In the Journals

ACS model predicts pancreatic fistula risk before surgery

Researchers have developed a modified Fistula Risk Score based on data from the American College of Surgeons National Surgical Quality Improvement Program, or NSQIP, which can preoperatively predict the risk for anastomotic leak after pancreas resection and help to prevent the development of pancreatic fistula.

The original pancreatic Fistula Risk Score was developed in 2011 and has been validated in several large studies, but was developed from single-institutional data and uses operative variables that are difficult to measure.

“While the original risk model focused on intraoperative factors surgeons couldn’t determine ahead of time, this model employs several variables that can be used to stratify risk preoperatively so surgeons can determine which patients are at high risk and take steps in the OR that prevent or modify risk of leak,” Marshall S. Baker, MD, MBA, FACS, an associate clinical professor of surgery at NorthShore University HealthSystem in Evanston, Ill., said in a press release.

To overcome the original risk score’s limitations, Baker and colleagues modified it by incorporating variables that are “more readily determined before resection” using NSQIP data.

“NSQIP gives us a much broader base from which to develop risk-predictive models,” as it provides “a tightly controlled data set with well-defined variables and it accrues information from patient populations treated in low, medium and high-volume centers across the country and around the world,” Baker said in the press release.

Further, Clifford Ko, MD, MS, MSHS, FACS, director of the ACS Division of Research and Optimal Patient Care, which administers NSQIP, added that the program provides “highly accurate and reliable patient, clinical, and surgical data [that] is also generalizable and can be applied to patients who are treated for pancreatic cancer in many types of surgery centers in the community.”

The investigators evaluated NSQIP Pancreatic Demonstration Project (PDP) data on 1,731 patients who underwent pancreaticoduodenectomy resection at 43 hospitals between 2011 and 2012, then identified and validated five predictors of clinically relevant postoperative pancreatic fistula. These included male sex (OR = 2.15), BMI at or above 25 (OR = 1.55), preoperative total bilirubin below 2 mg/dL (OR = 2.07), small pancreatic ductal diameter (OR = 4.63 for < 3 mm; OR = 2.89 for 3-6 mm) and soft gland texture (OR = 2.71).

Then they developed a 10-point model and further validated it using PDP data from 2014. The C-statistic of the receiver operating characteristic curve was 0.7 for the testing cohort, 0.7 for the internal validation cohort and 0.62 for the external validation cohort.

In addition to identifying at-risk patients before surgery, the investigators noted that the modified score can serve as a benchmark to help surgeons better understand the quality of care they are providing.

“If surgeons take care of high-risk patient populations but their fistula rates are low, they can substantiate that they are achieving the quality of care they should. If they are caring for low-risk patient populations but fistula rates are high, they can develop local strategies to improve fistula prevention and treatment,” Baker said in the press release.

Further, the model will inform future studies of pancreatic fistula, he added.

“We are able to use the risk score to assess how well new interventions and treatment strategies minimize the risk of fistula. Such clinical trials will be easier to do because the NSQIP platform gives us a dataset that accrues study populations quickly and broadly across a wide variety of surgical practices,” he concluded. – by Adam Leitenberger

Disclosures: The researchers report no relevant financial disclosures.

Researchers have developed a modified Fistula Risk Score based on data from the American College of Surgeons National Surgical Quality Improvement Program, or NSQIP, which can preoperatively predict the risk for anastomotic leak after pancreas resection and help to prevent the development of pancreatic fistula.

The original pancreatic Fistula Risk Score was developed in 2011 and has been validated in several large studies, but was developed from single-institutional data and uses operative variables that are difficult to measure.

“While the original risk model focused on intraoperative factors surgeons couldn’t determine ahead of time, this model employs several variables that can be used to stratify risk preoperatively so surgeons can determine which patients are at high risk and take steps in the OR that prevent or modify risk of leak,” Marshall S. Baker, MD, MBA, FACS, an associate clinical professor of surgery at NorthShore University HealthSystem in Evanston, Ill., said in a press release.

To overcome the original risk score’s limitations, Baker and colleagues modified it by incorporating variables that are “more readily determined before resection” using NSQIP data.

“NSQIP gives us a much broader base from which to develop risk-predictive models,” as it provides “a tightly controlled data set with well-defined variables and it accrues information from patient populations treated in low, medium and high-volume centers across the country and around the world,” Baker said in the press release.

Further, Clifford Ko, MD, MS, MSHS, FACS, director of the ACS Division of Research and Optimal Patient Care, which administers NSQIP, added that the program provides “highly accurate and reliable patient, clinical, and surgical data [that] is also generalizable and can be applied to patients who are treated for pancreatic cancer in many types of surgery centers in the community.”

The investigators evaluated NSQIP Pancreatic Demonstration Project (PDP) data on 1,731 patients who underwent pancreaticoduodenectomy resection at 43 hospitals between 2011 and 2012, then identified and validated five predictors of clinically relevant postoperative pancreatic fistula. These included male sex (OR = 2.15), BMI at or above 25 (OR = 1.55), preoperative total bilirubin below 2 mg/dL (OR = 2.07), small pancreatic ductal diameter (OR = 4.63 for < 3 mm; OR = 2.89 for 3-6 mm) and soft gland texture (OR = 2.71).

Then they developed a 10-point model and further validated it using PDP data from 2014. The C-statistic of the receiver operating characteristic curve was 0.7 for the testing cohort, 0.7 for the internal validation cohort and 0.62 for the external validation cohort.

PAGE BREAK

In addition to identifying at-risk patients before surgery, the investigators noted that the modified score can serve as a benchmark to help surgeons better understand the quality of care they are providing.

“If surgeons take care of high-risk patient populations but their fistula rates are low, they can substantiate that they are achieving the quality of care they should. If they are caring for low-risk patient populations but fistula rates are high, they can develop local strategies to improve fistula prevention and treatment,” Baker said in the press release.

Further, the model will inform future studies of pancreatic fistula, he added.

“We are able to use the risk score to assess how well new interventions and treatment strategies minimize the risk of fistula. Such clinical trials will be easier to do because the NSQIP platform gives us a dataset that accrues study populations quickly and broadly across a wide variety of surgical practices,” he concluded. – by Adam Leitenberger

Disclosures: The researchers report no relevant financial disclosures.