Recurrent Burkholderia Infection in Patients with Chronic Granulomatous Disease: 11-Year Experience at a Large Referral Center

Abstract

The epidemiology of Burkholderia infection in persons with chronic granulomatous disease is poorly understood. We used species-specific polymerase chain reaction–based assays and genotyping analyses to identify 32 strains representing 9 Burkholderia species among 50 isolates recovered from 18 patients with chronic granulomatous disease. We found that recurrent pulmonary infection with distinct Burkholderia strains is common in chronic granulomatous disease.

Phagocytes from persons with chronic granulomatous disease (CGD) are defective in the production of toxic oxygen metabolites and have impaired the killing of microbes, rendering patients with CGD susceptible to a variety of infections, including recurrent pneumonia [1]. Despite the defect in microbial killing, the spectrum of pathogens found in CGD is relatively narrow and includes Staphylococcus aureus, Serratia marcescens, Nocardia species, and Aspergillus species [2]. Infection with certain uncommon human pathogens, such as Chromobacterium violaceum [3], Granulibacter bethesdensis [4], and Francisella philomiragia [5], is almost pathognomonic for CGD.

For reasons that are not entirely clear, patients with CGD are also particularly susceptible to infections due to certain Burkholderia species, including Burkholderia gladioli and species within the Burkholderia cepacia complex (Bcc). The Bcc group consists of at least 17 phenotypically similar but genetically distinct species that rarely cause disease in healthy individuals [6]. However, in patients with cystic fibrosis (CF), respiratory tract infection with Bcc species is associated with increased rates of morbidity and mortality [7]. Although most Bcc species have been recovered from patients with CF, Burkholderia cenocepacia and Burkholderia multivorans account for the great majority of infections [8]. Recent work has also suggested that the clinical outcomes of infection due to Bcc species in CF may vary considerably among these species [9]. Furthermore, genotyping analyses of serial isolates from chronically infected patients with CF indicate that the great majority of patients harbor the same strain for prolonged periods of time [10]. Similar studies describing the natural history and epidemiology of Burkholderia infection in patients with CGD are lacking. We performed a retrospective analysis of Bcc isolates recovered from patients with CGD who received care at a national referral center from 1995 through 2006.

Methods

The culture collection of the Microbiology Laboratory at the National Institutes of Health (NIH) Clinical Center (Bethesda, MD) was reviewed to identify all Burkholderia isolates recovered from January 1995 through October 2006. Patients from whom these isolates were obtained were cross-referenced against a database of patients with CGD who received care at the NIH Clinical Center under protocol 93-I-0119. Isolates from patients with confirmed CGD were recovered from frozen stock and were identified to the species level by using species-specific 16S ribosomal DNA (rDNA)– and recA-targeted polymerase chain reaction (PCR) assays, as described elsewhere [8, 11]. For isolates that could not be definitively identified by PCR-based analyses, complete 16S rDNA PCR amplification, sequencing, and editing were performed as described elsewhere [11]. Isolates with 16S rDNA sequences that showed similarity to sequences of Bcc species in the National Center for Biotechnology Information GenBank database but that could not be placed confidently into 1 of the 17 described species in this group were designated as Bcc species “indeterminate.” Bacterial genotyping was performed using repetitive extragenic palindromic PCR with a BOX A1R primer, as described elsewhere [12]. For the purposes of this study, distinct strains were assigned a letter designation (A, B, C, etc.).

Results

Greater than 90% of the Burkholderia isolates identified in the NIH Clinical Center Microbiology Laboratory culture collection during the study period were recovered from patients with CGD. In cases in which multiple isolates from the same patient were available, only isolates recovered on different days were included in the analysis. A total of 50 isolates recovered from 18 patients with CGD were identified and confirmed as either Bcc (45 isolates) or B. gladioli (5 isolates) by PCR-based analyses. All isolates except 1 (which was cultured from blood) were cultured from lung biopsy specimens. For 8 patients, only a single isolate was available for analysis (table 1). In 5 of these 8 patients, B. multivorans was identified. In the remaining 3 patients, B. cepacia, B. gladioli, and an indeterminate Bcc species were identified. For each of the remaining 10 patients, 2–11 isolates were available. The species and strain distributions of the 42 isolates serially cultured from these patients are shown in table 1. For all patients from whom isolates were recovered during a period >1 year (i.e., all except patient 8 and patient 15), multiple infecting species and/or strains were identified. In total, 9 Burkholderia species (7 named species and 2 “indeterminate” species) and 32 distinct strains were identified among the 50 isolates from the 18 patients.

Table 1.

Burkholderia species and strains cultured from specimens obtained from 18 patients with chronic granulomatous disease.

