Model improves risk prediction for distal gastrectomy

Last Updated: 2015-03-10

By Will Boggs MD

NEW YORK (Reuters Health) - A risk model based on a score of risk factors accurately predicts 30-day and operative mortality after distal gastrectomy, researchers from Japan report.

Dr. Nobuhiro Kurita, from the Japanese Society of Gastroenterological Surgery, Tokyo, and colleagues recently reported that 30-day mortality estimates were lower than actual perioperative mortality after gastrectomy.

To address this problem, they formulated risk models for 30-day and operative mortality associated with distal gastrectomy using data from 33,917 gastric cancer cases in Japan's National Clinical Database.

Thirty-day mortality came to 0.52%, in-hospital mortality came to 1.16%, and operative mortality came to 1.2%, according to the February 25 Annals of Surgery online report.

Based on variables that were significantly more frequent in the 30-day and operative mortality groups than in the nonmortality group, the researchers developed a 17-variable model for 30-day mortality and a 21-variable model for operative mortality.

C-indices of 0.785 for 30-day mortality and 0.798 for operative mortality supported the good predictive abilities of the models, the researchers say.

"This system will contribute to an improved quality control in surgical practice and it should also be useful for counseling and for obtaining informed consent from patients," the authors conclude.

"Although our analysis used the nationwide database, the study population was limited to a single race," they cautioned. "Therefore, our results should be evaluated on the basis of comparisons with patients from other countries using the same variables and definitions. Thus, we are currently planning a mutual collaboration with the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP)."

Dr. Enver Ilhan, from zmir Bozyaka Training and Research Hospital, zmir, Turkey, who recently reviewed the efficacy of laparoscopy-assisted distal gastrectomy for gastric cancer, told Reuters Health by email, "Using this model may be difficult in daily practice. But this study is valuable to understand there are a lot of parameters related to surgical complications."

"If we determine some of them, we may be able to rectify them preoperatively," he said. "For example, nutritional status, laboratory variables, and comorbidity status. In addition, hospital and surgeon volume are relevant to morbidity and mortality."

"Surgical outcomes will improve when we take into account patient-related factors, not just stage of the disease," Dr. Ilhan concluded.

Dr. Kurita did not respond to a request for comments.

The authors reported no external funding or disclosures.

SOURCE: http://bit.ly/1HwmxsD

Ann Surg 2015.