J Gastrointest Surg (2014) 18:1917–1928 DOI 10.1007/s11605-014-2634-x
2014 SSAT POSTER PRESENTATION
Defining Perioperative Risk after Hepatectomy Based on Diagnosis and Extent of Resection Christopher R. Shubert & Elizabeth B. Habermann & Mark J. Truty & Kristine M. Thomsen & Michael L. Kendrick & David M. Nagorney
Received: 19 May 2014 / Accepted: 18 August 2014 / Published online: 9 September 2014 # 2014 The Society for Surgery of the Alimentary Tract
Abstract Outcomes after hepatectomy have been assessed incompletely and have not been stratified by both extent of resection and diagnosis. We hypothesized that operative risk is better assessed by stratifying diagnoses into low- and high-risk categories and extent of resection into major and minor resection categories to more accurately evaluate the outcomes after hepatectomy. ACS-NSQIP was reviewed for 30-day operative mortality and major morbidity after partial hepatectomy (PH), left hepatectomy (LH), right hepatectomy (RH), and trisectionectomy (TS). Mortality was reviewed per diagnosis. “High Risk” was defined as the diagnoses associated with the greatest mortality. Major and minor resections were defined by comparison of outcomes for extent of resection by univariate analysis. Chi-square tests, t tests, Fisher’s exact tests, and multivariable logistic regression were utilized to compare the outcomes across groups. Among the 7,043 patients, the greatest mortality was observed with hepatocellular carcinoma (5.2 %) and cholangiocarcinoma (8.2 %), either intra- or extrahepatic, which were classified “High Risk”. Metastatic disease, benign neoplasms, and gallbladder cancer had a mortality rate of 1.3, 0.5, and 1.0 %, respectively, and were classified “Low Risk”. PH and LH were similar statistically for operative mortality and major morbidity within respective diagnosis risk groups (Low Risk: PH vs. LH and High Risk: PH vs. LH; all p>0.05) and were defined as “Minor Resections”. Similarly, RH and TS had similar operative mortality and major morbidity within respective diagnosis risk groups (Low Risk: RH vs. TS and High Risk: RH vs. TS; all p>0.05) and were defined as “Major Resections”. Risks of major morbidity and mortality increased for both diagnoses and the extent of resection. With minor resections, mortality and major morbidity were 5 and 1.6 times greater respectively for high-risk diagnosis than for low-risk diagnosis. With major resections, mortality and major morbidity were 4 and 1.6 times greater, respectively, for high-risk diagnoses than low-risk diagnoses. With low-risk diagnoses, mortality and major morbidity were 2.9 and 1.7 times greater, respectively, for major resections than minor resections (p48 h, pneumonia, acute renal failure, progressive renal insufficiency, DVT/thrombophlebitis, pulmonary embolism, and return to OR. Complications
J Gastrointest Surg (2014) 18:1917–1928
defined as “Minor Morbidity” included superficial SSI and urinary tract infection. Thirty-day operative mortality per diagnosis was reviewed. Diagnoses associated with the greatest mortality were defined “High Risk”. Similarly, diagnoses with the lowest mortality were defined “Low Risk”. Operative mortality and major morbidity were reviewed for each operation within each respective diagnosis-risk group. Resections with statistically similar outcomes were grouped to define the extent of resection-risk groups. For the purposes of this study, “prolonged length of stay” was defined as being equal to or greater than the seventy-fifth percentile (fourth quartile) of the number of days from operation to discharge within the respective extent of resection category. All mortality data reported is 30-day mortality. Univariate and multivariable analyses were then performed for each combination of the extent of resection and diagnosisrisk groups. Multivariable analysis controlled for variables that were statistically significant and clinically relevant, while stratified by the extent of resection and diagnosis-risk groups. Multivariable models were adjusted for diagnosis-risk group, age, BMI, diabetes, functional status, preoperative chemotherapy within 30 days of operation, preoperative serum albumin, preoperative platelet count, preoperative total bilirubin, preoperative INR, preoperative serum creatinine, ASA class, and operative time quartile for respective extent of resection.
1919
cancer (1,626), cholangiocarcinoma (452), metastasis (3,876), benign neoplasms (787), and gallbladder cancer (302). About 51.1 % of all patients were male. Average age was 59 years. Definition of Diagnosis-Risk Groups The 30-day operative mortality rate for all patients was 2.6 % (n=181). Diagnoses associated with the greatest mortality rates was cholangiocarcinoma (8.2 %) and hepatocellular carcinoma (5.2 %). Metastases, benign neoplasms, and gallbladder cancer 30-day operative mortality were 1.3, 0.5, and 1.0 %, respectively (all p0.05). Therefore, PH and LH were defined as “Minor Hepatectomy”, and RH and TS were defined as “Major Hepatectomy” (Table 1).
