Indium ill-labeled white blood cell scans after vascular prosthetic reconstruction

Transcription

1 Indium ill-labeled white blood cell scans after vascular prosthetic reconstruction Marc M. Sedwitz, M.D., Richard J. Davies, M.D., Harold T. Pretorius, M.D., Ph.D., and Tony E. Vasquez, M.D., San Diego, Calif. The clinical value of indium Ill-labeled white blood cell (WBC) scanning done after vascular graft procedures was investigated to differentiate noninfectious postoperative inflammation associated with graft incorporation from early prosthetic graft infection. Indium Ill-labeled WBC scans were initially obtained in 30 patients before discharge from the hospital and during the subsequent follow-up period (334 days). Fourteen of 30 patients (47%) had normal predischarge scans that included all 10 patients who had grafts confined to the abdomen and 4 of 20 patients (20%) who had grafts arising or terminating at the femoral arteries (p < 0.05). Sixteen of 30 patients (53%) discharged with abnormal initial indium 111 WBC scans underwent serial scanning until the scan normalized or a graft complication developed. All of the 16 patients had grafts involving the groin region. Abnormal indium 111 uptake in the femoral region continued for a mean 114 days without the development of prosthetic graft infections. The sensitivity of indium 11 I-labeled WBC scans for detecting wound complications was 100%, whereas the specificity was 50%. Thus, the accuracy of the test was only 53%. We conclude that (1) abnormal indium 111 WBC scans are common after graft operations involving the groin region but are unusual after vascular procedures confined to the abdomen, and (2) in the absence of clinical suspicion, the indium 11 I-labeled WBC scan does not reliably predict prosthetic graft infection because of the low specificity of the test in the early postoperative period. (J VAse SURG 1987;6: ) Prosthetic graft infection is one of the most serious and potentially catastrophic complications in vascular surgery. Morbidity and mortality rates are high, the latter ranging between 25% and 75%. 1"4 Early and accurate detection ofoccuk graft infection is an important determinant of successful treatment of this lethal complication. In recent years indium 111-labeled white blood cell (WBC) scanning has been established as a valuable adjunct in the diagnosis of occult graft infection. The accuracy of indium 1 ll-labeled WBC scanning in patients suspected of having occult graft infection is 75% to 94%. s'# Indium ill-labeled leukocytes localize in noninfectious as well as infectious areas of inflammation. 7 Whether localized indium 111 uptake in the early postoperative period is a manifestation From the Vascular Surgery Service (Dr. Sedwitz), the Oncology Surgery Service (Dr. Davies), and the Department of Nuclear Medicine (Drs. Pretorius and Vasquez), Veterans Administration Medical Center, San Diego, and the University of California, San Diego School of Medicine. Presented at the Second Annual Meeting of the Western Vascular Society, Tucson, Ariz., Jan , Reprint requests: Dr. Marc M. Sedwitz, Vascular Surgery Service (112), VA Medical Center, 3350 La JoUa Village Dr., San Diego, CA of the postoperative inflammation associated with graft incorporation or a harbinger of subsequent graft infection has not been clarified. False positive indium 11 l-labeled WBC scans have been reported to be more common when the interval between operation and scan was short. 6 The purpose of the present prospective study was to determine the efficacy of the early postoperative indium 111 WBC scanning to better characterize the inflammatory process of normal graft incorporation and its relevance to accurate interpretation of the diagnostic test. PATIENTS AND METHODS The study population consisted of 30 male patients with a mean age of 72 years (range 55 to 79 years) who underwent vascular reconstructive surgery at the San Diego Veterans Administration Medical Center between September 1984 and December All patients voluntarily agreed to participate in the study and signed an informed consent. The study was approved by the Medical Center's Human Subjects Committee. All patients underwent preoperative noninvasive Doppler examination and arteriography. Aortic aneurysms were also evaluated by abdominal CT scan or ultrasound examination. All patients with active systemic or localized infections

