Low predictive value of positive transplant perfusion fluid cultures for diagnosing postoperative infections in kidney and kidney–pancreas transplantation
- Meaghan P Cotter1,2,
- Elizabeth Smyth2,
- Joanne O'Gorman1,2,
- Sarah Browne3,
- David P Hickey4,
- Hilary Humphreys1,2
- 1Department of Microbiology, Beaumont Hospital, Dublin, Ireland
- 2Department of Clinical Microbiology, The Royal College of Surgeons in Ireland, Dublin, Ireland
- 3Division of Nephrology, University of British Columbia, Vancouver, Canada
- 4Department of Urology and Transplantation, Beaumont Hospital, Dublin, Ireland
- Correspondence to Dr Meaghan Cotter, Department of Clinical Microbiology, Mater Misericordiae Hospital, Eccles Street, Dublin 3, Ireland;
- Accepted 26 July 2012
- Published Online First 22 August 2012
Aims Infection following transplantation is a cause of morbidity and mortality. Perfusion fluid (PF) used to preserve organs between recovery and transplantation represents a medium suitable for the growth of microbes. We evaluated the relevance of positive growth from PF sampled before the implantation of kidney or kidney–pancreas (KP) allografts.
Methods Between January 2007 and January 2011, 548 kidney/KP transplants were performed in our centre. A retrospective review of patient records with culture-positive PF was performed.
Results PF was received from 483 (88%) patients, of which 35 (7%, 95% CI 5.3% to 9.9%) were positive for bacteria (31/483, 6.4%, 95% CI 4.6% to 9.8%) and fungi (4/483, 0.8%, 95% CI 0.3% to 2.1%). Thirty-two of the 35 culture-positive PF (91.4%, 95% CI 77.6% to 97%) were considered insignificant. The remaining three patients developed sepsis postoperatively, which was considered to be possibly related to growth in PF; Escherichia coli in one and Klebsiella pneumoniae in two. Of the non-skin flora bacteria cultured from PF, six were resistant to the prophylactic antibiotic given intraoperatively, but only one developed infection postoperatively (E coli, resistant to the co-amoxiclav).
Conclusions Significant attributable morbidity associated with PF-positive culture results was relatively rare. Culture of organisms other than Enterobacteriaceae or fungi are likely to represent contamination.
Transplant recipients are particularly susceptible to infection due to the underlying disease and the immunosuppressive drugs administered to promote allograft survival. Much attention is focussed on opportunistic infection prophylaxis and post-transplantation surveillance in such patients.
Perfusion fluid (PF) is a preservation solution used following the recovery of organs for the transportation of organs to recipient centres where implantation occurs. The PF does not per se contain antibiotics and thus represents a medium in which microorganisms can grow. This fluid preserves the functional and anatomical integrity of cells until reperfusion occurs with transplantation. Despite best practice, contamination of organs can occur.
Many transplant centres routinely culture PF before implantation and the isolation of Staphylococcus aureus and Pseudomonas aeruginosa has been described as being associated with sepsis, anastomotic failure and graft loss.1–5 Furthermore, Candida spp. from PF are thought to predispose to mycotic arteritis/aneurysms.6 ,7
In view of the above, we performed a retrospective chart review of kidney and kidney–pancreas (KP) transplant recipients who had positive PF cultures.
Between January 2007 and January 2011, 548 adult kidney/KP were performed in Beaumont Hospital, Dublin, Ireland. Paediatric transplantation is performed elsewhere. Beaumont Hospital is a tertiary referral centre with approximately 700 inpatient beds and provides acute services to over 250 000 patients in north County Dublin. It is the national centre for kidney and pancreas transplantation, and is the only adult kidney/KP transplantation centre in the country.
Patients undergoing re-transplantation (for rejection/graft failure) during the 4-years period were included and both the initial and subsequent re-transplantation procedure were assessed separately. Patient details stored electronically in the encrypted transplant database, laboratory records and clinical notes in our institution were accessed to assess the significance or otherwise of positive PF results. Two of the authors reviewed these notes to validate the clinical significance of isolates identified by culture. Positive growth on PF specimens had been routinely communicated to clinical teams in Beaumont Hospital at the time to determine the patient's clinical status and to assess the possible significance. Written case notes for each patient were reviewed, as were microbiology notes recorded at the time, to determine if these patients had clinical signs or symptoms of infection. Clinical parameters such as temperature, white cell count, C-reactive protein and continued antibiotic therapy were noted, as were any relevant radiological reports (eg, chest radiograph reports). As there are no agreed definitions for post-transplant infections per se we adapted established definitions for surgical site infection (SSI), urinary tract infection, hospital-associated respiratory tract infection/ventilator-associated pneumonia and catheter-related bloodstream infection to determine if patients had definite or possible infections.8 The PF used in Beaumont Hospital is the University of Wisconsin preservation solution and is maintained at 4°C. In cases of KP transplantation, 50 mg of liposomal amphotericin B reconstituted with 50 ml of sterile water was instilled into the duodenum of the recovered pancreas. Approximately 2 ml of PF was sent to the microbiology laboratory for culture.
