INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: Burkholderia cepacia; cystic fibrosis; lung transplantation; genomovar; infection

Cystic fibrosis (CF) is the most common inherited respiratory disease that culminates in premature respiratory failure and death in ~ 400 children and young adults annually in the United States alone (1). Despite advances in our understanding of CF and incremental improvements in longevity over the past two decades, lung transplantation remains the only therapy that offers the opportunity for prolonged survival for patients with end-stage disease. Thus, ~ 125 patients with CF undergo this operation annually in the United States. Lung transplantation has provided successful results for CF patients who would have certainly succumbed from their underlying disease (2, 3). In fact, the 1-, 3-, and 5-yr survival rates, 81%, 58%, and 49%, respectively (3), for lung transplantation for CF patients are equivalent or superior to that for other indications (e.g., emphysema and pulmonary fibrosis) and are unequivocally better than for nontransplanted CF patients with end-stage disease (4). Even patients infected with pan-resistant Pseudomonas have done well (5).

However, patients infected with Burkholderia cepacia complex before or after transplant have suffered higher mortality rates. Snell and colleagues (6) demonstrated a markedly increased postoperative mortality in CF patients who harbored B. cepacia complex preoperatively (30% versus 0% mortality, p < 0.05) compared with those without B. cepacia complex and higher again for patients who acquired B. cepacia complex postoperatively (80%). Chapparo and coworkers recently updated the results from the Toronto Program and drew a similar conclusion that CF patients infected with B. cepacia complex had a much higher 5-yr mortality (64% versus 19%), predominantly because of B. cepacia complex-associated infection (7). As the prevalence of B. cepacia complex infection in CF patients has increased to 8% in the United States and 15% in Canada and the United Kingdom without a simultaneous increase in the availability of suitable lung donors, a growing controversy has evolved over whether lung grafts should be utilized for B. cepacia complex-infected patients (1, 7). Many transplant centers currently consider the presence of B. cepacia complex an absolute contraindication to lung transplantation (11), but some centers transplant patients with B. cepacia complex if antibiotic susceptibility is present (8). The official international guidelines for lung transplantation selection take a more moderate approach listing B. cepacia as a relative contraindication (14).

We conducted a retrospective review of all CF patients (n = 121) transplanted at the UNC since the inception of our program to determine the effects of preoperative infections with a genetically heterogeneous distribution of B. cepacia complex (n = 21) isolates to determine their impact on posttransplant outcome. All B. cepacia complex isolates underwent genomovar analysis and pulsed-field gel electrophoresis, which were considered during the analysis of immediate survival and other outcome measures. These results have implications for patient selection and management and the microbiological evaluation of B. cepacia complex-infected CF patients prior to lung transplantation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Population

The medical records of all pediatric (n = 19) and adult (n = 102) CF patients who underwent conventional (n = 113) or living-related (n = 8) double lung transplant from 1990 to March 2001 were reviewed. Retransplants were excluded. Donor and recipient selection criteria and the transplant operation have been previously described (2).

Microbiological Studies

All respiratory tract, blood, and pleural fluid cultures from the time of transplant evaluation until the time of death were reviewed. Preoperative sputum and/or intraoperative bronchial wash cultures were used to define the airway microbiology at the time of transplant (5). All respiratory specimens were cultured on horse blood, colistin-nalixidic acid, MacConkey, Mannitol salts, and PC agars. Bacteria were initially identified by conventional methods (15). Organisms suspected of belonging to the B. cepacia complex were stored at 70° C and subsequently identified to a specific genomovar using a combination of polymerase chain reaction (PCR) and conventional biochemical techniques (16, 17). Storage allowed all the isolates to be typed within the past 3 yr in the national B. cepacia Research Laboratory and Repository directed by Dr. John LiPuma at the University of Michigan (26). Susceptibility testing was done on all isolates using disc diffusion (18). Since 1993, all multidrug-resistant isolates were sent for antibiotic synergy results to the laboratory of Dr. Lisa Saiman at the Columbia University (19).

