Lower extremity bypass graft revision in diabetics.

Revision of lower extremity bypass graft stenoses identified by surveillance duplex scanning is frequently required in diabetic patients. The authors evaluated (1) the value of routine angiography before graft revision in diabetics, (2) factors that predict patients in whom angiography alters management, and (3) the incidence of recurrent stenosis and factors that might predict it. Forty-two infrainguinal primary vein bypasses undergoing primary revision were retrospectively studied. The initial graft stenosis was detected at a mean of 11.5 +/-3.6 months after the original bypass. Angiograms were obtained in 38 cases, revealing additional findings in 29 of 38 cases (76%), with a resultant alteration of the operative plan in 27 cases (71%). The most frequent additional angiographic finding was the identification or localization of a lesion in the inflow or outflow tracts (18 of 27 cases). Cases where the angiogram altered the management plan had a mean systolic velocity ratio across the stenosis (Vr) of 7.3 +/-6.1, versus a Vr of 4.8 +/-1.3 for cases where the angiogram did not alter the management plan (p<0.04). Duplex scanning identified 4 lesions that were not seen on angiography; 3 of 4 were confirmed as webs at surgery. Twenty of 42 grafts (48%) developed recurrent stenoses at a mean of 4.9 +/-3.8 months from initial revision. Restenosis occurred in 69% of female limbs as compared to 38% of male limbs (p=0.06). Recurrent stenosis was not a predictor of ultimate graft failure, unless left untreated. Four of 10 untreated grafts ultimately failed. A total of 9 of the 42 grafts eventually failed (21%), leading to 3 amputations (7%). The authors conclude that failing infrainguinal bypass grafts identified by duplex in diabetics should undergo a detailed angiographic evaluation. This frequently leads to an alteration in the management plan, especially in the presence of a high Vr across stenoses. High rates of limb salvage (93%) and assisted primary graft patency (79%) despite a high recurrent stenoses rate (48%) justify routine duplex surveillance, preoperative angiography, and aggressive graft revision in diabetic patients with infrainguinal grafts.

Ureteral injury during aortic aneurysm repair by the retroperitoneal approach.

Three ureteral injuries (two proximal, one middle) associated with retroperitoneal repair of aortic abdominal aneurysms are reported. The authors believe these represent traction injuries that are related to the use of stationary retractors and suggest that complete anterior mobilization of the left kidney from its posterior fossa will decrease the odds of such an injury.

Changing patterns of access site complications with the use of percutaneous closure devices.

The authors report their experience with 15 cases of groin complications associated with the use of percutaneous closure devices following femoral arterial catheterization over a 2-year period. The complication rate was 1.7% for catheterizations in which a closure device was used. The 15 cases included 7 uncomplicated pseudoaneurysms (PSA), 3 infected pseudoaneurysms, 4 nonarterial groin infections (infected hematomas and/or abscesses), and 1 case of femoral artery occlusion. These complications presented at an average of 5 +/- 4 days postcatheterization. One patient with an infected PSA required a below-the-knee amputation. During the same time interval, there were no infectious complications in patients not receiving closure devices. We conclude that groin complications associated with such devices tend to present late and include a higher percentage of infections as opposed to complications occurring in patients not receiving closure devices. An aggressive surgical approach to these problems appears warranted.

Noninvasive localization of infrainguinal arterial occlusive disease in diabetics.

A retrospective review of 101 diabetics without aortoiliac disease was carried out to analyze the ability of various noninvasive tests to predict the level of significant (>50% stenosis) infrainguinal arterial disease. Patients were studied with anklebrachial indices (ABI), toebrachial indices (TBI), segmental pulse volume recordings (PVR), segmental pressures (SEGP), segmental Doppler waveforms (DWF), and arteriography. Results were classified as normal, disease at the femoropopliteal level, infrapopliteal level, or both levels (multilevel), or noninterpretable. Our findings for the entire study showed that, as a single test, DWF appears to have the best angiographic correlation, although the summed diagnosis of combined DWF and PVR data is superior in distinguishing multilevel disease from isolated tibial disease. SEGP are of very limited use in diabetics, even in multimodality testing, and we no longer include that test in our routine evaluation of diabetics.

The cysteine-free fibroblast growth factor 1 mutant induces heparin-independent proliferation of endothelial cells and smooth muscle cells.

