|
A correction has been published: N Engl J Med 2004;351(19):2025.
A correction has been published: N Engl J Med 2004;351(4):397.
|
Background Allografts are commonly used in orthopedic reconstructive surgery. In 2001, approximately 875,000 musculoskeletal allografts were distributed by U.S. tissue banks. After the death from Clostridium sordellii sepsis of a 23-year-old man who had received a contaminated allograft from a tissue bank (Tissue Bank A), the Centers for Disease Control and Prevention initiated an investigation, including enhanced case finding, of the methods used for the recovery, processing, and testing of tissue.
Methods A case of allograft-associated clostridium infection was defined as a culture-proven infection of a surgical site within one year after allograft implantation, from January 1998 to March 2002. We traced tissues to tissue banks that recovered and processed these tissues. We also estimated the rates of and risk ratios for clostridium infections for tissues processed by the implicated tissue bank and reviewed processing and testing methods used by various tissue banks.
Results Fourteen patients were identified, all of whom had received allografts processed by Tissue Bank A. The rates of clostridium infection were 0.12 percent among patients who received sports-medicine tissues (i.e., tendons, femoral condyles, menisci) from Tissue Bank A and 0.36 percent among those who received femoral condyles in particular. The risk-ratio estimates for clostridium infections from tissues processed by Tissue Bank A, as compared with those from other tissue banks, were infinite (P<0.001) for musculoskeletal allografts, sports-medicine tissues, or tendons. Because Tissue Bank A cultured tissues only after treating them with a nonsporicidal antimicrobial solution, some test results were probably false negatives. Tissues from implicated donors were released despite the isolation of clostridium or bowel flora from other anatomical sites or reports of infections in other recipients.
Conclusions Clostridium infections were traced to allograft implantation. We provide interim recommendations to enhance tissue-transplantation safety. Tissue banks should validate processes and culture methods. Sterilization methods that do not adversely affect the functioning of transplanted tissue are needed to prevent allograft-related infections.
Processed tissue allografts are not necessarily sterile and may result in viral (hepatitis B,3 hepatitis C,4,5,6 or human immunodeficiency virus7 ) or bacterial (gram-negative organisms or enterococcus) infections.2,8,9,10 Clostridium septic arthritis remains a rare complication of orthopedic procedures.11,12,13,14,15,16,17 After the death from Clostridium sordellii sepsis of a 23-year-old man who had received a cadaveric musculoskeletal allograft,8 the Centers for Disease Control and Prevention (CDC) enhanced its efforts to find cases of allograft-associated clostridium infections and conducted an investigation of the processing and testing methods used by the tissue bank (hereafter referred to as Tissue Bank A) that had processed this patient's allograft. Herein we report our findings and provide interim recommendations for enhanced tissue-transplantation safety.
Methods
Definition and Identification of Cases
A case was defined as a surgical-site infection18 within 12 months after an allograft transplantation in an otherwise healthy U.S. patient without known risk factors for surgical-site infections19 (e.g., diabetes) or clostridium infection (e.g., injection-drug use or hematologic or colon cancer) at the site of allograft implantation during January 1998 through March 2002. All cases were confirmed by culture of blood or knee aspirates.
We solicited case reports by examining published reports8,9,10 and electronic-mail lists and by contacting the Food and Drug Administration (FDA); state regulatory authorities in California, Florida, and New York; and tissue banks. Our investigation consisted of epidemiologic studies, a review of tissue-bank procedures, and laboratory investigations.
Epidemiologic Studies
We used standardized forms to record data from potential case patients' medical and surgical records. We contacted tissue banks that recovered tissue from implicated donors and used standardized forms to record information, such as the time from the death of the donor to refrigeration of the body. We traced tissues from implicated donors to the tissue banks that recovered the tissue and to any other tissue banks that processed tissue from the same donor. We requested data on all available culture methods and results, tissue-processing methods, and patient outcomes.
