According to the American Association of Tissue Banks, over 1.5 million allografts are distributed annually in the United States1. Recent incidents involving the distribution of human tissues from donors not properly screened for infectious diseases have highlighted concerns for disease transmission through transplanted tissues2-4. Although rarely reported, allograft-associated infections have been associated with a variety of organisms and tissue types4. In all of the previously published reports, organisms were transmitted from the donor, because of either an unrecognized infection or contamination during tissue recovery. However, in the fall of 2006, we investigated the cases of two patients who had allograft-associated surgical site infections caused by Elizabethkingia meningoseptica (formerly Chryseobacterium meningosepticum), detected by orthopaedic surgeons, that may have been the result of environmental contamination at one firm during processing of tissues from different donors.
Elizabethkingia meningoseptica is a waterborne, gram-negative rod widely distributed in nature that rarely infects humans. It has been reported as a cause of neonatal meningitis, pneumonia in patients on ventilator support, peritonitis in a patient receiving peritoneal dialysis, and community-acquired necrotizing fasciitis5-9. The species is usually resistant to multiple antibiotics, including extended-spectrum beta-lactam agents, aminoglycosides, and vancomycin.
To investigate the cases of these two patients, we reviewed the medical records and the medical and/or social screening histories and serologic test results for their tissue donors. The Centers for Disease Control and Prevention (CDC) worked with the tissue processor and state and local health departments to contact health-care facilities receiving large numbers of soft tissues from this processor to look for additional cases of Elizabethkingia meningoseptica infection in allograft recipients.
The CDC received undistributed, processed tissues of both implicated donors from the processor. Tissues were processed in a tissue grinder and then were cultured in trypticase soy broth and incubated at 35°C. The following day, the trypticase soy broth was subcultured to blood and MacConkey agars and incubated at 35°C. Organisms were identified with use of a Vitek GNI+ card (bioMerieux, Durham, North Carolina). Isolates obtained from tissue samples and the one available patient isolate were compared by pulsed-field gel electrophoresis with use of a modified PulseNet Yersinia pestis protocol10. Patient and tissue isolates were compared by pulsed-field gel electrophoresis following the digestion of Elizabethkingia meningoseptica and universal standard Salmonella serotype Braenderup H9812 chromosomal DNA with the restriction endonucleases Apa I and Xba I, respectively. Restriction fragments were separated with a CHEF Mapper XA pulsed-field electrophoresis system (Bio-Rad Laboratories, Hercules, California). Pulsed-field gel electrophoresis conditions were a gradient of 6.0 V/cm, 120° angle, temperature of 14°C, and switch times of five to thirty-five seconds for a total run time of thirty hours. Gels were photographed with use of ultraviolet illumination. Gel images were analyzed with use of BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium). The percentage of similarities of the test isolates were identified on a dendrogram derived from the unweighted pair-group method with use of arithmetic averages and based on Dice coefficients. Band position tolerance and optimization were set at 1.3% and 0.5%, respectively.
CASE 1. A forty-six-year-old man had arthroscopic surgery on the right knee to reconstruct the anterior cruciate ligament with implantation of a bone-patellar tendon-bone allograft. On postoperative day 4 at home, he had worsening knee swelling and a fever of up to 104°F developed. He noticed no drainage from the arthroscopic portal incisions. Physical examination of the right knee on postoperative day 10 was notable for a moderate effusion with warmth but no erythema or incision drainage. A diagnostic arthrocentesis revealed dark yellow, slightly turbid synovial fluid. Fluid analysis demonstrated a white blood-cell count of 275,225 cells/mm3 (275.2 ×109/L) (92% neutrophils and 4% lymphocytes). An initial Gram stain showed no bacteria; however, growth at twenty-four hours in thioglycolate media revealed gram-negative rods identified by the state department of health as Elizabethkingia meningoseptica.