BACKGROUND: The structure/function relationships of fibroblast growth factor 1 (FGF-1) are being investigated using site mutation, yielding novel structures with potential clinical applicability for modulating tissue responses to vascular interventions. We generated a mutant FGF-1 in which all three cysteines were converted to serines and then tested the relative mitogenic activities on endothelial cells (ECs) and smooth muscle cells (SMCs) and the molecular stability of the protein to thrombin-induced degradation. METHODS: The dose responses of wild-type FGF-1 and the Cys-free mutant in the absence or presence of heparin were tested on ECs and SMCs. Cell proliferation was measured by [(3)H]thymidine incorporation. Data were normalized by positive control (20% fetal bovine serum) and expressed as percentage of positive control for comparison. The molecular stability was examined by exposure of the cytokines to thrombin at 37 degrees C for 0.5-24 h and then analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. RESULTS: Unlike wild-type FGF-1 which induced only minimal DNA synthesis at concentrations as high as 100 ng/ml, the Cys-free mutant induced a dose-dependent proliferation starting at 1 ng/ml on both ECs and SMCs in the absence of heparin. At 100 ng/ml, Cys-free mutant induced 4-fold more proliferation than wild-type FGF-1 on ECs (76.64 +/- 13.39% vs 14.58 +/- 1.38%, P < 0.01) and 12-fold more proliferation on SMCs (143.52 +/- 9.96 vs 11.25 +/- 3.32, P < 0.01). Heparin 5 U/ml potentiated the mitogenic activity of the Cys-free mutant at low dose range. Both proteins were degraded by thrombin progressively. But the Cys-free mutant showed more susceptibility with accelerated appearance of lower-molecular-weight fragment bands after incubation with thrombin. CONCLUSIONS: Conversion of cysteine residues to serine changed the heparin dependency of the growth factor and increased its mitogenic activity and its susceptibility to thrombin-induced degradation.

Mitogenicity and release of vascular endothelial growth factor with and without heparin from fibrin glue.

PURPOSE: Fibrin glue (FG) has been used for local cytokine delivery on both vascular grafts and angioplasty sites. We measured the diffusive release of vascular endothelial growth factor (VEGF) and heparin from FG and the mitogenic activity of VEGF with and without heparin in FG on canine endothelial cells (ECs) and smooth muscle cells (SMCs). METHODS: Release of VEGF labeled with iodine 125 and tritiated heparin from FG into the overlying media was serially measured over 96 hours, and the data are reported as the mean percent released +/- SD. Proliferation assays measuring tritiated thymidine incorporation were performed for ECs and SMCs plated in media with 10% serum on FG containing various concentrations of VEGF and heparin. Media was placed on the FG for 24 hours and removed before plating cells to minimize the effect of the released, soluble VEGF and heparin. RESULTS: At 24 hours, 54% +/- 1% and 58% +/- 1% of the radioactive VEGF and heparin were released, respectively, with minimal release thereafter (58% +/- 1% and 66% +/- 1% at 96 hours). The ECs, SMCs, or media only (no cells) was plated on FG containing radioactive VEGF in an immediate or 24-hour delayed fashion for 72 hours to determine the percent release of VEGF into the media with the two different methods of plating. Cell type and the presence or absence of cells did not affect VEGF release, but there was three times more VEGF in the media for the immediate versus delayed plating (P <.001). Without heparin, VEGF at 100 ng/mL or more in the FG was needed to induce EC proliferation. Heparin at 5 U/mL enhanced EC proliferation at the VEGF dose of 100 ng/mL as compared wtih no heparin (P <.001), but not at the VEGF dose of 1000 ng/mL, which likely represents a maximal response. With heparin at 500 U/mL, the ECs died. In contrast, VEGF, in the presence or absence of heparin, did not affect SMC proliferation. CONCLUSIONS: We conclude that FG with VEGF at 1000 ng/mL and heparin at 5 U/mL is the optimal concentration for in vivo use because this may encourage EC, but not SMC, proliferation. The VEGF at 1000 ng/mL should leave mitogenic concentrations of VEGF intact after the initial, diffusive loss, and the addition of heparin at 5 U/mL may enhance VEGF mitogenic activity.