PCR detection of the cable pilin gene (cblA) was done according to the method of Sajjan and coworkers (20). Pulsed-field gel electrophoresis (PFGE) was performed by initially growing cells overnight in trypticase soy broth and resuspended in TEN (100 mM Tris [pH 7.9], 100 EDTA, 150 mM NaCl) to an OD₆₀₀ of 1. The cells were treated with 10% formalin for 1 h (21) to inactivate nucleases, washed three times, and then resuspended in TEN. Cells (150 µl) were mixed with 150 µl of 2% melted agarose and allowed to solidify in 300 µl insert molds at 4° C. Plugs were incubated sequentially with lysozyme and lysis buffer for 1 h and placed in ESP buffer overnight. A 3 × 5 mm portion of each plug underwent restriction enzyme digestion with SpeI (New England Biolabs) overnight followed by PFGE (22). Gels were analyzed visually according to the criteria of Tenover and coworkers (23).

Clinical Outcome Measures

Six-month and 1-, 3-, and 5-yr survival rates were computed. We classified B. cepacia-related mortality as patients who died of septic shock with B. cepacia complex growing in blood cultures or respiratory cultures.

Factors possibly affecting the virulence of B. cepacia complex were assessed including host factors (augmented immunosuppression, poor nutritional status, severe graft dysfunction, diabetes, or non-lung organ failure) and organism factors (genomovar differences, antibiotic susceptibility patterns, and the presence of the cable pilin gene).

Immunosuppression and Medical Management

The immunosuppressive regimen, medical management, surveillance protocol, and antibiotic regimen have been previously described (24). In the past year, we have been using sustained, multidrug, antimicrobial therapy in the immediate posttransplant period and have avoided empiric methylprednisolone for all our B. cepacia complex patients.

Statistical Analysis

Unpaired t tests were used for continuous variables and Chi square tests for sex differences and survival comparisons (25). The Chi square test was used to compare mortality attributable to B. cepacia genomovar III with the other genomovars; the result was confirmed with Fisher's exact test. A two-sided alpha level of <0.05 was defined as significant. Using the Bonferroni correction for multiple analyses of the outcome variables (survival and B. cepacia deaths), a more conservative p value (i.e., 0.017) may be appropriate.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pretransplant Bacterial Flora

Patient characteristics, categorized by type of pretransplant flora are listed in Table 1. The age, pulmonary function (%FEV₁), nutritional status (body mass index [BMI]), and number of patients transplanted off a ventilator were similar for both groups. There were slightly more females in the B. cepacia group compared with the non-B. cepacia complex group. Twenty-one CF patients were infected with B. cepacia complex before lung transplantation. Among patients who were B. cepacia complex-positive before transplant, only three were infected with B. cepacia alone. Sixteen of 21 were coinfected with P. aeruginosa, 7 with oxacillin-sensitive Staphylococcus aureus, and 5 with all three bacteria. Antibiotic synergy study results were kindly provided on 20 of 21 B. cepacia complex isolates by Dr. Saiman. Five of the isolates were susceptible to minocycline, one to tobramycin, and one to multiple antibiotics when administered as single medications. Twelve isolates were resistant to all antibiotics when tested singly. Nine isolates were sensitive to synergistic combinations of antibiotics in vitro, but 11 isolates were not susceptible to any synergistic combination. Therefore, for 11 of the 20 isolates, no single antibiotic or combination of antibiotics provided any in vitro efficacy. Most isolates were inhibited by doses of tobramycin that could be achieved only by aerosolization. The single-antibiotic susceptibility results from UNC were concordant with the data from Dr. Saiman's laboratory.

When comparing genomovar III patients to those with other genomovars, the only significant difference was that the other genomovar group was composed of more women (Table 1). Additional characteristics such as functional status as determined by physical therapy record review and the prevalence of insulin-treated diabetes were similar in both groups.