On postoperative day 11, the patient had arthroscopic irrigation, débridement, and complete synovectomy. The reconstructed anterior cruciate ligament was intact. The Gram stain of the synovial fluid was negative, but it grew sparse Staphylococcus epidermidis on culture. He received no anti-infective agents before this surgery. Postoperatively, he received 3.375 g of piperacillin-tazobactam intravenously every six hours. One gram of vancomycin every twelve hours and 750 mg of oral ciprofloxacin every twelve hours were added soon afterward. Two sets of blood cultures drawn one day later were negative. He underwent one final irrigation and débridement forty-eight hours later (postoperative day 13). Repeat culture of synovial fluid was negative. After a nine-day hospitalization, the patient was discharged on a regimen of 2 g of intravenous ceftriaxone once daily for twenty-eight days as well as trimethoprim-sulfamethoxazole double strength orally twice daily and 750 mg of ciprofloxacin orally twice daily for forty-two days. One year after the last operation, he was well with no evidence of recurrent infection.
The tissue used in this patient had come from a fifty-eight-year-old man (Donor 1) who had a witnessed collapse and could not be resuscitated. His death was attributed to a massive myocardial infarction. He was morbidly obese (a body mass index of 44 kg/m2) and smoked cigarettes but was otherwise without substantial medical history. No autopsy was performed. One postmortem blood culture, obtained at the time of tissue recovery, grew coagulase-negative Staphylococcus. Of eighteen preprocessing cultures performed on the tissues recovered from this donor, one grew Propionibacterium acnes and one grew a fungus that was not further identified.
CASE 2. A previously healthy thirty-four-year-old man underwent open and arthroscopic reconstruction of the posterior cruciate ligament with use of an Achilles tendon allograft secured with Calaxo biointerference screws (Smith and Nephew, Andover, Massachusetts). On postoperative day 69, he presented to the orthopaedic clinic complaining of a “knot” at the inferior incision site with associated sharp, shooting pains. Five milliliters of serosanguineous arthrocentesis fluid was negative on Gram stain but grew Elizabethkingia meningoseptica, sensitive only to ciprofloxacin and trimethoprim-sulfamethoxazole. Prior to the availability of the culture results, the patient began a five-day course of clindamycin.
Two days after the arthrocentesis (postoperative day 71), he had a fever of 101.7°F develop and a repeat arthrocentesis revealed 60 mL of clear fluid with a normal “string sign” (the fluid was not sent for cell count or culture). The erythrocyte sedimentation rate was 76 mm/hr (normal, 0 to 15 mm/hr), and the peripheral white blood-cell count was 9600 cells/μL (9.6 × 109/L), with 84% segmented neutrophils. Methylprednisolone was injected into the knee along with lidocaine and bupivacaine with epinephrine. A magnetic resonance imaging scan showed a 1.5-cm fluid collection at the tibial end of the reconstructed tendon, extensive bone-marrow edema, and a large joint effusion; 500 mg of ciprofloxacin given orally twice daily was added to the clindamycin.
On postoperative day 79, a third arthrocentesis resulted in synovial fluid that was negative on Gram-staining and culture but had numerous crystals on microscopy. The uric acid level was elevated at 8.6 mg/dL (512 μmol/L) (normal, 2.4 to 8.4 mg/dL; 143 to 500 μmol/L); he received indomethacin for suspected gout. Over the next nineteen days, he underwent three incisional wound aspirations and one additional arthrocentesis, and the ciprofloxacin was increased to 750 mg twice daily, but the knee pain and swelling persisted. On postoperative day 98, the patient was returned to the operating room for irrigation and wound débridement. Revision meniscectomy, synovectomy, and chondroplasty were also performed. Knee shavings sent for histologic examination showed inflamed granulation tissue. All repeat wound and joint cultures were negative. Approximately six months after the original posterior cruciate ligament reconstruction and ninety-two days after the irrigation and débridement, the patient continued to have mild pain and swelling of the knee but had gradually improving function.
The tissue used in the second patient had come from a forty-five-year-old man (Donor 2) who had had chest pain and a witnessed collapse with asystole. He had no history of substantial medical problems; his only prescription medication was an antidepressant. An autopsy revealed cardiomegaly with pulmonary edema and mild hepatomegaly; macrovesicular steatosis of the liver was evident on histologic examination. One postmortem blood culture, obtained at the time of tissue recovery, showed no growth at five days. Of twenty-six preprocessing cultures performed on the tissues, eight grew coagulase-negative staphylococci alone, four grew coagulase-negative staphylococci and Propionibacterium species, and one grew group-F streptococcus.