The S130K fibroblast growth factor-1 mutant induces heparin-independent proliferation and is resistant to thrombin degradation in fibrin glue.

OBJECTIVE: Site-directed mutagenesis is an important technique that can alter cytokine function, thereby eliciting desired responses. S130K is a mutation of fibroblast growth factor-1 (FGF-1), with lysine replacing serine in the heparin-binding site. We measured molecular stability and mitogenic activity of FGF-1 and S130K, both in the media and when suspended in fibrin glue (FG), on smooth muscle cells (SMCs) and endothelial cells (ECs) to determine if the mutation altered the function and potential clinical applicability. METHODS: EC and SMC proliferation of soluble FGF-1 or S130K at 0, 0. 1, 1, 10, or 100 ng/mL with heparin at 0, 5, 50, or 500 units (U)/mL was measured on growth-arrested cells in serum-free media. EC and SMC proliferation assays with cells on FG containing either FGF-1 or S130K at 0, 1, 10, 100, or 1000 ng/mL in combination with heparin at 0, 5, 50 or 500 U/mL were also performed during the exponential growth phase. Molecular degradation by thrombin was measured by sodium dodecylsulfate-polyacrylamide gel electrophoresis. RESULTS: S130K induces greater EC and SMC proliferation in the absence of heparin than FGF-1 does (P <.0001 for both the 10 and 100 ng/mL doses). S130K is also significantly more potent than FGF-1 in the presence of heparin. Heparin in the media enhances cytokine-induced SMC and EC proliferation at doses of 5 U/mL, but inhibits SMC proliferation at concentrations of 500 U/mL. For the FG data, unlike FGF-1, S130K induces EC and SMC proliferation in the absence of heparin. The addition of 5 U/mL of heparin enhances the proliferation induced by S130K. For ECs, as the heparin dose increases to 50 U/mL, proliferation decreases, as compared with the 5 U/mL concentration when either FGF-1 or S130K in the FG was compared at concentrations of 10, 100, and 1000 ng/mL (P <.01). S130K is more potent in FG than is FGF-1 both with and without heparin and exhibits maximal EC and SMC proliferation at 10 ng/mL, whereas FGF-1 activity is maximal at 100 ng/mL. Gel electrophoresis demonstrated that S130K was relatively more resistant to thrombin degradation than FGF-1. CONCLUSIONS: Site-directed mutagenesis changed the potency and the heparin dependency on cellular proliferation of FGF-1 in vitro. These techniques should allow the delivery of mutant growth factors to areas of vascular intervention to induce specific, desired responses. We believe that these studies will enhance our knowledge of the function of various regions of the FGF-1 molecule, allowing us to more precisely design increasingly more useful FGF-1 mutants.

Modulation of vascular cell growth kinetics by local cytokine delivery from fibrin glue suspensions.

PURPOSE: Fibrin glue (FG) has been used as a delivery system for bioactive agents on grafts and angioplasty sites. Reports from two different institutions suggest that heparin concentrations of 500 U/mL in FG inhibit smooth muscle cell (SMC) proliferation, but do not effect endothelial cell (EC) proliferation. The purposes of this study were to (1) quantify the diffusive release of fibroblast growth factor-1 (FGF-1) and heparin from FG; (2) determine the effect of heparin and FGF-1 on SMC proliferation when the cells are immediately plated on the FG; and (3) by means of the diffusive release data, design a new in vitro model that may differentiate the effect of FG-incorporated FGF-1 and heparin, rather than the released, solubilized components of these two factors, on SMC and EC proliferation. METHODS: 125I-FGF-1 or 3H-heparin release from FG into the overlying media was measured serially in a 96-hour period, either with or without cells. SMCs were immediately plated on FG containing various concentrations of FGF-1 and heparin. SMCs or ECs were plated on identical groups of FG containing FGF-1 and heparin 24 hours after the FG was made to exclude the effect on cell growth of the initial release of FGF-1 into the media. RESULTS: In the first 24 hours, 70% +/- 1% of the FGF-1 and 59% +/- 2% of the heparin in the FG was released into the overlying media, with minimal release occurring thereafter. The cell type or absence of cells did not affect release, but there was five times more FGF-1 and four times more heparin in the media at 72 hours for the immediate plating versus the delayed plating because of a diffusive release primarily in the first 24 hours. A heparin concentration of 500 U/mL inhibited SMC proliferation, as compared with 5 U/mL heparin, only when immediate plating of SMCs was used. Comparing immediate versus delayed SMC plating, at equivalent FGF-1 and heparin doses, immediate plating induced greater proliferation than delayed plating; this was likely caused by the higher soluble FGF-1 concentration. Heparin doses as high as 500 U/mL had little effect on SMC proliferation. In contrast, ECs died with delayed plating on FG containing 500 U/mL heparin, and their growth was inhibited at 50 U/mL heparin, as compared with 5 U/mL heparin. CONCLUSION: The differences in SMC proliferation when comparing immediate versus delayed plating are likely caused by diffusive release of heparin and FGF-1 into the media. Our ongoing work uses an optimized in vitro FG system that minimizes the effects of soluble factors. This is an important distinction, because the cytokines that are released in vivo will be removed by blood flow and, thus, may not exert an effect unless they are contained within the FG.