Since reporting of clostridium infections was not mandatory and the number of allografts of specific types of tissue was unknown, we used available data to estimate the rates of clostridium infection. The numerator consisted of the number of cases reported to the CDC after allograft implantation in 2001. For the denominator, we obtained the total number of musculoskeletal allografts shipped by Tissue Bank A in 2001 using publicly available data filed with the Securities and Exchange Commission. Also, we used data from New York State's licensing program to obtain the number of so-called sports-medicine tissues (i.e., tendons used for reconstruction of the anterior cruciate ligament, fresh femoral condyles, and menisci) distributed nationally by Tissue Bank A in 2001. Next, we calculated the rates of clostridium infection according to the type of tissue transplanted.
Using New York State's licensing-program data, we calculated the proportion of musculoskeletal tissues processed and distributed by Tissue Bank A, as compared with that of other tissue banks in the period between January 1998 and March 2002 (i.e., the market share). We estimated risk ratios of clostridium infections associated with tissues distributed nationally by Tissue Bank A from January 1998 through March 2002. Tissue-specific denominator data for 1998 through 2001 were obtained from New York State. We estimated rates for January 2002 through March 2002, extrapolating 2001 data, assuming the same frequency of allograft distribution.
Procedural Review
We visited three tissue banks, including Tissue Bank A, and observed how tissue was recovered from cadavers, processed, and cultured; reviewed operating-procedure manuals; and interviewed tissue-bank personnel. In addition, we reviewed inspection reports from New York State and the FDA, warning letters, and published regulations and standards.20,21,22,23
Laboratory Studies
We obtained nonimplanted tissues from implicated donors from Tissue Bank A and cultured the specimens. We conducted in vitro experiments to evaluate the tissue bank's method of detecting clostridium spores. To determine whether residual antimicrobial agents on tissue previously suspended in antimicrobial solution could cause bacteriostasis,24 we determined the expected antimicrobial concentrations in culture medium in anaerobic, standard blood-culture bottles that did not contain charcoal (BacT/Alert, Organon Teknika) used by Tissue Bank A. We conducted in vitro experiments to determine whether these antimicrobial concentrations caused bacteriostasis when exposed to low inoculums (100 spores per blood-culture bottle) of C. sordellii (American Type Culture Collection strain 14337); we used three different amounts of antimicrobial solution (0.5 ml, 1.0 ml, and 2.0 ml) to represent residual carryover. Blood-culture bottles were incubated at 35°C for 14 days; bottles showing growth underwent Gram's staining, were plated aerobically and anaerobically, and were confirmed as C. sordellii. Bottles having no growth after 14 days were held for an additional 7 days and then plated; plates were incubated for 10 days. We also examined whether the use of bottles containing activated charcoal helped in spore recovery.25
Statistical Analysis
Data were analyzed with the use of Epi Info software (CDC, version 6.04). The chi-square or Fisher's exact test, where appropriate, was used to compare categorical variables. Risk ratios were calculated. All P values are two-sided.
Results
Epidemiologic Studies
Characteristics of the Patients
Between January 1998 and March 2003, 14 patients (including the index patient) met the case definition (Table 1). C. septicum was isolated in 12 (86 percent). The median age was 32 years (range, 15 to 52); 10 patients (71 percent) were male. Symptoms began a median of 4.5 days (range, 2 to 85) after allograft implantation. The sole death was that of the index patient, who had C. sordellii infection.9 All other patients required hospitalization, intravenous antimicrobial therapy, joint irrigation, and débridement. Ten patients (71 percent) required allograft removal; three were scheduled for knee-replacement surgery for intractable pain, and three could not maintain full-time employment owing to postinfection disability.
|
Tissues implanted into the 14 patients came from 9 donors by way of seven tissue banks. Donors had no overt signs of infection before death; six died from trauma. The median interval from death to refrigeration of the body was 5.9 hours (range, 4.1 to 19.0). Industry standards state that tissue may be recovered up to 24 hours after death as long as the interval between death and refrigeration of the body does not exceed 12 hours23; this limit was exceeded in the case of two donors (15 hours and 19 hours).