Transplant Outcomes for Patients Infected with B. cepacia Complex before Transplant

Mean (SD) follow-up was 3.5 ± 3.1 yr for the group as a whole (n = 121), 1.5 ± 2.0 yr for the B. cepacia complex patients (n = 21), and > 4 mo for all patients. Actuarial survival for all CF transplant recipients was 77%, 64%, and 57% at 1, 3, and 5 yr, respectively. Actuarial survival for patients infected with B. cepacia complex before transplant is contrasted with patients who were not infected with B. cepacia complex before transplant in Figure 1. A significant decrease in survival was noted for those infected with B. cepacia complex before transplant (n = 21) in the first 6 mo (64% versus 88%, p = 0.03) and 12 (50% versus 83%, p = 0.006) postoperative mo. The four unequivocal and one contributory B. cepacia complex-related deaths were all within the first 5 posttransplant mo with a median time to death of 2.3 mo. The four unequivocal and one contributory deaths resulted in a B. cepacia complex-infection death rate of 24% (5 of 21). Antibiotic susceptibility patterns were not helpful in discriminating B. cepacia complex survivors from nonsurvivors. Two deaths occurred in 1992-1993 and three deaths occurred in 1998-1999. In the interim, we did not transplant any B. cepacia complex patients, in part, because of the early experience. In 1997, after we demonstrated that panresistant P. aeruginosa patients could be transplanted without unnecessary risk, we liberalized our selection criteria to again include B. cepacia complex patients.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Posttransplant survival percentages plotted at the 1, 6, 12, 24, 36, and 60 mo intervals in CF patients infected (n = 21) and not infected (n = 100) with B. cepacia complex before transplant, demonstrating significantly lower survival in those infected with B. cepacia complex at 1 yr (50% versus 83%, p = 0.006), 3 yr (30% versus 70%, p = 0.008), and 5 yr (30% versus 58%, p = 0.03). The data were generated using Kaplan-Meier theory (Kaplan and Meier. Nonparametric estimation from incomplete observations. J Am Statistical Assoc 1958; 53:457-481).

New Acquisitions of B. cepacia Complex Infection after Transplant

Three patients acquired B. cepacia complex after transplantation, but all were successfully treated. One patient (genomovar unknown) had left lower lobe pneumonia that was treated with multiple antibiotics for a month and the patient survived 6 yr ultimately to die of obliterative bronchiolitis (OB). A second patient acquired genomovar III 1 mo after transplant, required no therapy, never recultured this organism, and is well 2 yr later (genotype M, Figure 2). A third patient who had mild OB 8.5 yr after transplant acquired B. vietnamensis and presented with bronchitis (genotype P, Figure 2). He required several courses of empiric oral antibiotics and died of renal failure 2.5 yr later.

View larger version (90K): [in this window] [in a new window]   Figure 2.   PFGE analysis of SpeI-digested genomic DNA from B. cepacia complex isolated from 22 different transplant patients. Lane 1, 48.5-kb ladder; lane 2, genomovar I (genotype A); lanes 3-8, genomovar II (B. multivorans, genotypes B-G) isolates; lanes 9-12, cluster 1 genomovar III isolates (genotype H); lanes 13-15, cluster 2 genomovar III isolates (genotype I); lanes 16-21, unique genomovar III isolates (genotypes J-O); lane 22, genomovar V (B. vietnamiensis, genotype P) isolate.

Genomovar III-associated Mortality

All five patients who died from B. cepacia complex infection after transplantation were infected with genomovar III, resulting in a statistically increased mortality in genomovar III patients compared with those infected with other genomovars (mostly B. multivorans) (5 of 12 versus 0 of 8, one isolate not typed; p = 0.035). The short-term survival for B. multivorans patients was similar to that of patients infected with organisms other than B. cepacia complex (i.e., predominantly P. aeruginosa) (6 mo: 88% versus 88%). Insufficient long-term data exist for the group infected with other genomovars due to the short follow-up on recent transplants.

Two of the five patients dying from genomovar III-related sepsis suffered from neutropenia. None of the five isolates had single antibiotic susceptibilities, but two had in vitro synergistic combinations. However, there were no significant overall differences in age, sex, BMI, antibiotic susceptibility patterns, percentage with leukopenia or neutropenia, mean cyclosporine level, percentage receiving high-dose corticosteroid infusions, other organ dysfunction, and the presence of diabetes comparing the B. cepacia complex-infected patients who survived with those who died from B. cepacia complex infection. Seven genomovar III patients have successfully been transplanted. Although management protocol changes have occurred over the years, they have had no impact on success or failure in the genomovar III group.

