August 2021

Synopsis: Atrial fibrillation is common, particularly in the elderly, and is responsible for a quarter of ischemic strokes. Many of these strokes are cardioembolic in nature, with emboli originating in the left atrial appendage. Oral anticoagulation, when therapeutic, is effective at preventing ischemic stroke. However, achieving therapeutic anticoagulation is associated with challenges, including nonadherence, underdosing, difficulty maintaining INR, and having to discontinue use due to bleeding. It’s been theorized that occluding the left atrial appendage would limit the origin of emboli and would reduce ischemic stroke. This had never been tested in randomized fashion. The Left Atrial Appendage Occlusion Study III is a multicenter, randomized control trial that randomized patients to either atrial appendage occlusion or no occlusion at the time of cardiac surgery for another indication, in addition to standard of care anticoagulation. The primary outcome was the occurrence of ischemic stroke (including transient ischemic attack with positive neuroimaging) or systemic embolism.

Patients greater than 18 years of age, with a history of atrial fibrillation with a CHA2DS2 -VASc score of at least 2 (scale of 0 to 9, with higher scores indicating higher risk) who were undergoing first time cardiac surgery on cardiopulmonary bypass were included. Exclusions were made for patients undergoing mechanical valve implantation, cardiac transplant, complex congenital surgery, LVAD, and those who had a previous atrial closure device implanted. The primary analysis population included 2379 participants (68% men) in the occlusion group and 2391 (67% men) in the no-occlusion group, with a mean age of 71 years and a mean CHA2DS2 -VASc score of 4.2 in both groups. Aside from the surgical team, caretakers were blinded to the randomization. Left atrial appendage occlusion was performed during cardiac surgery with the use of any of the following techniques: amputation and closure (55.7%) (preferred), stapler closure (11.2%), double-layer linear closure from within the atrium (13.8%) in participants undergoing mini-thoracotomy or closure with an approved surgical occlusion device (4.1%). Neither percutaneous closure nor purse-string closure was permitted.

The participants were followed for a mean of 3.8 years. A total of 92.1% of the participants received the assigned procedure, and at 3 years, 76.8% of the participants continued to receive oral anticoagulation. Stroke or systemic embolism (including TIA occurred in 114 participants (4.8%) in the occlusion group and in 168 (7.0%) in the no-occlusion group (hazard ratio, 0.67; 95% confidence interval, 0.53 to 0.85; P=0.001). During the first 30 days after surgery, a primary-outcome event occurred in 53 participants (2.2%) in the occlusion group and in 65 (2.7%) in the no-occlusion group (hazard ratio, 0.82; 95% CI, 0.57 to 1.18). After 30 days, a primary-outcome event occurred in 61 participants (2.7%) in the occlusion group and in 103 (4.6%) in the no-occlusion group (hazard ratio, 0.58; 95% CI, 0.42 to 0.80). The incidence of perioperative bleeding, heart failure, or death did not differ significantly between the trial groups, bypass time was 6 mins longer in the closure group, with a mean of 4 additional minutes cross-clamped.

The authors conclude that left atrial appendage ligation at the time of concomitant cardiac surgery is protective against ischemic stroke when compared to a similar cohort, particularly beyond 30 days postoperatively. The estimated number of patients needed to treat to prevent one stroke is 37. The group also concludes that it is safe to perform appendage closure at the time of surgery, given the lack of a difference in the risk of heart failure or major bleeding. Utilization of left atrial appendage ligation is not advocated by this study to be a primary replacement for systemic anticoagulation, only to augment this treatment and to help in prevention when anticoagulation is not being used effectively or is contraindicated.

This study suggests that ligation of the left atrial appendage at the time of cardiac surgery is safe and effective at preventing ischemic stroke when used in addition to systemic anticoagulation.

Synopsis: The coronavirus (COVID-19) pandemic has significantly impacted all aspects of healthcare, including surgical residency training. To evaluate its effect on resident education, this study assessed changes to the volume of surgical cases performed by general surgical residents during the pandemic as a primary outcome. The secondary outcomes assessed were changes to the educational curriculum, the effect of COVID-19 on resident preparedness for the next phase of their career, and resident wellness/burnout.

