Abstract
Background
Resection of the bile duct is required for the treatment of cholangiocarcinoma and is sometimes indicated in resections of liver and gallbladder malignancies. The goal of this retrospective review was to characterize surgical outcomes in patients submitted to bile duct resection for malignancy when additional procedures, specifically hepatic or vascular resections, were performed.
Methods
The American College of Surgeons National Surgical Quality Improvement Program database was searched to identify a total of 747 patients who underwent: (i) biliary-enteric anastomosis (BEA) only (Group 1, n = 266); (ii) BEA with hepatic resection (Group 2, n = 439), or (iii) BEA with hepatic and vascular resection (Group 3, n = 42). Postoperative outcomes were compared and regression-adjusted risk factors were analysed to produce observed and expected (O/E) morbidity and mortality ratios.
Results
The performance of hepatic and vascular resections significantly increased rates of overall morbidity (P < 0.001) and mortality (P = 0.021). Risk-adjusted O/E mortality ratios in Groups 1, 2 and 3 were 1.44 [95% confidence interval (CI) 0.84–2.30], 2.16 (95% CI 1.51–2.98) and 5.92 (95% CI 2.54–11.66), respectively. Multivariate analysis identified Group 2 (P < 0.001) and Group 3 (P = 0.001) status as independent predictors of morbidity, and Group 3 status (P = 0.008) as independently associated with mortality. More than 30% of deaths were associated with pulmonary complications and septic shock.
Conclusions
The addition of hepatic and vascular resections to bile duct resection significantly increased morbidity and mortality. The high O/E mortality ratios for patients in Groups 2 and 3 suggest these outcomes can be improved.
Introduction
The surgical treatment of biliary and gallbladder malignancies continues to evolve as more extensive procedures are performed in efforts to provide complete tumour extirpation with negative margins and promote longterm survival or potential cure. In the case of hilar cholangiocarcinoma, bile duct resection is required for treatment, but the addition of hepatic resection has been shown to improve oncologic outcomes with reasonable rates of morbidity and mortality.1–5
More recently, various authors have reported the use of vascular resection in the treatment of cholangiocarcinoma and gallbladder malignancy with mixed results.6–12 Vascular resection is usually carried out when the primary tumour has invaded the portal vein and/or hepatic artery, necessitating the resection and reconstruction of these structures in order to obtain a negative-margin (R0) resection. These reports derive from specialized units and surgeons with expertise in the management of the reported malignancies and describe both perioperative outcomes and longterm overall survival. With the exception of one report,11 information is limited to single-institution studies with small sample sizes and is subject to biases associated with the reporting institution. In addition, there have been several recent meta-analyses of the outcomes of vascular resection in hilar cholangiocarcinoma.13–15 As a group, these studies show vascular resection to be a feasible operative strategy, but some have reported postoperative outcomes similar to those of less extensive procedures, whereas others have shown increased mortality.
The aims of this study were to examine clinical outcomes in patients undergoing biliary surgery for hepatobiliary and gallbladder malignancies and to determine the effects of additional procedures such as hepatic and vascular resection on postoperative outcomes using data from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP).16 Unlike data derived from specialized hepatobiliary centres, the NSQIP database represents outcomes from a variety of hospital settings with varying levels of volume and expertise and thus its data are more representative of actual practice across the USA. In addition, the results are risk-adjusted based on preoperative clinical information to facilitate the calculation of expected rates of morbidity and mortality. The secondary aim of this study was to identify independent risk factors for increased morbidity and mortality in patients undergoing bile duct resection and reconstruction with or without hepatic or vascular resection. It was hypothesized that hepatic and vascular procedures combined with bile duct reconstruction will increase the postoperative incidences of morbidity and mortality associated with surgery for biliary and gallbladder malignancies. This study sought to quantify the decrement in outcomes in a large population-based dataset and to determine if these outcomes can be improved.
Materials and methods
This study was approved by the Wake Forest University Health Sciences Institutional Review Board and the Protocol Review Committee of the Comprehensive Cancer Center of Wake Forest University.
Data source
The ACS NSQIP is a data-driven, risk-adjusted, outcomes-based programme designed to measure and improve the quality of surgical care.16 Details regarding sampling strategy, data collection protocol, the variables collected and organization have been previously reported.17 This study uses the 2005–2012 Participant Use Files (PUF), which contain 295 Health Insurance Portability and Accountability Act (HIPAA) complaint variables on 543 885 surgical patients.
Study design
Patients older than 16 years of age were identified using Current Procedural Terminology (CPT) codes for biliary-enteric anastomosis (BEA), hepatic resection and vascular resection (Appendix S1, online). The codes were then correlated with International Classification of Diseases, Ninth Revision (ICD-9) codes for biliary and gallbladder malignancies (Appendix S1). Patients were then grouped into three categories based on their CPT codes according to whether they underwent BEA only (Group 1), BEA with hepatic resection (Group 2), or BEA with hepatic and vascular resection (Group 3). All patients in these three groups had a diagnosis equivalent to one of the ICD-9 codes described above.