Prophylactic antimicrobial agents were administered to patients before transplantation, ie, co-amoxiclav 1.2 g intravenously (IV) single dose in the case of kidney transplantation, and three doses for KP transplantation. In the case of penicillin allergy either cefuroxime 1.5 g IV or ciprofloxacin 400 mg IV were administered. In patients who were known to be colonised with multidrug-resistant organisms, for example, methicillin-resistant S aureus, or vancomycin-resistant enterococci antimicrobial prophylaxis was tailored specifically for each per patient. When surgery was prolonged over 6 h, or if significant blood loss occurred (>1.5l), an additional dose of prophylactic antibiotic(s) was administered. Each patient also received prophylaxis for Pneumocystis jiroveci with co-trimoxazole 480 mg once a day for a minimum of 6 months in kidney transplant recipients, or lifelong in KP transplant recipients. Antifungal prophylaxis with fluconazole 200 mg once a day was commenced if anti-thymocyte globulin (ATG) was used in kidney transplantation and in all KP recipients for the duration of ATG therapy. Before organ recovery, donors received prophylactic cefotaxime 2 g as a single dose.
Immunosuppression in all transplant patients included tacrolimus, mycofenolate mofetil and steroids. Pancreas transplant recipients received anti-thymocyte globulin (ATG) for 10–14 days post-transplant. Kidney transplant recipients considered low immunological risk received basiliximab on day 1 and day 4 after surgery.
Donor details, including age, length of stay in intensive care unit before recovery of organs and cause of death of donor were not collected. Statistical analysis was performed. Categorical variables were expressed using proportions and group differences were tested using Fisher's exact test.
During the 4-year period, 548 transplant procedures (514 kidney and 34 KP) were performed, and 483 (88%) PF were received for culture; 29/34 (85.3%) pancreas and 454/514 (88.3%) kidney.
Of the 483 fluids, 35 (7%, 95% CI 5.3% to 9.9%) were culture positive (table 1); 27 from kidney allografts and eight from KP allografts. Thirty-one (6.4%, 95% CI 4.6% to 8.9%) grew bacteria and four grew Candida albicans, three of which were from patients undergoing KP transplantation (3/29, 10.3%) and one from kidney PF (1/454, 0.2%).
One patient developed oral candidiasis 69 days following transplantation, unrelated to the PF culture. The three other patients were asymptomatic following transplantation, but were treated for a minimum of 1 week pre-emptively for possible candida infection with oral fluconazole. Three of these patients received fluconazole prophylaxis preoperatively (all KP allografts and thus received ATG immunosuppression). One patient developed a pancreatic artery graft pseudoaneurysm 3 months postoperatively but the PF at the time of transplant was sterile. The patient underwent successful angioplasty and stenting of the pseudoaneurysm, and culture of tissue was negative for fungi.
Thirty-two positive fluids were felt to represent contamination based on a review of all records. Fourteen fluids grew organisms consistent with contamination by skin flora, for example, coagulase-negative staphylococci and Propionibacterium spp. and there were no infections caused by these bacteria. Thirteen fluids grew enteric bacteria (Gram-negative rods or Enterococcus spp.), 10 of which were felt to represent contamination.
Three cases of infection caused by enteric bacteria postoperatively were considered to be related to growth from the PF—one with E coli and two with K pneumoniae (table 1). Two of these patients underwent KP transplantation. In one KP patient E coli resistant to the co-amoxiclav prophylaxis administered perioperatively was cultured. This patient was admitted to the intensive care unit and successfully treated for pneumonia with ciprofloxacin. E coli (similar to the PF isolate) was cultured from a bronchoalveolar lavage. Other clinical specimens were negative for E coli.
In both patients with K pneumoniae isolates were co-amoxiclav susceptible. One patient was febrile postoperatively and co-amoxiclav prophylaxis was continued. Blood cultures remained sterile but K pneumoniae grew from a central line vascular catheter tip. The patient recovered fully following 2 weeks of co-amoxiclav therapy. The second patient, who underwent KP transplantation, remained afebrile, but was commenced on ciprofloxacin 2 days postoperatively when the white cell count was raised. However, antibiotic treatment was discontinued when blood cultures and other clinical specimens were sterile.
