Abstract
This article analyzes the complication rates of diagnostic arteriographies performed by a single vascular surgeon and compares them to those previously published by interventional radiologists. Five hundred fifty-eight consecutive patients who underwent diagnostic arteriographies were analyzed. A modification of one study's criteria was used to compile perioperative complications. The technical success rate was 99%. These included 345 aortoiliofemoral arteriograms with runoff, 64 aortoiliofemoral arteriograms for abdominal aortic aneurysms, 83 aortoiliofemoral arteriograms with contralateral selective iliacs, 35 aortoiliofemoral arteriograms with carotids, and 27 aortoiliofemoral arteriograms with selective visceral/renal. Femoral artery puncture was used in 93%, and left brachial artery in 7%. The mean amount of contrast was 97 cc and the mean operative time was 25 minutes. The overall complication rate was 3.8% (1.3% major), which was comparable to what was published previously (1.9% and 2.9%) but superior to what we published previously as performed by our radiologists (7%, p <.001). A logistic regression could not find any variables that were significant for the prediction of a major complication. However, increased age, a longer operating time (≥ 30 minutes), and smoking were associated with an increase in overall complications. It was determined that diagnostic arteriography can be done safely by experienced vascular surgeons with low complication rates that compare favorably with what was published by interventional radiologists.
Endovascular therapy is increasingly becoming more common in the field of vascular intervention. Although interventional radiologists still perform the majority of these percutaneous procedures, many vascular surgeons are beginning to become more involved.
The complication rates of diagnostic arteriography performed by radiologists have been well documented. 1–4 However, there has been only one report of these rates as performed by vascular surgeons. 5 Another study was reported by vascular surgeons analyzing predictors of complications after diagnostic and therapeutic endovascular procedures. 6 This study analyzed the results of 558 consecutive diagnostic arteriographies performed by a single vascular surgeon (A.F.A.) at Charleston Area Medical Center/West Virginia University, Charleston, West Virginia. This study was undertaken not only to study the complication rates but also to determine if certain patient comorbidities affect these rates. These results were compared with published complication rates, including a similar report from the same institution for diagnostic arteriographies performed by radiologists.
Patient Population and Methods
The medical records of 558 consecutive patients who underwent diagnostic arteriography performed by a board-certified vascular surgeon during a recent 4-year period in the Circulatory Dynamic Laboratory (a modern endovascular suite), using the GE LCA/DLX Digital Imaging System, were reviewed. The surgeon's background in endovascular therapy includes training in diagnostic arteriography during his vascular fellowship training in 1977 and 1978 and several courses in endovascular therapy, including several endovascular workshops. The operator also passed radiation safety monitoring protocol that was established at our institution, which requires reading material and passing an examination. Diagnostic arteriography was routinely done by the operator during his early career in vascular surgery, which became less frequent over the years owing to the extra load of vascular surgery cases; however, diagnostic arteriography and endovascular therapy became a major part of the practice during the last 10 years owing to the evolution of endovascular therapy.
The medical records of all 558 consecutive patients who underwent diagnostic arteriography were reviewed, including the demographic data. The records were examined for any indication of a local problem, specifically postprocedure bleeding, hematoma, infection, pseudoaneurysm formation, or thromboembolic events distal to the puncture site. Special attention was given to the indication for the arteriogram and the presence of risk factors (including hypertension, diabetes mellitus, renal disease, coronary artery disease, and congestive heart failure). The preoperative and postoperative blood urea nitrogen and serum creatinine levels were recorded. Other recorded factors included procedure approach (transfemoral, transbrachial, or graft puncture), the number of arteries punctured, the number and type of selectively injected arteries, the amount of contrast material used, and catheterization time. Only patients undergoing diagnostic procedures (excluding sole carotid arteriography) were included in this analysis.
Reporting Criteria for Complications
The following definitions were used for interpretation of the data. 2,3 Complications were defined as any unanticipated event deleterious to the health of the patient occurring within 24 hours of the initiation of the arteriographic procedure that, in the judgment of a trained clinical observer, resulted from the procedure.
Minor: A complication that does not significantly alter the health or activity of the patient, require extra hospitalization, or necessitate extraordinary therapy (pertains primarily to complaints by the patient), for example, a minor puncture-site hematoma (< 5 cm), a minor hypotensive episode (few minutes), skin rash, seeing spots for a few minutes after injections, or headaches of less than a few hours' duration.
Major: A complication that significantly alters the health or activity of the patient and requires extraordinary therapy or hospitalization. This includes all other complications, for example, local complications other than minor hematomas (major hematomas necessitating blood transfusions or surgical evacuation, arterial thrombosis, embolization, perforation or dissection, and wound infection), cardiac complications (angina, myocardial infarction), nervous system complications (transient ischemic attack, stroke, and brachial plexus and femoral nerve injury), renal failure, and major allergic reactions (shock and laryngeal spasm).
The following terms were used for this study:
Preoperative renal disease was defined as an elevation of the preoperative serum creatinine by at least 20% beyond the normal range in our laboratory, which is 0.5 to 1.2 mg/dL.
