Implications of Intraoperative Vascular Surgery Assistance for Hospitals and Vascular Surgery Trainees

Key Points

Question  What are the implications of intraoperative vascular surgery assistance for patients, hospitals, and trainees?

Findings  In this review of nonvascular surgeries requiring intraoperative vascular surgery assistance, vascular surgeons assisted a wide spectrum of surgical specialists. All operations required open surgical exposure by the vascular surgeon, and all vascular repairs were done using open technique, and the need for vascular repair was associated with higher incidence of death, myocardial infarction, or unplanned return to the operating room.

Meaning  Vascular surgeons frequently provide intraoperative assistance to other surgeons in procedures that require expertise in open anatomic exposure and vascular reconstruction.

Importance  Vascular surgeons possess a skill set that allows them to assist nonvascular surgeons in the operating room. Existing studies on this topic are limited in their scope to specific procedures or clinical settings.

Objective  To describe the broad spectrum of cases that require intraoperative vascular surgery assistance.

Design, Setting, and Participants  A retrospective medical record review of patients undergoing nonvascular surgery procedures that required intraoperative vascular surgery assistance between January 2010 and June 2014 at a single urban academic medical center (Northwestern Memorial Hospital, Chicago, Illinois). Trauma patients and inferior vena cava filter placements were excluded.

Exposures  Intraoperative vascular surgery assistance stratified by need for vascular reconstruction, anatomic location, urgency of consultation, and timing of consultation.

Main Outcomes and Measures  A composite primary end point of death, myocardial infarction, or unplanned return to the operating room within 30 days of the index operation.

Results  We identified 299 patients involving 12 different surgical subspecialties that met the study criteria. The cohort included 148 men (49.5%) and had a mean (SD) age of 56.4 (15) years. Most consultations occurred preoperatively (n = 224; 74.9%; odds ratio, 0.04; 95% CI, 0.02-0.08; P < .001) and were elective (n = 212; 70.9%; odds ratio, 0.06; 95% CI, 0.03-0.12; P < .001 ). The indications for vascular surgery assistance were 156 spine exposure (52%), 43 vascular control without hemorrhage (14.4%), 43 control of hemorrhage (14.4%), and 57 vascular reconstruction (19%). Vascular repairs consisted of 13 bypasses (4.3%), 18 patch angioplasties (6.0%), and 79 primary repairs (26.4%). All procedures required open surgical exposure by the vascular surgeon. The incidence of death, myocardial infarction, or unplanned return to the operating room was 11.4% for the cohort with a mortality rate of 1.7%. Patients who required vascular repair had a higher incidence of death, myocardial infarction, or unplanned return to the operating room (17.4% vs 7.9%; P = .01). These cases resulted in an additional 1371.46 work relative value units per year.

Conclusions and Relevance  Vascular surgeons provide crucial operative support across multiple specialties. Although vascular reconstruction is not needed in most patients, it may be associated with increased risk of death, myocardial infarction, or unplanned return to the operating room. The high proportion of emergent cases that require vascular repair demonstrates the importance of having vascular surgeons immediately available at the hospital. To continue providing this valuable service, vascular surgery trainees need to continue to learn the full breadth of open anatomic exposures and vascular reconstruction.

Surgeons occasionally request assistance in the operating room (OR) from surgeons in a different specialty when their expertise is required. Intraoperative vascular surgery assistance has been used by nonvascular surgeons for support with anatomic exposure,^1-4 oncologic resection,^5-8 and repair of vascular injury.^9,10 The description of circumstances in which vascular surgeons are asked to assist other surgeons in the OR has been limited to specific procedures^1-8 or clinical settings.^9-11 Therefore, the full scope of emergent and elective cases with which vascular surgeons assist nonvascular surgeons in the OR is lacking.

Understanding this role has implications for resident training and hospital staffing. In this study, we aim to detail the spectrum of cases in which vascular surgeons assist other nonvascular surgeons in the OR and identify which cases are more likely to require vascular repair. We hypothesize that the need for vascular repair is associated with worse short-term outcomes.