2 Volume 6 Number 5 November 1987 Indium Ill-labeled white blood cell scanning 477 POSTOPERATIVE DISCHARGE INDIUM SCAN ABNORMAL I NORMAL REPEAT SCAN CLINICAL (30-90 DAY) FOLLOW-UP I NORMAL CLINICAL GRAFT INFECTION DEVELOPS SUSPECTED INFECT ION REPEAT INDIUM SCAN Fig. 1. Scanning algorithm after vascular reconstructive surge~. were excluded from the study. The operation was the first vascular operation for 29 of the 30 patients. The indications for operation were aortoiliac occlusive disease in 18 patients, abdominal aortic aneurysm in 10 patients, and occlusive disease of the superficial femoral artery in two patients. Each patient in the study had an indium 111- labeled WBC scan performed before discharge from the hospital (within 14 days after operation) and subsequently at 30- to 90-day intervals if their initial predischarge indium 11 i-labeled WBC scan was abnormal. All patients were evaluated closely and underwent serial indium ill-labeled WBC scanning until either the test normalized or clinically apparent graft infection developed (Fig. 1). Patients with normal postoperative scans were evaluated in the clinic and any such patient who required readmission to the hospital underwent repeat scanning routinely. Three patients underwent preoperative indium ill-labeled WBC scanning immediately after angiography. Knitted Dacron grafts (Bard USCI Vasculour II) were used in 27 surgical procedures. One femorofemoral bypass was performed with a ringed polytetrafluoroethylene (PTFE) graft (Impra). Reversed autogenous saphenous veins were used in two femoropopliteal bypasses. All surgical procedures included systemic heparinization before clamping and heparin reversal was not used. All patients received standard perioperative intravenous prophylactic antibiotics, which were continued until all invasive monitoring lines were removed. All patients received a minimum of three postoperative doses of antibiotics. Indium 111-labeled WBC scanning was performed with the approval of the Medical Center's Human Subjects Committee. The leukocyte labeling technique was a modification of the original description by Thakur et al.8 Sixty milliliters of the patient's whole venous blood was collected and added to 10 ml of sodium heparin and hydroxyethyl starch with adequate mixing. Red blood cells were allowed to sediment and the leukocyte-rich supernatant was treated with centrifugation at 150 g for 10 minutes. The leukocyte-poor plasma was decanted and the leukocyte button was resuspended in 5 ml of leukocytepoor plasma. One millicurie of indium 111 was incubated with the leukocyte suspension for 40 minutes. During the incubation period the suspension was gently agitated. The suspension was then centrifuged at 150 g for 10 minutes. The supernatant was removed and saved for evaluation of labeling efficiency. The WBC pellet was resuspended in 5 ml leukocyte-poor plasma and was reinjected intravenously into the patient. The activity of the final product was limited to 0.5 mci. The maximum time from blood drawing to reinjection was 3 hours. Eighteen to 24 hours later imaging was begun. A largefield gamma camera (General Electric 400) with a medium-energy collimator set for two photo peaks at 173 and 247 kev with 20% windows aided instrumentation. Areas of imaging included the thorax, abdomen, and lower extremities. All scans were performed and reviewed by two skilled radiologists experienced in the performance and interpretation of indium l 1 l-labeled WBC scans. Scans were interpreted as being either normal or abnormal. Abnormal indium 111 uptake was specifically sought in the region of graft placement and in synchronous regions that were unrelated to the graft itself. Scans were correlated with clinical status of the patient to corroborate the presence or absence of graft infection. Statistical analysis was performed with Fisher's