Tissue Trace-Back Investigation
The tissue processor for both donors is a nonprofit organization that provides orthopaedic tissue allografts (including bone and soft tissues) from cadaveric donors. Tissues are recovered with use of aseptic surgical techniques and then are sent to one of two tissue-processing facilities. Preprocessing cultures of recovered tissues are performed at the tissue-processing facilities, and tissues contaminated with Clostridium species or group-A streptococcus are discarded in accordance with American Association of Tissue Banks standards12. Tissue preparation is performed in a clean-room environment and includes dissection of the soft tissue from bone and removal of all extraneous tissues, blood, and lipids.
Soft-tissue processing is performed in physically separate areas from bone-tissue processing. The tissues used in the two patients (Cases 1 and 2) were pretreated with low-dose gamma irradiation prior to exposure to a combination of antibiotics for two hours at 37°C with constant agitation and two sterile buffer rinses. Final sterility testing is performed by means of culturing small samples. In this method, a small sample of the tissue is removed, processed along with the parent tissue, and then cultured by immersion in culture medium (postprocessing culture). If postprocessing cultures are positive for any growth, then all soft tissues from that donor are rejected. The processor has an established protocol for environmental monitoring of the processing facilities with regular microbiologic and/or particle sampling of air, surfaces, water, and personnel.
Between March and September 2006, the tissue bank noticed a threefold increase in the number of postprocessing cultures growing bacteria among soft tissues at one of the two processing facilities. Organisms recovered were predominantly gram-negative bacteria generally found in water and included Elizabethkingia meningoseptica. The positive postprocessing cultures were widely distributed among donors, days, times of day, and tissue types. The positive cultures prompted more extensive environmental sampling, including cultures of all staging and processing areas, raw materials, processing equipment, and the water system. The only positive cultures came from clean-room drains and traps, which grew Elizabethkingia meningoseptica. After receiving the information that the isolates recovered from the recipient (Case 1) and the donor were a genetic match, the tissue bank voluntarily recalled all soft-tissue units that had been processed between February 21, 2006, when the rate of positive postprocessing cultures had started to increase, and September 8, 2006. The tissue bank also quarantined all units remaining in inventory and voluntarily suspended soft-tissue processing at the facility. According to the tissue bank, over 4700 tissue units had been distributed to roughly 750 health-care facilities during the time period covered by the voluntary recall, and approximately 3800 had been implanted. A media article published shortly after the processor initiated the recall described the organism involved and prompted the clinicians involved with the second patient (Case 2) to report the case to the tissue processor.
The tissue processor contacted the clinicians who had implanted the other tissues recovered from Donor 1 (five tissues) and Donor 2 (seven tissues). The clinicians indicated that all other patients were doing well with no evidence of infection. Additionally, the CDC called fifty-seven facilities that had received the largest number of recalled tissue allografts (867) to search for other potential Elizabethkingia meningoseptica infections. No additional cases were identified.
In response to their investigative findings, the tissue processor replaced all clean-room sink drains and traps at the processing facility, installed check valves in the drains, began routine sanitization of the drains, and revised processing procedures to include samples of the final buffer rinsate solution. Tissue-processing resumed following these changes, and sterility failure rates returned to baseline levels with no identification of Elizabethkingia meningoseptica or other waterborne gram-negative bacteria.
Tissue Culture and Molecular Typing
Culture of the one remaining tissue product from Donor 1 grew an Elizabethkingia meningoseptica isolate that was genetically related (92.3% similarity) to the clinical isolate obtained from the first patient (Case 1). Cultures of one of the two remaining tissues from Donor 2 also grew an Elizabethkingia meningoseptica isolate that was genetically indistinguishable (100% similarity) from the isolate obtained from Donor 1 (Fig. 1). There was no growth on culture of the second tissue from Donor 2. No clinical isolate from the second patient (Case 2) or environmental isolates from cultures obtained at the processing facility were available for analysis.