An anonymous survey was sent to all 260 general surgery program directors and coordinators for distribution. Over 1,000 general surgery residents across multiple states, sites, and PGY years responded to the survey. A large proportion of residents (42.3%) reported feeling that they could not meet the traditional case number graduation criteria set by the Accredited Council for Graduate Medical Education (ACGME). Consistent with this finding, resident operative experience declined significantly with 80.7% logging ≤3 cases/week, up from 23.7% prior to the pandemic (p<0.01). Additionally, 77% reported performing fewer floor duties and 80.67% attended no clinic during the pandemic. In place of the reduced direct clinical experience, many reported an increase in didactic teaching (41.3%) and time for self-directed reading (49.5%) with most residents reporting an educational curriculum that was entirely online (80.6%). Although this study didn’t evaluate how this was perceived by residents, other studies have demonstrated that most surgical residents viewed these changes to the educational curriculum in a positive light¹. This highlights the benefit of online curricula and the flexibility and accessibility that come with them. This will likely become a staple of many educational programs.

Another relevant finding of this study is that the negative consequences of the pandemic were not limited to the clinical experiences of trainees, but also included an impact on mental health. Despite more time off overall, 33.1% of residents reported more burnout than usual, suggesting that burnout is not solely associated with duty hours. Cited sources of stress included finances, work stress, fear of contracting COVID, and fear of transmitting COVID to friends and family. This highlights the importance of a multipronged approach to resident wellness. Duty hour compliance is certainly important but focusing on the quality of the hours spent at work and other aspects of resident life may prove more fruitful.

¹Hope C, Reilly J, Griffths G, Lund J and Humes D. The impact of COVID-19 on surgical training: a systematic review. Techniques in Coloproctology 2021 May;25(5):505-520.

Synopsis: Surgical resection, in addition to chemo- and radiotherapies, is the standard of care for localized esophageal cancer. Minimally invasive esophagectomy (MIE) has gained popularity in the last decade, for which there are multiple surgical techniques involving variable entry of the abdomen, chest and neck. There has historically been debate over the merits of a neck vs chest anastomosis – and to date the data largely consists of prospective case series and lacks robust randomized controlled trials.

This was a randomized control trial involving 9 hospitals in the Netherlands conducted between 2016 and 2020. They included patients with localized squamous cell or adenocarcinomas of the mid or distal esophagus or GE-junction (Siewert I, II). All patients received neoadjuvant or perioperative chemotherapy. Patients either underwent total MIE or hybrid MIE with lymph node dissection and were randomized to either cervical or thoracic anastomosis. Anastomotic technique was up to the discretion of the surgeon. All patients had an omental wrap around the anastomosis. Additionally, operative videos in 20% of patients were reviewed for surgical quality. The primary outcome was anastomotic leakage within 30 days of esophagectomy requiring endoscopic, radiologic or surgical reintervention. Secondary outcomes included incidence of post-operative functional outcomes, recurrent laryngeal nerve palsy, tumor-free resection margin rate, hospital and ICU length of stay, ICU readmission rate and death. 262 patients were enrolled (211 who underwent total MIE and 50 who underwent hybrid MIE), 130 of which had intrathoracic anastomoses and 132 of which had cervical anastomoses.

Overall, anastomotic leak occurred in 15 patients (12.3%) in the intrathoracic anastomosis group and 39 patients in the cervical anastomosis group (31.7%) (risk difference -19.4%, [95% CI -29.5 to -9.3%], p < 0.001). The subgroup of hybrid MIE had no significant difference in rates of leak requiring intervention between the two types of anastomoses. Most interventions were performed endoscopically in both groups. In terms of secondary outcomes, recurrent laryngeal nerve palsy (risk difference, −7.3% [95% CI, −12.1% to −2.5%], p 0.003), incidence of severe complications (risk difference, −11.3% [−20.4% to −2.2%], p 0.02) and pleural effusion requiring drainage (risk difference risk different -11.3%, [-20.2 to -2.4], p=0.01) were more common in the cervical anastomosis group Hospital length of stay was also longer in the cervical group (p=0.003). In the subgroup of patients with anastomotic leakage, the severity of leakage between cervical and intrathoracic anastomoses were similar (as measured by Esophagectomy Complications Consensus Group classification, Clavien Dindo classification, hospital length of stay, ICU admission and death).