Fibrin sealant in vascular surgery: a review.

Fibrin sealant (FS) is a mixture of concentrated fibrinogen and thrombin that creates a fibrin matrix that is slowly degraded by the body's fibrinolytic system. FS is currently being used in the clinical arena for many applications. Perhaps the most relevant indication for vascular surgeons concentrates on FS's hemostatic properties. Current research in many centers is investigating FS's capability to incorporate drugs and cytokines into the fibrin matrix for slow release as a drug delivery system for future clinical use. This review will focus on three main uses of FS: as an anastomotic sealant, as an antibiotic coating, and as an agent for endothelialization of grafts.

Treatment of venous malformations by direct injection with ethanol.

PURPOSE: Venous malformations (VMs) may be discrete or extensive, and larger lesions may be difficult to remove with surgery. Incompletely removed lesions have a tendency to recur. We report our experience with ethanol ablation of VMs. METHODS: All 12 patients (seven women; mean age, 37 years) were evaluated with magnetic resonance imaging before treatment. A total of 19 prior surgical excisions had been performed for seven of the patients. Symptoms were present in all 12 patients and included bleeding, pain, swelling, and limitation of exercise. The VMs were present in the lower extremities of seven patients, in the upper extremities of three patients, and in the flank and buttocks in two patients. RESULTS: The 12 patients have undergone 30 injection procedures, with six patients requiring one, three patients requiring two, two patients requiring three, and one patient having undergone 12 treatments. General anesthesia was used in 11 patients. Blood loss was minimal for all procedures, and 28 of the 30 procedures were technically successful. Skin ulceration was seen in approximately half of the treated VMs, all of which healed with local wound care. Focal VMs were injected in six patients and resolved with a single treatment in five patients. Patients were free of symptoms at a mean follow-up of 10 months. Extensive VMs were injected for discrete, symptomatic areas in five patients. These lesions all regressed and were asymptomatic at a mean follow-up of 23 months in all but one patient. However, these lesions required multiple treatments as additional areas became problematic. CONCLUSIONS: Ethanol sclerosis is a well-tolerated, safe, and effective adjunct to the management of VMs. Advantages of ethanol injection include the ability to treat a very localized area without an incision. Conversely, extensive lesions may be palliated as symptoms occur.

Elevations of tissue-type plasminogen activator and differential expression of urokinase-type plasminogen activator in diseased aorta.