This study was approved by the Northwestern University institutional review board. Informed consent was waived for this retrospective medical record review. Records of patients undergoing nonvascular surgery procedures who required intraoperative vascular surgery assistance between January 2010 and June 2014 were reviewed. All operations were performed at Northwestern Memorial Hospital, Chicago, Illinois, a single academic medical center that also serves as a level I trauma center. Cases were identified for review using intraoperative attendance records that listed a vascular surgeon as an assisting surgeon to a nonvascular surgeon.

Trauma patients and patients having an inferior vena cava filter placement as the only vascular surgery portion of the case were excluded. Trauma patients were excluded because vascular injuries are exclusively treated by vascular surgeons at our institution, and as such, these cases did not fit into the construct of a vascular surgeon as an assisting surgeon. Along this line, inferior vena cava filter placement done during the same anesthetic as another procedure did not represent surgical assistance.

Along with patient demographics and comorbidities, the consulting surgical service, index operation, indication for vascular surgery assistance (spine exposure [SE], vascular control without hemorrhage [VC], control of hemorrhage, and vascular reconstruction), timing of consultation (preoperative and intraoperative), urgency of consultation (elective and emergent), anatomic location (truncal and peripheral), and operative details were recorded. When vascular reconstruction was required, the type of repair (primary repair, patch angioplasty, or bypass) and arterial and/or venous involvement were noted. Work relative value units (wRVUs) were recorded for the vascular surgery portion of the procedure.

The cohort was analyzed for differences in outcomes for patients with and without vascular repair, differing anatomic locations, arterial or venous involvement, urgency of consultation, and timing of consultation. A composite primary end point of death, myocardial infarction (MI), or unplanned return to the OR (DMUOR) within 30 days of the index operation was used. Any cause of mortality within the 30-day postoperative period was included as a death. We defined MI as elevated troponin with symptoms consistent with myocardial ischemia or concurrent changes in the patient’s electrocardiogram from baseline. Unplanned return to the OR was defined as any return to the OR within 30 days of the index operation that was not planned at the time of the index operation.

The patient variables, case variables, and 30-day outcomes were compared using the 2-tailed t test for continuous data and the Fisher exact test for categorical data. Univariate and multivariate analysis were conducted to determine variables associated with need for vascular repair. P values of less than .05 were considered to be statistically significant.

During the study period, 553 patients underwent nonvascular surgery procedures that required intraoperative vascular surgeon assistance. Of these, 28 were trauma patients and 226 had inferior vena cava filter placements and were excluded, leaving a cohort of 299 unique patients for analysis. This represents 6.9% of the operative volume of the vascular surgery service during the study period. The mean (SD) age for the cohort was 56.4 (15) years and 148 were men (49.5%) (Table 1). Intraoperative assistance was requested by 12 different surgical subspecialties (Figure). The most consultations came from neurosurgery (n = 101; 33.8%), orthopedic surgery (n = 79; 26.4%), urology (n = 47; 15.7%), and surgical oncology (n = 20; 6.7%). The breakdown for vascular surgeon assistance by indication was 156 SE (52%), 43 VC (14.4%), 43 control of hemorrhage (14.4%), and 57 vascular reconstruction (19%). There were statistically significant differences in baseline characteristics of the patients with and without vascular repair (Table 1). One hundred two patients (45.1%) underwent surgery by the nonvascular surgeon for resection of a mass or tumor.

For the entire cohort, most vascular surgery consultations were elective (n = 212; 70.9%) and requested preoperatively (n = 224; 74.9%). There were more vascular repairs done in the emergent setting (n = 68; 78.1%) than in the elective setting (n = 42; 19.3%; P < .001). Nearly all cases (295, 98.7%) started before 6 pm with only 4 cases (1.3%) starting after 6 pm. There were 16 cases (5.3%) in which the vascular portion of the case started after 6 pm.

For the entire cohort, 110 patients (36.8%) required vascular repairs with 13 bypasses (4.3%), 18 patch angioplasties (6.0%), and 79 primary repairs (26.4%) performed. Additionally, there were 8 major vascular ligations and 3 common iliac artery conduits. The distribution of these cases was 64 venous (21.4%), 43 arterial (14.7%), and 14 combined arterial and venous (5%). In the patients with SE, there were no arterial injuries and only 4 venous injuries requiring repair (2.6%).