Postoperative azotemia (acute renal failure) was defined as an elevation of the serum creatinine by 1.0 mg/dL or more above the preoperative level associated with oliguria, requiring diuretic management and restriction of fluids or with hyposthenuria in the presence of normal urine volume (nonoliguric renal failure).
Statistical Methods
SAS, version 8.02 (SAS Institute, Cary, NC), was used for statistical analysis. Categorical variables, such as patients' demographics, clinical characteristics, risk factors, and the location of the procedure, were compared with perioperative outcomes using Fisher's exact test or χ 2 test, and Student's t-test was used for continuous response variables. Logistic regression was implemented to determine the risk factors associated with complications (major or minor). Complication rates were compared with what has been reported by others, including Hessel and colleagues, 1 Egglin and colleagues, 4 Balduf and colleagues, 5 and past series by the same lead author 3 using Fisher's exact test or χ 2 test. Statistical tests with a resultingp value < .05 were considered significant.
Results
A summary of various comorbidities in our present series is compared with previously published series in Table 1. The mean amount of contrast was 97 cc (range 6–350 cc), and the mean catheterization time (the time that the catheter was in place during diagnostic arteriography) was 25 minutes (range 4–119 minutes). The types of arteriography included 345 aortoiliofemoral (AIF) arteriograms with runoff; 64 AIF arteriograms were performed to define preoperative aneurysmal anatomy, 83 AIF arteriograms with contralateral selective iliac arteriograms, 35 AIF arteriograms with four-vessel arch aortograms/carotid arteries, and 27 AIF arteriograms with selective vessel/renal arteriograms for patients with renovascular hypertension and/or visceral ischemia, and four miscellaneous (trauma and arteriovenous malformations). The right femoral artery access was used in 68%, the left femoral artery in 25%, and the left brachial artery in 7%. Overall, 11 (2%) of these punctures were through femoral grafts. Five hundred sixty (91.3%) necessitated one arterial puncture versus 48 (8.7%) that necessitated two punctures.
Comorbidity Comparison
The technical success rate in this series was 99.3% (554 of 558). We were unable to gain access in four patients, and there was only one complication, a minor hematoma.
Overall Complications
Table 2 summarizes the overall complication rates. There were 14 minor complications (2.5%), which included eight minor bleedings (minor hematomas), one case of nausea and vomiting, two skin rashes, and three arrhythmias that did not require additional hospitalization. There were also seven major complications (1.3%), including one major bleeding that required blood transfusion, one case of acute renal failure, one patient who sustained myocardial infarction and stroke and eventual death, one patient who had local neurologic complications that lasted less than 10 days, which was felt to be femoral nerve injury, and one patient who sustained an acute thromboembolic event, underwent thromboembolectomy, and eventually expired.
Perioperative Complications
*This is the same patient.
Table 3 summarizes the correlation of risk factors and complications. A logistic regression could not find any variables that were significant for the prediction of a major complication. However, increased age, a longer operative time (≥ 30 minutes), and smoking were associated with an increase in overall complications, minor and major complications combined (Table 4). Smokers were 5.7 times more likely to have a minor or major complication. For every year increase in age, the patient was 1.06 times more likely to have a minor or major complication. For arteriograms lasting longer than 30 minutes, the patient was 3.5 times more likely to have a minor or major complication.
Perioperative Complications and Risk Factors
Logistic Regression Analysis of Risk Factors and Complications
Major Complications and Selective Catheterization
Table 5 summarizes the correlation between adding selective catheterization plus the AIF arteriogram and the incidence of major complications. As noted in this table, there was no association between selective catheterization plus the AIF arteriogram and any major complications.
Effect of Adding Selective Catheterization on Major Complications
AIF = aortoiliofemoral.
Comparison of Our Complications with Those of Other Series Published in the Recent Literature
As noted in Table 6, our overall major complication rate of 1.3% was comparable to what was reported by Hessel and colleagues, 1 Egglin and colleagues, 4 and Balduf and colleagues. 5 The overall major complication rate in our present series was statistically significantly superior to what was published in our past series in 1993, 3 which included complications performed by our radiologists at our institution during that time period (p < .001).
Comparison of the Incidence of Major Complications in Relevant Series
Discussion
Levin and colleagues reported that of more than 750,000 noncoronary therapeutic vascular procedures performed in 1996 on Medicare patients, 21% of these had a percutaneous approach, and interventional radiologists performed 75.5% of these procedures (mainly angioplasty), cardiologists performed 12.6%, and vascular surgeons performed only 6.3%. 7 They concluded that only a small minority of the diagnostic angiographies were done by vascular surgeons.