PURPOSE: Elevations of plasmin have been implicated in the pathogenesis of abdominal aortic aneurysms (AAA) because of its ability to digest extracellular matrix proteins. Plasminogen activators regulate the conversion of plasminogen to plasmin. Tissue-type plasminogen activator (tPA) is more important in modulation of fibrinolysis, and urokinase-type plasminogen activator (uPA) is predominant in tissue remodeling. The purpose of this study was to determine the levels of plasminogen activators in diseased aorta because they may be responsible for the increased plasmin levels previously described in AAA. METHODS: Levels of tPA and uPA in AAA, occlusive, and normal (organ donor) aorta were studied in tissue explant supernatants. Supernatant tPA and uPA levels were measured with an enzyme-linked immunosorbent assay. Northern analysis was used to quantitate uPA messenger RNA (mRNA) levels in aortic tissue. RESULTS: Levels of tPA in the supernatants were similar in occlusive (20 +/- 4 ng/ml) and AAA (23 +/- 8) aorta, but threefold higher than in normal aorta (7 +/- 5; p < 0.005 for normal vs occlusive and p < 0.001 for normal vs AAA). In contrast, uPA supernatant levels were differentially expressed, with the highest level existing in AAA (9.7 +/- 2.7 ng/ml), followed by occlusive (4.9 +/- 3.5), and the lowest levels in normal aorta (1.2 +/- 0.7; p < 0.05 for normal vs occlusive, p < 0.001 for normal vs AAA, and p < 0.005 for occlusive vs AAA). Inhibition of protein or RNA synthesis by addition of cyclohexamide or actinomycin D, respectively, revealed no significant difference between treated and control supernatants, suggesting that the increases were caused by protein release rather than active synthesis. Levels of uPA mRNA followed the same trend as the supernatant uPA levels (AAA 1.07 +/- 0.54, occlusive 0.54 +/- 0.08, and normal aorta 0.01 +/- 0.01). CONCLUSIONS: Levels of tPA were similar in aneurysmal and occlusive aorta, but exhibited a threefold increase over normal aorta, suggesting that the elevations of tPA are associated with the arteriosclerosis present in both aneurysmal and occlusive disease. Differences in uPA levels were significant between all three groups, with the highest levels in AAA and the lowest levels in normal specimens. Northern analysis of uPA mRNA followed the same trend, suggesting that the increase in uPA may be regulated at the level of transcription. As uPA plays an important role in tissue remodeling, our findings may also reflect the relative tissue repair activities in these three types of specimens and may explain the previously reported increased levels of plasmin seen in AAA.

Plasminogen activator levels are influenced by location and varicosity in greater saphenous vein.

PURPOSE: The plasminogen system, which includes tissue type plasminogen activator (tPA), urokinase type plasminogen activator (uPA), and their main inhibitor, plasminogen activator inhibitor type 1 (PAI-1), plays a major role in both fibrinolysis and tissue remodeling. This study compares the levels of tPA, uPA, and PAI-1 at the groin and ankle in normal and varicose greater saphenous vein (GSV). METHODS: GSV was collected from patients undergoing varicose vein (VV) removal and from normal vein (NV) from arterial bypass procedures. Portions of the GSV at the groin and the ankle were minced and placed in serum-free media for 48 hours. Assays of the supernatants were obtained for tPA, uPA, and PAI-1 protein by enzyme-linked immunosorbent assay. Cyclohexamide and actinomycin D were also added to the media of the VV tissue explant supernatants to inhibit protein and RNA synthesis, respectively. RESULTS: Levels of tPA were significantly higher at the groin (11 +/- 2) than the ankle (5 +/- 1) in the VV (p < 0.005), and this trend was also seen in the NV (groin 10 +/- 2 and ankle 7 +/- 3). Levels of uPA were significantly higher in the groin VV (14 +/- 4.3) than in NV (3.0 +/- 0.8, p < 0.05). This difference, although not statistically significant, applied to the ankle as well (VV 14.5 +/- 6.3 and NV 5.3 +/- 2.7). No significant difference was seen between NV and VV for PAI-1 (NV, groin 155 +/- 73 and ankle 113 +/- 53, VV, groin 161 +/- 20 and ankle 142 +/- 38) or tPA. Inhibitor studies revealed no significant difference among control, cyclohexamide, and actinomycin D supernatants for tPA, suggesting release of protein rather than active synthesis. In contrast, inhibitor supernatants were significantly lower for uPA and PAI-1 than control supernatants (p < 0.05), suggesting that uPA and PAI-1 were actively synthesized. CONCLUSIONS: In the tissue explant supernatant model uPA and PAI-1 are actively synthesized, but tPA is not. Levels of PAI-1 were comparable in all four groups. Levels of uPA in the varicose GSV were higher than in NV, suggesting a role for uPA in the pathologic makeup of VV. Levels of tPA were higher at the groin versus the ankle position, potentially explaining the previously described increased fibrinolytic activity seen at the groin.

Elevated levels of plasminogen-activator inhibitor type 1 in atherosclerotic aorta.