The 13 bypasses consisted of a brachial artery reconstruction with basilic vein, a subclavian artery reconstruction with greater saphenous vein, an aortoaxillary bypass with polyethylene terephthalate, a thoracic aorta interposition graft with polyethylene terephthalate, an aortomesenteric bypass with polytetrafluoroethylene, 2 portal vein reconstructions with jugular vein, 2 external iliac artery bypasses with polytetrafluoroethylene, a femoral to posterior tibial artery bypass with polytetrafluoroethylene, a common femoral artery reconstruction with polytetrafluoroethylene, an external iliac artery interposition with homograft, and a femoral-femoral bypass with homograft. Patch angioplasty was performed in 18 patients consisting of 15 bovine pericardial patches (7 arterial and 8 venous) and 3 vein patches (1 arterial and 2 venous). Primary repair was done in 79 patients consisting of 17 arterial, 53 venous, and 9 combined arterial and venous. In addition to these, 8 patients underwent the following ligations: subclavian artery and vein, branch of the superior mesenteric artery, external iliac vein, external iliac artery and vein, profunda femoris artery, greater saphenous vein, renal artery and vein, and superficial femoral artery.

Univariate analysis determined a variety of variables that were associated with need for vascular repair (Table 2). The largest correlate was intraoperative consultation (odds ratio [OR], 25.9; 95% CI, 12.3-54.3; P < .001) followed by urologic procedure (OR, 18.1; 95% CI, 7.4-44.5; P < .001). The leading variable with negative association with need for vascular repair was spine surgery (OR, 0.02; 95% CI, 0.01-0.03; P < .001). The multivariate analysis demonstrated SE and VC with a negative association with need for vascular repair with ORs of 0.001 (95% CI, 0.0001-0.01; P < .001) and 0.02 (95% CI, 0.002-0.194; P < .001), respectively.

Operations were distributed throughout the body (Table 3). For the cohort, 255 patients (85.3%) were truncal and 44 patients (14.7%) were peripheral. The vascular surgeon in all cases performed open surgical exposure, and open repair was performed in all cases that required vascular repair. Only a single case used endovascular technique to obtain proximal control of the external iliac artery. The anatomic distribution in patients requiring vascular repair was 80 truncal (72.7%) and 30 peripheral (27.3%).

For the cohort, the primary end point of DMUOR was observed in 34 patients (11.4%) with 5 deaths (1.7%) in the 30-day postoperative period. In this time frame, 16 patients (5.3%) had an MI, and 20 patients (6.6%) had an unplanned return to the OR. Of these, 8 reoperations occurred in the same operative site as the vascular portion of the case, and none of them were directly related to the vascular repair.

Patients who required any vascular repair had a higher incidence of DMUOR when compared with patients who did not require vascular repair (17.4% vs 7.9%, P = .01). The incidence for DMUOR was not different for venous vs arterial involvement, anatomic location, timing of consultation, or urgency of consultation (Table 4). A subgroup analysis was performed on the cohort after exclusion of the patients with SE. For this group, as with the full cohort, the incidence for DMUOR was not different for venous vs arterial involvement (22.7% vs 13.3%, P = .29), anatomic location (truncal 13.1% vs peripheral 18.2%, P = .45), timing of consultation (preoperative 20.5% vs intraoperative 9.3%, P = .06), or urgency of consultation (emergent 13.7% vs elective 15.9%, P = .81). Interestingly, for this subgroup, there was no difference in DMUOR between patients with and without vascular repair (16.0% vs 10.5%, P = .59).

Data on wRVUs were available for cases starting in fiscal year 2011. Data were available for 249 of the 299 patients. The mean wRVUs were 21.11 (range, 5.00-79.86) per case for the vascular surgery portion. This corresponds to an average yearly increase of 1371.47 wRVUs for the Division of Vascular Surgery.

In this study, we describe a broad spectrum of nonvascular operations, both elective and emergent, at a single urban academic center that required vascular surgery assistance. Prior reports of intraoperative vascular assistance with nonvascular operations have been limited to procedure-specific series^1-8,12,13 and major intraoperative vascular complications.^9,10 A 2015 report characterizing intraoperative vascular surgery consultations increased the scope of cases but did not include preoperative consultations for intraoperative assistance.¹¹ Our study provides a more complete view of the role assisting vascular surgeons play in the OR to understand the implications for resident training and hospital staffing as well as the risks associated with these procedures.

