Pandoraea sp infection in a lung transplant patient and the critical role of MALDI-TOF in accurate bacterial identification

Abstract

Diagnosis and clinical management of pulmonary infections in lung transplant patients are challenging. The increased diversity of bacterial species identified from clinical samples with novel proteomics-based systems can further complicate clinical decision making in this highly vulnerable population. Whether newly recognized organisms are colonizers or true pathogens often remains controversial since symptoms causality and impact on lung function is often unknown. We present the case of a 48-year-old female lung transplant patient with Pandoraea sp infection. We review and discuss the role of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) for accurate bacterial identification. We report on therapeutic management and clinical outcome.

Case Presentation

A 48-year-old woman was admitted with pulmonary sepsis in June 2017. She had had bilateral pulmonary transplantation 7 years earlier for severe pulmonary veno-occlusive disease following hematopoietic stem cell transplant for Hodgkin lymphoma. Her medical history also included stage 2 chronic kidney disease and corticosteroids-induced diabetes mellitus. Her regular immunosuppressive therapy included systemic corticosteroids, a calcineurin inhibitor (tacrolimus) and mofetil mycophenolate. She was receiving inhaled colistin for lung infection prophylaxis. A type IV hypersensitivity reaction to meropenem had previously been confirmed by provocation testing and she also reported possible allergies to trimethoprim–sulfamethoxazole (TMP–SMX), azithromycin, cloxacillin, clindamycin and amphotericin-B. Microorganisms previously and recurrently identified from her respiratory tract samples included Aspergillus sp, methicillin-susceptible Staphylococcus aureus (MSSA), Pseudomonas aeruginosa, and Achromobacter sp.

Prior to admission, she had already received 4 days of ceftriaxone for possible bronchitis, with the most recent sputum culture showing Achromobacter sp (Table 1—Organism 1). On admission, based on drug susceptibility testing (DST) results showing resistance to third-generation cephalosporins, therapy was modified to imipenem; this was discontinued after 3 days because of a maculopapular rash suspected to be a type IV allergic reaction. Combined tigecycline and ciprofloxacin therapy was then introduced, leading to rapid clinical improvement. Based on her immunocompromised status and repetitive identification of Aspergillus sp in sputum cultures, but without classical radiological signs, she was also initiated on voriconazole therapy with a diagnosis of probable invasive aspergillosis per the European Organisation for Research and Treatment of Cancer (EORTC) definition. Antibiotic therapy was continued for 3 weeks.

Table 1:

Microbial identification and drug susceptibility testing of sequentially isolated bacterial organisms

Identification methods	Organism 1		Organism 2		Organism 3		Organism 4		Organism 5
ID	Confidence	ID	Confidence	ID	Confidence	ID	Confidence	ID	Confidence
Phenotype	Achromobacter sp	–	NF-GNB	–	NF-GNB	–	NF-GNB	–	Achromobacter sp	–
Vitek automated system	Achromobacter sp	–	no ID	–	Chryseobacterium indologenes	–	Bordetella hinzii	–	Achromobacter sp	–
Vitek MALDI-TOF MS	–	–	Pandoraea sputorum	99.9%	–	–	Pandoraea sputorum	99.9%	–	–
16Sr RNA sequencing	–	–	Pandoraea sputorum	99.9%	–	–	Pandoraea pnomenusa	99.3%	–	–
Consensus ID	Achromobacter sp		Pandoraea sputorum		Chryseobacterium indologenes		Pandoraea sp		Achromobacter sp
Drug susceptibility testing	ETEST* MIC	Interpretation†	ETEST* MIC	Interpretation†	ETEST* MIC	Interpretation†	ETEST* MIC	Interpretation†	ETEST* MIC	Interpretation†
Aztronam			>256	R
Ceftazidime	≥256	R	>256	R	1	S	≥256	R	≥256	R
Ceftazolame–tazobactam			≥256
Chloramphenicol			64	R
Ciprofloxacin	≥32	R	>32	R	0.25	S	≥32	R	≥32	R
Colistin	≥256	R
Gentamicin	≥256	R	>256	R		R	≥256	R	≥256	R
Imipenem	1	S	1	S	0.25	S	1	S	1	S
Meropenem						R
Piperacillin–tazobactam	≥256	R	32	I	2	S	16	S	64	I
Tigecycline	16		16		2
TMP–SMX	≤0.03	S	0.03	S	0.125	S	0.06	S	0.06	S

Notes: ID confidence refers to percentage of confidence of expert systems for VITEK (bioMérieux, Marcy-Létoile, France), VITEK MALDI-TOF MS, and to percentage of nucleotide bases concordance on targeted 16S rRNA sequencing. Drug susceptibility interpretation criteria from Clinical Laboratory Standards Institutes (CLSI) were used (17).