In order to reduce confounding and create a study population that represented the inherent risks associated with the procedures specifically selected based on CPT codes for this study, patients with CPT codes that referred to major surgical procedures other than those covered in the present three procedure groups were excluded (n = 40). These included patients submitted to partial gastrectomy, partial colectomy or pancreatectomy.
Primary endpoints for this study were 30-day postoperative mortality rates, overall morbidity rates, and specific complication rates stratified by the type of procedure performed. Specific complications analysed included: cardiac arrest requiring cardiopulmonary resuscitation (CPR); deep incisional surgical site infection (SSI); organ space SSI; sepsis or septic shock; unplanned intubation; mechanical ventilation for >48 h; pneumonia; acute renal failure; progressive renal insufficiency; deep vein thrombosis (DVT); pulmonary embolus; any need for return to the operating room; superficial SSI, and urinary tract infection. Definitions of complications are based on the NSQIP Operations Manual.17 Other covariates of interest included age, body mass index (BMI), sex, race, American Society of Anesthesiologists (ASA) physical status class,18 smoking status (current smoker within 1 year), comorbidities (diabetes mellitus, ascites), recent weight loss (>10% of body weight in the 6 months prior to surgery), preoperative laboratory values (serum albumin and total bilirubin), and functional status (independent, partially dependent, totally dependent, unknown).
Statistical analysis
Demographic, preoperative and surgical characteristics were compared by surgical procedure group using analyses of variance (anovas), chi-squared tests and Fisher's exact tests. All additional analyses were then stratified by surgical procedure. Observed and expected (O/E) morbidity and mortality ratios using validated risk-adjustment models were calculated for the overall group and by surgical procedure.19 Observed and expected ratios were calculated using the mortality and morbidity observed, and the expected probabilities of mortality and morbidity for each patient. The expected probabilities of mortality and morbidity are included in the NSQIP database and are determined ‘for all surgical patients based on a hierarchical regression analysis using the patient's preoperative characteristics as the independent or predictive variables’.17 To predict whether each patient would experience an event, sampling from a Bernoulli distribution using the NSQIP probability was performed. The predicted events for the entire sample were summed to obtain an expected number of events. This process was repeated 500 times. The final expected event rate was calculated using the mean of the 500 sampled expected number of events.17 Univariate and multivariate logistic regression models were used to assess the relationships between covariates of interest and mortality and morbidity, controlling for variables that were statistically significant and clinically relevant. Additionally, a backward elimination model strategy was used to determine which surgical complications were associated with mortality in the multivariate model. The criterion to be met in order to remain in the model was a value of 0.2 and P-values of <0.05 were considered to indicate statistical significance. All analyses were performed using sas Version 9.4 (SAS Institute, Inc., Cary, NC, USA).
Results
Demographics
A total of 747 patients who met the CPT code criteria and had ICD-9 diagnoses of hepatobiliary or gallbladder malignancies were identified in the 2012 NSQIP PUF. These included 266 (35.6%) patients who underwent BEA only (Group 1), 439 (58.8%) patients who underwent BEA and hepatic resection (Group 2), and 42 (5.6%) patients who underwent BEA, hepatic resection and vascular resection (Group 3). Table1 presents demographic and preoperative data for the study cohort stratified by procedure. One hundred sixty-three patients in this study cohort did not have race identified. Therefore the denominator for each column of Table1 under ethnicity is 584, 341, 208, and 35, respectively. Significant P-values indicate a difference across all groups.
Table 1.