In our experience, vascular surgeons were asked to assist surgeons from 12 different surgical specialties in the OR. This parallels the experience of others who have described major intraoperative vascular injuries^9,10 and intraoperative consultations.^10,11 The indication for consultation in our series was also heterogeneous, with 156 SE (52%), 43 VC (14.4%), 43 control of hemorrhage (14.4%), and 57 vascular reconstruction (19%). Every case required open surgical exposure by the vascular surgeon. Because the patients in the cohort were identified using the name of the treating vascular surgeon and not billing or diagnosis codes, it is less likely that cases using endovascular or hybrid techniques were not captured.

As in other similar studies, we saw a predominance of open surgical exposures in a variety of anatomic positions and vascular beds.^10,11 With operations occurring throughout the body (Table 3), the importance for trainees to continue to learn the full spectrum of open anatomic exposures is highlighted. The implementation of the integrated vascular surgery residency program reduces the total number of years spent in training and may limit the opportunity to learn open surgical technique.¹⁴ This concern was expressed by a survey of surgery department chairs.¹⁴ Even traditional vascular surgery fellows’ experience in open surgical exposures is being curtailed during their general surgery training.^15-17 With decreasing opportunities for open surgery, development of simulation may be needed to fill this gap.¹⁸ However, nothing can replace performing open surgery, and assisting with anatomic exposure during training is a good educational opportunity.¹

An argument could be made that some anatomic exposures do not require a vascular surgeon; however, when a vascular injury is encountered, vascular surgery expertise is required. Spine exposure is an example of this type of procedure, and SEs made up most of the cohort. We observed a low rate of vascular injury (n = 4; 2.6%) that is comparable with the rates of vascular injury (1.9%-6.1%) in collaborative studies between spine and vascular surgeons.^1-4 Series without the use of an “access” surgeon had similar rates of vascular injury (2%-6.2%).^12,13 However, these tended to be highly selected patients with single-level interventions. Our vascular repairs in the patients with SE were limited to 4 venous injuries, which is consistent with the low incidence of arterial injury observed in the literature.^1-4,12,19 Resection of a tumor or mass made up 135 patients (45.1%) and was associated with need for vascular repair with an OR of 11.8 (95% CI, 6.7-20.8; P < .001). However, with multivariate analysis, SE and VC were the only variables associated with vascular repair and were inversely related. In our series, all vascular repairs were conducted using open surgical techniques, demonstrating the need to continue to teach open vascular reconstruction. One hundred ten vascular repairs were conducted with most of them (71.8%) being primary repair. Yoo et al¹⁰ conducted 27 open vascular repairs for intraoperative vascular injuries and had a similar experience (primary repair done in 63% of patients). In contradistinction, Danczyk et al¹¹ did 120 vascular reconstructions for intraoperative vascular consultations, but only 23 (19.1%) were primary repairs. In this series, they used autologous tissue in 63% of their reconstructions. Most bypasses performed in our study were done using synthetic bypass grafts. Similarly, other studies that evaluated vascular bypass in retroperitoneal malignancy used prosthetic bypass more frequently.^5,6 In our series, patch angioplasties were performed using biologic grafts without the use of synthetic material. Because of the heterogeneous population within our study and differences between these other studies, it is difficult to interpret the significance of differences in materials and methods of revascularization. Many of these could be explained by surgeon preference and the unique circumstances of each clinical setting. In either case, these studies and our own indicate that continued training in the spectrum of open vascular reconstructions is important. However, the opportunity to receive this training may be limited with the rise of endovascular therapies.²⁰

Timely availability of vascular surgery assistance is important because the variables that were most associated with the need for vascular repair were intraoperative consultation and emergent consultation (Table 2). Multivariate analysis did not demonstrate this, likely because of the heterogeneous nature of the cases. Most cases started during normal OR hours with only 16 cases (5.3%) starting after 6 pm. This timing was also the experience of Danczyk et al.¹¹ All of these factors illustrate the need to have vascular surgeons readily available during the workday. Some of this cost could be offset with the additional source of revenue. Although Danczyk et al¹¹ had more RVUs per case of 30.9, our series had more wRVUs per year with the broader case selection. These cases represent 6.9% of the operative volume of the vascular surgery service, which is a good proportion. Although not included in this analysis, we have previously reported on iatrogenic injuries outside the OR that required vascular surgery expertise.²¹ When you include these, complications after cardiac catheterization, and trauma, the usefulness of having vascular surgeons available is further expanded.