PURPOSE: Plasminogen activator inhibitor type I (PAI-1) inhibits the plasminogen activators that convert plasminogen to plasmin. In addition to initiating fibrinolysis, plasmin activates tissue matrix metalloproteinases, which cause degradation of the extracellular matrix (ECM) in the arterial wall. Elevated levels of PAI-1 ultimately decrease plasmin formation and may lead to an accumulation of ECM and arteriosclerosis. METHODS: PAI-1 was studied by four methods in atherosclerotic (aneurysmal and occlusive) and normal (organ donor) aorta: (1) PAI-1 secretion by tissue explant supernatants, including time course and inhibition studies; (2) tissue PAI-1 by protein extraction; (3) PAI-1 mRNA was quantitated by Northern analysis using glyceraldehyde-3-phosphate dehydrogenase to normalize for RNA loading; and (4) in situ hybridization was used to localize the cells that produced PAI-1 mRNA. RESULTS: Supernatant PAI-1 levels at 48 hours were 776 +/- 352, ng/ml in 11 atherosclerotic aortas and 248 +/- 98 ng/ml in 8 normal aortas (p < 0.005). Tissue PAI-1 levels per 100 mg of tissue were 99 +/- 58 ng in 11 atherosclerotic aortas and 38 +/- 20 ng in 5 normal aortas (p < 0.05). PAI-1 mRNA levels by Northern analysis were 0.91 +/- 0.49 in seven atherosclerotic aortas and 0.44 +/- 0.27 in five normal aortas. Supernatant time-course experiments revealed that PAI-1 increased over time. Inhibitor studies revealed that PAI-1 decreased to approximately one third of control values when cycloheximide or actinomycin D were added to the media, indicating that active synthesis of PAI-1 had occurred. In-situ hybridization localized PAI-1 mRNA predominately to endothelial cells and a few scattered vascular smooth muscle and inflammatory cells. Subgroup analysis revealed no statistically significant differences between aneurysmal and occlusive PAI-1 levels in any of the experiments. CONCLUSION: PAI-1 secretion, as measured by tissue explant supernatants, and total tissue PAI-1 in the protein extracts were significantly increased in atherosclerotic aorta. This elevation was also observed in the mRNA, which suggests that the increase is controlled at the level of transcription. PAI-1 mRNA was localized to endothelial, vascular smooth muscle, and inflammatory cells. We conclude that elevated levels of PAI-1 exist in diseased aorta. These elevated levels may lead to an accumulation of ECM, thereby contributing to the arteriosclerosis found in aortic occlusive and aneurysmal disease.

Endothelial cell function: biologic and physiologic functions in health and disease.

Endothelial cells (EC) form a monolayer with a strategic role in the control of many physiologic and biologic pathways. Although the endothelium initially was thought to be an inert and passive barrier of the vasculature, the diverse functions of EC have been better defined in the past decade. EC play an important role in the control of coagulation, vasomotor tone, growth of vascular smooth muscle cells, white cell trafficking, ischemic and reperfusion injuries, and the development of arteriosclerosis. Owing to their location in the arterial tree, EC also are considered a potential target for gene therapy. The introduction of foreign DNA into EC by in vitro transfection with viral vectors has produced encouraging results when using genes that encode for proteins such as insulin and urokinase. The seeding of EC onto vascular grafts appears to convey the thromboresistant properties of saphenous vein to an artificial surface. This review focuses on the important biologic and physiologic functions of EC in health and disease.

Surgical management of atheroembolization.

