PHILADELPHIA — Some patients with abdominal aortic aneurysms are simply too sick to safely undergo aneurysm repair, be it by open surgery or with an endovascular procedure. Evidence is now starting to accumulate on how to identify patients for whom aneurysm repair is too risky.
New data suggest that only a few patients, perhaps less than 3%, fall into the high-risk group that shouldn't undergo repair of an abdominal aortic aneurysm (AAA). And new findings also indicate that, contrary to prior belief, it's the fittest patients who gain the biggest advantage from undergoing endovascular aneurysm repair (EVAR) instead of open surgery.
“The trend is toward the fitter the patient, the more EVAR may benefit over open repair,” Dr. Roger M. Greenhalgh said at the Vascular Annual Meeting, sponsored by the Society for Vascular Surgery. “At the other end, in extremely unfit patients, you eventually get to a point where the unfitness is so great that EVAR won't help at all,” added Dr. Greenhalgh, professor of surgery and head of vascular surgery at Imperial College, London. “It's sensible to use a scoring method to assess fitness.”
One formula for measuring preoperative fitness was derived empirically by looking at all of the AAA repair patients in the more than 41 million patients of the Medicare data set from 2000 to 2004. In this group were 39,792 patients who underwent an elective, first-time AAA repair using EVAR. Overall in this group, the rate of death during the first 30 days after treatment was 1.73%, said Dr. K. Craig Kent, chief of vascular surgery at New York-Presbyterian Hospital.
Use of EVAR rose throughout the 5 years examined, from 1,500 patients in 2000 to 12,000 in 2004. And the 30-day mortality rate was cut in half, from 2.5% in 2000 down to 1.25% in 2004.
Based on a multivariate analysis that assessed the role of a variety of comorbidities in 30-day mortality, Dr. Kent and his associates calculated a very preliminary scoring system to assess the risk of perioperative death that a patient faces from EVAR.
Renal failure emerged as the single most important comorbidity, scoring five points in the system Dr. Kent presented. (See upper part of box.)
When this comorbidity scoring system was applied to the Medicare cohort, scores could be associated with specific perioperative mortality rates. (See lower part of box.)
Finally, Dr. Kent and his associates selected a representative high-risk score of 9, linked with a 9.3% risk of death, to determine how many of the Medicare patients met or exceeded this arbitrarily selected high-risk threshold. They found that 2.3% of the more than 39,000 patients in the group had a score of 9 or more. The remaining 97.7% of the AAA patients had a lower score and hence a lower risk of 30-day death.
Until now, “we had thought that high-risk patients might be 20%–30%” of all patients with an AAA, Dr. Kent said.
A different approach to risk assessment was tested by Dr. Greenhalgh and his associates using data they had collected on patients in the landmark EVAR trial 1 (Lancet 2005;365:2179–86) and EVAR trial 2 (Lancet 2005;365:2187–92). EVAR trial 1 randomized nearly 1,100 patients to EVAR or open surgical repair; EVAR trial 2 included 338 patients judged unfit for open surgery who were then randomized to EVAR or to no repair.
The risk assessment tool they used was a modified form of a previously reported scoring system that had been developed to assess perioperative mortality in patients undergoing vascular surgery by researchers at Erasmus Medical Center, Rotterdam, Netherlands (Arch. Intern. Med. 2005;165:898–904).
The modified Rotterdam formula is called the customized probability index (CPI), and uses seven clinical characteristics: renal dysfunction (defined as a serum creatinine level of 2 mg/dL or greater) adds 16 points, uncontrolled heart failure adds 14 points, ischemic heart disease adds 13 points, hypertension adds 7 points, and chronic pulmonary disease (defined as a forced expiratory volume of less than 60% of predicted) also adds 7 points. Treatment with either of two medications was considered to cut the mortality risk and therefore subtracts points. Treatment with a β-blocker subtracts 15 points, and treatment with a statin pares 10 points off the total.
The result is a CPI point total that can range from −25 to +57, said Louise C. Brown, a statistician who works with Dr. Greenhalgh at Imperial College.
When the researchers applied the CPI to 1,174 patients in EVAR Trial 1 studies, they found that 47% fell into a category with good fitness for surgery, having a CPI score ranging from −25 to 0. Another 26% had a moderate CPI score of 1–10, and 27% had poor fitness, with a score of 11 to 36. The average CPI score for all patients was 3.7 in EVAR trial 1, and it was 10.1 in EVAR trial 2.