To measure outcomes, we used a composite end point of DMUOR. The overall incidence of DMUOR was 11.4%. Although most patients did not require vascular repair, those who did had higher rates of DMUOR (17.4% vs 6.5%, P <.001). However, this difference was no longer statistically significant in the subgroup analysis with the patients with SE excluded. Although vascular repair may add some risk to the patient, the patients who required vascular repair had significantly more comorbidities at baseline, which also contributes to this outcome. We expected emergent and intraoperative consultations to have a higher incidence of poor outcome than elective cases, but this was not observed. Small numbers and type II error could explain this.

The overall perioperative mortality rate of 1.7% for the cohort is similar to other studies evaluating intraoperative vascular surgery consultations with mortality rates ranging from 1.6% to 6.2%.^5,6,10,11 Oderich et al,⁹ who evaluated intraoperative abdominal and pelvic venous injuries, had a much higher perioperative morality of 18%. We had a single death in the SE group, which is consistent with the low perioperative mortality seen in these patients.^1-4,12,13

The other components of the composite end point were chosen as indicators of quality and long-term mortality. The addition of the unplanned return to the OR to the end point was based on other studies of intraoperative vascular consultations and its use as a quality marker. It is a global quality indicator and can be used across a variety of cases, as in our study.¹⁵ The observed rate of unplanned return to the OR after intraoperative vascular assistance in the literature is broad (4.3% to 46%), likely reflecting procedure-specific risks and the definition of unplanned return to the OR.^5,9,10 Our study had an overall unplanned return to the OR rate of 6.6%, with 10% in the VR group. We used a broad definition to capture these occurrences and assess outcomes across a heterogeneous group of patients and procedures. Myocardial infarction was included as part of the composite end point because it has been shown to be associated with long-term mortality after vascular surgery.^22,23 Our rate of perioperative MI (5.3%) corresponds to what has been observed in the literature after vascular surgery.

As with all retrospective reviews, our study has inherent limitations. We did not have a control group of patients that did not receive assistance from a vascular surgeon. The patients in our cohort were very heterogeneous in terms of both baseline comorbidities and index operation. Patients who required vascular repair had a higher incidence of baseline comorbidities, which placed them at higher risk of postoperative events and introduced bias into the study. Additionally, the patients were undergoing a variety of index operations, each associated with a unique risk profile. Finally, the practices at our institution may not be generalizable to all vascular surgeons because this experience is largely dependent on the case selection and practice pattern of the nonvascular surgeons at our institution.

Vascular surgeons are often called on by nonvascular surgeons for assistance in the OR in a variety of clinical situations and anatomic locations. Open surgical techniques are required in nearly all cases. Although unplanned vascular reconstruction is not needed in most patients, it may be associated with increased risk of DMUOR when performed. While most consultations occurred preoperatively, a high proportion of emergent cases that are more likely to require vascular repair demonstrates the importance of having vascular surgeons immediately available at the hospital. To continue providing this valuable service, vascular surgery trainees will need to continue to learn the full breadth of anatomic exposures and open vascular reconstruction.

Corresponding Author: Tadaki M. Tomita, MD, Division of Vascular Surgery, Northwestern University Feinberg School of Medicine, 676 N St Clair St, Ste 650, Chicago, IL 60611 (ttomita1@nm.org).

Accepted for Publication: May 10, 2016.

Published Online: August 3, 2016. doi:10.1001/jamasurg.2016.2247.

Author Contributions: Drs Tomita and Eskandari had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Tomita, Rodriguez, Hoel, Eskandari.

Acquisition, analysis, or interpretation of data: Tomita, Rodriguez, Ho, Pearce, Eskandari.

Drafting of the manuscript: Tomita, Ho.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Tomita, Ho.

Administrative, technical, or material support: Tomita, Rodriguez, Pearce, Eskandari.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This paper was presented at the Vascular and Endovascular Surgery Society Spring Meeting; June 17, 2015; Chicago, Illinois.