3 478 Sedwitz et al. Journal of VASCULAR SURGERY Fig. 2. Indium 111 uptake localized in region of superficial groin wound cellulitis. Fig. 3. Abnormal indium 111 uptake after needle catheterization of both groin regions after angiography. exact test. A p value of less than 0.05 was taken to represent a significant difference. RESULTS Fourteen of 30 patients (47%) had a normal initial postoperative indium ill-labeled WBC scan (group A). Those 14 patients included all 10 patients who had intra-abdominal prosthetic grafts placed because of aneurysmal disease. The remaining four patients included three patients with aortobifemoral grafts placed because of occlusive disease and one patient who underwent femoropopliteal bypass. Follow-up of group A patients (mean 305 days, range 35 to 504 days) revealed that 13 patients were alive without any postoperative evidence of graft infection or other complications. One patient with an initial normal indium 111-labeled WBC scan required readmission because of abdominal pain and leukocytosis 2 weeks after discharge. A subsequent indium 111-labeled WBC scan was abnormal with uptake in the distribution of the entire large bowel as a result of documented pseudomembranous colitis. There was no clinical suggestion of graft infection on the indium 111-klbeled WBC scan. A subsequent scan done at discharge from the hospital was normal after appropriate antibiotic treatment. Sixteen patients (53%) were discharged with abnormal indium 111-labeled WBC scans (group B). This group comprised 80% of the 20 patients with grafts arising or terminating at the femoral arteries with overlying groin incisions; these included 11 aor- tobifemoral bypass grafts (one also had bilateral renal artery thromboendarterectomy), two femorofemoral bypasses, one femoropopliteal bypass, one thoracoabdominal aorta-bifemoral bypass graft, and one axillobifemoral bypass graft. All 16 group B patients had uptake in the groin regions. Twelve patients (75%) had uptake in both groins. Two patients (12%) had groin wound complications (one groin hematoma and one superficial unilateral groin cellulitis), after aortobifemoral grafting and both had abnormal scans with uptake in these areas (Fig. 2). Four of the 20 patients (20%) with groin incisions had normal initial indium 111-labeled WBC scans. There were no postoperative wound problems or other complications among these four patients. The sensitivity of the initial scan for acute wound complications was thus 100% (two of two patients), but the specificity was only 50% (14 of 28 patients). Thus, the accuracy of the test for clinically apparent operative site complications was only 53% (16 of 30 patients). Abnormal indium 111 uptake was common after prosthetic graft placement in the groin (16 of 20 patients) and was rare with grafts positioned exclusively in the abdomen (none of 10 patients) (p < 0.05). Uptake in the femoral region was bilateral in 12 patients and unilateral in four. Synchronous areas of abnormal uptake were also common in the gastrointestinal tract, bone, thorax, and soft tissue. These additional areas of uptake clearly did not reflect active infection; all 16 group B patients are

4 Volume 6 Number 5 November 1987 Indium 11 i-labeled white blood cell scanning I 4!-@""---'O 51 "O'--"O - g I- z 10 a. - -C - "C A w 2 A w A w g. Q Scans, -, =, 7,: DAYS Fig. 4. Duration of abnormal indium 111 uptake in all patients with abnormal scans after discharge. alive and none had postoperative graft infection during long-term follow-up (mean 359 days, range 140 to 485 days). Three patients underwent indium l 1 l-labeled WBC scanning after arteriography. In two cases focal uptake appeared in the area of the femoral catheterization and the scans remained abnormal in the early postoperative period (Fig. 3). The 16 group B patients with initial abnormal scans had repeat scans done at their convenience at approximately 60-day intervals (range 45 to 90 days) until the indium 111-labeled WBC scan normalized (Fig. 4). The scans remained abnormal for a mean of 114 days (range 33 to 360 days). All follow-up scans showed progressively decreasing uptake over time. One patient continued to have indium 111 uptake in both groins for 360 days. That patient has had no clinical or other laboratory evidence to suggest graft infection. A subsequent scan at 420 days was normal. Areas of synchronous uptake completely disappeared in all patients except for persistent uptake in bony areas in two patients. DISCUSSION The early diagnosis of prosthetic