Cumulative risks for reoperation due to bleeding after carotid endarterectomy and the associated clinical impact of bleeding events.

Abstract

Objective: The purpose of this study was to identify all preoperative and intraoperative variables in the Vascular Quality Initiative (VQI) carotid endarterectomy (CEA) module that have a statistically significant association with reoperation for bleeding. A weighted risk score was developed and validated to predict this event, with assessment of its impact on 30-day mortality and other adverse perioperative events.

Methods: The VQI CEA module was queried between January 2003 and October 2023. Overall, 192,547 CEA procedures met study inclusion. An internal VQI validation cohort was created with the same exclusion criteria utilizing CEAs performed between November 2023 and October 2024, over which time period 17,449 procedures met inclusion criteria.

Results: The following variables had a statistically significant multivariable association (P < .05) with reoperation for bleeding after CEA: Black race (adjusted odds ratio [aOR], 1.53); body mass index <20 kg/m² (aOR, 1.40); hypertension (aOR, 1.19); history of coronary artery disease revascularization (aOR, 1.16); congestive heart failure (CHF) (aOR, 1.37); chronic obstructive pulmonary disease (aOR, 1.19); dual antiplatelet at time of surgery (aOR, 1.51); on anticoagulation baseline (aOR, 1.23); preoperative Rankin score 2 or higher (aOR, 1.41); urgent/emergent CEA (aOR, 1.36); eversion CEA technique (aOR, 1.33); surgeon selection for drain placement (aOR, 1.17); and lack of protamine utilization intraoperatively (aOR, 2.08). The following variables had a significant (P < .05) protective effect vs reoperation for bleeding after CEA: female sex (aOR, 0.84); body mass index >35 kg/m² (aOR, 0.85); and active smoking status (aOR, 0.85). Patients with risk scores of zero or less had an only 0.006% risk of return to the operating room for bleeding. There was significant elevation in risk for return to the operating room for bleeding with escalating risk sores. Patients with risk scores 11 and higher had an absolute reoperation for bleeding event rate of 3.6%, which was a total event rate 600 times higher than individuals with scores of 0 or less and 3.6 times as high as individuals with scores as high as 5. The internal VQI validation cohort experienced the event of return to the operating room for bleeding at very similar rates to the primary study source cohort with no statistically significant difference at any of the risk score points, indicating consistency for the risk score. Patients who experienced return to the operating room for bleeding after CEA experienced a statistically significant increased rate of 30-day mortality (OR, 1.59); cranial nerve injury (OR, 2.03); perioperative neurologic event (OR, 5.80); myocardial infarction (OR, 6.56); cardiac dysrhythmia (OR, 4.20); perioperative CHF (OR, 5.26); and skin-soft tissue infection postoperatively (OR, 12.61) with P < .001 for all.

Conclusions: A validated quantitative risk score has been developed to predict reoperation for bleeding after CEA. The most impactful variables, which are also largely modifiable, include intraoperative protamine utilization and avoidance of dual antiplatelet therapy. Patients who experience reoperation for bleeding after CEA experience significantly higher rates of 30-day mortality, myocardial infarction, CHF, cranial nerve injury, skin-soft tissue infection, and adverse perioperative neurologic events.