Legionella sainthelensi pneumonia associated with aquatic traumatic injury in New South Wales, Australia.

Legionella species are important water-associated pathogens of public health interest and are classically associated with pneumonia in humans. In Australia, the most common species isolated from clinical samples are L. pneumophila and L. longbeachae. Legionella sainthelensi was first reported in 1980 from water near Mount St Helens, USA,1Campbell J. Bibb W.F. Lambert M.A. et al.Legionella sainthelensi: a new species of Legionella isolated from water near Mt. St. Helens.Appl Environ Microbiol. 1984; 47: 369-373Crossref PubMed Scopus (30) Google Scholar and is a rarely isolated species which has been associated with severe pulmonary disease, as well as outbreaks.2Kamus L. Roquebert B. Allyn J. et al.Severe bilateral pleuropneumonia caused by Legionella sainthelensi: a case report.BMC Infect Dis. 2021; 21: 966Crossref PubMed Scopus (1) Google Scholar, 3Slow S. Anderson T. Biggs P. Kennedy M. Murdoch D. Cree S. Complete genome sequence of Legionella sainthelensi isolated from a patient with Legionnaires’ disease.Genome Announc. 2018; 6: e01588-e01617Crossref PubMed Scopus (4) Google Scholar, 4Benson R.F. Thacker W.L. Fang F.C. Kanter B. Mayberry W.R. Brenner D.J. Legionella sainthelensi serogroup 2 isolated from patients with pneumonia.Res Microbiol. 1990; 141: 453-463Crossref PubMed Scopus (19) Google Scholar, 5Loeb M. Simor A.E. Mandell L. et al.Two nursing home outbreaks of respiratory infection with Legionella sainthelensi.J Am Geriatr Soc. 1999; 47: 547-552Crossref PubMed Scopus (37) Google Scholar Distribution appears to be broad, with clinical isolates described from Reunion Island, New Zealand, and Canada, and environmental isolates from Portugal, China, USA and the Netherlands.6Costa J. Tiago I. da Costa M.S. Veríssimo A. Presence and persistence of Legionella spp. in groundwater.Appl Environ Microbiol. 2005; 71: 663-671Crossref PubMed Scopus (56) Google Scholar,7Veríssimo A. Marrão G. da Silva F.G. da Costa M.S. Distribution of Legionella spp. in hydrothermal areas in continental Portugal and the island of São Miguel, Azores.Appl Environ Microbiol. 1991; 57: 2921-2927Crossref PubMed Scopus (38) Google Scholar Here, we report the first described case of L. sainthelensi pneumonia in NSW, Australia, and the challenges in diagnosing infections caused by uncommon Legionella species using routine laboratory methods. Formal informed consent was obtained from the patient. A 48-year-old woman was admitted to a tertiary trauma referral hospital following a fall onto submerged rocks. She sustained significant injuries including bilateral pneumothoraces, multiple fractures, and soft tissue lacerations, and required intubation on the scene. Her background history was significant for depression and chronic pain only. On day 9 of admission, she developed mixed respiratory failure, with worsening bilateral lung infiltrates on chest X-ray. Routine sputum culture from endotracheal aspirate isolated light growth of Haemophilus influenzae and Acinetobacter baumannii complex. Empirical intravenous piperacillin/tazobactam 4 g/0.5 g Q8H was commenced, but was later switched to cefepime 2 g Q8H to treat a presumed A. baumannii ventilator associated pneumonia. On day 12, intravenous erythromycin 250 mg Q6H was commenced for prokinetic effect due to ongoing colonic pseudo-obstruction affecting her ventilation. A bronchoscopic evaluation and lavage was performed on day 14 due to progressive pulmonary disease and antibiotic treatment was escalated to intravenous meropenem 1 g Q8H. As legionella and nocardia culture was specifically requested, the specimen was also inoculated onto buffered charcoal yeast (BCYE) agar, with and without selective antibiotics, and incubated at 37°C in aerobic conditions for 14 days. On day 14, a Gram-negative bacillus was isolated from the right upper lobe washings. This was presumptively identified as L. sainthelensi on matrix-assisted laser desorption/ionisation time of flight (MALDI-TOF) mass spectroscopy (Bruker Daltonik, Germany). No other significant pathogens were isolated. The isolate was referred to the state reference laboratory. Latex agglutination was performed as per manufacturer's instructions against L. pneumophila serogroup 1, L. pneumophila serogroups 2–14 and other Legionella spp. including L. bozemanii, L. dumoffii, L. gormanii, L. longbeachae, L. micdadei, L. jordanis and L. anisa (Legionella Latex Test; ThermoFisher, USA). No agglutination was observed. The isolate also underwent 16S rRNA U1 and U3 polymerase chain reaction (PCR) and amplicon sequencing.8James G. Universal bacterial identification by PCR and DNA sequencing of 16S rRNA gene.in: Schuller M. Sloots T. James G. Halliday C. Carter I. PCR for Clinical Microbiology. Springer Netherlands, Dordrecht2010: 209-214Crossref Scopus (68) Google Scholar A basic local alignment search tool against the NCBI GenBank database identified Legionella genus, but did not discriminate to species level. The isolate underwent whole genome sequencing to confirm species level identification. Genomic DNA was extracted with a Qiagen kit, quality