Statistical Analysis
Minor Hepatectomy
Mortality and major morbidity were compared across individual diagnoses using Chi-square tests. Stratifying by highversus low-risk diagnosis group, Chi-square tests and Fisher’s exact test were then performed to compare mortality and major morbidity between the extents of resections. Chisquare tests, Fisher’s exact tests, t tests, and multivariable logistic regression were performed to assess the impact of diagnosis and the extent of resection on outcomes. Chisquare tests were used to assess the relationship of complications on mortality. Univariate and multivariate regression statistics were considered significant at p10 % loss of body weight in last 6 months, bleeding disorders, elevated creatinine, low serum albumin, elevated bilirubin, elevated transaminase, elevated alkaline phosphatase, preoperative platelet count less than 100,000, and ASA class 3 or 4. High-risk patients were less likely obese or had received preoperative chemo- or radiotherapy (Table 1).
Descriptive Statistics
Postoperative Outcomes
Ten thousand three hundred (10,300) patients undergoing hepatectomy were identified in NSQIP 2005–2011 PUF. Three thousand two hundred fifty-seven (3,257) patients were excluded. Seven thousand forty-three (7,043) patients met the inclusion criteria: hepatocellular
High-risk diagnoses undergoing minor hepatectomy had significantly greater mortality and major morbidity compared to their low-risk diagnoses undergoing minor hepatectomy. Specifically, high-risk diagnoses had 5 times the mortality and 1.6 times the major morbidity as low-risk diagnoses.
1920
J Gastrointest Surg (2014) 18:1917–1928
Fig. 1 Major morbidity and mortality by the extent of resection and diagnosis-risk groups: Defining “Major” and “Minor” resection categories
High-risk patients were more likely to receive intraoperative blood transfusion and twice as likely to receive 5+ units. They were also significantly more likely to experience superficial and organ space surgical site infections, unplanned intubation, ventilator >48 h, acute renal failure, cardiac arrest requiring CPR, DVT, sepsis or septic shock, return to OR, and prolonged length of stay (7+ days). High-risk patients stayed on average 1.4 days longer than their low-risk counterparts. High-risk patients were also less likely to be discharge to home or home facility and more likely
to be discharge to a rehabilitation or skilled care facility. There were no statistically significant differences in readmission rate (Table 2). Multivariable Analysis Multivariable analysis of patients undergoing minor hepatectomy proved high-risk diagnoses as an independent risk factor for mortality, major morbidity, and prolonged length of stay (Table 3).
Characteristic/feature, no. (column %) Gender* Male Female Race/ethnicity (all years), no. (col %) Non-hispanic white Hispanic white Black or African American Asian, Native Hawaiian or Pacific Islander American Indian or Alaska Native Other or Unknown Age Mean (SD)
2014 SSAT POSTER PRESENTATION
Defining Perioperative Risk after Hepatectomy Based on Diagnosis and Extent of Resection Christopher R. Shubert & Elizabeth B. Habermann & Mark J. Truty & Kristine M. Thomsen & Michael L. Kendrick & David M. Nagorney
Received: 19 May 2014 / Accepted: 18 August 2014 / Published online: 9 September 2014 # 2014 The Society for Surgery of the Alimentary Tract
Abstract Outcomes after hepatectomy have been assessed incompletely and have not been stratified by both extent of resection and diagnosis. We hypothesized that operative risk is better assessed by stratifying diagnoses into low- and high-risk categories and extent of resection into major and minor resection categories to more accurately evaluate the outcomes after hepatectomy. ACS-NSQIP was reviewed for 30-day operative mortality and major morbidity after partial hepatectomy (PH), left hepatectomy (LH), right hepatectomy (RH), and trisectionectomy (TS). Mortality was reviewed per diagnosis. “High Risk” was defined as the diagnoses associated with the greatest mortality. Major and minor resections were defined by comparison of outcomes for extent of resection by univariate analysis. Chi-square tests, t tests, Fisher’s exact tests, and multivariable logistic regression were utilized to compare the outcomes across groups. Among the 7,043 patients, the greatest mortality was observed with hepatocellular carcinoma (5.2 %) and cholangiocarcinoma (8.2 %), either intra- or extrahepatic, which were classified “High Risk”. Metastatic disease, benign neoplasms, and gallbladder cancer had a mortality rate of 1.3, 0.5, and 1.0 %, respectively, and were classified “Low Risk”. PH and LH were similar statistically for operative mortality and major morbidity within respective diagnosis risk groups (Low Risk: PH vs. LH and High Risk: PH vs. LH; all p>0.05) and were defined as “Minor Resections”. Similarly, RH and TS had similar operative mortality and major morbidity within respective diagnosis risk groups (Low Risk: RH vs. TS and High Risk: RH vs. TS; all p>0.05) and were defined as “Major Resections”. Risks of major morbidity and mortality increased for both diagnoses and the extent of resection. With minor resections, mortality and major morbidity were 5 and 1.6 times greater respectively for high-risk diagnosis than for low-risk diagnosis. With major resections, mortality and major morbidity were 4 and 1.6 times greater, respectively, for high-risk diagnoses than low-risk diagnoses. With low-risk diagnoses, mortality and major morbidity were 2.9 and 1.7 times greater, respectively, for major resections than minor resections (p48 h, pneumonia, acute renal failure, progressive renal insufficiency, DVT/thrombophlebitis, pulmonary embolism, and return to OR. Complications