Tissue Tracing
Allograft tissues from all 14 patients had been processed by Tissue Bank A. These included nine hemipatellar tendons used for reconstruction of the anterior cruciate ligament (64 percent), four femoral condyles (29 percent), and one meniscus (7 percent). The four femoral condyles were fresh; the other 10 allografts were frozen (71 percent). Tissues from three donors were processed by Tissue Bank A alone, whereas tissues from five donors were processed and distributed by Tissue Bank A as well as other tissue banks. Complete information on tissue processing and distribution was not available for one donor.
Some tissue allografts from the five donors that were sent to other tissue banks were processed with the use of sterilization methods. Gamma irradiation was used for 17 bone, 6 tendon, and 2 fascia lata allografts. Sixty-five bone allografts were processed with the use of a low-temperature chemical-sterilization method (BioCleanse, Regeneration Technologies). No reports of infection were identified among recipients of tissues processed by these other tissue banks with the use of sterilization with gamma irradiation or BioCleanse; tissues at Tissue Bank A did not undergo sterilization.
Rates and Estimated Risk of Clostridium Infection for Tissues Processed by Tissue Bank A
The rates of clostridium infection in 2001 among patients receiving tissues processed by Tissue Bank A were 0.12 percent for all sports-medicine tissues, 0.09 percent for tendons, 0.15 percent for menisci, and 0.36 percent for femoral condyles. Tissue Bank A was estimated to have 0.1 percent of the market share for all musculoskeletal tissue processed for transplantation, 25 percent for sports-medicine tissues, 23 percent for tendons, 78 percent for femoral condyles, and 94 percent for menisci in the period between January 1998 and March 2002. The risk of clostridium infection was significantly higher after the receipt of musculoskeletal tissues processed by Tissue Bank A than after the receipt of tissue from other tissue banks (P<0.001) (Table 2).
|
Tissue-Processing Methods
Tissue Bank A used aseptic techniques during the processing of recovered tissues. We observed no cross-contamination between tissues from different donors. Tissue was decontaminated by suspension in a proprietary solution containing imipenem, vancomycin, amikacin, amphotericin B, and fluconazole. Tissue Bank A did not provide validation of the efficacy of this solution in killing spore-forming organisms, such as clostridium.
Testing Methods
Companion tissue (e.g., a cartilage sliver from a femoral condyle) was processed in parallel with the allograft. After suspension of allografts and companion tissue in antimicrobial solution, companion tissue was placed in cell-culture medium and briskly agitated for 30 seconds. Aliquots of this medium were then cultured aerobically with the use of BacT/Alert blood-culture bottles containing activated charcoal and anaerobically with the use of standard anaerobic bottles (Organon Teknika) and on non-prereduced blood agar.
All cultures were incubated at 35°C; the bottles were incubated for seven days in a BacT/Alert 3D automated system introduced in September 2001. Nonautomated broth-culture methods were used before September 2001. Identification of isolates was performed by Tissue Bank A; not all isolates were fully identified. Tissue Bank A did not routinely measure the microbial burden (quantitative cultures) of incoming tissues and, in 1995, had stopped performing qualitative cultures of tissue before exposing them to antimicrobial solution. When Tissue Bank A identified a positive culture, only tissue associated with the culture was discarded, not other tissues from the same donor. Also, tissues from implicated donors were released for implantation despite reports of infections in other recipients. Tissue Bank A had not conducted bacteriostasis or fungistasis studies to evaluate their culture methods.
Laboratory Studies
Cultures of Non-Implanted Tissues from Tissue Bank A
Tissue Bank A provided 27 processed allografts that had not been implanted from four of nine implicated donors (44 percent). We isolated clostridium (C. sordellii or C. septicum) in cultures from 4 of 17 musculoskeletal tissues (24 percent; from three of four implicated donors). We grew C. sordellii from 2 of 11 musculoskeletal tissues from Donor 4, although all 12 specimens of companion tissue from Donor 4 cultured after antimicrobial treatment by Tissue Bank A for the purpose of quality assurance were negative. Tissue Bank A did not provide corresponding culture results for Donors 3 and 6.
Bacteriostasis
The antimicrobial solution used for processing tissue at Tissue Bank A included two agents with potential activity against clostridium — imipenem and vancomycin. Approximately 1 ml of fluid was carried over with each companion tissue, according to reports obtained from personnel at the tissue bank. Our in vitro experiment demonstrates the potential effect on bacterial growth, especially with a low inoculum, of carrying over antimicrobial solution (Table 3). The use of anaerobic blood-culture bottles that contained charcoal improved the ability to detect C. sordellii, especially with increasing volumes of antimicrobial carryover.
|
Allografts can substantially improve the quality of life.26 Our investigation, however, demonstrates that infection acquired through bacterial contamination of allografts may result in substantial complications or death. Furthermore, our findings suggest that current federal regulations and industry standards for processing and testing allograft tissue need to be enhanced to prevent allograft-associated infections.2,10
Our investigation highlighted several factors that contributed to clostridium infections. First, implanted tissues were not processed with the use of methods that achieved sterility or that were sporicidal. However, current regulations do not require tissue banks to eliminate bacteria present on tissues at the time of recovery or to use processing methods that guarantee tissue sterility. Second, the concentration of bacteria ("bioburden") before processing was unknown, because no tissues were cultured before being exposed to antimicrobial agents. Third, cultures of specimens obtained after processing were probably false negative as a result of carryover of antimicrobial solution. Fourth, evidence of clostridium or bowel flora at other anatomical sites or reports of infections in other allograft recipients were not used as criteria for determining the suitability of donor tissues for transplantation.
At the CDC, cultures of two nonimplanted tissues from one donor yielded clostridium. In contrast, cultures of all 12 companion tissues from the donor were reported to be negative by Tissue Bank A. There are at least two potential reasons for this discrepancy. Because tissues were cultured only after their suspension in antimicrobial solution, residual antimicrobial agents on the tissues may have caused false negative culture results through bacteriostasis. Furthermore, companion tissues used for quality-assurance testing were small and had a larger surface-area-to-volume ratio than the allografts themselves, permitting better antimicrobial penetration of companion tissue.
We hypothesized that donor tissue became hematogenously contaminated by bowel flora, including clostridium and spores, at or before tissue recovery. Factors that might contribute to such contamination include a prolonged interval between the donor's death and tissue recovery, delays in refrigeration, or death from trauma.27,28 With respect to 2 of the 14 patients with clostridium infections (14 percent, or 2 of 9 donors [22 percent]), the interval between the donor's death and refrigeration of the body exceeded the limit recommended by voluntary industry standards.23
Aseptic processing of tissue minimizes bacterial contamination but will not eradicate contamination with organisms or spores, especially in tissue that is heavily contaminated at the time of recovery.2 To reduce bacterial contamination of allografts, some tissue banks, including Tissue Bank A, suspend tissue in antimicrobial solutions. However, these solutions may not eradicate spores, as demonstrated by our in vitro studies (CDC: unpublished data). Two sterilization methods that would eliminate spores — gamma irradiation and treatment with ethylene oxide — have technical problems, limiting their use in tissue processing.2 High doses of gamma irradiation may adversely affect the biomechanical properties of allografts.29,30,31,32 Ethylene oxide has a limited ability to penetrate tissue and has been associated with adverse outcomes such as synovitis33 or damage to musculoskeletal tissue, resulting in an unacceptably high rate of mechanical failure.34
One tissue bank has developed and implemented a low-temperature chemical-sterilization approach (BioCleanse) that kills spores but preserves the biomechanical integrity and function of some allografts.35,36,37 The efficacy of this sterilization method is supported by the absence of reports of bacterial or viral allograft-associated infections in tissue processed by this method (CDC: unpublished data). In contrast, tissues processed with all other disinfection and sterilization methods, including gamma irradiation, have been associated with reports of allograft-associated infections (CDC: unpublished data). Currently, BioCleanse and other sterilization methods cannot be used to process fresh femoral condyles, since it is thought that chondrocytes must be viable to maintain articular cartilage function.
There were several limitations to our investigation. First, we identified a relatively small number of cases. To meet the case definition, cases had to be culture-positive. However, according to a recent survey of U.S. infectious-disease specialists, only 22 percent of respondents always cultured joint aspirates for anaerobic microorganisms; 39 percent rarely or never did.38 Second, determination of the true rate of clostridium infection is challenging, because patients who become symptomatic may not present to the institution (often ambulatory surgical centers) where the surgery was performed. Furthermore, because clinicians presume that allografts are sterile, allografts are not usually considered a potential source of infection.
Third, most clostridium isolates obtained at the time of tissue recovery, before processing, packaging, or implantation, were not available to confirm species identification. This may partially explain the discrepancy between the clostridium species isolated from blood or surgical sites and culture results of tissues obtained before implantation. However, the most likely reason for this discrepancy is that agonal or postmortem bacteremia is frequently polymicrobial, and tissue contamination is not uniform.2
Fourth, there are no reliable data on the number of specific types of tissue distributed by all U.S. tissue banks for use in estimating the incidence of clostridium infections. We used data collected by New York State, which licensed 17 of the largest U.S. tissue banks that processed musculoskeletal tissue in 2001, including Tissue Bank A. Thus, we may have overestimated the market share of Tissue Bank A and consequently underestimated the risk ratio of clostridium infections associated with this tissue bank.
On the basis of our investigation, we made recommendations to tissue banks, the American Association of Tissue Banks, and the FDA to reduce the risk of allograft-associated infection (Table 4). The use of a validated sporicidal process will confer the greatest protection for patients; the other proposed measures are less effective. For example, the sensitivity of cultures of specimens obtained before processing is low (10 to 22 percent)39,40; thus, relying on the results of cultures alone to identify and discard tissues potentially contaminated with clostridium spores is problematic. However, for tissues that are not amenable to sterilization (e.g., fresh femoral condyles), the use of donor screening and preprocessing cultures will most likely reduce, but not eliminate, the risk of infection.
|
In conclusion, improved guidelines for tissue processing and testing, together with monitoring of allograft-associated adverse events, should enhance tissue-transplantation safety. However, the best way to reduce the risk of pathogen transmission is to develop and use sterilization methods that do not adversely affect the functioning of the tissue after transplantation.
We are indebted to the clinicians, patients, and public health staff who reported cases, and to tissue-bank personnel who assisted in performing trace-backs and trace-forwards of tissues from implicated donors and shared data on microbiologic cultures and processing methods.
Source Information
From the Epidemic Intelligence Service, Division of Applied Public Health Training, Epidemiology Program Office (M.A.K.), and the Division of Healthcare Quality Promotion (M.A.K., D.N.W., H.T.H., W.R.J., D.B.J., L.K.A.), Centers for Disease Control and Prevention, Atlanta; the Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia (M.A.K.); and Wadsworth Center, New York State Department of Health, Albany (J.V.L.).
Address reprint requests to Dr. Kainer at the Tennessee Department of Health, 4th Fl., Cordell Hull Bldg., 425 5th Ave. N., Nashville, TN 37247, or at marion.kainer{at}state.tn.us.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Infections Associated with Musculoskeletal-Tissue Allografts
De Andrade J. R., Ray G. S., Pruss A., Göbel U. B., Pauli G., Polder J. A., Wood J. J., Braun M. M., Kainer M. A., Jarvis W. R.
Extract |
Full Text |
PDF
N Engl J Med 2004;
351:1358-1360, Sep 23, 2004.
Correspondence
Correction: Infections and Musculoskeletal-Tissue Allografts
Kainer M. A., Archibald L. K.
Extract |
Full Text |
PDF
N Engl J Med 2004;
351:397-398, Jul 22, 2004.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