Recipients of kidney/KP transplantation are at risk of infective complications ranging from SSI to graft-threatening complications, ie, septicaemia, pyelonephritis and vascular anastomotic infection.5 ,9 Microbes in organ preservation solutions may be a source of infection postoperatively.10 In March 2012 Bristol Meyers Squibb issued a precautionary recall of Viaspan PF due to possible contamination of their product with Bacillus cereus. Of note, one of our PF cultured Bacillus species; however, no clinical evidence of infection was recorded in the patient notes on review.11
Contamination of PF may occur at a number of stages during the process of deceased donor transplantation, and there may be a delay of several hours before engraftment allowing contaminating organisms to multiply. With current organ preservation techniques pancreatic grafts can be safely transplanted up to 30 h after procurement.12
The proportion of transplant procedures from which fluid was received, ie, 88%, compares favourably with other series.13 ,14 The overall contamination rate was 7.2%, similar to the 4.2–28% reported in previous studies.15 Several publications report yeasts in PF, 1.7–2.75%; our rate was lower at 0.8%.16 ,17 This may be due to a number of reasons; we did not inoculate into fungal agar media, specimens were not enriched or centrifuged, and we cultured 2 ml and not 10–40 ml as in other studies.6 ,14 ,16 We added liposomal amphotericin B to the donor duodenum in KP transplantation, which would reduce the recovery rate of fungi.
None of the four patients from whom Candida spp. were isolated developed the complications reported by other authors, ie, SSI and pseudoaneurysms.2 ,6 ,13 All patients undergoing KP transplantation in our centre receive ATG immunosuppression, and thus are given fluconazole prophylaxis, which is continued for the duration of ATG therapy. In addition, liposomal amphotericin B is instilled into the duodenum of the recovered pancreas, as in other centres, but no reference to this practice is cited in studies that report complications caused by yeasts.2 ,6 ,11 ,14 Although one of our patients developed a pseudoaneurysm, Candida spp. was not cultured from this patient.
The proportion of positive culture results in this study was 6.4%, similar to the 7–24% reported by others.2 ,17 ,18 Skin commensals were responsible for 14 of 31 culture-positive fluids, and a review of clinical notes indicated that these were not clinically significant. Two cases of K pneumonia-positive PF were considered significant; in one patient postoperative cultures were sterile and a central line vascular catheter tip was the sole positive culture result. The only other case was a KP patient who developed postoperative pneumonia with an isolate of similar antibiotic susceptibility profile to the E coli isolated from the PF. Both isolates were resistant to co-amoxiclav, the perioperative prophylactic antibiotic administered.
In multiorgan recovery the kidneys are almost always the last organ to be recovered. Organisms from a variety of sources (air, skin and bowel flora) can settle in the peritoneal cavity and its contents. Gut ischaemia in the donor may also contribute to translocation of bowel flora from the gut to the bloodstream and thence to other organs.19 Antibiotic prophylaxis, ie, cefuroxime to our donors is likely to minimise contamination of organs and infection in recipients. As renally excreted antibiotics in donors may persist in renal tissue following nephrectomy, it may explain the finding that contamination before implantation was less frequent than at the time of procurement (17.4% vs 45%).17
The protective effect of prophylactic antibiotics was demonstrated in the canine model in 1979 by Weber et al.20 Canine kidneys were perfused with E coli-contaminated PF and transplanted. The untreated recipients died from either anastomotic rupture or sepsis but when antibiotics were commenced 1 day postoperatively 60% of canine recipients died.
The limitations of our study include its retrospective nature, the absence of specialised fungal culture systems or broth enrichment, the failure to obtain a PF from all transplants undertaken, the unavailability of molecular typing to compare isolates from PF and clinical specimens in patients with suspected postoperative infections and the absence of a postoperative review of those patients with negative PF cultures but possible infection. Nonetheless, unlike Woeste and colleagues,21 who focussed on KP only, we did not see an association between positive PF culture results and subsequent intra-abdominal infections but we may be missing potential fungal infections in the absence of the use of specific fungal culture media.
Previous studies are difficult to compare due to non-standardised specimen type (PF/swabs), different culture methods and insufficient numbers to show significant differences in contamination rates and complications. Many authors feel that routine culture of PF can be justified to ensure that potentially high virulence organisms are detected and treatment instituted.17 In our institution we assess and treat only if clinically indicated. However, it may be preferable to target certain categories of transplants for follow-up when the PF grows Gram-negative bacilli, fungi or S aureus.
There was no clear association between culture positivity and attributable infection.
Unlike previous studies, the presence of Gram-negative bacilli did not always correlate with infection, although infection was more likely in these patients (3/15, 20%) than those with skin flora (0/14, 0%), Enterococcus spp. (0/2, 0%), or Candida spp. (0/4, 0%) from PF (p=0.44).
The recovery of Candida spp. was lower than reported elsewhere, possibly due to different culture techniques.
Prospective evaluation of PF culture significance is required to assess fully the significance of Gram-negative bacilli and yeasts. However, in this study we document the low utility of routine PF cultures.
The authors wish to acknowledge the assistance of the transplant coordinators and the microbiology department in Beaumont Hospital for their contributions to the conduct of this study.
Contributors MPC, ES and JO'G were involved in data collection. MPC, DPH and HH contributed significantly to the writing, editing and preparation of the manuscript. SB edited the manuscript and performed analysis.
Competing interests None.
Ethics approval At the time of accessing data study ethics approval was not deemed necessary as the study was considered audit rather than research.
Provenance and peer review Not commissioned; externally peer reviewed.