Patient, year of culture (no. of isolatesa)	Burkholderia species	Strain
Patient 1  2000	B. multivorans	A
Patient 2  1996	B. ambifaria	B
1997 (2)	B. ambifaria	B
1998	B. multivorans	C
2001	B. multivorans	D
2001	B. metallica	E
2001	B. cenocepacia	F
2002	B. cepacia	G
2003	B. cenocepacia	H
2005	B. cenocepacia	H
2005	B. gladioli	I
Patient 3  1996	B. multivorans	J
2003	B. gladioli	K
Patient 4  2003	B. cenocepacia	H
2004	B. multivorans	L
2006	Indeterminate	M
Patient 5  2002	B. multivorans	N
Patient 6  2000	B. multivorans	A
2006 (2)	B. multivorans	O
Patient 7  2000	B. cepacia	P
Patient 8  1995 (8)	B. multivorans	Q
Patient 9  1995 (2)	B. multivorans	R
2005	B. cenocepacia	S
Patient 10  2005	B. multivorans	T
Patient 11  1998	B. multivorans	U
1999	B. multivorans	U
2003	B. cepacia	V
2004	B. multivorans	W
2005	B. cepacia	V
Patient 12  1995	B. multivorans	X
Patient 13  2001	B. multivorans	Y
Patient 14  2000	Indeterminate	Z
Patient 15  1997 (2)	B. cenocepacia	AA
Patient 16  1996 (2)	B. multivorans	BB
2000	B. vietnamiensis	CC
Patient 17  2000	B. gladioli	DD
Patient 18  2005	B. gladioli	EE
2006	B. gladioli	FF

^aThe number of isolates of the designated strain type recovered from this patient during the year indicated.

Discussion

Apart from occasional nosocomial outbreaks, human infections with B. gladioli and Bcc species are generally restricted to patients with CF or CGD. In both of these patient populations, infection primarily affects the respiratory tract. However, Burkholderia infection in these 2 patient groups differs with respect to pathology and, as our study demonstrates, epidemiology and natural history.

In CF, pulmonary infection with Burkholderia species primarily causes endobronchial disease, and diagnosis is made by identification of Burkholderia species in sputum culture. In contrast, Burkholderia lung infection in CGD typically results in a pneumonic process associated with nodular infiltrates that frequently requires lung biopsy for a definitive diagnosis. Longitudinal genotyping studies in CF indicate that, for the majority of patients, chronic Burkholderia infection is caused by the persistence, sometimes for several years, of a single strain [10]. In contrast, although we observed the persistence of the same strain for periods of up to 2 years in 2 patients with CGD (patient 2 and patient 11), we found that pulmonary infection with Burkholderia species in CGD typically involves recurrent episodes with distinct strains. Indeed, we observed repeat lung infections with distinct Burkholderia strains in every patient from whom isolates serially recovered during periods >1 year were available. This is best illustrated by patient 2 (table 1), in whom 8 distinct strains representing 6 Burkholderia species were recovered from lung biopsy specimens during a 10-year period.

In CF, so-called epidemic Bcc strains that infect multiple patients have been identified [7]. In the present study, we did not find strains similarly shared by several patients with CGD. We identified only 2 strains that were each cultured from 2 patients. B. cenocepacia strain H was recovered in 2003 from patient 2 and patient 4, who are siblings. This strain was again recovered from patient 2 in 2005. B. multivorans strain A was recovered in 2000 from patient 1 and patient 6. These patients are not related and, to our knowledge, were not otherwise epidemiologically related. We also did not identify any of the previously described CF epidemic strains (e.g., ET12, PHDC, and Midwest strain) among the 50 isolates included in our study. Finally, in contrast to CF, in which B. multivorans and B. cenocepacia account for the great majority of Burkholderia infections, we found a broader representation of Burkholderia species in this relatively small set of patients with CGD. This included Burkholderia ambifaria and Burkholderia metallica, species that have been rarely recovered in human infection, including in patients with CF. Of note, the first reported case of human infection due to another Burkholderia species, the rice pathogen Burkholderia glumae, was also described in a patient with CGD [13].

The reasons for these striking differences in Burkholderia infection epidemiology between patients with CGD and patients with CF are not readily apparent. The severely impaired airway clearance that is characteristic of CF and the formation of intraluminal antimicrobial-resistant biofilms likely contribute to persistent, antimicrobial-refractory Burkholderia infection in CF. The lung parenchymal disease more typical of CGD may not afford, at least to the same extent, the opportunity for similarly sustained bacterial communities. Furthermore, whereas Burkholderia isolates recovered from CF respiratory secretions very often display broad-spectrum antimicrobial resistance, the majority of the CGD isolates in our study for which antimicrobial susceptibility data were available were susceptible to trimethoprim-sulfamethoxazole (30 of 31 isolates) and ceftazidime (29 of 31 isolates). Thus, in contrast to CF, conventional antimicrobial therapy can succeed in eradicating Burkholderia lung infection in CGD. However, because of the continued susceptibility of patients with CGD to Burkholderia infection and the ongoing exposure to these species in the natural environment, recurrent infection is likely to continue to occur.

Along with previous observations showing that repeat infection, rather than persistent infection, with various species appears to be the rule in CGD [14], these data support a role for aggressive antimicrobial prophylaxis to reduce the risk of severe infections in CGD. In this regard, adherence to long-term treatment with trimethoprim-sulfamethoxazole, to which most strains in our study were susceptible, should continue to be emphasized to control the incidence of Burkholderia infection in this vulnerable patient population.