and concentration impairment, and psychological sequelae. This can have medicolegal implications in those infected through negligent exposure to poorly maintained water systems.

Pontiac fever

This is the acute nonpneumonic form of legionella infection that presents as an influenza-like illness. The attack rate is extremely high, with an incubation period of usually 36 to 48 h. Investigations and chest radiograph are normal, and the illness improves spontaneously, usually within 5 days.

Laboratory diagnosis

A variety of laboratory methods can be used to diagnose legionella infection. In order of usefulness, they are:

1. 1 Antigen detection:
   
   1. a Urinary antigen detection
   2. b Direct immunofluorescence
2. 2 Culture
3. 3 Serology
4. 4 Molecular methods (e.g. polymerase chain reaction (PCR))

Urine is a readily available clinical sample. Diagnosis by urinary antigen detection has now become the mainstay of diagnosis in many centres, usually becoming positive at an early stage of infection and remaining positive for several weeks. Several well-validated commercial enzyme immunoassays and an immunochromatography test are available with excellent specificity and good sensitivity. Immunochromatography can give results in as little as 15 min. It is recommended that legionella urine antigen tests are performed for all patients with severe community-onset pneumonia. Their principal drawback is that only L. pneumophila SG1 infection is detected. This is an important limitation, particularly in immunocompromised patients in whom every effort should be made to obtain a positive culture. A negative urine antigen test does not exclude legionella infection and the test should be repeated as clinically indicated.

Direct immunofluorescence with a monoclonal antibody specific for L. pneumophila can be used to detect bacteria in suitable respiratory specimens. This technique can provide a diagnosis early in the course of the infection, but is relatively time consuming and relies on the availability of a good respiratory tract sample (e.g. bronchoalveolar lavage (BAL) fluid).

Legionellae can be cultured from suitable respiratory samples (e.g. sputum, endotracheal aspirates, and BAL fluid) using appropriately enriched and permissive agar such as buffered charcoal yeast extract. Culture is diagnostic of infection, as colonization without infection has not been demonstrated, but it is time consuming, expensive, slow (culture can take up to 10 days), and relatively insensitive (especially once legionella active antibiotic therapy has been started). Culture does allow detection of species and serogroups other than L. pneumophila SG1, and comparison with isolates from suspected environmental sources. Such speciation and typing, using well-validated methods such as a DNA sequence-based typing scheme, is normally done in reference laboratories.

In the past, serology (i.e. the detection of an antibody response to legionella) was the mainstay of diagnosis, and this is still of value. Properly evaluated serological assays (especially those with sufficient specificity) are based on detecting antibodies to L. pneumophila SG1. Antibody responses can be delayed or absent in some patients, but about 40% of patients admitted to hospital will have raised antibodies on admission. A confirmed serological diagnosis involves demonstrating a fourfold or greater rise in antibody titre in suitably timed paired sera. A single high titre suggests infection. False-positive results can occur in some patients with recent campylobacter infection. In these cases, serology should be repeated in the presence of a campylobacter blocking fluid.

The detection of species and subtype-specific legionella DNA in clinical samples by PCR is available in reference laboratories and offers good sensitivity. However, the lack of commercially available assays limits widespread diagnostic value.

Treatment

There are no randomized controlled trials of antibiotic therapy for legionellosis and evaluation of agents has been based on relatively small case studies as well as in vitro and animal experiments. Macrolides (erythromycin and more recently clarithromycin) have for many years been regarded as the antibiotics of first choice, mainly based on clinical experience and retrospective analysis of outcome data. One of the key factors is the ability of an antibiotic to reach therapeutic concentrations within alveolar macrophages where the legionella bacteria multiplies. There have been increasing reports of the successful use of fluoroquinolone antibiotics, notably levofloxacin, and even some suggestion that, when combined with early diagnosis using urine antigen detection, they can reduce mortality and morbidity compared to traditional macrolide therapy. Fluoroquinolones demonstrate excellent bioavailability, bactericidal activity against legionellae, and very good intracellular penetration.

For nonsevere community-acquired legionella infection, our practice is to use an oral fluoroquinolone, with a macrolide as an alternative.

For the management of severe or life-threatening legionella pneumonia, specialists in this area consider the use of a combination of antibiotics including a fluoroquinolone and a macrolide, especially during the crucial first few days, with rifampicin as an alternative if one of these agents cannot be used. Clinicians should be alert to the potential small risk of prolongation of the QT interval on the ECG with the recommended combination, particularly in the presence of other proarrhythmic risk factors. Parenteral rifampicin has a risk of hyperbilirubinaemia, which usually resolves on stopping the drug. There are also reports that the azalide antibiotic azithromycin and tetracyclines (e.g. doxycycline) may be useful.

Prognosis

The patient’s previous health and appropriate early therapy are the two most important factors determining outcome. Mortality in previously fit patients is 5 to 15%. It is lower when early diagnosis by urine antigen detection allows prompt treatment. The mortality is approximately 30% in those requiring assisted ventilation, but in immunosuppressed individuals it can approach 75%.

Areas of uncertainty

Three main areas of uncertainty are the pathophysiology of multisystem involvement, (particularly of neuropsychological symptoms), optimal antibiotic management, and the long-term prognosis.

Likely future developments

Advances in bedside urine antigen testing create the possibility of early diagnosis and properly controlled trials of antibiotic therapy, together with controlled follow-up studies to assess long-term sequelae. Near-source rapid testing of water may enhance surveillance of water systems and outbreak investigations, especially if there is progress in detecting species other than L. pneumophila SG1.