Demographic, preoperative and postoperative data for patients submitted to biliary-enteric anastomosis (BEA) with and without hepatic or vascular resection
| Characteristic | All patients (n = 747) | BEA only (n = 266) | BEA with hepatic resection (n = 439) | BEA with hepatic and vascular resection (n = 42) | P-value | ||||
|---|---|---|---|---|---|---|---|---|---|
| Preoperative characteristics | |||||||||
| Age, years, median (range) | 67 | (20–89) | 68 | (33–89) | 66 | (20–86) | 63 | (33–82) | <0.001 |
| BMI, kg/m2, median (range) | 26 | (11–54) | 26 | (11–54) | 26 | (11–52) | 25 | (18–42) | 0.538 |
| Male sex, n (%) | 433 | (58%) | 162 | (61%) | 244 | (56%) | 27 | (64%) | 0.210 |
| White ethnicity, n (%) | 446/584 | (76%) | 161/208 | (77%) | 256/341 | (75%) | 29/35 | (83%) | 0.533 |
| ASA class 3 or 4, n (%) | 557 | (75%) | 201 | (76%) | 325 | (74%) | 31 | (74%) | 0.897 |
| DM with oral agents or insulin, n (%) | 134 | (18%) | 58 | (22%) | 73 | (17%) | 3 | (7%) | 0.038 |
| Current smoking, n (%) | 124 | (17%) | 43 | (16%) | 71 | (16%) | 10 | (24%) | 0.434 |
| Loss of >10% of body weight in last 6 months, n (%) | 113 | (15%) | 45 | (17%) | 63 | (14%) | 5 | (12%) | 0.546 |
| Independent functional health status prior to surgery, n (%) | 723 | (97%) | 251 | (94%) | 430 | (98%) | 42 | (100%) | 0.016 |
| Ascites, n (%) | 22 | (3%) | 8 | (3%) | 11 | (3%) | 3 | (7%) | 0.236 |
| Preoperative serum albumin, g/dl, median (range) | 4 | (2–7) | 3 | (2–5) | 4 | (2–7) | 4 | (2–5) | <0.001 |
| Preoperative total bilirubin, mg/dl, median (range) | 1 | (0–15) | 2 | (0–15) | 1 | (0–15) | 2 | (0–9) | 0.036 |
| Chemotherapy for malignancy within 30 days preoperatively, n (%) | 11 | (2%) | 4 | (2%) | 7 | (2%) | 0 | 0.766 | |
| Radiotherapy for malignancy in last 90 days, n (%) | 6 | (1%) | 4 | (2%) | 2 | (1%) | 0 | 0.326 | |
| Postoperative data | |||||||||
| Total operation time, min, median (range) | 362 | (11–1053) | 294 | (11–864) | 394 | (111–1053) | 486 | (249–803) | <0.001 |
| Length of total hospital stay, days, median (range) | 9 | (0–94) | 9 | (2–62) | 9 | (0–94) | 12 | (1–42) | 0.171 |
| Return to operating room, n (%) | 75 | (10%) | 20 | (8%) | 47 | (11%) | 8 | (19%) | 0.053 |
| Superficial SSI, n (%) | 78 | (10%) | 32 | (12%) | 44 | (10%) | 2 | (5%) | 0.325 |
| Deep incisional SSI, n (%) | 24 | (3%) | 8 | (3%) | 16 | (4%) | 0 | 0.4289 | |
| Organ space SSI, n (%) | 134 | (18%) | 24 | (9%) | 99 | (23%) | 11 | (26%) | <0.001 |
| Pneumonia, n (%) | 43 | (6%) | 13 | (5%) | 26 | (6%) | 4 | (10%) | 0.474 |
| Unplanned intubation, n (%) | 62 | (8%) | 16 | (6%) | 43 | (10%) | 3 | (7%) | 0.203 |
| Pulmonary embolism, n (%) | 10 | (1%) | 2 | (1%) | 8 | (2%) | 0 | 0.360 | |
| On ventilator for >48 h, n (%) | 63 | (8%) | 14 | (5%) | 43 | (10%) | 6 | (14%) | 0.041 |
| Progressive renal insufficiency, n (%) | 21 | (3%) | 8 | (3%) | 12 | (3%) | 1 | (2%) | 0.963 |
| Acute renal failure, n (%) | 19 | (3%) | 1 | (0.4%) | 15 | (3%) | 3 | (7%) | 0.007 |
| Urinary tract infection, n (%) | 30 | (4%) | 10 | (4%) | 20 | (5%) | 0 | 0.344 | |
| Cardiac arrest requiring CPR, n (%) | 19 | (3%) | 3 | (1%) | 15 | (3%) | 1 | (2%) | 0.173 |
| Intra- or postoperative bleeding or transfusion, n (%) | 154 | (21%) | 26 | (10%) | 112 | (26%) | 16 | (38%) | <0.001 |
| DVT requiring therapy, n (%) | 22 | (3%) | 3 | (1%) | 16 | (4%) | 3 | (7%) | 0.009 |
| Sepsis, n (%) | 119 | (16%) | 29 | (11%) | 81 | (18%) | 9 | (21%) | 0.018 |
| Septic shock, n (%) | 61 | (8%) | 13 | (5%) | 43 | (10%) | 5 | (12%) | 0.046 |
| At least one surgery complication, n (%) | 402 | (54%) | 109 | (41%) | 264 | (60%) | 29 | (69%) | <0.001 |
| 30-day mortality, n (%) | 61 | (8%) | 17 | (6%) | 36 | (8%) | 8 | (19%) | 0.021 |
ASA, American Society of Anesthesiologists; BMI, body mass index; CPR, cardiopulmonary resuscitation; DM, diabetes mellitus; DVT, deep vein thrombosis; SSI, surgical site infection.
