Development and preliminary assessment of a machine learning model to predict myocardial infarction and cardiac arrest after major operations.

INTRODUCTION: Accurate prediction of complications often informs shared decision-making. Derived over 10 years ago to enhance prediction of intra/post-operative myocardial infarction and cardiac arrest (MI/CA), the Gupta score has been criticized for unreliable calibration and inclusion of a wide spectrum of unrelated operations. In the present study, we developed a novel machine learning (ML) model to estimate perioperative risk of MI/CA and compared it to the Gupta score. METHODS: Patients undergoing major operations were identified from the 2016-2020 ACS-NSQIP. The Gupta score was calculated for each patient, and a novel ML model was developed to predict MI/CA using ACS NSQIP-provided data fields as covariates. Discrimination (C-statistic) and calibration (Brier score) of the ML model were compared to the existing Gupta score within the entire cohort and across operative subgroups. RESULTS: Of 2,473,487 patients included for analysis, 25,177 (1.0%) experienced MI/CA (55.2% MI, 39.1% CA, 5.6% MI and CA). The ML model, which was fit using a randomly selected training cohort, exhibited higher discrimination within the testing dataset compared to the Gupta score (C-statistic 0.84 vs 0.80, p < 0.001). Furthermore, the ML model had significantly better calibration in the entire cohort (Brier score 0.0097 vs 0.0100). Model performance was markedly improved among patients undergoing thoracic, aortic, peripheral vascular and foregut surgery. CONCLUSIONS: The present ML model outperformed the Gupta score in the prognostication of MI/CA across a heterogenous range of operations. Given the growing integration of ML into healthcare, such models may be readily incorporated into clinical practice and guide benchmarking efforts.

Association of Laparoscopic Approach With Improved Short-Term Outcomes in Patients With Hematologic Malignancy Undergoing Splenectomy.

BACKGROUND: Patients with hematologic malignancies undergo splenectomy for both diagnostic and therapeutic purposes. Although minimally invasive surgery continues to be increasingly utilized for a variety of abdominal operations, no large-scale data has compared the postoperative outcomes for laparoscopic vs open splenectomy in patients with hematologic malignancy. METHODS: Patients with a diagnosis of hematologic malignancy who underwent laparoscopic and open splenectomy between 2015 and 2020 were queried using the ACS-NSQIP database. 30-day outcomes of laparoscopic vs open splenectomy were compared. RESULTS: Out of 430 patients included in the study, 52.6% were male, with a mean age of 63.4 ± 13.1 years. 233 patients (54.2%) underwent laparoscopic splenectomy. On bivariate analysis, laparoscopic surgery was associated with lower rates of 30-day mortality [2.1% vs 11.7% (P < .001)] and morbidity [9.0% vs 24.4% (P < .001)]. On multivariate regression, elective operations (OR .255, 95%CI: 0.084-.778, P = .016) and laparoscopic surgery (OR .239, 95%CI: 0.075-.760, P = .015) were independently associated with lower mortality, while history of metastatic cancer (OR 3.331, 95%CI: 1.144-9.699, P = .027) was associated with higher mortality. Laparoscopic surgery (OR .401, 95%CI: 0.209-.770, P = .006) and steroid use (OR 2.714, 95%CI: 1.279-5.757, P = .009) were the only two factors independently associated with 30-day morbidity. Laparoscopic surgery was also associated with shorter hospital length of stay (median 3 [IQR:3] vs 6 [IQR:7] days). CONCLUSION: Laparoscopic splenectomy was associated with lower 30-day mortality and morbidity, and shorter length of stay in patients with hematologic malignancies. These data suggest that laparoscopic approach, when feasible, may be preferred for splenectomy in this patient population.

Factors Associated with Amputation after Popliteal Vascular Injuries.

BACKGROUND: Popliteal artery trauma has the highest rate of limb loss of all peripheral vascular injuries. The objectives of this study were to evaluate outcomes after popliteal vascular injury and to identify predictors of amputation. METHODS: Retrospective data over a 14-year period were collected for patients with popliteal artery with or without vein injuries. Patient demographics, mechanism of injury, Injury Severity Score (ISS), Mangled Extremity Severity Score (MESS), and physiologic parameters were extracted. Time to operative intervention, operative time, type of vascular repair, need for concomitant orthopedic procedures, and outcomes including amputation rate, and in-hospital mortality were recorded. RESULTS: Fifty-one patients were found to have popliteal artery injuries, with a median age of 25 (range 10-70 years). The median ISS was 9, and the mean extremity Abbreviated Injury Severity score was 3. The mechanism of injury was blunt for 43% and penetrating for 57%. Fasciotomies were performed in 74% of patients and 64% of patients underwent combined orthopedic and vascular procedures. Overall, 66% of these patients had their vascular procedure performed first. Ten patients required amputation: 1 immediate and 9 after attempted limb salvage (20%). We found that those patients requiring amputation had a higher incidence of blunt trauma (80% vs. 35%, P = 0.014) and higher MESS score (7.1 vs. 4.7, P = 0.02). There was no difference in the incidence of amputation for those who underwent orthopedic fixation before vascular repair (P = 0.68). CONCLUSIONS: Popliteal vascular injuries continue to be associated with a high risk of amputation. Those patients undergoing attempted limb salvage should be revascularized expediently, but selected patients may undergo orthopedic stabilization before vascular repair without increased risk of limb loss.

The Addition of Ultrasound Arterial Examination to Upper Extremity Vein Mapping before Hemodialysis Access.

BACKGROUND: Routine upper extremity vein mapping by ultrasound (Ven-US) is recommended by current National Kidney Foundation/Kidney Disease Outcomes Quality Initiative guidelines before arteriovenous fistula (AVF) creation. However, the impact of concomitant arterial US (Art-US) examination is not clear. METHODS: The Ven-US protocol at our institution was modified to include Art-US starting January 2013. Therefore, retrospective review of patients who received Ven-US with Art-US between January 2013 and July 2014 was performed. The Art-US component included distal brachial and radial artery diameters, level of brachial bifurcation, and Doppler Allen's test. A plan for hemodialysis (HD) access was proposed by 2 attending vascular surgeons (VS1 and VS2) and based on a set of criteria for fistula creation (CFC) using Ven-US findings alone. The Art-US findings were subsequently reviewed, and the plan was changed based on either vascular surgeon judgment (VS1 and VS2) or predetermined arterial anatomic criteria (CFC). RESULTS: In total, 163 patients (326 arms) were included. The mean age was 53 years, most patients were male (60%), and most were HD dependent at the time of US evaluation (67%). The initial plan based on Ven-US was: 17-19% radiocephalic (RC) AVF, 33-48% brachiocephalic AVF, 20-27% brachiobasilic AVF, and 14-23% grafts. The Art-US revealed 159 radial arteries (49%) with diameter <2 mm, 16 brachial arteries (5%) with high bifurcation, 93 (29%) incomplete palmar arches, and 7 arms (2%) with arterial waveform blunting. Review of Art-US findings resulted in an overall change to the operative plan from 4% to 12% of patients. Those with an initially planned RC AVF were more likely to have a change in operative approach (21-57%) compared with all other types of planned access (1-3%, P < 0.001). CONCLUSIONS: Preoperative Art-US may significantly change the operative plan, particularly when planning a RC AVF, and should be performed before HD access surgery at the wrist.

Predictors of Severity in Diabetic Foot Infections.

BACKGROUND: Diabetic foot infections (DFIs) constitute a large burden of the morbidity of diabetes, with more than 70,000 lower-extremity amputations (LEA) performed annually in the United States. A necrotizing infection signifies the most severe form of infection and is a key factor in the decision to proceed to LEA for source control. Key clinical and laboratory variables can assist in the identification of necrotizing infections; however, the effect of diabetes on these variables is unknown. Given the increased level and complexity of metabolic derangements in diabetic patients, we sought to examine characteristics predictive of necrotizing infection in patients with DFI who underwent LEA. METHODS: We performed a single-institutional retrospective analysis of diabetic patients who underwent a LEA for DFIs over an 18-month period. Patients with necrotizing infection on final pathology were compared with patients without this pathologic finding. Multivariate analysis was performed to identify independent predictors of necrotizing infection. RESULTS: Of 183 patients, 57 (31%) had evidence of necrotizing infections. Factors associated with necrotizing infection on univariate analysis were the presence of bullae (11% vs. 2%; odds ratio [OR] = 4.8, P = 0.03), a higher mean admission white blood cell count (WBC; 15 vs. 12, P = 0.002), a lower mean absolute sodium (132 vs. 134, P = 0.01), a higher hemoglobin A1C (11.3 vs. 10.3, P = 0.05), hyperglycemia (289 vs. 248, P = 0.04), elevated C-reactive protein (20 vs. 11, P = 0.02), and the presence of Pseudomonas aeruginosa on final tissue culture (12.3 vs. 1.6, P = 0.004). These patients were taken to surgery more rapidly (22.5 vs. 31 hr, P = 0.04), and they had a longer postoperative stay (7 vs. 4 days, P = 0.02). On multivariate analysis, an elevated WBC was predictive of necrotizing infection (OR = 1.1, P = 0.01), whereas alcohol use was found to be protective (OR = 0.3, P = 0.04). CONCLUSIONS: Clinical and laboratory variables known to be associated with necrotizing infections among the general population appear to be predictive of disease severity among patients undergoing amputation for DFIs. Identification of these abnormalities preoperatively may allow for improved operative planning, shared decision making, and resource management. Prospective validation of these findings is potentially warranted.

General Surgery Resident Vascular Operative Experience in the Era of Endovascular Surgery and Vascular Fellowships.

Advances in endovascular surgery have resulted in a decline in major open arterial reconstructions nationwide. Our objective is to investigate the effect of endovascular surgery on general surgery resident experience with open vascular surgery. Between 2004 and 2014, 112 residents graduated from two academic institutions in Southern California. Residents were separated into those who graduated in 2004 to 2008 (period 1) and in 2009 to 2014 (period 2). Case volumes of vascular procedures were compared using two-sample t test. A total of 43 residents were in period 1 and 59 residents were in period 2. In aggregate, there was no significant difference in open cases recorded between the two periods (84 vs 87, P = 0.194). Subgroup analysis showed period 2 recorded significantly fewer cases of open aneurysm repair (5 vs 3, P < 0.001), cerebrovascular (14 vs 10, P = 0.007), and peripheral obstructive procedures (16 vs 13, P = 0.017). Dialysis access procedures constituted the largest group of procedures and remained similar between the two periods (35 vs 42, P = 0.582). General surgery residents experienced a significant decline in several index open major arterial reconstruction cases. This decline was offset by maintenance of dialysis access procedures. If the trend continues, future general surgeons will not be proficient in open vascular procedures.

General Anesthesia Is Not Necessary for Hemodialysis Access Surgery.

Chronic kidney disease has been identified as a risk factor for mortality after procedures under general anesthesia (GA). However, a recent study showed that 85 per cent of arteriovenous fistulas in the United States are performed under GA. Our aim was to demonstrate that GA can be avoided in patients with chronic kidney disease and end-stage renal disease by using local anesthesia (LA) with monitored anesthesia care or brachial plexus block (BPB) during hemodialysis access surgery. A retrospective review was performed at a single institution. Outcome measures included need for conversion to GA, major perioperative complications, and 30-day mortality. Four hundred and fourteen access procedures were performed by seven vascular surgeons between 2011 and 2014. Arteriovenous fistulas were placed in 379 (92%), arteriovenous grafts were placed in 31 (7%), and four (1%) received unsuccessful extremity exploration. Anesthetic approach was LA in 344 (83%) and BPB in 64 (15%). GA was initially induced in three (0.7%) and three (0.7%) additional patients required conversion to GA from LA. There were no cardiopulmonary events or perioperative deaths. Of the 32 patients who received an arteriovenous graft, only three (10%) required GA. In conclusion, LA and BPB are safe and conversion to GA is rare. GA should be avoided in hemodialysis access surgery.

Predictors of Mortality in Patients with Penetrating Inferior Vena Cava Injuries Surviving to the Operating Room.

Inferior vena cava (IVC) injuries are associated with significant morbidity and mortality. To identify clinical factors associated with mortality in patients undergoing operative intervention for penetrating IVC injuries, a retrospective review of 98 patients was performed, excluding blunt injuries (n = 20) and deaths before surgery (n = 16). The overall mortality was 58 per cent. Nonsurvivors more commonly presented with hypotension (50% vs 23%, P = 0.03) and underwent resuscitative thoracotomy more frequently (42% vs 4%, P = 0.01). Retrohepatic injuries were more common among nonsurvivors (P = 0.04). There was no difference in the use of ligation (7% vs 17%, P = 0.29) or the massive transfusion protocol (35% vs 25%, P = 0.41). On multivariate analysis, after controlling for mechanism of injury, admission hypotension, Glasgow Coma Scale score , preoperative cumulative fluids, resuscitative thoracotomy , absence of spontaneous tamponade, and location of IVC injury, the only independent predictor of mortality was the absence of spontaneous tamponade at the time of laparotomy (odds ratio = 5.4, 95% confidence interval: 1.11-25.95; P = 0.04). Penetrating IVC injuries continue to be associated with a high mortality, particularly among patients with free intraabdominal hemorrhage at laparotomy. Large multicenter studies are required to define the optimal resuscitative and operative management techniques in these severely injured patients.

