New Tool Gives Surgeons Realistic Risk Assessments for Laparoscopic Sleeve Gastrectomy Patients

Physicians treating individuals affected by obesity who are contemplating laparascopic sleeve gastrectomy (LSG) have a new risk assessment tool available. The user-friendly online risk calculator uses a robust statistical model and can offer an accurate preoperative risk assessment, based on updated outcome data.

Drawing from the American College of Surgeons’ National Surgical Quality Improvement Program (ACS-NSQIP), Cleveland Clinic Clinical Scholar Ali Aminian, MD and collaborators developed the risk calculator, presenting findings on November 4, 2014 in Boston, MA during Obesity Week.

The LSG risk calculator uses a composite primary endpoint that factors in 14 serious adverse events, as well as 30-day mortality. Earlier models, Aminian noted, were limited by drawing on old data, reflecting a higher complication rate than currently exists for bariatric procedures. Further, older calculators combined open and laparoscopic procedures and did not include a large number of variables.

In contrast, using the ACS-NSQUIP data allowed investigators to draw on 5872 LSG cases performed in 2012 at 374 participating sites. Initially, 52 baseline variables were included in univariate and multivariate regression analysis that examined relationships between baseline patient characteristics and 14 serious adverse events. These included such serious complications as cardiac events, stroke, organ failure, significant infection, pulmonary embolism, sepsis, and death. Less serious adverse events, such as uncomplicated urinary tract infection, were not included in the model.

Initial examination of the LSG procedures performed in 2012 showed an overall mortality rate of 0.05% and a 2.38% composite adverse event rate. When odds ratios (ORs) for the 52 initial variables were analyzed, 7 factors were found to be most associated with increased risk of complications and were included in the final risk calculator.

Of these 7 variables, symptomatic or new-onset congestive heart failure and steroid use for a chronic condition accounted for the highest increased risk. These had ORs of 6.23 (95% CI 1.25 — 31.07) and 5.00 (95% CI 2016 – 1.05), respectively. Male sex, diabetes, elevated preoperative serum total bilirubin, BMI, and preoperative hematocrit level all were also associated with greater risk of adverse events or death, with much lower ORs. These ranged from 1.68 for male sex down to 0.95 for hematocrit level, and were all included in the final calculator.

The model was then tested for goodness of fit, accuracy, and specificity by application against the 2011 data. The final risk scoring calculator has good accuracy (c statistic 0.68), and is now available online through the Cleveland Clinic’s public risk calculator portal.

Questioning after the presentation, led by Michigan State University’s Wayne English, MD, highlighted the difficulty of constructing a user-friendly model that uses robust data but still accounts for variables likely to impact surgical outcomes. For example, the ACS-NSQUIP data do not specify the reason for steroid use, nor do they account for severity of comorbidities. Both Aminian and questioners noted that neither previous history of thromboembolism nor history of obstructive sleep apnea are available from ACS-NSQUIP. Moreover, Aminian emphasized that characteristics of the surgical center and the surgeon were not captured here and should be taken into account when assessing risk.