Artery, periarterial adipose tissue, and blood microbiology during vascular reconstructive surgery: Perioperative and early postoperative observations Thomas W. Wakefield, M D , Carl L. Pierson, PhD, Dennis R. Sehaberg, M D , Louis M. Messina, M D , S. M a r t i n Lindenauer, M D , Lazar J. Greenfield, MI), Gerald B. Zelenock, M D , and James C. Stanley, M D , A n n Arbor, Mich. The presence and clinical significance of bacteria in the arterial wall, periarterial adipose tissue, and blood samples acquired during elective vascular operations were assessed in this study. Specimens were obtained from a random series of 84 patients (56 men, 28 women) undergoing 75 primary and 9 secondary arterial reconstructions. Operations performed most frequently included abdominal aortic aneurysmectomy (42), aortofemoral bypass reconstruction (15), and lower extremity bypass surgery for occlusive disease(7). Perioperative antibiotics were administered to all but one patient. A sample of artery, adjacent adipose tissue, and blood were obtained for microbial analysis during the vascular anastomosis or arteriotomy. This yielded a total of 152 artery, 139 adipose tissue, and 129 blood samples for study. Each specimen was divided and placed in blood agar plates, thioglycolate broth, and brain-heart infusion broth. Tissues yielding growth of the same organism(s) in two or more different media were considered positive for the presence of bacteria. Bacteria were present in at least one of the three tissues studied in 32/84 patients (38%). The frequency of positive cultures in primary (29/75, 39%) and secondary procedures (3/9, 33%) were similar. One positive culture site occurred in 26/32 (81%) patients, two positive culture sites existed in 5/32 (16%) patients, and three positive sites were found in 1/22 (3%) patients. Eighteen individual artery (18/152, 12%) and 19 adipose tissue samples (19/139, 14%) harbored bacteria, whereas only two blood cultures were positive (2/129, 2%). Organisms identified included coagulase-negative staphylococci (71%), y-streptococci (7%), diphtheroids (7%), Micrococcus (5%), ~streptococci (5%), Staphylococcus aureus (2%), and Pseudomonaspicketti (2%). Coagulasenegative staphylococci accounted for 60% of positive arterial samples, 79% of positive adipose tissue samples, and both positive blood cultures. Brief follow-up of I to 29 months (mean 15 months) revealed no clinical evidence of arterial or graft infection. These results document the relatively common presence of bacteria in the arterial wall and adjacent adipose tissue during vascular reconstructive procedures, and support the contention that acute infectious complications do not necessarily accompany positive arterial wall or adipose tissue cultures. (J VAsc SURG 1990;11:624-8.) Vascular graft infection affects between 1% and 6% of patients receiving synthetic conduits during arterial reconstructions, with the average reported
From the Section of Vascular Surgery, Department of Surgery, the Department Of Pathology, and the Division of Infectious Diseases, Department of Internal Medicine, University of Michigan. Presented at the Thirteenth Annual Meeting of the Midwestem Vascular SurgicalSociety,Chicago,IlL, Sept. 29-30, 1989. Reprint requests: Thomas W. Wakefield, MD, University Hospital-2210 THCC, 1500 E. MedicalCenter Dr., Ann Arbor, MI 48109-0329. 24/6/19422 624
incidence being 0.7% for aortoiliac bypass surgery and 1.6% for aortofemoral bypass surgery) The seriousness of this complication is evident in the near 50% mortality with infected aortofemoral bypass grafts and associated high rate of lower extremity amputation in survivors. Factors contributing to potential graft infection include contact of the conduit with the skin during placement, contaminated lymphatics, breaks in sterile technique during operation, wound contamination or sepsis, and transient bacteremias before the conduit has developed a stable pseudointimal surface. In addition, a recent hypothesis suggests that placement of a conduit to an artery
Volume i1 Number 5 May 1990
or its contents that harbor bacteria may contribute to vascular graft infection. In support of this hypothesis is the recent recognition that more than 40% of arterial walls examined during clean vascular reconstructive procedures exhibit bacterial growth. 2,3 The present investigation was designed to establish the precise incidence of bacteria in arterial samples, periarterial adipose tissue, and blood acquired during elective vascular operations. MATERIAL AND METHODS
Specimens for bacteriologic analyses were obtained from a series of 84 patients chosen at random in a nonselective manner among individuals undergoing elective vascular surgical procedures at the University of Michigan Hospital between November 1986 and January 1989. The patient population included 56 men and 28 women, with a mean age of 65 years. Informed consent was obtained in accordance with the protocol of the University of Michigan Institutional Review Board. Operative procedures included abdominal aortic aneurysmectomy (42), aortofemoral bypass reconstruction for occlusive disease (15), lower extremity bypass grafting (7), femoral pseudoaneurysm resection (6), femorofemoral bypass surgery (3), profundoplasty and thrombectomy (2), carotid-subclavian bypass grafting (2), carotid endarterectomy (2), and one instance each of a thoracoabdorninal aneurysmectomy, subclavianaxillary bypass grafting, femoral artery aneurysmectomy, popliteal artery aneurysmectomy, and a thoracofemoral Dacron graft thrombectomy. There were 75 primary procedures and nine secondary procedures in this experience. Perioperative antibiotic prophylaxis was used in 83 of the 84 patients, including cefazolin in 74 cases. Preoperative skin preparation was accomplished by povidone iodine (Solo Prep, Deseret, Becton-Dickinson, Franklin Lake, N.J.) followed by placement of an iodophor drape (Ioben II, 3M Co., St. Paul, Minn.). In no instance was graft or artery infection suspected or diagnosed before operation or recognized during operation. Small samples of artery from the site of an anastomosis or arteriotomy, periarterial adipose tissue immediately adjacent to the blood vessel, and blood aspirated from the blood vessel were obtained for microbial analyses during the arterial anastomosis or arteriotomy. Because of the contingencies of the operative procedure, samples from all three sources were not always obtained on every occasion. One hundred fifty-two artery segments, 139 adipose tissue segments, and 129 blood samples were examined. Excised arterial and adjacent
Vascular surgery microbiology 625
adipose tissues were placed separately into sterile, dry transport tubes. At the same time 20 ml of blood was collected in a sterile syringe and divided equally into two Isolator (Dupont Co., Wilmington, Del.) blood tubes. Specimens were usually processed within 30 minutes, with occasional tissue samples kept for a maximum of 2 hours on crushed ice before processing. All specimens were handled with clean gloves in a class II biological safety cabinet. Arterial and adipose tissues were minced with iris scissors and placed into scintered glass grinders containing 2 ml of brain-heart infusion broth. These tissues were ground by hand with frequent vortexing until a uniform suspension was achieved. Subsequently, 0.5 ml of the suspension was inoculated into (1) two 5% sheep blood agar (SBA) plates, (2) one 20 ml thioglycolate broth, and (3) one 10 ml brain-heart infusion broth. Isolator tubes containing blood were centrifuged at 3000 g for 30 minutes, and the supernatant was discarded. The resuspended pellets were used to inoculate four chocolate SBA plates. Two of these plates and one of the tissue sample SBA plates were incubated under anerobic conditions at 35 ° C. All other plates and tubed media were incubated in 5% CO2 at 35 ° C. The cultures were observed daily for 7 days for evidence of microbial growth. Recovered organisms were isolated and identified by use of techniques standardized by the American Society for Microbiology. 4 Tissues yielding growth of the same organism(s) in two or more different media were considered positive for the presence of bacteria. Tissues yielding growth of an organism recovered in only one type of medium were considered contaminated during the processing of the tissue. Blood cultures were considered positive if the same organism was recovered on more than one plate. Quantification of the organisms on all plates was performed and reported as colony forming units (CFU). Complete patient follow-up in 76 patients, including eight who died, was obtained by office visit or chart review (43) or telephone interview (33). Eight patients were lost to followup. Statistical analysis of data was by chi-square analysis where appropriate. RESULTS Thirty-two of 84 patients (38%) exhibited bacterial growth from one or more of the three (artery, adipose tissue, or blood) sample sites (Table I). The overall presence of bacteria in primary operative procedures (29/75, 39%) was similar to that in secondary procedures (3/9, 33%). Considering the actual number of positive culture sites, 26 of 32 patients
626
journal of VASCULAR SURGERY
Wakefield et al.
Table I. Number and percent of patients in total series of 84 exlfibiting positive and negative cultures~ Positive culture Negative culture
Arterial wall
Adipose tissue
Blood
Overall~
18 (21%) 66 (79%)
17 (20%) 67 (80%)
2 (2%) 82 (98%)
32 (38%) 52 (62%)
~No difference in frequency of positive cultures in primary procedures (29/75, 39%) or secondary procedures (3/9, 33%). ?Solitary positive cultures occurred in 26 patients, two positive cultures in 5 patients, three positive cultures in one patient, and negative cultures in 52 patients.
Table II. Number and percent of tissues exhibiting positive cultures Total specimens Positive specimens
Arterial wall
Adipose tissue ~
Blood
Total
152 18 (12%)
139 19 (14%)
129 2 (2%)
420 39 (9%)
~Two patients revealed organisms in two separate adipose tissue specimens.
exhibiting bacterial growth (81%) had only one positive site, whereas five of 32 (16%) had two positive sites, and one of 32 (3%) exhibited three positive sites. Of the total number of samples studied, 18 arterial (12%), I9 adipose tissue (14%), and two blood samples (2%) revealed bacterial growth (Table II). Only three patients had open lower extremity ulcers or gangrene, and of these three cases two had positive adipose tissue cultures. Organisms cultured (Table III) involved primarily gram-positive varieties, including a predominance of coagulase-negative staphylococci (71%). Coagulase-negative staphylococci accounted for 60% of all positive arterial wall samples, 79% of all positive adipose tissue samples, and 100% of the two positive blood cultures. Less commonly encountered organisms were ,/-streptococci and diphtheroids (7% each), Micrococcus and a-streptococci (5% each), and Staphylococcus aureus and Pseudomonas picketti (2% each). The numbers of organisms cultured ranged from one to 24 colony forming units (CFUs)/platc, with an average of 5 CFUs/plate. Among primary operative procedures, 21 of 29 (72%) cultures grew coagulase-negative staphylococci, whereas in secondary procedures this organism was noted twice (67%), and 7-streptococci and M i crococcus organisms together were noted once (33%). No statistically significant difference was found between the 18 of 42 (43 %) patients with abdominal aortic aneurysmectomy who exhibited positive cultures and the four of 15 (27%) patients undergoing aortofemoral bypass surgery having positive culture results. Three of the latter positive cultures in aortofemoral procedures involved the femoral site.
Three of the seven lower extremity bypass grafts exhibited bacteria in the femoral adipose tissue samples, One of the three patients undergoing femorofemoral bypass grafting exhibited a positive femoral adipose tissue culture. There appeared to be no correlation between the timing of the antibiotic administration and presence of a positive bacteriologic culture. Patient follow-up averaged 15 months (range, 1 to 29 months) with no clinical evidence of arte U or graft infection, including examination of the eight patients who died during the course of this study. DISCUSSION
Prosthetic vascular graft infection remains a serious complication of arterial reconstructive surgeU encompassing an overall 34% mortality rate and 20% amputation rate, with up to a 59% to 70% mortality in the case of infected aortoiliac-aortofemoral prostheses. 5,6 The best contemporary outcome with aortic graft infection includes a 27% mortality rate and amputation rate of 25%, most at an above-theknee level. 7 Among those factors considered critical to development of graft sepsis, the presence of bacteria in the arterial wall or arterial contents has recently received considerable attention. The contents of abdominal aortic aneurysms often contain bacteria, but the importance of this finding is unknown. Ernst et al.8 reported a 15 % yield of organisms in aneurysm contents and an 11% yield in intestinal bag fluid. When these two sources were combined in this series, it was found that 20% of patients exhibited bacterial growth. Among 60 surviving patients followed at least 6 months in this former report, one of 10 pa-
Volume 11 Number 5 May 1990
Vascular surgery microbiology
627
Table III. Number and percent of organisms identified in 41 positive cultures ~ Arterial wall Coagulase-ne.gative staphylococci y-streptococcx Diphtheroids
Micrococcus s-streptococci
Staphylococcus aureus Pseudomonaspicketti
Adipose tissue
Blood
15 (37%) 2 (5%) -1 (2%) -1 (2%) --
2 (5%) -------
12 1 3 1 2
(29%) (2%) (7%) (2%) (5%) -1 (2%)
Total 29 3 3 2 2 1 1
(71%) (7%) (7%) (5%) (5%) (2%) (2%)
Average no. of colonyforming units~plate 5 20 14 8 2 10 10
~Two arterial wall samples revealed two different organisms.
tients with positive cultures from either the aneurysm contents or intestinal bag fluid developed graft sepsis, as did one of the 50 survivors having negative cultures. In another study, Williams and Fisher 9 demonstrated bacterial growth, including S. aureus, Escberecbia coli, Pseudomonas and hemolytic streptococci, in seven of 68 cultures of ancurysm contents. Buckels et al.10, in a more recent evaluation, documented a 16.7% incidence of positive cultures from aneurysmal contents of ruptured aneurysms, 9.1% from symptomatic expanding aneurysms, and 4.2% from asymptomatic ancurysms. Gram-positive organisms predominated in this latter experience, and subsequent graft sepsis was significantly greater (p < 0.001) in those with positive cultures (7/22) compared to those with negative cultures (6/253). Two additional reports have documented similar incidences (14% and 23%) of positive cultures from abdominal aortic aneurysm contents. 1~a2 Macbeth et al.z were the first to emphasize an unexpectedly high incidence of bacteria in the artery wall itself. They undertook bacteriologic studies on 88 arterial samples from patients undergoing clean elective arterial reconstructions between 1981 and 1982. Control cultures were obtained from adjacent adipose tissue or lymph nodes. Bacteria were present in 43% of the arterial walls cultured but in none of the control tissues. S. epidermidis was encountered in 71% of their positive specimens. All three graft infections in patients undergoing primary operation in their own practice occurred in individuals having positive arterial wall cultures at the time of the initial operation. In addition, arterial and graft wall cultures were obtained from 20 patients treated for 22 graft infections over the preceeding 13 years at their medical center, and positive arterial wall cultures at the aortic closure site carried a 57% incidence of eventual suture line disruption versus no disruptions in those having negative cultures. This report clearly estab-
lished the relevance of bacteria residing in the vessel wall as a cause of serious complications. The former University of Arizona study was subsequently extended to include 3 years of operative cultures involving 172 patients? The overall incidence of positive arterial wall cultures in this larger series was 44% (75 / 172). That portion of the arterial wall more prone to organism growth appeared to be the intima. However, a limited number of samples prevented any firm conclusions regarding this matter. There were six infections (3.5%) among the 172 patients during an 18.7 month follow-up. All six infections occurred in patients having prior arterial wall cultures that were positive. Subsequent graft infections affected 8% (6/75) of patients with positive cultures, compared to no infections (0/97) in those without positive cultures. This difference proved significant (p < 0.025). In this extended experience, there were 132 primary operations and 40 secondary procedures. Positive cultures were noted in 43% and 45% of primary and secondary procedures, respectively. S. epidermidis was the most common organism encountered, affecting 60% and 53% of primary and secondary operations. Among primary operations, positive arterial cultures were not predictive of subsequent graft infection (1/57) when compared to negative culture (0/75). However, in secondary procedures positive arterial wall cultures were predictive (2o < 0.05) of later graft infection (5/18) when compared to negative cultures (0/22). The Arizona group suggested that routine cultures need only be obtained during secondary operative procedures? Concern for the potential of subsequent graft infection in patients with positive cultures led them to recommend long-term administration of antibiotics. In their experience with this subgroup of patients, the incidence of eventual graft infection was 20% in those taking long-term antibiotics versus 38% in those not receiving long-term antibiotics. A
628
)%urnaiof VASCULAR SURGERY
Wakefield et al.
review o f recent reports revealed the risk o f graft infection associated with a positive arterial wall culture to be 10.5% compared to a 1.3% risk with a negative culture, thus attesting to the potential importance o f this finding. 3 S. aureus was previously considered the most common pathogen in vascular graft infection, 6a3 although recent emergence ofS. epidermidis as the dominant pathogen deserves note.14 This latter organism is by far the most frequently encountered o f the coagulase-negative staphylococci species. The basis for the more common recognition o f S. epidermidis organisms may be a realization that it must be recovered by culturing the graft or artery in broth for an extended period o f time and the need to free the bacteria from their glycocalyx slime in which they live. is Tissue grinding, as performed in the present investigation, and ultrasound sonication have been found equal in their ability to maximize recovery o f this organism26 Despite the recently reported high incidence o f positive cultures from either the arterial wall or adjacent adipose tissue, clinical graft infections are not common. Differences between our results and those from previous investigations including the large number o f positive adipose tissue cultures in our series may relate to differences in biopsy sites, culture technique, specimen size, or the definition o f infection. The fact that only two patients had positive blood samples suggests that the high frequency o f positive cultures does not relate to hematologic crossseeding o f the samples at the time they were obtained. Possible explanations for the lad< o f clinical graft infection include the low virulence o f coagulasenegative staphylococci compared to other bacteria, the low concentration o f bactcria present, the use o f prophylactic antibiotics in all but one patient studied, and the potential that these low-virulent organisms may cause covert infections that only become manifest much later.~4 Nevertheless, the presence o f artery and periartcrial tissue bacteria and serious consequences o f overt graft infection have caused appropriate clinical concern. In this regard, our study indicates that acute infectious complications are not a frequent early accompaniment o f elective vascular reconstructive procedures associated with positive cultures o f either the arterial wall or periarterial adipose tissue. However, final conclusions regarding the im-
portance o f bacterial presence in these circumstanccs must await longer follow-up. The authors acknowledge the technical assistance of Cathy A. PFaus for manuscript preparation. REFERENCES
1. SzilagyiDE, Smith RE, Elliott JP, VrandecicMD. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-33. 2. Macbeth GA, Rubin JR, McIntyre KE, GoldstoneI, Malone IM. The relevaa~ceof arterial wall microbiologyto the treatment of prosthetic graft infections: graft infectionvs. arterial infection. J VAse SURG1984;1:750-6. 3. Durham JR, MaloneJM, BernhardVM. The impact of multiple operations on the importance of arterial wall cultures. J VAse SURG1987;5:160-9. 4. Lennette EH, BalowsA, Hausler WJ Jr, Shadomg HI, eds. Manual of clinical microbiology,4th ed. Washington DC, American Societyfor Microbiology, 1985. 5. Fry WI, Lindenauer SM. Infection complicating the use of', plastic arterial implants. Arch Surg 1967;94:600-9. 6. LiekwegWG, GreenfieldLJ. Vascular prosthetic infections: collectedexperienceand results of treatment. Surgery 1977; 81:335-42. 7. ReillyLM, StoneyRJ, GoldstoneJ, EhrenfeldWK. Improved management of aortic graft infection: the influenceof operation sequence and staging. J VASeSUV,G 1987;5:421-31. 8. Ernst CB, Campbell HC, Daugherty ME, Sachatello CR, GriffinWO. Incidenceand significanceof intra-operativebacterial culturesduring abdominalaortic aneurysmectomy.Ann Surg 1977;185:626-33. 9. Williams RD, Fisher FW. Aneurysmcontents as a source of graft infection. Arch Surg 1977;112:415-6. 10. Buckels JAC, Fielding IWL, Black J, Ashton F, Slaney G. Significance of positive bacterial cultures from aortic aneurysm contents. Br J Surg 1985;72:440-2. 11. McAuleyCE, Steed DL, Webster MW. Bacterialpresencein aortic thrombus at electiveaneurysmresection: is it clinically significant? Am J Surg 1984;147:322-4. 12. ScobieK, McPhailN, BarberG, Elder R. Bacteriologicmonitoring in abdominalaortic surgery.Can J Surg 1979;22:36871. la. Bunt TJ. Synthetic vascular graft infections. I. Graft infections. Surgery 1983;93:733-46. 14. BandykDF, Berni GA, Thiele BL, Towne lB. Aortofemoral graft infection due to Staphylococcusepidermidis. Arch Surg 1984;I 19:102-8. 15. Tollefson DF, Bandyk DF, Kaebnick HW, Seabrook GR, Towne JB. Surfacebiofih'ndisruption: enhanced recoveryof microorganisms from vascular prostheses. Arch Surg 1987; 122:38-43. 16. BerganainiTM, BandykDF, Govostis D, Vetsch R, Townc lB. Identification of Staphylococcus epidermidis vascular graft infections: a comparisonof culture techniques. I VASCSURG 1989;9:665-70.
From the Section of Vascular Surgery, Department of Surgery, the Department Of Pathology, and the Division of Infectious Diseases, Department of Internal Medicine, University of Michigan. Presented at the Thirteenth Annual Meeting of the Midwestem Vascular SurgicalSociety,Chicago,IlL, Sept. 29-30, 1989. Reprint requests: Thomas W. Wakefield, MD, University Hospital-2210 THCC, 1500 E. MedicalCenter Dr., Ann Arbor, MI 48109-0329. 24/6/19422 624
incidence being 0.7% for aortoiliac bypass surgery and 1.6% for aortofemoral bypass surgery) The seriousness of this complication is evident in the near 50% mortality with infected aortofemoral bypass grafts and associated high rate of lower extremity amputation in survivors. Factors contributing to potential graft infection include contact of the conduit with the skin during placement, contaminated lymphatics, breaks in sterile technique during operation, wound contamination or sepsis, and transient bacteremias before the conduit has developed a stable pseudointimal surface. In addition, a recent hypothesis suggests that placement of a conduit to an artery
Volume i1 Number 5 May 1990
or its contents that harbor bacteria may contribute to vascular graft infection. In support of this hypothesis is the recent recognition that more than 40% of arterial walls examined during clean vascular reconstructive procedures exhibit bacterial growth. 2,3 The present investigation was designed to establish the precise incidence of bacteria in arterial samples, periarterial adipose tissue, and blood acquired during elective vascular operations. MATERIAL AND METHODS
Specimens for bacteriologic analyses were obtained from a series of 84 patients chosen at random in a nonselective manner among individuals undergoing elective vascular surgical procedures at the University of Michigan Hospital between November 1986 and January 1989. The patient population included 56 men and 28 women, with a mean age of 65 years. Informed consent was obtained in accordance with the protocol of the University of Michigan Institutional Review Board. Operative procedures included abdominal aortic aneurysmectomy (42), aortofemoral bypass reconstruction for occlusive disease (15), lower extremity bypass grafting (7), femoral pseudoaneurysm resection (6), femorofemoral bypass surgery (3), profundoplasty and thrombectomy (2), carotid-subclavian bypass grafting (2), carotid endarterectomy (2), and one instance each of a thoracoabdorninal aneurysmectomy, subclavianaxillary bypass grafting, femoral artery aneurysmectomy, popliteal artery aneurysmectomy, and a thoracofemoral Dacron graft thrombectomy. There were 75 primary procedures and nine secondary procedures in this experience. Perioperative antibiotic prophylaxis was used in 83 of the 84 patients, including cefazolin in 74 cases. Preoperative skin preparation was accomplished by povidone iodine (Solo Prep, Deseret, Becton-Dickinson, Franklin Lake, N.J.) followed by placement of an iodophor drape (Ioben II, 3M Co., St. Paul, Minn.). In no instance was graft or artery infection suspected or diagnosed before operation or recognized during operation. Small samples of artery from the site of an anastomosis or arteriotomy, periarterial adipose tissue immediately adjacent to the blood vessel, and blood aspirated from the blood vessel were obtained for microbial analyses during the arterial anastomosis or arteriotomy. Because of the contingencies of the operative procedure, samples from all three sources were not always obtained on every occasion. One hundred fifty-two artery segments, 139 adipose tissue segments, and 129 blood samples were examined. Excised arterial and adjacent
Vascular surgery microbiology 625
adipose tissues were placed separately into sterile, dry transport tubes. At the same time 20 ml of blood was collected in a sterile syringe and divided equally into two Isolator (Dupont Co., Wilmington, Del.) blood tubes. Specimens were usually processed within 30 minutes, with occasional tissue samples kept for a maximum of 2 hours on crushed ice before processing. All specimens were handled with clean gloves in a class II biological safety cabinet. Arterial and adipose tissues were minced with iris scissors and placed into scintered glass grinders containing 2 ml of brain-heart infusion broth. These tissues were ground by hand with frequent vortexing until a uniform suspension was achieved. Subsequently, 0.5 ml of the suspension was inoculated into (1) two 5% sheep blood agar (SBA) plates, (2) one 20 ml thioglycolate broth, and (3) one 10 ml brain-heart infusion broth. Isolator tubes containing blood were centrifuged at 3000 g for 30 minutes, and the supernatant was discarded. The resuspended pellets were used to inoculate four chocolate SBA plates. Two of these plates and one of the tissue sample SBA plates were incubated under anerobic conditions at 35 ° C. All other plates and tubed media were incubated in 5% CO2 at 35 ° C. The cultures were observed daily for 7 days for evidence of microbial growth. Recovered organisms were isolated and identified by use of techniques standardized by the American Society for Microbiology. 4 Tissues yielding growth of the same organism(s) in two or more different media were considered positive for the presence of bacteria. Tissues yielding growth of an organism recovered in only one type of medium were considered contaminated during the processing of the tissue. Blood cultures were considered positive if the same organism was recovered on more than one plate. Quantification of the organisms on all plates was performed and reported as colony forming units (CFU). Complete patient follow-up in 76 patients, including eight who died, was obtained by office visit or chart review (43) or telephone interview (33). Eight patients were lost to followup. Statistical analysis of data was by chi-square analysis where appropriate. RESULTS Thirty-two of 84 patients (38%) exhibited bacterial growth from one or more of the three (artery, adipose tissue, or blood) sample sites (Table I). The overall presence of bacteria in primary operative procedures (29/75, 39%) was similar to that in secondary procedures (3/9, 33%). Considering the actual number of positive culture sites, 26 of 32 patients
626
journal of VASCULAR SURGERY
Wakefield et al.
Table I. Number and percent of patients in total series of 84 exlfibiting positive and negative cultures~ Positive culture Negative culture
Arterial wall
Adipose tissue
Blood
Overall~
18 (21%) 66 (79%)
17 (20%) 67 (80%)
2 (2%) 82 (98%)
32 (38%) 52 (62%)
~No difference in frequency of positive cultures in primary procedures (29/75, 39%) or secondary procedures (3/9, 33%). ?Solitary positive cultures occurred in 26 patients, two positive cultures in 5 patients, three positive cultures in one patient, and negative cultures in 52 patients.
Table II. Number and percent of tissues exhibiting positive cultures Total specimens Positive specimens
Arterial wall
Adipose tissue ~
Blood
Total
152 18 (12%)
139 19 (14%)
129 2 (2%)
420 39 (9%)
~Two patients revealed organisms in two separate adipose tissue specimens.
exhibiting bacterial growth (81%) had only one positive site, whereas five of 32 (16%) had two positive sites, and one of 32 (3%) exhibited three positive sites. Of the total number of samples studied, 18 arterial (12%), I9 adipose tissue (14%), and two blood samples (2%) revealed bacterial growth (Table II). Only three patients had open lower extremity ulcers or gangrene, and of these three cases two had positive adipose tissue cultures. Organisms cultured (Table III) involved primarily gram-positive varieties, including a predominance of coagulase-negative staphylococci (71%). Coagulase-negative staphylococci accounted for 60% of all positive arterial wall samples, 79% of all positive adipose tissue samples, and 100% of the two positive blood cultures. Less commonly encountered organisms were ,/-streptococci and diphtheroids (7% each), Micrococcus and a-streptococci (5% each), and Staphylococcus aureus and Pseudomonas picketti (2% each). The numbers of organisms cultured ranged from one to 24 colony forming units (CFUs)/platc, with an average of 5 CFUs/plate. Among primary operative procedures, 21 of 29 (72%) cultures grew coagulase-negative staphylococci, whereas in secondary procedures this organism was noted twice (67%), and 7-streptococci and M i crococcus organisms together were noted once (33%). No statistically significant difference was found between the 18 of 42 (43 %) patients with abdominal aortic aneurysmectomy who exhibited positive cultures and the four of 15 (27%) patients undergoing aortofemoral bypass surgery having positive culture results. Three of the latter positive cultures in aortofemoral procedures involved the femoral site.