checked, and genomic libraries constructed using Nextera XT Library Preparation kit (Illumina, USA). Sequencing was performed on the Illumina NextSeq500 platform. The sequence data were first passed through an in-house quality control procedure, including assessment of read quality and contamination (Trimmomatic version 0.36, FastQC version 0.11.3 and Centrifuge version 1.0.4). The core-genome alignment was produced by Roary version 3.6.1. The trimmed reads were first assembled using SKESA version 2.3.0, followed by genome annotation using Prokka version 1.13.3. This confirmed the genome identity as L. sainthelensi. No known mutations which could confer in vitro resistance of bacteria to fluoroquinolones were identified by analysis of the gyrA, gyrB and parC genes in the QRDR regions of the L. sainthelensi genome. There was no evidence of macrolide resistance mutations in the reads of the 23s rRNA (either A2058G or C2611G) nor any mutations in rplD or rplV genes. Mutations in the lpeAB efflux pump promoter region, which have been associated with increased minimum inhibitory concentration (MIC) to macrolides in L. pneumophila, were not present as the L. sainthelensi genome appears to lack this operon.9Massip C. Descours G. Ginevra C. Doublet P. Jarraud S. Gilbert C. Macrolide resistance in legionella pneumophila: the role of LpeAB efflux pump.J Antimicrob Chemother. 2017; 72: 1327-1333PubMed Google Scholar MIC of selected antimicrobial agents against both L. pneumophila ATCC35152 and L. sainthelensi were performed by gradient test strip (ETEST; bioMérieux, France), as per the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidance document on susceptibility testing of L. pneumophila10European Committee on Antimicrobial Susceptibility TestingGuidance Document on Antimicrobial Susceptibility Testing of Legionella pneumophila.May 2021https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Guidance_documents/Legionella_guidance_note_-_20210528.pdfGoogle Scholar (Table 1). While no clinical breakpoints exist for L. sainthelensi, the tested MICs were below the tentative highest MIC for wild-type L. pneumophila isolates.Table 1Minimum inhibitory concentration (MIC) of selected antimicrobial agents against Legionella pneumophila ATCC35152 and Legionella sainthelensi, performed by gradient test stripAntibioticL. pneumophila ATCC 35152 MIC (μg/mL)L. sainthelensi MIC (μg/mL)Erythromycin0.50.5Ciprofloxacin0.50.5Doxycycline82.0 Open table in a new tab Legionella nucleic acid amplification testing (NAAT) was retrospectively performed on bronchial washing samples using an in-house PCR method on the Lightcycler II (Roche Diagnostics, Switzerland) targeting a genus-conserved 16S target and a L. pneumophila specific target.11Reischl U. Linde H.J. Lehn N. Landt O. Barratt K. Wellinghausen N. Direct detection and differentiation of Legionella spp. and Legionella pneumophila in clinical specimens by dual-color real-time PCR and melting curve analysis.J Clin Microbiol. 2002; 40: 3814-3817Crossref PubMed Scopus (88) Google Scholar The genus specific target only was detected; melt-curve showed a similar profile to L. longbeachae (Fig. 1). The samples were also tested on the Atypical Pneumonia PCR 8-well assay (AusDiagnostics, Australia); neither L. pneumophila nor L. longbeachae were detected. Legionella immunofluorescence serology for L. longbeachae serogroup 1 and L. pneumophila serogroup 1 (Euroimmun, Germany) collected on days 45 and 51 were non-reactive. Unfortunately, no urine samples were available for legionella urinary antigen testing. At the time of isolation of L. sainthelensi, our patient had received a total of 6 days of treatment with erythromycin and given her overall respiratory improvement, she was not prescribed further specific therapy. No other Legionella-active antimicrobials were administered during her admission. While the optimum therapy for L. sainthelensi pneumonia is unclear, the isolate had a low erythromycin MIC on susceptibility testing. Her remaining ICU stay was complicated by peritonitis and slow respiratory recovery from ARDS. She was discharged home after a prolonged inpatient admission including rehabilitation. This case demonstrates the difficulty in establishing the diagnosis of uncommon Legionella species infection with current diagnostic techniques. Similar to other Legionella species, L. sainthelensi has a strong association with water.1Campbell J. Bibb W.F. Lambert M.A. et al.Legionella sainthelensi: a new species of Legionella isolated from water near Mt. St. Helens.Appl Environ Microbiol. 