Comparison of Neurocognitive Outcomes after Carotid Endarterectomy and Carotid Artery Stenting.

Cognitive and emotional outcomes after carotid endarterectomy (CEA) and carotid artery stenting with embolic protection device (CAS + EPD) are not clear. Patients were entered prospectively into a United States Food and Drug Administration-approved single-center physician-sponsored investigational device exemption between 2004 and 2010 and received either CEA or CAS + EPD. Patients underwent cognitive testing preprocedure and at 6, 12, and 60 months postprocedure. Cognitive domains assessed included attention, memory, executive, motor function, visual spatial functioning, language, and processing speed. Beck Depression and anxiety scales were also compared. There were a total of 38 patients that met conventional indications for carotid surgery (symptomatic with ≥50% stenosis or asymptomatic with ≥70% stenosis)-12 patients underwent CEA, whereas 26 patients underwent CAS + EPD. Both CEA and CAS + EPD patients showed postprocedure improvement in memory and executive function. No differences were seen at follow-up in regards to emotional dysfunction (depression and anxiety), attention, visual spatial functioning, language, motor function, and processing speed. Only two patients underwent neuropsychiatric testing at 60 months-these CAS + EPD patients showed sustained improvement in memory, visual spatial, and executive functions. In conclusion, cognitive and emotional outcomes were similar between CEA and CAS + EPD patients.

A Rare Case of Acroangiodermatitis Associated with a Congenital Arteriovenous Malformation (Stewart-Bluefarb Syndrome) in a Young Veteran: Case Report and Review of the Literature.

Acroangiodermatitis (AD) is a rare angioproliferative disease manifesting with cutaneous lesions clinically similar to Kaposi's sarcoma. AD is a benign hyperplasia of preexisting vasculature and may be associated with acquired or congenital arteriovenous malformations (AVM), or severe chronic venous insufficiency (because of hypostasis, elevated venous pressure, arteriovenous shunting). Stewart-Bluefarb syndrome is the rare syndrome in which AD is associated with a congenital AVM. We present the case of a young veteran with a painful, chronic nonhealing ulcer and ipsilateral popliteal artery occlusion likely because of trauma, who elected transmetatarsal amputation for symptomatic relief. A 24-year-old male veteran presented with a 5-year history of a nonhealing dorsal left foot ulcer, resulting from a training exercise injury. He ultimately developed osteomyelitis requiring antibiotics, frequent debridements, multiple trials of unsuccessful skin substitute grafting, and severe unremitting pain. He noted a remote history of left digital deformities treated surgically as a child, and an AVM, previously endovascularly treated at an outside facility. Arterial duplex revealed somewhat dampened left popliteal, posterior tibial (PT), and dorsalis pedis (DP) artery signals with arterial brachial index of 1.0. CT angiography showed occlusion of the proximal to mid popliteal artery with significant calcifications felt initially to be a result of prior trauma. Pedal pulses were palpable and transcutaneous oxygen measurements revealed adequate oxygenation. Because of unremitting pain, the patient opted for amputation. Pathology revealed vascular proliferation consistent with AD. This case illustrates an unusual diagnosis of acroangiodermatitis, and a rare syndrome when associated with his underlying AVM (Stewart-Bluefarb syndrome). This resulted in a painful, chronic ulcer and was further complicated by trauma-related arterial occlusive disease. AD disease can hinder wound healing even in the presence of clinically evident blood flow. Although rare, such unusual diagnoses should be entertained particularly in the unusually young vascular surgical patient.

Intravascular ultrasound.

Intravascular ultrasound (IVUS) has an interesting history that parallels that of many of the advancements that have led to the endovascular era. The use of IVUS in conjunction with standard cross-sectional imaging and three-dimensional reconstructions offers a powerful tool in both the diagnosis and treatment of complex vascular pathology. The use of IVUS has increased over the years and is currently in the process of being incorporated into several modalities that will offer more in the way of real-time information in both the aortic arena and the treatment of increasingly complex peripheral vascular disease. Currently, we use IVUS as a powerful adjunct in combination with other modalities to increase our understanding of vessel architecture and assist in the management of complex vascular pathology.