Morbidity
The overall morbidity rate was 53.8%. Complications significantly increased in the unadjusted setting as hepatic and vascular resections were performed in conjunction with bile duct resection (Table1).
Predictors of morbidity for the entire study cohort are shown in Table2. Risk-adjusted O/E ratios for morbidity and mortality by group are shown in Table3.
Table 2.
Univariate and multivariate predictors using logistic regression of 30-day morbidity in patients submitted to biliary-enteric anastomosis (BEA) with and without hepatic or vascular resection
| Univariate models | Multivariate modela | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Ageb | 0.99 | 0.98–1.01 | 0.356 | |||
| BMI | 1.01 | 0.98–1.03 | 0.539 | |||
| Operation timec | 1.02 | 1.01–1.03 | 0.003 | 1.01 | 1.00–1.02 | 0.234 |
| Preoperative serum albumin | 0.68 | 0.54–0.84 | 0.005 | 0.78 | 0.60–1.02 | 0.072 |
| Preoperative total bilirubin | 1.09 | 1.03–1.16 | 0.002 | 1.07 | 1.00–1.14 | 0.041 |
| Female sex | 1.05 | 0.79–1.41 | 0.723 | |||
| Current smoking | 1.09 | 0.74–1.61 | 0.655 | |||
| Loss of >10% body weight in last 6 months | 1.55 | 1.03–2.34 | 0.038 | 1.44 | 0.91–2.28 | 0.125 |
| ASA class 3 or 4 status | 1.46 | 1.05–2.03 | 0.026 | 1.27 | 0.88–1.85 | 0.207 |
| Independent functional status | 0.22 | 0.66–0.08 | 0.007 | 0.71 | 0.32–1.56 | 0.389 |
| Diabetes | 1.24 | 0.85–1.81 | 0.261 | |||
| Ascites | 4.00 | 1.34–11.92 | 0.013 | 2.28 | 0.71–7.31 | 0.167 |
| Group | ||||||
| BEA only | Ref | |||||
| BEA with hepatic resection | 2.17 | 1.59–2.96 | <0.001 | 2.84 | 1.95–4.12 | <0.001 |
| BEA with hepatic and vascular resection | 3.21 | 1.60–6.46 | 0.001 | 4.38 | 1.98–9.68 | 0.003 |
Variables with P-values of <0.1 were selected for the multivariate model.
Per 10-year increase.
Per 10-min increase.
95% CI, 95% confidence interval; ASA, American Society of Anesthesiologists; BMI, body mass index; OR, odds ratio.
Table 3.
Observed and expected (O/E) ratios overall and by group in patients submitted to biliary-enteric anastomosis (BEA) with and without hepatic or vascular resection
| All patients (n = 747) | Group 1: BEA only (n = 266) | Group 2: BEA with hepatic resection (n = 439) | Group 3: BEA with hepatic and vascular resection (n = 42) | |||||
|---|---|---|---|---|---|---|---|---|
| O/E index | 95% CI | O/E index | 95% CI | O/E index | 95% CI | O/E index | 95% CI | |
| Morbidity | 1.52 | 1.38–1.68 | 1.20 | 0.98–1.45 | 1.71 | 1.51–1.93 | 1.92 | 1.29–2.76 |
| Mortality | 2.03 | 1.55–2.61 | 1.44 | 0.84–2.30 | 2.16 | 1.51–2.98 | 5.92 | 2.54–11.66 |
95% CI, 95% confidence interval.
Mortality
The overall mortality rate was 8.2%. Rates of 30-day mortality were 6.4%, 8.2% and 19.1% in Groups 1, 2 and 3, respectively (P = 0.021).
In Group 1, there were 17 deaths and the most commonly associated complications were re-intubation (n = 7), mechanical ventilation for >48 h (n = 6) and septic shock (n = 5). Similarly, in Group 2, there were 36 deaths and the most commonly associated complications were re-intubation (n = 21), mechanical ventilation for >48 h (n = 19) and septic shock (n = 20). In Group 3, there were eight deaths and the most commonly associated complications were mechanical ventilation for >48 h (n = 3) and acute renal failure (n = 3).
Predictors of mortality for the entire study cohort are shown in Table4.
Table 4.