Genomovar Types and Cable Pilin Gene Analysis (Figure 2)

Of patients infected with B. cepacia complex before transplant (n = 21), 1, 7, and 12 had isolates of genomovar I, B. multivorans, and genomovar III, respectively, and 1 had an unknown isolate (see Figure 2). Of patients who newly acquired B. cepacia complex after transplant, one each had genomovar III (lane 18), B. vietnamenesis (lane 23), and an unknown isolate. Screening for the cblA gene indicated that cable pilin-positive organisms were not present in any of our patients. Pulsed-field gel electrophoresis showed that all B. multivorans (genomovar II) isolates were unique genotypes (Figure 2, lanes 3-9). With the genomovar III isolates, we found two clusters, one of four patients with a strain designated cluster 1 (lanes 10-13) and a second cluster of three patients with a strain designated cluster 2 (lanes 14-16). In addition, five genomovar III isolates were unique genotypes (lanes 17-21). One patient was infected with B. vietnamenesis (genomovar V) (lane 21). The five genomovar III-related deaths were each caused by an organism with a unique genotype (lanes 11, 15, 17, 18, and 20).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The most important finding of this study is that CF patients infected with genomovar III of the B. cepacia complex before transplantation have markedly increased B. cepacia complex- related early mortality after transplantation. The survival disadvantage for our B. cepacia complex group, when viewed as a whole, was attributable solely to the influence of genomovar III. No B. cepacia complex-associated deaths occurred in patients infected with other genomovars. Although certain genotypes of genomovar III seem to be associated with higher posttransplant mortality (e.g., the ET-12 strain), our genomovar III isolates were heterogeneous, suggesting that virulence factors within genomovar III are not genotype specific. This information is important to the CF lung transplant community because of the intense controversy over transplantation for B. cepacia complex-infected patients and, simultaneously, the increasing numbers and referrals of such patients to transplant centers. Our results with CF patients infected with a heterogeneous distribution of B. cepacia complex isolates probably reflect the type of CF patients referred to U.S. transplant centers (26) in contrast to the more homogeneous distributions of isolates seen in the largest Canadian and European centers. In particular, centers in Toronto and Manchester may experience patients predominantly infected with the ET-12 strain (27), which is suspected to be highly transmissible, but is far less common in the United States (26). Our experiences with strains other than ET-12 suggests that genomovar III organisms, as a whole, are more aggressive and result in higher mortality in the posttransplant environment than B. mulitvorans, Pseudomonas aeruginosa, or other CF pathogens.

Our result of increased early mortality due to B. cepacia builds on previous reports from our institution (5) and others. Snell and coworkers were the first to report enhanced mortality after transplant due to B. cepacia complex with 3 of 10 of their B. cepacia complex-infected patients dying in the first postoperative year (6). Egan and coworkers reported a more auspicious experience with 14 patients with B. cepacia complex from the United Kingdom a year later, bringing into question the general practice of considering B. cepacia complex as an absolute contraindication to lung transplantation (9). Nonetheless, the pooled U.K. results demonstrated a trend toward higher posttransplant mortality due to sepsis in the B. cepacia complex-infected patients compared with patients infected with P. aeruginosa (21% versus 11%, p = 0.38) and a trend toward lower actuarial survival (p = 0.16) (9). In the interim, most U.S. lung transplant centers have declined to transplant B. cepacia complex patients on the basis of the results of Snell and coworkers. Small series, such as those of Kanj and coworkers and Metras and coworkers with two B. cepacia complex-infected patients and no mortality has prompted a more inclusive listing strategy at those institutions (28, 29).