J Gastrointest Surg (2014) 18:1917–1928
defined as “Minor Morbidity” included superficial SSI and urinary tract infection. Thirty-day operative mortality per diagnosis was reviewed. Diagnoses associated with the greatest mortality were defined “High Risk”. Similarly, diagnoses with the lowest mortality were defined “Low Risk”. Operative mortality and major morbidity were reviewed for each operation within each respective diagnosis-risk group. Resections with statistically similar outcomes were grouped to define the extent of resection-risk groups. For the purposes of this study, “prolonged length of stay” was defined as being equal to or greater than the seventy-fifth percentile (fourth quartile) of the number of days from operation to discharge within the respective extent of resection category. All mortality data reported is 30-day mortality. Univariate and multivariable analyses were then performed for each combination of the extent of resection and diagnosisrisk groups. Multivariable analysis controlled for variables that were statistically significant and clinically relevant, while stratified by the extent of resection and diagnosis-risk groups. Multivariable models were adjusted for diagnosis-risk group, age, BMI, diabetes, functional status, preoperative chemotherapy within 30 days of operation, preoperative serum albumin, preoperative platelet count, preoperative total bilirubin, preoperative INR, preoperative serum creatinine, ASA class, and operative time quartile for respective extent of resection.
1919
cancer (1,626), cholangiocarcinoma (452), metastasis (3,876), benign neoplasms (787), and gallbladder cancer (302). About 51.1 % of all patients were male. Average age was 59 years. Definition of Diagnosis-Risk Groups The 30-day operative mortality rate for all patients was 2.6 % (n=181). Diagnoses associated with the greatest mortality rates was cholangiocarcinoma (8.2 %) and hepatocellular carcinoma (5.2 %). Metastases, benign neoplasms, and gallbladder cancer 30-day operative mortality were 1.3, 0.5, and 1.0 %, respectively (all p0.05). Therefore, PH and LH were defined as “Minor Hepatectomy”, and RH and TS were defined as “Major Hepatectomy” (Table 1).
Statistical Analysis
Minor Hepatectomy
Mortality and major morbidity were compared across individual diagnoses using Chi-square tests. Stratifying by highversus low-risk diagnosis group, Chi-square tests and Fisher’s exact test were then performed to compare mortality and major morbidity between the extents of resections. Chisquare tests, Fisher’s exact tests, t tests, and multivariable logistic regression were performed to assess the impact of diagnosis and the extent of resection on outcomes. Chisquare tests were used to assess the relationship of complications on mortality. Univariate and multivariate regression statistics were considered significant at p10 % loss of body weight in last 6 months, bleeding disorders, elevated creatinine, low serum albumin, elevated bilirubin, elevated transaminase, elevated alkaline phosphatase, preoperative platelet count less than 100,000, and ASA class 3 or 4. High-risk patients were less likely obese or had received preoperative chemo- or radiotherapy (Table 1).
Descriptive Statistics
Postoperative Outcomes
Ten thousand three hundred (10,300) patients undergoing hepatectomy were identified in NSQIP 2005–2011 PUF. Three thousand two hundred fifty-seven (3,257) patients were excluded. Seven thousand forty-three (7,043) patients met the inclusion criteria: hepatocellular
High-risk diagnoses undergoing minor hepatectomy had significantly greater mortality and major morbidity compared to their low-risk diagnoses undergoing minor hepatectomy. Specifically, high-risk diagnoses had 5 times the mortality and 1.6 times the major morbidity as low-risk diagnoses.
1920
J Gastrointest Surg (2014) 18:1917–1928
Fig. 1 Major morbidity and mortality by the extent of resection and diagnosis-risk groups: Defining “Major” and “Minor” resection categories
High-risk patients were more likely to receive intraoperative blood transfusion and twice as likely to receive 5+ units. They were also significantly more likely to experience superficial and organ space surgical site infections, unplanned intubation, ventilator >48 h, acute renal failure, cardiac arrest requiring CPR, DVT, sepsis or septic shock, return to OR, and prolonged length of stay (7+ days). High-risk patients stayed on average 1.4 days longer than their low-risk counterparts. High-risk patients were also less likely to be discharge to home or home facility and more likely
to be discharge to a rehabilitation or skilled care facility. There were no statistically significant differences in readmission rate (Table 2). Multivariable Analysis Multivariable analysis of patients undergoing minor hepatectomy proved high-risk diagnoses as an independent risk factor for mortality, major morbidity, and prolonged length of stay (Table 3).
Characteristic/feature, no. (column %) Gender* Male Female Race/ethnicity (all years), no. (col %) Non-hispanic white Hispanic white Black or African American Asian, Native Hawaiian or Pacific Islander American Indian or Alaska Native Other or Unknown Age Mean (SD)