1984; 47: 369-373Crossref PubMed Scopus (30) Google Scholar,12NNDSS Annual Report Writing GroupAustralia’s notifiable disease status, 2012: Annual report of the National Notifiable Diseases Surveillance System.Commun Dis Intell Q Rep. 2015; 39: E46-E136PubMed Google Scholar In our patient's case, we hypothesise that acquisition occurred from water exposure during the initial trauma, although we were unable to definitively establish a link despite detailed history and public health investigation. Legionella sainthelensi has been reported only once before in Australia's Notifiable Diseases Surveillance System, although several cases have been reported in New Zealand.12NNDSS Annual Report Writing GroupAustralia’s notifiable disease status, 2012: Annual report of the National Notifiable Diseases Surveillance System.Commun Dis Intell Q Rep. 2015; 39: E46-E136PubMed Google Scholar It is possible that L. sainthelensi infection is underdiagnosed in Australia and worldwide, as usual techniques are largely biased towards L. pneumophila and L. longbeachae. Traditional diagnostic algorithms for Legionella pneumonia include urinary antigen testing, culture, convalescent serology, and NAAT. L. sainthelensi can be isolated after prolonged culture on selective media such as BCYE, although sensitivity of culture is reduced if the patient has received prior antibiotics. Not all respiratory samples are routinely cultured for Legionella species, and often optimal yield requires more invasive respiratory sampling such as bronchoalveolar lavage. Identification can be accurately achieved using MALDI-TOF but due to the slow growth rate, diagnosis is often made later in the disease course. Historically, Legionella serology has been a cornerstone in the diagnosis of L. pneumophila and L. longbeachae infection as culture has suboptimal sensitivity, but serology is limited by the need of sampling acute and convalescent sera, with the diagnosis only being confirmed retrospectively. Our case demonstrated that L. sainthelensi infection did not induce antibodies that cross-react with of L. pneumophila and L. longbeachae antigens by immunofluorescence. In cases where Legionella infection is highly suspected but initial serology is non-reactive, further testing against additional antigens may assist with non-L. pneumophila/L. longbeachae infections.5Loeb M. Simor A.E. Mandell L. et al.Two nursing home outbreaks of respiratory infection with Legionella sainthelensi.J Am Geriatr Soc. 1999; 47: 547-552Crossref PubMed Scopus (37) Google Scholar However, the kinetics of the antibody response to L. sainthelensi infection are also not well documented. It is unfortunate that urine samples were not tested for legionella urinary antigen contemporaneously, and that specimens were discarded before such testing could be added on retrospectively. Further studies to investigate the antigen and antibody response in L. sainthelensi infection should be considered. Specific L. pneumophila and L. longbeachae NAAT targets did not show amplification on either the commercial AusDiagnostics Atypical Pneumonia PCR 8-well assay or the in-house assay. While the pan-Legionella target on the in-house PCR demonstrated expected amplification, melt-curve analysis of PCR product was indistinguishable from that of a control L. longbeachae strain (Fig. 1). We note melt-curve analysis is a crude method of typing and the probes used in this assay do not provide sufficient sequence heterogeneity to separate non-pneumophila species.11Reischl U. Linde H.J. Lehn N. Landt O. Barratt K. Wellinghausen N. Direct detection and differentiation of Legionella spp. and Legionella pneumophila in clinical specimens by dual-color real-time PCR and melting curve analysis.J Clin Microbiol. 2002; 40: 3814-3817Crossref PubMed Scopus (88) Google Scholar 16S rRNA sequencing was also unable to identify to species level due to high similarity of amplified sequence of L. sainthelensi, L. longbeachae and L. oakridgensis species. The genetic homology between L. longbeachae and L. sainthelensi make differentiation difficult and it is possible that previous cases in Australia have been misidentified as being caused by L. longbeachae. Our case demonstrates the difficulty in diagnosis of L. sainthelensi pneumonia. Clinicians should consider other Legionella spp. as a potential cause of atypical pneumonia, especially in cases associated with recent water exposure. Legionella culture should be performed, ideally prior to antibiotic therapy, as routine molecular and serology methods may miss these infections. It is also important for clinicians to understand the targets included in diagnostic platforms used in their laboratories and liaise closely with microbiologists if considering the diagnosis of uncommon pathogens. This case showed that the inclusion of a pan-Legionella 16S target as opposed to specific L. pneumophila or L. longbeachae targets in molecular platforms may rapidly and sensitively diagnose infections caused by other Legionella species. While L. sainthelensi is uncommonly isolated, it has the potential to cause both severe clinical disease and outbreaks and clinicians should be aware of its pathogenic potential.2Kamus L. Roquebert B. Allyn J. et al.Severe bilateral pleuropneumonia caused by Legionella sainthelensi: a case report.BMC Infect Dis. 2021; 21: 966Crossref PubMed Scopus (1) Google Scholar,5Loeb M. Simor A.E. Mandell L. et al.Two nursing home outbreaks of respiratory infection with Legionella sainthelensi.J Am Geriatr Soc. 1999; 47: 547-552Crossref PubMed Scopus (37) Google Scholar Further investigations are warranted to define the epidemiology, clinical manifestations, and optimal treatment of L. sainthelensi infection. The authors state that there are no conflicts of interest to disclose.