Univariate and multivariate predictors using logistic regression of 30-day mortality in patients submitted to biliary-enteric anastomosis (BEA) with and without hepatic and vascular resection
| Univariate models | Multivariate modelsa | |||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Ageb | 1.04 | 1.01–1.06 | 0.008 | 1.06 | 1.02–1.10 | 0.003 |
| BMI | 1.04 | 1.00–1.08 | 0.062 | 1.05 | 0.99–1.11 | 0.079 |
| Preoperative serum albumin | 0.41 | 0.28–0.61 | <0.001 | 1.01 | 0.94–1.09 | 0.729 |
| Preoperative total bilirubin | 1.14 | 1.06–1.22 | 0.005 | 0.98 | 0.90–1.07 | 0.627 |
| Female sex | 0.68 | 0.39–1.18 | 0.168 | |||
| Current smoking | 0.63 | 0.28–1.42 | 0.266 | |||
| Independent functional status | 0.32 | 0.12–0.89 | 0.028 | 0.97 | 0.18–5.17 | 0.972 |
| ASA class 3 or 4 status | 3.36 | 1.42–7.94 | 0.006 | 1.84 | 0.59–5.72 | 0.293 |
| Operation timec | 1.00 | 1.00–1.00 | 0.864 | |||
| Return to operating room | 5.09 | 2.77–9.34 | <0.001 | 3.06 | 1.05–8.90 | 0.040 |
| Deep incisional SSI | 0.48 | 0.06–3.6 | 0.477 | |||
| Organ space SSI | 2.06 | 1.15–3.69 | 0.016 | 1.18 | 0.40–3.46 | 0.762 |
| Pneumonia | 2.34 | 1.00–5.51 | 0.051 | 0.04 | 0.01–0.22 | 0.003 |
| Unplanned intubation | 19.78 | 10.70–36.57 | <0.001 | 10.23 | 3.09–33.90 | 0.001 |
| Pulmonary embolism | 2.87 | 0.60–13.84 | 0.188 | |||
| Progressive renal insufficiency | 7.82 | 3.10–19.69 | <0.001 | 16.23 | 3.53–74.61 | 0.003 |
| Acute renal failure | 30.69 | 11.17–84.32 | <0.001 | 20.81 | 4.13–104.85 | 0.002 |
| Cardiac arrest requiring CPR | 80.95 | 22.75–288.09 | <0.001 | 35.84 | 5.62–228.76 | 0.002 |
| Intra- or postoperative bleeding or transfusion | 1.55 | 0.86–2.79 | 0.147 | |||
| DVT requiring therapy | 1.13 | 0.26–4.95 | 0.872 | |||
| Sepsis | 0.66 | 0.30–1.50 | 0.324 | |||
| Septic shock | 15.23 | 8.26–28.07 | <0.001 | 11.44 | 3.79–34.52 | <0.001 |
| Group | ||||||
| BEA only | Ref | |||||
| BEA with hepatic resection | 1.31 | 0.72–2.38 | 0.378 | 0.937 | 0.39–2.24 | 0.884 |
| BEA with hepatic and vascular resection | 3.45 | 1.38–8.59 | 0.008 | 5.255 | 1.43–19.33 | 0.013 |
Variables with P-values of <0.1.
Per 10-year increase.
Per 10-min increase.
95% CI, 95% confidence interval; ASA, American Society of Anesthesiologists; BMI, body mass index; CPR, cardiopulmonary resuscitation; DVT, deep vein thrombosis; OR, odds ratio; SSI, surgical site infection.
Discussion
Bile duct resection is a required component in the surgical treatment of cholangiocarcinoma; however, it may also be performed as part of the management of certain hepatic and gallbladder neoplasms depending on the clinical presentation. The focus of this analysis was to determine the effects of hepatic and vascular resections in addition to bile duct resection for hepatobiliary and gallbladder malignancies. Given the nature of the procedures examined and the selection criteria used, the vast majority of these tumours were most likely those of hilar cholangiocarcinoma and much of the discussion will focus on the outcomes of surgical treatment for this disease. Although the ICD-9 codes used also specify primary liver cancer and gallbladder malignancies, the use of specific CPT codes starting with BEA was selected to distinguish any malignancy with extrahepatic biliary involvement that may have been included in these diagnosis groupings.
Over the last two decades, several studies have demonstrated improved survival in patients with hilar cholangiocarcinoma treated with bile duct resection and major hepatic resection as a result of increased R0 resection rates.1–5 A recent analysis of hilar cholangiocarcinoma resected with combined major hepatic resection demonstrated a positive correlation with the tumour-free resection margin rate.20 Reported rates of postoperative morbidity and mortality ranged from 6% to 52% and from 2% to 12%, respectively.20–26 In addition, a large population-based analysis examined national trends in the management of gallbladder carcinoma and found radical resection with hepatectomy to be associated with improved survival on multivariate analysis.27
The literature is unclear with respect to postoperative outcomes asso ciated with bile duct resection combined with both hepatic and vascular resection for hilar cholangiocarcinoma. A recent multi-institution analysis of 305 patients with hilar cholangiocarcinoma examined the impact of portal vein resection on outcomes.11 In this group, 51 patients (16.7%) underwent combined bile duct resection, hepatectomy and vascular resection. The incidence of 90-day mortality in patients undergoing bile duct resection alone was lower (1.2%) than that in patients submitted to bile duct resection with hepatic resection (10.6%), and bile duct resection with hepatic and vascular resection (17.6%) (P < 0.001). A meta-analysis by Abbas and Sandroussi reviewed the role of vascular resection in the treatment of hilar cholangiocarcinoma.14 Its search criteria obtained 24 articles that referred to a total of 2457 patients, 669 (27.2%) of whom underwent vascular resection. The meta-analysis showed no significant difference in morbidity between the two groups. With reference to mortality rates, the authors found significantly higher mortality among patients undergoing vascular resection (odds ratio 2.07, 95% confidence interval 1.21–3.57; P = 0.008).14 Another recent meta-analysis included 13 studies with a total of 1921 patients with hilar cholangiocarcinoma, of whom 458 were submitted to vascular resection.15 No significant differences in postoperative morbidity or mortality were found between the non-vascular resection and vascular resection groups.