Recently, Chaparro and coworkers have reported the updated experience at Toronto, easily the center with the largest B. cepacia complex (n = 28) experience in the world (7). They found decreased 1-, 2-, and 3-yr survival (67, 49, and 45%) in the B. cepacia complex-infected patients compared with 28 patients infected with P. aeruginosa (96, 92, and 86%) (p < 0.01). Similar to our findings, 14 of the 15 deaths in the B. cepacia group resulted directly from B. cepacia complex infection, with the majority occurring in the first 3 postoperative mo. Only 1 of 28 in the non-B. cepacia complex group died from infection. Previous reports have indicated that the predominant strain in the Toronto center is ET-12 (27). Whether the differences in organ allocation in the Toronto program (differences that favor allocation to P. aeruginosa-infected patients over B. cepacia complex-infected patients) play a role in the posttransplant outcomes for their B. cepacia complex- infected patients is not known. A preliminary report from De Soyza and coworkers from the center at Newcastle-Upon-Tyne, U.K. lends support to our findings. They reported three early deaths from B. cepacia complex sepsis attributable to genomovar III and no deaths among the four patients with genomovars II or V (30).

Our knowledge of B. cepacia complex virulence factors is limited. Studies have shown that cable pilin gene is associated with transmissability, but it remains to be determined if this factor increases the virulence of B. cepacia complex (19, 25). Cable pilin-positive organisms belong to genomovar III (31). Little is know about the comparative virulence of genomovar III versus the other members of the B. cepacia complex. In our study, none of the strains was cable-pilin positive but all the patients with B. cepacia-associated deaths were infected with genomovar III organisms. This suggests that other virulence factors are central to the pathogenesis of B. cepacia complex. B. cepacia complex has been shown to be invasive in tissue culture cell lines (32) and invasiveness may be an important factor in the development of bacteremia in this immunocompromised patient population.

A previous study performed on the initial isolates of B. cepacia complex recovered from our transplant population showed that each of five patients was infected with a unique genotype (22). As the number of B. cepacia complex transplant patients has increased, we have found two clusters of genomovar III strains. A genomovar III stain designated cluster 1 was found in four patients, three of whom were from Michigan, and the fourth who received his care in Cleveland prior to transplant. A previous study has shown that 90% of strains from five Michigan CF centers were of the same or similar genotype suggesting that B. cepacia complex cross-infection was common among patients in this region of the country (33). The second cluster of three patients (cluster 2) infected with a genomovar III strain had all received their care in New York prior to transplant. These data suggest that clustering of B. cepacia complex in CF patients in a transplant center may not be a result of acquisition at that center but may be due to cross-infection in the past at another CF center. Because we were able to genotype only two isolates from patients infected posttransplant (or their organism was first detected then), little can be said about the epidemiology of posttransplant acquisition of B. cepacia complex. The one patient who acquired genomovar III posttransplant had a unique genotype. Nevertheless, strict infection control guidelines to prevent the spread of this organism in this vulnerable patient population should be enforced.

A study of this type is most limited by small numbers. After 11 yr of performing lung transplantation with a focus on CF, we were able to accumulate only 24 cases of B. cepacia complex infection. Our single-center experience needs to be affirmed in a multicenter data analysis to diminish the risk of transplanting CF patients who are at very high risk of posttransplant death. Our study is also limited by our poor understanding of putative B. cepacia complex virulence factors within genomovar III and, therefore, we do not understand why some patients with genomovar III isolates enjoy long-term survival whereas others succumb to an early infectious death. Last, the rapid dissemination of information on B. cepacia complex and transplant over the past 5 yr has changed the antibiotic management of patients in recent years.

In conclusion, our results indicate that CF patients who are infected with genomovar III of the B. cepacia complex before transplant are at higher risk for postoperative sepsis and death than patients infected with B. multivorans or other genomovars. In our series, patients infected with B. multivorans and other genomovars had a similar early survival when compared with CF patients infected with Pseudomonas species. A retrospective, multicenter data analysis of B. cepacia complex patients who have already undergone transplantation would be the fastest way to confirm our results. We recommend testing isolates to determine the species within the B. cepacia complex and performing synergy testing at the time of transplantation, although synergy results have not been compared with patient outcomes in this setting. Even though genomovar III patients suffer increased early mortality, long-term survival has not been well-studied and may approximate that experienced with other high-risk populations (e.g., IPF and PPH) (7, 34). More comprehensive studies, both retrospective and prospective, pooling data from multiple institutions will be needed to provide more power to draw definitive conclusions on genomovar typing of B. cepacia complex patients so that clinical practice guidelines can be developed.