Tendon, bone, and ligament allografts are commonly used in orthopaedic surgical procedures, which are most often successful and life-enhancing. However, rarely, allograft-associated infection can result in substantial complications for the recipient. Previous reports have described allograft-associated infections from undetected donor infections. However, to our knowledge, this is the first report of an allograft-associated infection caused by contamination of tissues during processing. The findings of this investigation indicate that tissues from two different donors were likely contaminated with Elizabethkingia meningoseptica during processing and were the likely sources of infection in two patients who received tissues from these donors.
Elizabethkingia meningoseptica was recovered from sink drains and traps in the clean rooms where the tissues were processed, suggesting that the organism was present at least intermittently in the facility’s water supply. Cultures of water from the sinks did not grow Elizabethkingia meningoseptica, perhaps because the organism was present intermittently or at levels too low to detect with standard culture methods. The tissues themselves or the instruments used during processing might have become contaminated by coming into contact with organisms in the tap water or organisms that were splashed up from the drain. Elizabethkingia meningoseptica was recovered from two of three recovered but unimplanted tissues from the two donors. Molecular typing revealed that isolates from unimplanted tissues were genetically indistinguishable from each other and were closely related to the one patient isolate that was available for analysis. These findings suggest a common source of the Elizabethkingia meningoseptica recovered from the two donors and the first patient (Case 1).
This episode demonstrates the importance of careful analysis of postprocessing culture results and indicates that these results can provide crucial information about tissues from a specific donor, as well as overall tissue-processing at a given facility. In this case, the tissue processor recognized that the increase in positive postprocessing cultures might reflect contamination of tissues during processing. This recognition, combined with the knowledge of the types of organisms recovered, helped to direct a root-cause analysis and additional microbiologic sampling, which ultimately identified a probable source of contamination.
These findings also underscore previously recognized limitations in the accuracy of postprocessing cultures of allograft tissues. All soft-tissue products from the donors in this investigation had negative postprocessing cultures; however, organisms were later recovered from processed, unimplanted tissues. A potential explanation is the residual inhibitory effect of antimicrobials used for processing; a previous report has identified a similar discrepancy in culture results caused when carryover of an antimicrobial solution used for processing led to false-negative cultures11. Another potential explanation for the different results might be the fact that the CDC performed destructive cultures of the tissues, which would facilitate the recovery of organisms beneath the surface of the tissue. Further research on postprocessing culture protocols, including consideration of feasible methods to incorporate destructive sampling methods, is needed. When evaluating infections after allograft implantation, transmission by means of the allograft should not be ruled out solely on the basis of negative postprocessing cultures because of the limitations of these cultures. Testing retained tissues with use of destructive culturing methods that incorporate steps to neutralize residual antimicrobials can assist in evaluating suspected allograft-associated transmissions.
Our findings also demonstrate the important limitations of routine environmental monitoring programs. In this case, the facility was conducting regular assessments of the processing environment, including microbiologic sampling. No important abnormalities in these tests had been noted, despite the increase in positive postprocessing cultures. Ultimately, the source of the organism appeared to be an area (sink drains) that is not considered part of the critical processing environment and hence is generally not monitored. Although these types of environmental monitoring programs are generally in place, there are few data to guide their performance and interpretation. Hence, the results of environmental monitoring programs can only be an adjunct to, but not a replacement for, a thorough microbial assessment of processed allografts.
This episode also highlights some important advances that have been made in investigating allograft-associated infections. The initial case was reported quickly to both the tissue processor and public health officials, indicating that clinicians are increasingly aware of the potential for allografts to transmit infections, especially when unusual pathogens are recovered. Likewise, it appears that health-care providers and facilities are increasingly aware of the importance of tracking allograft tissues. The tissue processor had received information on approximately 90% of the implanted tissues through return of tissue-implant cards. Furthermore, health-care facilities were able to help the tissue bank to determine the final disposition of all tissues involved in this episode, usually within a few days of the initial notification. Finally, this investigation provides an excellent example of how clinicians, tissue processors, and public health officials are working collaboratively to improve the safety of allograft tissues.