Three of the seven lower extremity bypass grafts exhibited bacteria in the femoral adipose tissue samples, One of the three patients undergoing femorofemoral bypass grafting exhibited a positive femoral adipose tissue culture. There appeared to be no correlation between the timing of the antibiotic administration and presence of a positive bacteriologic culture. Patient follow-up averaged 15 months (range, 1 to 29 months) with no clinical evidence of arte U or graft infection, including examination of the eight patients who died during the course of this study. DISCUSSION
Prosthetic vascular graft infection remains a serious complication of arterial reconstructive surgeU encompassing an overall 34% mortality rate and 20% amputation rate, with up to a 59% to 70% mortality in the case of infected aortoiliac-aortofemoral prostheses. 5,6 The best contemporary outcome with aortic graft infection includes a 27% mortality rate and amputation rate of 25%, most at an above-theknee level. 7 Among those factors considered critical to development of graft sepsis, the presence of bacteria in the arterial wall or arterial contents has recently received considerable attention. The contents of abdominal aortic aneurysms often contain bacteria, but the importance of this finding is unknown. Ernst et al.8 reported a 15 % yield of organisms in aneurysm contents and an 11% yield in intestinal bag fluid. When these two sources were combined in this series, it was found that 20% of patients exhibited bacterial growth. Among 60 surviving patients followed at least 6 months in this former report, one of 10 pa-
Volume 11 Number 5 May 1990
Vascular surgery microbiology
627
Table III. Number and percent of organisms identified in 41 positive cultures ~ Arterial wall Coagulase-ne.gative staphylococci y-streptococcx Diphtheroids
Micrococcus s-streptococci
Staphylococcus aureus Pseudomonaspicketti
Adipose tissue
Blood
15 (37%) 2 (5%) -1 (2%) -1 (2%) --
2 (5%) -------
12 1 3 1 2
(29%) (2%) (7%) (2%) (5%) -1 (2%)
Total 29 3 3 2 2 1 1
(71%) (7%) (7%) (5%) (5%) (2%) (2%)
Average no. of colonyforming units~plate 5 20 14 8 2 10 10
~Two arterial wall samples revealed two different organisms.
tients with positive cultures from either the aneurysm contents or intestinal bag fluid developed graft sepsis, as did one of the 50 survivors having negative cultures. In another study, Williams and Fisher 9 demonstrated bacterial growth, including S. aureus, Escberecbia coli, Pseudomonas and hemolytic streptococci, in seven of 68 cultures of ancurysm contents. Buckels et al.10, in a more recent evaluation, documented a 16.7% incidence of positive cultures from aneurysmal contents of ruptured aneurysms, 9.1% from symptomatic expanding aneurysms, and 4.2% from asymptomatic ancurysms. Gram-positive organisms predominated in this latter experience, and subsequent graft sepsis was significantly greater (p < 0.001) in those with positive cultures (7/22) compared to those with negative cultures (6/253). Two additional reports have documented similar incidences (14% and 23%) of positive cultures from abdominal aortic aneurysm contents. 1~a2 Macbeth et al.z were the first to emphasize an unexpectedly high incidence of bacteria in the artery wall itself. They undertook bacteriologic studies on 88 arterial samples from patients undergoing clean elective arterial reconstructions between 1981 and 1982. Control cultures were obtained from adjacent adipose tissue or lymph nodes. Bacteria were present in 43% of the arterial walls cultured but in none of the control tissues. S. epidermidis was encountered in 71% of their positive specimens. All three graft infections in patients undergoing primary operation in their own practice occurred in individuals having positive arterial wall cultures at the time of the initial operation. In addition, arterial and graft wall cultures were obtained from 20 patients treated for 22 graft infections over the preceeding 13 years at their medical center, and positive arterial wall cultures at the aortic closure site carried a 57% incidence of eventual suture line disruption versus no disruptions in those having negative cultures. This report clearly estab-
lished the relevance of bacteria residing in the vessel wall as a cause of serious complications. The former University of Arizona study was subsequently extended to include 3 years of operative cultures involving 172 patients? The overall incidence of positive arterial wall cultures in this larger series was 44% (75 / 172). That portion of the arterial wall more prone to organism growth appeared to be the intima. However, a limited number of samples prevented any firm conclusions regarding this matter. There were six infections (3.5%) among the 172 patients during an 18.7 month follow-up. All six infections occurred in patients having prior arterial wall cultures that were positive. Subsequent graft infections affected 8% (6/75) of patients with positive cultures, compared to no infections (0/97) in those without positive cultures. This difference proved significant (p < 0.025). In this extended experience, there were 132 primary operations and 40 secondary procedures. Positive cultures were noted in 43% and 45% of primary and secondary procedures, respectively. S. epidermidis was the most common organism encountered, affecting 60% and 53% of primary and secondary operations. Among primary operations, positive arterial cultures were not predictive of subsequent graft infection (1/57) when compared to negative culture (0/75). However, in secondary procedures positive arterial wall cultures were predictive (2o < 0.05) of later graft infection (5/18) when compared to negative cultures (0/22). The Arizona group suggested that routine cultures need only be obtained during secondary operative procedures? Concern for the potential of subsequent graft infection in patients with positive cultures led them to recommend long-term administration of antibiotics. In their experience with this subgroup of patients, the incidence of eventual graft infection was 20% in those taking long-term antibiotics versus 38% in those not receiving long-term antibiotics. A
628
)%urnaiof VASCULAR SURGERY
Wakefield et al.
review o f recent reports revealed the risk o f graft infection associated with a positive arterial wall culture to be 10.5% compared to a 1.3% risk with a negative culture, thus attesting to the potential importance o f this finding. 3 S. aureus was previously considered the most common pathogen in vascular graft infection, 6a3 although recent emergence ofS. epidermidis as the dominant pathogen deserves note.14 This latter organism is by far the most frequently encountered o f the coagulase-negative staphylococci species. The basis for the more common recognition o f S. epidermidis organisms may be a realization that it must be recovered by culturing the graft or artery in broth for an extended period o f time and the need to free the bacteria from their glycocalyx slime in which they live. is Tissue grinding, as performed in the present investigation, and ultrasound sonication have been found equal in their ability to maximize recovery o f this organism26 Despite the recently reported high incidence o f positive cultures from either the arterial wall or adjacent adipose tissue, clinical graft infections are not common. Differences between our results and those from previous investigations including the large number o f positive adipose tissue cultures in our series may relate to differences in biopsy sites, culture technique, specimen size, or the definition o f infection. The fact that only two patients had positive blood samples suggests that the high frequency o f positive cultures does not relate to hematologic crossseeding o f the samples at the time they were obtained. Possible explanations for the lad< o f clinical graft infection include the low virulence o f coagulasenegative staphylococci compared to other bacteria, the low concentration o f bactcria present, the use o f prophylactic antibiotics in all but one patient studied, and the potential that these low-virulent organisms may cause covert infections that only become manifest much later.~4 Nevertheless, the presence o f artery and periartcrial tissue bacteria and serious consequences o f overt graft infection have caused appropriate clinical concern. In this regard, our study indicates that acute infectious complications are not a frequent early accompaniment o f elective vascular reconstructive procedures associated with positive cultures o f either the arterial wall or periarterial adipose tissue. However, final conclusions regarding the im-
portance o f bacterial presence in these circumstanccs must await longer follow-up. The authors acknowledge the technical assistance of Cathy A. PFaus for manuscript preparation. REFERENCES
1. SzilagyiDE, Smith RE, Elliott JP, VrandecicMD. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972;176:321-33. 2. Macbeth GA, Rubin JR, McIntyre KE, GoldstoneI, Malone IM. The relevaa~ceof arterial wall microbiologyto the treatment of prosthetic graft infections: graft infectionvs. arterial infection. J VAse SURG1984;1:750-6. 3. Durham JR, MaloneJM, BernhardVM. The impact of multiple operations on the importance of arterial wall cultures. J VAse SURG1987;5:160-9. 4. Lennette EH, BalowsA, Hausler WJ Jr, Shadomg HI, eds. Manual of clinical microbiology,4th ed. Washington DC, American Societyfor Microbiology, 1985. 5. Fry WI, Lindenauer SM. Infection complicating the use of', plastic arterial implants. Arch Surg 1967;94:600-9. 6. LiekwegWG, GreenfieldLJ. Vascular prosthetic infections: collectedexperienceand results of treatment. Surgery 1977; 81:335-42. 7. ReillyLM, StoneyRJ, GoldstoneJ, EhrenfeldWK. Improved management of aortic graft infection: the influenceof operation sequence and staging. J VASeSUV,G 1987;5:421-31. 8. Ernst CB, Campbell HC, Daugherty ME, Sachatello CR, GriffinWO. Incidenceand significanceof intra-operativebacterial culturesduring abdominalaortic aneurysmectomy.Ann Surg 1977;185:626-33. 9. Williams RD, Fisher FW. Aneurysmcontents as a source of graft infection. Arch Surg 1977;112:415-6. 10. Buckels JAC, Fielding IWL, Black J, Ashton F, Slaney G. Significance of positive bacterial cultures from aortic aneurysm contents. Br J Surg 1985;72:440-2. 11. McAuleyCE, Steed DL, Webster MW. Bacterialpresencein aortic thrombus at electiveaneurysmresection: is it clinically significant? Am J Surg 1984;147:322-4. 12. ScobieK, McPhailN, BarberG, Elder R. Bacteriologicmonitoring in abdominalaortic surgery.Can J Surg 1979;22:36871. la. Bunt TJ. Synthetic vascular graft infections. I. Graft infections. Surgery 1983;93:733-46. 14. BandykDF, Berni GA, Thiele BL, Towne lB. Aortofemoral graft infection due to Staphylococcusepidermidis. Arch Surg 1984;I 19:102-8. 15. Tollefson DF, Bandyk DF, Kaebnick HW, Seabrook GR, Towne JB. Surfacebiofih'ndisruption: enhanced recoveryof microorganisms from vascular prostheses. Arch Surg 1987; 122:38-43. 16. BerganainiTM, BandykDF, Govostis D, Vetsch R, Townc lB. Identification of Staphylococcus epidermidis vascular graft infections: a comparisonof culture techniques. I VASCSURG 1989;9:665-70.