This study has merit in being able to achieve a randomized comparison of surgical techniques. It is well accepted that cervical anastomoses have higher leak rates compared to intrathoracic anastomoses. Notwithstanding, contrary to popular dogma, the severity of leaks was not higher in the intrathoracic group. There is some difficulty generalizing this data to our own practice, as the transhiatal MIE (which is most commonly performed at Penn) was not studied in this RCT, and perhaps avoiding entry into the chest accounts for more benefits.

Bottom line: In a randomized controlled trial of minimally invasive transthoracic esophagectomy, an intrathoracic anastomosis, as opposed to a cervical anastomosis, was associated with significantly lower rates of anastomotic leak requiring intervention, with similar leak severity, and improved overall morbidity.

Synopsis: The Milan criteria, first introduced by Mazzaferro et al. in 1996¹, is an important set of criteria used to determine who is eligible to receive liver transplantation as a treatment for hepatocellular carcinoma (HCC). Briefly, the criteria states that patients with one lesion less than 5 cm or up to 3 lesions smaller than 3 cm each without macroscopic vascular invasion can receive orthotopic liver transplantation. While that study had its limitations, including a shockingly small number of patients (less than 50), it nonetheless established a useful principle that guides the use of liver transplantation for HCC across the entire world today: larger pre-transplant HCC volume is inversely related to post-transplant recurrence-free and overall survival. Since then, there have been numerous other criteria developed (UCSF, Pittsburgh, Navarra, Valencia, Shanghai, etc.) used to decide which patients with HCC can receive a liver transplant. While many will argue which criteria is “best”, no one will argue the underlying assumption that more HCC volume means worse survival after transplant. Obtaining precise and accurate pre-transplant imaging is, therefore, paramount in making the best decision regarding allocation of scarce livers, regardless of the criteria used.

In this presented study, Dr. Krumeich and her colleagues begin with a simple observation and build up a concise, concrete, and convincing reason explaining that observation. It seems quite clear that the inception of the project was the simple observation, shown in Figure 1, that HCC patients with TIPS had significantly more occult disease on explant pathology (80% vs 43%). I hope the reader can appreciate from a clinical standpoint why this may be cause for concern – if a group of transplant patients are consistently being undersized with regards to tumor volume, transplant may be doing more harm than good. In fact, the authors make this exact point in showing the drastic difference in recurrence-free and overall survival based on Milan criteria status of explant pathology (Figure 2).

The clear strength of this study is the consideration and exhaustion of all potential theories to explain the observation that TIPS patients, with similar HCC disease status as non-TIPS patients, experience greater rates of occult disease. These theories include 1) differing pathologic evaluation of TIPS vs non-TIPS explant livers, 2) higher propensity of TIPS patients to develop more aggressive HCC, 3) more pre-transplant in non-TIPS patients, and 4) TIPS effects on imaging. The authors present this last theory as the most likely and provide two reasons from their data to support this. First, portal venous thrombosis (PVT), which would similarly alter blood flow, was also a statistically significant factor in having occult disease, even on multivariable analysis. Secondly, the presence of indeterminate lesions contributed to the rate of occult disease, further implicating suboptimal imaging. It would be of interest to extend this study to generate a model predicting the rate of occult HCC pathology (or rate of explant HCC pathology outside Milan criteria), which would likely include PVT, TIPS, and indeterminate lesions. When there is limited organ supply for transplantation, patient selection is vital, and I commend Dr. Krumeich and her co-authors in their identification of patients that may need more thoughtful attention when it comes to their pre-transplant workup and care.

¹Mazzaferro, Vincenzo, et al. "Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis." New England Journal of Medicine 334.11 (1996): 693-700.