* ETEST (bioMérieux, Marcy-Létoile, France)

† Interpretation: S = susceptible; I = intermediate; R = resistant

ID = Identification; MALDI-TOF MS = Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; NF-GNB = Non-fermenting gram-negative bacilli; MIC = Minimal inhibitory concentration; TMP–SMX = Trimethoprim–sulfamethoxazole

Six weeks after antibiotic discontinuation, the patient presented with increased fatigue and sputum production and pulmonary function tests showing a decreased FEV1 from baseline (37% from 43%). A chest X-ray showed no new infiltrate. Tigecycline and ciprofloxacin were reintroduced. We identified a glucose non-fermenting gram-negative bacilli (NF-GNB) with a highly drug-resistant profile (Table 1—Organism 2) and Chryseobacterium indologenes (Table 1—Organism 3) in the bronchoalveolar lavage.

Three weeks later, while still on tigecycline, ciprofloxacin, inhaled colistin and voriconazole, she was re-hospitalized with significant respiratory distress requiring non-invasive mechanical ventilation. Therapy was modified to include piperacillin–tazobactam, and control sputum cultures grew Bordetella hinzii (Table 1—Organism 4) and Achromobacter sp (Table 1—Organism 5). The patient slowly improved and was discharged with intravenous piperacillin–tazobactam therapy.

Bacterial isolates were retrospectively sent to Laboratoire de Santé Publique du Québec (LSPQ) for reference testing. The NF-GNB (Table 1—Organism 2) and B. hinzii (Table 1—Organism 4) were both identified as P. sputorum (99.9% confidence) by VITEK MS MALDI-TOF (bioMérieux, Marcy-Létoile, France) and were respectively identified as P. sputorum (99.9% confidence) and P. pnomenusa (99.3% confidence) by 16S rRNA sequencing. Detailed microbial identification method results and comprehensive drug susceptibility testing for all organisms are presented in Table 1.

Piperacillin–tazobactam and voriconazole therapy were continued for 4 and 12 weeks, respectively. Pulmonary function tests before and after infection showed no significant decrease in FEV1 (43% and 41% of predicted respectively). Follow-up pulmonary secretions cultures one year later showed the presence of Aspergillus fumigatus, Chryseobacterium gleum, MSSA, and Burkholderia cepacia complex. During that year, she was further treated for two episodes of pulmonary infections with rapid clinical improvement.

Discussion

Pandoraea sp colonization or infection is mostly reported in cystic fibrosis patients, among which its pathogenicity remains difficult to establish. Indeed, chronic colonization, bacteremia, acute respiratory distress syndrome, and sepsis with multi-organ failure in recently transplanted patients have been reported (1–4). Its presence has also been identified in urinary and upper airway specimens as well as blood suggesting invasive infections potential (2,3). Such virulence was demonstrated for P. apista and P. pulmonicola in non-human in vivo experiments through pulmonary epithelium invasion and translocation (5).

Challenging microbiological identification is likely to have contributed to underdiagnosis of cases of Pandoraea sp infections. Pandoraea sp (family–Burkholderiaceae, class–Betaproteobacteria) was initially described in 2000, and its full genome was first assembled in 2016 (6,7). The isolation and identification of Pandoraea sp is challenged by low biochemical activity, phenotype attenuation, and variability, and absence of (or poor representation in) available proteomic and genotypic databases built to date (8–11). Its slower growth also represents a barrier for identification, especially in pulmonary transplanted and CF patients with many other lung colonizers more rapidly found on culture media (11,12).

The increased use of MALDI-TOF technology in routine laboratories has set the ground for improvement in the identification of phenotypically challenging organisms, including NF-GNB (8,13). However, lack of discriminatory power was noted with Pandoraea sp, which can be incorrectly identified as Achromobacter sp and other NF-GNB, including Burkholderia cepacia complex, a close phylogenetic relative (2,9,13–15).

Isolates of the Pandoraea sp genus were shown to present variable susceptibility profiles to beta-lactams, being frequently resistant to meropenem and susceptible to imipenem. Among other classes, tetracycline and TMP–SMX were shown to have efficacy on a series of clinical isolates, including multi-drug resistant strains (4,15,16).

We have presented a case of Pandoraea sp lung infection in a pulmonary transplanted patient for which incomplete and inaccurate microbiological identification prevented the recognition of the microorganism’s persistence in time and causality in the patient’s respiratory symptoms. This delay in proper diagnosis led to suboptimal antibiotic choices. Indeed, imipenem therapy was only received for 3 days initially. With accurate microbiological identification, earlier allergy testing and antibiotic desensitization could have been performed to ensure optimal therapy. Although clinical improvement was finally observed, our patient’s Pandoraea sp isolates were shown to have intermediate susceptibility to the piperacillin–tazobactam therapy received.

Funding:

No funding was received for this work.

Disclosures:

The authors have nothing to disclose.

Informed Consent:

Informed consent was obtained from the patients.

Peer Review:

This manuscript has been peer reviewed.

Animal Studies:

N/A.