The overall complication rate in our 558 patient series that was performed by a board-certified vascular surgeon was 3.8%, and the mortality rate was 0.4%. Minor complications comprised 2.5%, whereas major complications accounted for 1.3%. The previously reported series from our same institution, performed by our radiologists (many were general radiologists), had an overall complication rate of 14.3%, 7.9% minor and 7.1% major. Identical criteria were used in both series to allow for easy comparison. There was a statistically significant difference between the major complication rates of these studies (p < .0001). The large discrepancy in the two series can be accounted for by several reasons. There has been an 11-year interim between the two studies. In that time, our facility has developed a dedicated circulatory dynamics laboratory, which performs only diagnostic and therapeutic endovascular procedures. This laboratory also has a dedicated and well-trained staff, as well as dedicated vascular specialists who are well trained in vascular diagnostic and endovascular therapeutic procedures. Balduf and colleagues reported that their worst complication of a cerebral air embolus occurred owing to the inexperience of the support staff with the injection system. 5 The technical advances of the equipment, as well as the development of better catheters, guidewires, etc, has surely assisted in lowering the complication rates also.
Our major complication rate of 1.3% is comparable to that of other reports in the radiology literature. Hessel and colleagues reported a major complication rate of 1.9%, 1 whereas Egglin and colleagues reported 2.9%. 4 Balduf and colleagues' recent study, with vascular surgeons performing the arteriography, reported a 2.1% major complication rate. 5 There was no statistical significance between these studies.
Recently, Young and colleagues reported on complications with outpatient angiography and interventional procedures; of 560 cases of aortofemoral arteriography, there were 124 complications (22%), which included 62 cases of hematoma and bruising (11%), local pain (8.6%), and others (contrast reactions, weakness in the legs, headache in 2.5%). 8 It should be noted that most of these complications were considered minor, except for some of the other complications. 8 In a later article by the same group, they reported an overall complication rate of 18.7% in 1,128 aortofemoral arteriograms: 9.7% hematoma and bruising, 7% local pain, and 2% others (allergy reactions, headaches, weakness in the legs). 9
Arterial thromboembolization during arteriography has received considerable attention. Fogarty and Krippaehane postulated that an intimal flap at the puncture site was a nidus for thrombus, which resulted in late arterial occlusions. 10 Siegelman and colleagues also demonstrated, convincingly, with “pullout” arteriograms, that there was a high incidence of thrombi on catheters at the completion of femoral arteriography. 11 They further defined several variables, including long catheterization times and thick catheters, associated with catheter clots. Intermittent frequent catheter flushing with heparinized saline solution is important in preventing intracatheter thrombus formation. Surprisingly, in our present study, the incidence of thromboembolism was only 0.4% (two patients), which contrasts with that reported by Siegelman and colleagues (2.3%). 11 Heparin (5,000 U in 500 mL of normal saline solution) was used routinely for catheter flushing in our patients. This discrepancy can also be explained by the small size of the selective catheter used in our series (5F).
Patient comorbidities, as well as age and gender, were evaluated to determine if any relationship to complication rates exists. Race was also documented but failed to provide any relevant information owing to the lack of diversity at our institution. Greater than 98% of the study population was Caucasian. No statistical significance was found between any of the listed comorbidities (hypertension, diabetes mellitus, coronary artery disease, hyperlipidemia, congestive heart failure, previous or current smoking, and renal failure with or without dialysis) and the occurrence of a major complication. However, it was noted that an increased age and a previous or current smoking history was associated with an increase in overall complications. An odds ratio analysis shows that a current or previous smoking history patient is 5.7 times more likely to have a major or minor complication. For every year increase in age, the patient is 1.06 times more likely to have a major or minor complication. Also, all complications (minor and major) from this study occurred in the 98% of the study patients that had a least one of the listed comorbidities.
The technical aspects of each procedure were also analyzed. The average operative time was 25 minutes (range 4–119 minutes) with an average injection of 97 mL of contrast material (range of 6–350 mL). Although neither of these factors proved to be statistically significant, an odds ratio analysis showed that a patient is 3.5 times more likely to have a major or minor complication if the procedure lasts longer than 30 minutes. Also, approximately 23 mL less of contrast material (97 vs 120 mL) was used on average for the procedures without complications as opposed to the ones with major or minor complications. Only one patient developed acute renal failure following the procedure.
The site of arterial puncture, the number of punctures, and whether a native vessel or graft was used were also evaluated. The transaxillary approach has been shown to have a higher complication rate owing to its location providing difficulty with hemostasis. 3 Therefore, we did not use this approach in this series. Approximately 93% of our procedures used a femoral approach, with the majority of those on the right side (68%), and 7% used the brachial approach. Two puncture sites had to be used in only 9% of the cases and showed no significant increase in the complication rates. Only 2% of our procedures had a graft puncture as opposed to a native vessel. Because of this small number, we are unable to elicit any significant information from these data.
The majority (62%) of our diagnostic angiographies were aortoiliofemoral angiograms for the evaluation of peripheral vascular disease. We also performed diagnostic studies for evaluation of abdominal aortic aneurysms, mesenteric vessels, four-vessel arch/carotid arteries, and renal arteries. No extra complications were demonstrated from adding four-vessel arch/carotid, renal, or mesenteric arteriograms.
In conclusion, we have demonstrated that vascular surgeons can safely perform diagnostic arteriographies, with complication rates similar to those of interventional radiologists.