Patient Selection Is Critical in Splenic Artery Embolization

ATLANTA -- Splenic artery embolization may be overpromoted for the management of blunt injury to the spleen, according to results of a study presented at the annual meeting of the American Association for the Surgery of Trauma.

"Nonoperative management became more widely adopted in the 1990s, and in the late 1990s, embolization failure rates of 5%-7% were reported. A review by the Western Trauma Association found a failure rate of 13%. Since our participation in the WTA review, we have noticed increased rates of splenic embolization failures," said Dr. Hadley E. Smith of Brown University, Providence, R.I.

Suspecting that trauma surgeons have gone "too far" in the use of embolization, Dr. Smith and her colleagues conducted a retrospective study of 221 patients admitted with blunt splenic injuries from January 2000 to June 2004. They used their trauma registry for splenic injuries and splenic embolization, which was then cross-referenced with their interventional radiology database. After a complete chart review of the embolization patients, all the admission CTs were read and graded according to the splenic injury scale of the AAST. This was done by a radiologist who was blinded to clinical outcome. Similarly, the embolization films were then reviewed by a single interventional radiologist, Dr. Smith said.

Of the 221 patients admitted with splenic injuries, 56 were directed to the operating room. Of the 165 managed nonoperatively, 124 were managed prospectively and 41 went to embolization. Most of the grade 1 and 2 injuries were managed nonoperatively, with the operation rate paralleling the increase in injury severity. Dr. Smith said that 40% of grade 3 injuries, 38% of grade 4 injuries, and 40% of grade 5 injuries ultimately required splenectomy. This gave a total nonoperative failure rate of 14%, with most of those involving embolization patients who had a total failure rate across the grades of 27%, she said.

Among patients with either no or small amounts of hemoperitoneum, the failure rate was 8% (a single patient with a grade 4 injury). Among patients with moderate or large amounts of hemoperitoneum, the failure rate was 36%; all had grade 3, 4, or 5 injuries, Dr. Smith reported. "In trying to predict failures clinically, we ... learned that four patients undergoing embolization experienced at least one bout of transient hypotension in the ED, and three of those eventually required splenectomy. CT scans predicted a high risk of failure among patients with grade 3, 4, or 5 injuries or moderate or large amounts of hemoperitoneum," she said.

The investigators also found that embolization was more likely to fail if extravasation was seen on the angiogram (59% vs. 4%). Coils were more often successful than particles, such as Gelfoam, and main artery embolization succeeded more often than the selective artery procedure.

"Embolization may have salvaged many spleens, but splenectomy was required in 27% of patients. Patient selection is critical to successful management. Any hypotension in the face of a contrast blush probably warrants laparotomy. The combination of high-grade injury and significant hemoperitoneum, or extravasation on angiogram, predicts a high risk of failure and thus warrants a low threshold for splenectomy if bleeding persists. Technical embolization considerations may impact success, but this requires investigation," Dr. Smith concluded.

"I would agree with the hypothesis that we probably have gone too far with splenic artery embolization, and we will swing back toward a central decision, but I think there are several important issues you haven't brought up," Dr. Carl. J. Houser of New Jersey Medical School, Newark, said during discussion. "First of all, interventional radiology is highly operator dependent as well as patient-selection dependent."

"Secondly, the differences in technique are very important. Central embolization may well tend to fail due to persistent hemorrhage, whereas peripheral embolization may fail late due to sepsis and infarction of the spleen," Dr. Houser added. Dr. Smith replied that she had no data relating early and late failures to persistent hemorrhage and complications of sepsis in the left upper quadrant.

Another discussant, Dr. Kimberly A. Davis, said that these data from Dr. Smith support a previous study led by Dr. Andrew B. Peitzman and other members of the Multi-Institutional Trials Committee of the Eastern Association for the Surgery of Trauma. In that study, 49% of patients with moderate hemoperitoneum and 73% of those with large hemoperitoneum ultimately had laparotomy, irrespective of injury grade (J. Trauma 2000;49:177-87).

"When one factors in the associated grade of injury, the failure rates were even higher, with 12% of grade 4 injuries and less than 5% of grade 5 injuries with large hemoperitoneum being successfully managed nonoperatively," said Dr. Davis of Loyola University Medical Center, Maywood, Ill. "Your management strategy of incorporating embolization has significantly improved upon those numbers with your failure rate of 36% in this group of patients."

"Based on your data and Dr. Peitzman's most recent paper, I think it's unequivocal that we have gone too far," said Dr. Donald D. Trunkey, of Oregon Health and Science University, Portland. "You're going to have to study spleen function after embolization to see if even those small numbers of successes really prevent late thromboembolization and splenic infarction."