graft infection remains an ongoing challenge for vascular surgeons. The clinical manifestations of the potentially lethal complication can be so minimal as to thwart attempts at accurate diagnosis and to delay lifesaving treatment. This is particularly true of prosthetic grafts residing entirely within the abdomen. Table I. Indium-labeled leukocyte imaging: Reasons for increased uptake ComlTlOn Abscess Idiopathic Intravenous sites, noninfected Phlegmon Pneumonia Wound infection Hematoma Uncommon Accessory spleen Bowel infarction Cdlulitis Colonic fistulas with pericolonic abscess Crohn's disease Empyema Graft infection Injection site Ischemic colitis Osteomyelitis Septic arthritis Transplant rejection Wound, noninfected Rare Acute cholecystitis Idiopathic false obstruction Nonseptic pulmonary embolism Pancreatitis Most authorities believe that graft infections begin with contamination of the prosthetic material at the time of implantation. However, Malone et al. 9 reported that a Dacron graft may be vulnerable to transient bacteremia if neointimal healing is incomplete. Either mechanism may be important in both early and late-appearing graft infections. Reilly and Goldstone 3 reported that only 36% of aortic graft infections were recognized during the first year after implantation, whereas Szilagyi et al. m noted that 65% manifested within the first 30 days after operation. It would be most advantageous to have a test that would accurately diagnose prosthetic grafts at risk for late infection before the appearance of clinical signs and symptoms so that follow-up could be closer and aggressive therapy could be started earlier. Since the introduction in the 1970s by Thakur et al.,8 the indium 111-labeled WBC scan has been used to accurately detect and localize areas of infection. n Serota et al. x2 first reported the sensitivity and specificity of the technique for detecting occult prosthetic graft infections in an animal model. The isotope appeared in the region of the graft for up to 72 hours after graft placement. However, uptake disappeared by 1 week unless infection was present. McKeon, Miller, and Jamieson 13 reported the accuracy of the indium 111-labeled WBC scan in patients

5 480 Sedwitz eta!. Journal of VASCULAR SURGERY Table II. Established accuracy of the indium 11 i-labeled leukocyte scans for detection of prosthetic graft infection No. of Authors Year patients Sensitivity Specificity Accuracy Brunner et al % 85% 88% Lawrence et al % 87% -- Mark et al % Current study % ~ 50% 53% *Wound complications without infection. with aortic infection. Other investigators have retrospectively reviewed large series and documented an overall accuracy of 88% to 92%.s'6 Although the diagnostic efficacy of this technique for prosthetic graft infection is well established, enthusiasm has not been unanimous. Mark, McCarthy, and Moss 14 reported that the CT scan is a more sensitive tool than the indium 111-labeled WBC scan in evaluating the extent of aortic graft involvement with infection. Others have reported that the indium 1 ll-labeled WBC scan is too sensitive and nonspecific to serve as the basis for verifying graft infection? With increased experience with indium illlabeled WBC scanning, it has become clear that noninfectious inflammatory processes may also produce focal uptake of the isotope. McCaffee and Somin 7 reported the inaccuracies of indium Ill-labeled WBC scan interpretation and cautioned that the scan be used as an adjunctive tool in conjunction with other diagnostic methods because many lesions are capable of producing an abnormal leukocyte image (Table I). Our study indicates that indium 111 commonly localizes after vascular prosthetic procedures and does not predict either early or late graft infection. The groin was the most common area of uptake (53% of all cases) and both groins were almost always involved. Although the sensitivity of the examination in identifying postoperative graft complications was 100% (two of two patients), the specificity was a disappointing 50% (14 of 28 patients), with an overall accuracy of 53% (Table II). No vascular graft infections developed during the follow-up period. There was a total absence of indium 111 localizing to the aortic region in all prosthetic