Our investigation was subject to limitations. No isolate was available from the second patient (Case 2) for molecular analysis, although the isolate recovered from one retained sample from the donor whose tissues were used in this patient was related to the isolates recovered from the tissue in the first patient (Case 1). Similarly, no environmental samples of Elizabethkingia meningoseptica obtained from the tissue-bank processing areas were available for molecular analysis. Finally, there is little published information on the genetic diversity of this organism to help to guide interpretation of the pulsed-field gel electrophoresis patterns. However, a previous study has documented eight different strains among seventeen Elizabethkingia meningoseptica infections8, which suggests that the genetic relatedness we observed is not likely explained by a lack of genetic diversity in the organism.
Allograft recovery and processing is a complex procedure that involves careful recovery of the tissues to ensure proper function of the tissues and to minimize the risk of contamination, processing of the tissues (in most instances) to further reduce the risk of contamination, proper technique during implantation, and postimplantation surveillance for adverse events. Ensuring the safety of allograft tissues requires safeguards and vigilance by those involved with all steps of tissue recovery and processing as well as by the tissue-banking industry and public health officials. While allograft tissues have been and remain very safe, it is important to emphasize that experiences with previous episodes of allograft contamination have led to important advances in tissue safety. With respect to tissue recovery, the American Association of Tissue Banks (AATB) has developed a list of organisms that are considered to pose a high risk of infection in the recipient, even with adequate tissue-processing12. The AATB standards require that all tissues from any donor with preprocessing cultures that grow one of these pathogens be discarded. Some have raised the question as to whether all tissue with a positive preprocessing culture should be discarded. However, such a strategy would likely have a substantial impact on tissue availability and unclear benefits to overall patient safety. Recent publicity over episodes concerning illegal activities related to tissue recovery also has generated concern2,3. Although these types of illegal activities pertaining to tissues are thought to be very uncommon, it is important to note that the U.S. Food and Drug Administration and the tissue-banking industry have taken steps to address this issue. For tissue processors, the experience described in this report provides relevant information on the importance of ongoing review and trending of the results of postprocessing cultures. Conducting this type of review and investigating abnormal results promptly will further enhance the safety of processed tissue allografts. Also, as mentioned earlier, ongoing research on improving techniques for postprocessing tissue cultures will also be helpful.
Surgeons who implant the tissue provide the final safeguards in the “chain” of allograft use. Surgical staff should carefully read and follow any instructions provided with the tissue. Likewise, to improve tissue-tracking and the ability to notify clinicians of potential problems with tissues, the staff should complete the allograft implant card included with all tissues and return it to the tissue bank. Episodes of allograft-associated infections have also emphasized the importance of considering the possibility of allograft contamination when postoperative infections occur. During a workshop on organ and tissue allograft safety in 2005, experts proposed that clinicians should suspect a possible allograft infection and report to tissue processors and public health officials any of the following outcomes in tissue recipients: (1) infection due to an unexpected pathogen, (2) removal of tissues due to possible infection, and (3) hospitalization within three months of tissue implantation for management of a possible infection13. In an effort to facilitate postimplantation surveillance of allograft tissues, the organ and tissue transplant community, through a cooperative agreement between the United Network for Organ Sharing and the CDC, is developing an adverse event recognition system called the Transplantation Transmission Sentinel Network to help to improve allograft safety. However, until such a system can be implemented nationally, clinicians and health-care facilities should continue to work to develop methods to ensure that possible allograft-associated infections are reported to the tissue processors and public health officials for further investigation, with the findings communicated back to clinicians. Tissue processors should establish procedures and protocols to closely monitor the results of environmental and postprocessing tissue cultures and should initiate investigations when unusual pathogens are recovered or when the number of positive cultures increases.
NOTE: The authors thank Ruth Lynfield, MD, and Jane Harper, BSN, MS, CIC, for their assistance with the investigation.
* The Food and Drug Administration Tissue Safety Team Investigators are Hang Dinh, Melissa Greenwald, MD, Laura St. Martin, MD, and Ruth Solomon, MD.
Disclaimer: The findings and conclusions in this report reflect those of the authors and do not necessarily reflect the official opinion of the Centers for Disease Control and Prevention or the Food and Drug Administration.
Disclosure: The authors did not receive any outside funding or grants in support of their research for or preparation of this work. Neither they nor a member of their immediate families received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity.