To the Editor—Legionella pneumophila (Lp) serogroup 1 is a well-known cause of hospital-acquired pneumonia. Extrapulmonary manifestations as well as Lp bacteremic pneumonia are rare and occur mainly in immune-compromised subjects.Reference Lowry and Tompkins ^{1 –} Reference Marrie and Martin ⁵

Potable water has been implicated in nosocomial cases of Lp infection via inhalation of contaminated aerosols from hot-water systems.Reference Bollin, Plouffe, Para and Hackman ⁶ Culture is the gold standard for the diagnosis of Legionella infection, even though culture-proven legionellosis cases are not frequently reported and documented cases of bacteria isolation from the blood are even less common.Reference Lee, Lai and Exner ⁷

On January 5, 2014, a 58-year-old woman with a clinical history of alcoholic liver cirrhosis was admitted for worsening dyspnea to the Emergency Department of IRCCS AOU San Martino–IST Hospital, Genoa, Italy. On admission, the patient presented peripheral edema and signs of portal hypertension. Chest and abdominal radiograph did not detect pleuro-parenchymal lung lesions.

The same day, the patient was transferred to the gastroenterologic unit where she stayed until February 6, 2014. She underwent 3 paracentesis procedures, a transjugular intrahepatic portosystemic shunt, and 3 angiography controls because of a progressive deterioration in liver function.

On February 6, 2014, the patient experienced a severe respiratory failure due to acute pulmonary edema. She was intubated and chest radiograph (Figure 1) detected pleuro-parenchymal lung lesions compatible with pneumonia.

FIGURE 1

Chest radiograph of the patient performed during the hospital stay in the intensive care unit 1 day before death.

Blood cultures (Bactec Plus aerobic; BD) were performed and the patient was empirically treated with piperacillin-tazobactam. No specific diagnostic test for Legionella (culture, urinary antigen, serology) was performed. On February 7, 2014, the patient died because of respiratory and multiorgan failure.

After 6 days, blood culture became positive and subculture on blood agar (Kima; Megalab) and chocolate-enriched agar plates (Kima) was performed, showing the growth of colonies that, unexpectedly, turned out to be Lp by matrix-assisted laser/desorption-ionization time-of-flight mass spectrometry identification technique (bioMérieux).

Extensive environmental investigation was performed by the hospital’s hygiene unit to identify the source of the infection.

No sources of infection could be detected in the operating room and no cooling towers or evaporative condensers were present in the neighborhood of the hospital. Invasive procedures, such as transjugular intrahepatic portosystemic shunt and paracentesis, were excluded as potential sources of infection because of use of disposable equipment and absence of water contact.

Hot- and cold-water samples were collected from the rooms (washbasins, showers, heaters) of all the wards where the patient was hospitalized, and analyzed at the Legionella Regional Reference Laboratory, according to national guidelines. All the samples had negative results (Lp<100 colony-forming units/L) as expected, as the hospital’s water system was continuously disinfected by a chlorine dioxide system and monitored for biocide level monthly.

Concomitantly, the clinical isolate and an aliquot of the collected water samples were sent to the Legionella National Reference Laboratory. A latex agglutination test (Oxoid) confirmed the identification of the clinical isolate as Lp serogroup 1. Typing assays performed by monoclonal antibodies and sequence-based typing typed the clinical strain as Olda and sequence type 154, respectively.Reference Scaturro, Fontana and Ricci ⁸ Furthermore, all water samples were analyzed by real-time polymerase chain reaction assay, using the iQ-Check quantification kit (BioRad), resulting in Legionella positive (7×10² to 2×10³ genomic units/L).

Further sampling was carried out by the hygiene unit, increasing the volume of analyzed water to 2 liters and the amount spread on selective medium to 0.5 mL. The increased volume of the water tested allowed the isolation of Lp serogroup 1 at very low concentration (20 colony-forming units/L) from the hot-water samples collected from one of the taps routinely used by the patient. Monoclonal antibody typing and sequence-based typing performed at Legionella National Reference Laboratory showed that the environmental and clinical isolates were related and were both Olda monoclonal antibody subgroup and sequence type154, thus revealing the source of infection.

In the case here described, Lp infection was diagnosed postmortem by positive blood culture, whereas neither urine antigen assay nor sputum culture were performed.

The occurrence of a Legionella infection 31 days after the patient’s hospital admission was consistent with a nosocomial case, thus justifying the extensive environmental investigation that eventually led to the isolation of Lp from the tap of the washbasin used by the patient. This route of transmission of Legionnaires’ disease, rarely described in literature,Reference Brûlet, Nicolle and Giard ⁹ was probably favored by the severe immune deficiency of the patient, a consequence of the alcoholic liver cirrhosis at its final stage. The isolation of Legionella from blood was also remarkable.

Lp growth in blood culture is seldom investigated, although this occurrence is not so infrequent especially in aerobic condition.Reference Kaku, Yanagihara and Morinaga ¹⁰ The scanty use of blood culture probably comes from the bias that Legionella is not able to grow in broth media utilized for blood culture. Indeed, these media, as well as blood and chocolate-enriched agar plates used for Legionella subculture, are lacking in ferric pyrophosphate and L-cysteine essential for Lp growth. However, we can speculate that small amount of the essential nutrients might be supplied by erythrocytes and by the metabolism of other blood cells, allowing the growth of Legionella in blood culture.

In conclusion, this case highlights that Legionnaires’ disease should be always suspected in severely immune-compromised patients when clinical and radiologic findings are suggestive for pulmonary infection, and the patient has been hospitalized for a long period, even in the absence or near-absence of Legionella water system contamination provided by a routine surveillance program. Another lesson learned by this case is that the adoption of blood culture, in addition to respiratory secretion culture, to isolate Legionella needs to be considered.