People with left ventricular assist devices (LVADs) often underwent non-cardiac surgery -- and faced higher-than-average morbidity and mortality in doing so, a study showed.
One in six Medicare patients who had LVADs implanted in 2012-2017 subsequently underwent non-cardiac surgery through the end of 2017, according to CMS records analyzed by Paulino Alvarez, MD, of Cleveland Clinic, and colleagues.
For these patients, perioperative MACE (the composite of all-cause mortality, ischemic stroke, or intracerebral hemorrhage) was "frequent" after non-cardiac surgery, the group .
Emergent or urgent surgeries resulted in a 16.9% incidence of perioperative major adverse cardiovascular events (MACE). These procedures were associated with more early mortality (adjusted HR 8.78, 95% CI 7.20-10.72) and late mortality alike (adjusted HR 1.71, 95% CI 1.53-1.90) compared with LVAD patients who did not have non-cardiac surgery.
Elective surgeries were associated with a 7.1% rate of perioperative MACE. Like emergent procedures, the elective ones were tied to higher mortality both in the early postoperative period (adjusted HR 2.65, 95% CI 1.74-4.03) and later during follow-up (adjusted HR 1.29, 95% CI 1.07-1.56).
Early mortality counted deaths less than 60 days after surgery, whereas late mortality included deaths after 60 days or more.
Overall, postoperative mortality reached 8.7% of people within 30 days of non-cardiac surgery, and 37.2% were readmitted. More than one in three surgical patients received blood transfusion.
In general, the study's results came as no surprise, commented Jennifer Cowger, MD, MS, of Henry Ford Health System in Detroit, who was not involved with the analysis.
"Patients on LVAD support often have multiple comorbidities and are prone to the same medical maladies of non-LVAD patients such as gallbladder/abdominal, orthopedic/trauma, and malignant issues," she told 51˶ in an email. She added that the flow delivered by the LVAD is often less pulsatile than normal and does not increase to the extent expected during surgery and other stressors.
Difficulties in surgery for LVAD patients include clinical monitoring of the person's oxygenation and blood pressure and the need for mindful surgical incisions that avoid damage to the LVAD pump and its components, according to Cowger.
"To this end, experts in our field strongly recommend any surgery on an LVAD patient be performed at their center of implant [with a cardiac anesthesiologist] who is familiar with the intricacies of LVAD support. Careful monitoring of anticoagulation before and after surgery is important, and flows may need to be adjusted by advanced heart failure specialists perioperatively based on patient volume status and clinical state," she said.
The cohort study was based on the CMS 100% Medicare Provider and Analysis Review Part A files.
Included were 8,118 patients (mean age 63, 80% men) with advanced heart failure who survived their index LVAD placement hospitalization. There was no difference in age between patients who underwent non-cardiac surgery and those who did not.
Median time from LVAD implantation to surgery was 309 days.
The number of LVAD recipients receiving non-cardiac surgeries increased over time (from 64 in 2012 to 304 in 2017).
This rise "likely reflects the growing number of LVADs implanted in the United States and the improved long-term survival of patients on devices," according to Cowger.
Of the 1,326 non-cardiac procedures identified, 75.4% were emergent or urgent and 24.6% elective. General surgery (abdominal, pelvic, and GI procedures) represented the most common procedures (46.2%), followed by thoracic surgery (16.5%) and orthopedic surgery (15.0%).
Vascular and thoracic surgeries were independent predictors of perioperative MACE after urgent and elective procedures. Other predictors of MACE were acute kidney injury and surgery within 6 months of LVAD implantation.
Notably, Alvarez's team had excluded neurological procedures (to avoid stroke as a confounder), breast and gynecological procedures (due to low prevalence), and tracheostomy and gastrostomy (because they are usually done to support the critically ill).
The study was limited by the lack of hemodynamic and echocardiographic data. Investigators were also unable to identify specific LVAD models from their dataset, though they noted that only continuous-flow LVADs were being implanted in the U.S. during the study period.
"Our findings highlight the need for collaborative research in the perioperative management of patients with LVAD to answer important questions, such as management of anticoagulation and antiplatelets in the perioperative period, and the optimum timing of elective non-cardiac surgery after LVAD placement," Alvarez and colleagues said.
The "sobering" MACE rates "should prompt serious discussion for clinicians caring for these complex patients," according to an by Charlotte Rajasingh, MD, and Sherry Wren, MD, both of Stanford University School of Medicine, California.
"Strategically translating [the study] into a conversation about surgical treatment is vital to facilitate shared decision-making in the acute setting for these patients with high risk. The best case/worst case framework for guiding these discussions is one way to engage the patient in decision-making while maintaining attention to the patient's overall values and goals," the editorialists suggested.
Moreover, communication is key for attending physicians, nurses, and medical trainees alike, who should all be on the same page regarding treatment decisions when caring for a patient with acute illness, according to Rajasingh and Wren.
Disclosures
Alvarez and Wren had no disclosures.
Rajasingh disclosed receiving NIH and institutional grants.
Primary Source
JAMA Network Open
Mentias A, et al "Trends, perioperative adverse events, and survival of patients with left ventricular assist devices undergoing noncardiac surgery" JAMA Network Open 2020; DOI: 10.1001/jamanetworkopen.2020.25118.
Secondary Source
JAMA Network Open
Rajasingh CM, Wren SM "Translating outcomes data into improved shared decision-making in patients with left ventricular assist devices" JAMA Network Open 2020; DOI: 10.1001/jamanetworkopen.2020.25673.