It is interesting to note that the mortality rate of 6.4% in patients undergoing BEA alone in the current analysis appears to be higher than equivalent rates reported in the literature. One explanation may refer to the fact that the NSQIP database does not indicate whether an operation for cancer is performed with curative or palliative intent; no designation of resection status is given. As BEA is the CPT code used as a surrogate for bile duct resection, it is possible that some patients in the study cohort may not have undergone curative resections, but, rather, palliative bypasses for obstructive jaundice in the setting of unresectable disease. Jarnagin and colleagues at Memorial Sloan–Kettering Cancer Center reported a postoperative mortality rate of 11% in patients undergoing palliative biliary-enteric bypass in the context of unresectable hilar cholangiocarcinoma or gallbladder carcinoma.28
Multivariate analyses demonstrated preoperative total bilirubin, BEA with hepatic resection, and BEA with hepatic and vascular resection to be independently associated with morbidity. The addition of hepatic and vascular resection was significantly correlated with increased postoperative morbidity. The multivariate analysis of mortality showed age, return to the operating room, pneumonia, unplanned intubation, progressive renal insufficiency, acute renal failure, cardiac arrest requiring CPR, septic shock, and BEA with hepatic and vascular resection to be significant predictors. The performance of hepatic resection alone with BEA did not correlate with increased postoperative mortality.
The O/E ratios in Groups 2 and 3 for postoperative 30-day morbidity and mortality suggest these outcomes can be improved. Recent studies have reported that 90-day postoperative outcomes after hepatobiliary procedures represent optimal measures of surgical quality and that 30-day outcomes are likely to underestimate postoperative occurrences for patients in this specific subset.29,30 This is a limitation of the NSQIP database, although it has been shown that hospitals participating in the NSQIP have been able to significantly improve their risk-adjusted complication and mortality rates by making interventions based on analysis of their 30-day outcomes.31 Although true rates of postoperative morbidity and mortality for patients in this analysis may be higher, the data presented here clearly demonstrate significant differences between procedure groups in 30-day outcomes, which potentially can be improved.
The structure of the NSQIP database precludes any highly detailed examination of the clinical information as a result of the standardized nature of data collection and efforts to protect patient privacy. A report by Loehrer and colleagues at Indiana University used the NSQIP database to examine outcomes in cholangiocarcinoma and mentioned the lack of hepatobiliary-specific variables that might provide greater clinical insight and risk adjustment for patients with hepatobiliary and gallbladder carcinomas.32 However, the current analysis does bring up topics for discussion that may guide future efforts to improve outcomes. Over 30% of deaths in all procedure groups were associated with pulmonary complications. The use of preoperative pulmonary rehabilitation and optimization, as well as evidence-based respiratory care pathways in the perioperative period may help to prevent patients from requiring extended time on ventilator support. Septic shock was also associated with mortality, and patients with biliary involvement by tumour can often develop cholangitis as a result of biliary obstruction. The multivariate morbidity analysis showed the total bilirubin level to be significant and indicated that the role of preoperative biliary drainage needs to be better defined. The increased rate of acute renal failure seen in the Group 3 patient deaths may reflect the end-stage result of fulminant postoperative hepatic failure with concomitant hepatorenal syndrome. The inclusion of data on the future liver remnant, the use of portal vein embolization, and postoperative liver function laboratory values would be extremely helpful in analyses of outcomes in these patients in future research studies.
Conclusions
This large database analysis clearly demonstrates that when hepatectomy and vascular resection are performed in conjunction with bile duct resection for malignancy, 30-day postoperative outcomes are significantly worse than those in patients submitted to bile duct resection alone. The performance of hepatectomy, or hepatectomy and vascular resection with BEA correlated with increased morbidity, whereas only combined hepatectomy and vascular resection was a significant predictor of mortality. More than a third of all deaths in each procedure group were associated with pulmonary complications and septic shock. The high O/E ratios in patients undergoing BEA and hepatectomy, as well as BEA with combined hepatectomy and vascular resection suggest these outcomes can be improved. The current results support further efforts to clearly define the specific drivers of these outcomes so that appropriate interventions can be performed.
Acknowledgments
The authors wish to acknowledge the support of the Biostatistics Shared Resource, the Comprehensive Cancer Center of Wake Forest University and the National Cancer Institute Cancer Center Support Grant P30 CA012197. The authors also wish to thank Bonny B. Morris, MSPH, Comprehensive Cancer Center of Wake Forest Baptist Medical Center, for her assistance with the review and preparation of this manuscript.
Conflicts of interest
None declared.
Supporting Information
Additional Supporting Information may be found in the online version of this article:
Appendix S1. List of Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD-9) codes used for patient study selection.
References
- Sugiura Y, Nakamura S, Iida S, Hosoda Y, Ikeuchi S, Mori S, et al. Extensive resection of the bile ducts combined with liver resection for cancer of the main hepatic duct junction: a cooperative study of the Keio Bile Duct Cancer Study Group. Surgery. 1994;115:445–451. [PubMed] [Google Scholar]
- Ogura Y, Mizumoto R, Tabata M, Matsuda S, Kusuda T. Surgical treatment of carcinoma of the hepatic duct confluence: analysis of 55 resected carcinomas. World J Surg. 1993;17:85–92. doi: 10.1007/BF01655714. ; discussion 92–93. [DOI] [PubMed] [Google Scholar]
- Neuhaus P, Thelen A. Radical surgery for right-sided Klatskin tumor. HPB. 2008;10:171–173. doi: 10.1080/13651820801992708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee SG, Lee YJ, Park KM, Hwang S, Min PC. One hundred and eleven liver resections for hilar bile duct cancer. J Hepatobiliary Pancreat Surg. 2000;7:135–141. doi: 10.1007/s005340050167. [DOI] [PubMed] [Google Scholar]
- Miyazaki M, Ito H, Nakagawa K, Ambiru S, Shimizu H, Okaya T, et al. Parenchyma-preserving hepatectomy in the surgical treatment of hilar cholangiocarcinoma. J Am Coll Surg. 1999;189:575–583. doi: 10.1016/s1072-7515(99)00219-7. [DOI] [PubMed] [Google Scholar]
- Ebata T, Nagino M, Kamiya J, Uesaka K, Nagasaka T, Nimura Y. Hepatectomy with portal vein resection for hilar cholangiocarcinoma: audit of 52 consecutive cases. Ann Surg. 2003;238:720–727. doi: 10.1097/01.sla.0000094437.68038.a3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hemming AW, Kim RD, Mekeel KL, Fujita S, Reed AI, Foley DP, et al. Portal vein resection for hilar cholangiocarcinoma. Am Surg. 2006;72:599–604. ; discussion 604–605. [PubMed] [Google Scholar]
- Miyazaki M, Kato A, Ito H, Kimura F, Shimizu H, Ohtsuka M, et al. Combined vascular resection in operative resection for hilar cholangiocarcinoma: does it work or not? Surgery. 2007;141:581–588. doi: 10.1016/j.surg.2006.09.016. [DOI] [PubMed] [Google Scholar]
- Song GW, Lee SG, Hwang S, Kim KH, Cho YP, Ahn CS, et al. Does portal vein resection with hepatectomy improve survival in locally advanced hilar cholangiocarcinoma? Hepatogastroenterology. 2009;56:935–942. [PubMed] [Google Scholar]
- Nagino M, Nimura Y, Nishio H, Ebata T, Igami T, Matsushita M, et al. Hepatectomy with simultaneous resection of the portal vein and hepatic artery for advanced perihilar cholangiocarcinoma: an audit of 50 consecutive cases. Ann Surg. 2010;252:115–123. doi: 10.1097/SLA.0b013e3181e463a7. [DOI] [PubMed] [Google Scholar]
- de Jong MC, Marques H, Clary BM, Bauer TW, Marsh JW, Ribero D, et al. The impact of portal vein resection on outcomes for hilar cholangiocarcinoma: a multi-institutional analysis of 305 cases. Cancer. 2012;118:4737–4747. doi: 10.1002/cncr.27492. [DOI] [PubMed] [Google Scholar]
- Tamoto E, Hirano S, Tsuchikawa T, Tanaka E, Miyamoto M, Matsumoto J, et al. Portal vein resection using the no-touch technique with a hepatectomy for hilar cholangiocarcinoma. HPB. 2014;16:56–61. doi: 10.1111/hpb.12067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wu XS, Dong P, Gu J, Li ML, Wu WG, Lu JH, et al. Combined portal vein resection for hilar cholangiocarcinoma: a meta-analysis of comparative studies. J Gastrointest Surg. 2013;17:1107–1115. doi: 10.1007/s11605-013-2202-9. [DOI] [PubMed] [Google Scholar]