grafts that were placed. This was the case in all 10 patients who had an aneurysm resection with an intra-abdominal graft. There may be a fundamental difference between the groin and the retroperitoneum with respect to the propensity for indium 111 uptake in the early postoperative period and in the absence of infection. The prolonged period of indium 111 uptake in the groin region is consistent with the noninfectious inflammatory process of wound healing and graft incorporation. Groin uptake eventually resolved in all patients. There were no graft compfications of any type seen during subsequent follow-up periods, which ranged from 1 to 16 months (mean 11 months). We conclude that the indium 111-labeled WBC scan is a sensitive indicator of wound healing and graft incorporation in the femoral (groin) locations. Gooding, Effeney, and Goldstone is reported that hematomas documented by ultrasound are quite common after graft implantation and may take months to resolve. In a similar review, Qvarfordt et a1.16 evaluated CT scans after aortic graft placement to assess the period for graft incorporation. They found that graft incorporation and hematoma resolution occurred more rapidly in the retroperitoneum than in the groin region but was virtually complete by 48 days. Our study supports this finding and suggests that noninfectious inflammation may continue for months after operation. The resolution of noninfected hematomas and the graft incorporation process continues to be a common cause of false positive indium 111-labeled WBC scans. This suggests that the indium Ill-labeled WBC scan should be performed before any blood drawing or catheterization in the groin. Finally, the low specificity of indium ill-labeled WBC scanning after vascular graft procedures appears to limit the value of this examination as a routine technique in the early postoperative period unless the process of graft incorporation has been completed. REFERENCES 1. Liekweg WG, Greenfield L]. Vascular prosthetic graft infections: collected experiences and results of treatment. Surgery 1977;81: Fry WJ, Lindenauer SM. Infection complicating the use of plastic arterial implants. Arch Surg 1967;94: Reilly L, Goldstone ]. The infected aortic graft. In: Bergan

6 Volume 6 Number 5 November 1987 Indium 11 I-labeled white blood ceil scanning 481 JJ, Yao JT, eds. Reoperative arterial surgery. New York: Grune & Stratton, Inc, 1986: Hoffwer PW, Gensler S, Haimovici H. Infection complicating the use of arterial grafts. Arch Surg 1965;90: Lawrence PF, Dries DJ, Alazraki N, Albo D Jr. Indium 11 l-labeled leukocyte scanning for detection of prosthetic vascular graft infection. J VASC SURG 1985; 2: Brunner MC, Mitchell RS, Baldwin JC, et al. Prosthetic graft infection: limitations of indium white blood cell scanning. J VASC SURG 1986;3: McCaffee JG, Somin A. Indium 111 labeled leukocytes: a review of problems in image interpretation. Radiology 1985;155: Thakur ML, Segal AW, Louis L, Welch MJ, Hopkins J, Peters TJ. Indium-ill-labeled cellular blood components: mechanism of labelling and intracellular location in human neutrophils. J Nucl Med 1977;18: Malone JM, Moore WS, Campagna G, Bean B. Bacteremic infectability of vascular grafts: the influence of pseudoinfmal integrity and duration of graft function. Surgery, 1975;78: Szilagyi D, Smith R, Elliot J, Vrandecic MP. Infbction in arterial reconstruction with infected grafts, Ann Surg 1972; 176: McDougall IR, Baumert JE, Lantieri RL. Evaluation of'qn leukocyte whole body scanning. AJR 1979;133: Serota AI, Williams RA, Rose JG, Wilson SE. Uptake of radiolabeled leukocytes in prosthetic graft infection. Snrgen, 1981;90: McKeon PP, Miller DC, Jamieson SW. Diagnosis of arterial prosthetic graft infection bv indium 111 oxine WBC scan. Circulation 1982;66(Suppl 11): Mark AS, McCarthy SM, Moss AA. Detection of abdominal aortic graft infection: comparison of CT and In-labeled white blood cell scans. AJR 1985;144: Gooding GA, Effeney DJ, Goldstone J. The aortofemoral graft: detection and identification of healing complications by ultrasonography. Surgery 1981;89: Qvarfordt PG, Reillv LM, Mark AS, et al. Computerized tomographic assessment of graft incorporation after aortic reconstruction. Am J Surg 1985;150:227-3l.