PURPOSE: Atheroembolization may cause limb loss or organ failure. Surgical outcome data are limited. We report the largest series of atheroembolization focusing on patterns of disease, surgical treatment and outcome. METHODS: One hundred patients (70 men), mean age 62 +/- 11 years, operated on for lower extremity, visceral, or nonthoracic outlet upper extremity atheroemboli were identified prospectively and monitored over a 12-year period. The atheroembolic source was localized by use of a combination of computed tomography scanning (n = 55), arteriography (n = 93), duplex scanning (n = 25), transesophageal echocardiography (n = 6), and magnetic resonance imaging (n = 4). Occlusive aortoiliac disease (47 patients) and small aortic aneurysms (20 patients; mean aneurysm size 3.5 +/- 0.8 cm) were the most common source of atheroemboli. Imaging studies revealed 12 patients with extensive suprarenal aortic thrombus. Correction of the embolic source was achieved with aortic bypass (n = 52), aortoiliac endarterectomy and patch (n = 11), femoral or popliteal endarterectomy and patch (n = 11), infrainguinal bypass (n = 3), extraanatomic reconstruction (n = 6), graft revision (n = 3), upper extremity bypass (n = 11), or upper extremity endarterectomy and patch (n = 3). RESULTS: All four deaths within 30 days and all seven deaths within the first 6 months after operation were among the 12 patients with suprarenal aortic thrombus. The cumulative survival probabilities for all patients at 1, 3, and 5 years were 89%, 83%, and 73%, respectively. After operation, nine patients required major leg amputations and 10 required toe amputations. Renal atheroemboli led to hemodialysis in 10 patients. Recurrent embolic events occurred in five of 97 patients monitored for a mean of 32 months. All five recurrences occurred in the first 8 months after operation. Three patients with recurrent emboli had suprarenal aortic disease, one of whom had undergone axillofemorofemoral bypass. Four of 15 patients receiving postoperative warfarin anticoagulation had development of recurrent embolism. Only one patient not receiving postoperative warfarin had a recurrent event (p < 0.05 by Fisher exact test). CONCLUSION: The atheroembolic source is the aorta or iliac arteries in two thirds of patients who underwent operation. Computed tomography scanning of the aorta is a useful diagnostic technique. The source of the emboli can be eliminated surgically with low mortality or limb loss rates except when the suprarenal aorta is involved.

Upper extremity bypass grafting. A 15-year experience.

OBJECTIVE: To analyze the perioperative morbidity and mortality, long-term patient survival, and patency characteristics of arterial bypass related to upper extremity ischemia. DESIGN: This is a retrospective review of sequential patients undergoing upper extremity arterial bypass during a 15-year period at a single tertiary-care teaching hospital. Data are expressed in a 5-year life-table format and interpreted using log-rank analysis. PATIENTS: Seventy-four patients with upper extremity ischemia undergoing arterial bypass, which included 95 separate operations. MAIN OUTCOME MEASURES: Operative morbidity and mortality, life-table survival, life-table bypass graft patency, and limb salvage are reported. RESULTS: There was no operative mortality, and there was a single major amputation. Survival rate was 86% at 5 years, and overall patency rate was 61.2% at 5 years, with autogenous conduits superior at all sites compared with prosthesis (70.9% vs 37.7%). Secondary operation was inferior to primary bypass (66% vs 48%) and associated with higher morbidity (22% vs 5%). All far distal forearm prosthetic bypass grafts failed within 1 year. CONCLUSIONS: Primary upper extremity bypass with venous conduit is a safe, durable procedure, after which prolonged patient survival can be expected.

Endometrial tuberculosis acquired by a health care worker in a clinical laboratory.

I describe a case of endometrial tuberculosis acquired by a microbiologist while she was working in a clinical laboratory. Granulomatous endometritis was found, and Mycobacterium tuberculosis was isolated. Respiratory tract exposure from a faulty exhaust hood is the likely source of infection. Endometrial tuberculosis and laboratory-acquired infections are briefly discussed.

In vivo and in vitro MR imaging of renal tumors: histopathologic correlation and pulse sequence optimization.

Magnetic resonance (MR) imaging has given mixed results in the detection of renal masses. To identify the reasons for this and to determine the optimal pulse sequences for evaluating renal tumors, the authors imaged 12 primary renal tumors in vivo and 17 in vitro at 0.35 T. Histopathologic findings for each specimen were closely correlated with the MR images. Four of seven solid tumors imaged in vivo were isointense with surrounding normal renal parenchyma at all pulse sequences. The other three tumors were hyperintense in vivo at T2-weighted sequences. At heavily T2-weighted sequences eight solid tumors were hyperintense in vitro and four were hypointense. There was no correlation between signal intensity and specific tissue type or histologic pattern for solid tumors. The five cystic tumors were well seen both in vivo and in vitro on T2-weighted images. However, the signal intensity of the cyst fluid was an unreliable indicator of benignancy. SE MR imaging at 0.35 T has significant limitations in the detection of solid renal masses.

Intramural cyst of the stomach.

A 61-year-old woman had a benign intramural gastric cyst found incidentally at autopsy. This rare lesion may be of foregut origin.