- Abbas S, Sandroussi C. Systematic review and meta-analysis of the role of vascular resection in the treatment of hilar cholangiocarcinoma. HPB. 2013;15:492–503. doi: 10.1111/j.1477-2574.2012.00616.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen W, Ke K, Chen YL. Combined portal vein resection in the treatment of hilar cholangiocarcinoma: a systematic review and meta-analysis. Eur J Surg Oncol. 2014;40:489–495. doi: 10.1016/j.ejso.2014.02.231. [DOI] [PubMed] [Google Scholar]
- American College of Surgeons. 2015. National Surgical Quality Improvement Program. Available at http://site.acsnsqip.org/ (last accessed 6 April 2015)
- American College of Surgeons National Surgical Quality Improvement Program. 2014. User Guide for the 2013 ACS NSQIP Participant Use Data File (PUF)
- Mak PH, Campbell RC, Irwin MG. The ASA Physical Status Classification: inter-observer consistency. American Society of Anesthesiologists. Anaesth Intensive Care. 2002;30:633–640. doi: 10.1177/0310057X0203000516. [DOI] [PubMed] [Google Scholar]
- Pitt HA, Kilbane M, Strasberg SM, Pawlik TM, Dixon E, Zyromski N, et al. ACS-NSQIP has the potential to create an HPB-NSQIP option. HPB. 2009;11:405–413. doi: 10.1111/j.1477-2574.2009.00074.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Xiang S, Lau WY, Chen XP. Hilar cholangiocarcinoma: controversies on the extent of surgical resection aiming at cure. Int J Colorectal Dis. 2015;30:159–171. doi: 10.1007/s00384-014-2063-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kawarada Y, Das BC, Naganuma T, Tabata M, Taoka H. Surgical treatment of hilar bile duct carcinoma: experience with 25 consecutive hepatectomies. J Gastrointest Surg. 2002;6:617–624. doi: 10.1016/s1091-255x(01)00008-7. [DOI] [PubMed] [Google Scholar]
- Rea DJ, Munoz-Juarez M, Farnell MB, Donohue JH, Que FG, Crownhart B, et al. Major hepatic resection for hilar cholangiocarcinoma: analysis of 46 patients. Arch Surg. 2004;139:514–523. doi: 10.1001/archsurg.139.5.514. ; discussion 523–525. [DOI] [PubMed] [Google Scholar]
- Rocha FG, Matsuo K, Blumgart LH, Jarnagin WR. Hilar cholangiocarcinoma: the Memorial Sloan–Kettering Cancer Center experience. J Hepatobiliary Pancreat Sci. 2010;17:490–496. doi: 10.1007/s00534-009-0205-4. [DOI] [PubMed] [Google Scholar]
- Neuhaus P, Thelen A, Jonas S, Puhl G, Denecke T, Veltze-Schlieker W, et al. Oncological superiority of hilar en bloc resection for the treatment of hilar cholangiocarcinoma. Ann Surg Oncol. 2012;19:1602–1608. doi: 10.1245/s10434-011-2077-5. [DOI] [PubMed] [Google Scholar]
- Cannon RM, Brock G, Buell JF. Surgical resection for hilar cholangiocarcinoma: experience improves resectability. HPB. 2012;14:142–149. doi: 10.1111/j.1477-2574.2011.00419.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kow AW, Wook CD, Song SC, Kim WS, Kim MJ, Park HJ, et al. Role of caudate lobectomy in type III A and III B hilar cholangiocarcinoma: a 15-year experience in a tertiary institution. World J Surg. 2012;36:1112–1121. doi: 10.1007/s00268-012-1497-0. [DOI] [PubMed] [Google Scholar]
- Mayo SC, Shore AD, Nathan H, Edil B, Wolfgang CL, Hirose K, et al. National trends in the management and survival of surgically managed gallbladder adenocarcinoma over 15 years: a population-based analysis. J Gastrointest Surg. 2010;14:1578–1591. doi: 10.1007/s11605-010-1335-3. [DOI] [PubMed] [Google Scholar]
- Jarnagin WR, Burke E, Powers C, Fong Y, Blumgart LH. Intrahepatic biliary enteric bypass provides effective palliation in selected patients with malignant obstruction at the hepatic duct confluence. Am J Surg. 1998;175:453–460. doi: 10.1016/s0002-9610(98)00084-1. [DOI] [PubMed] [Google Scholar]
- Hyder O, Pulitano C, Firoozmand A, Dodson R, Wolfgang CL, Choti MA, et al. A risk model to predict 90-day mortality among patients undergoing hepatic resection. J Am Coll Surg. 2013;216:1049–1056. doi: 10.1016/j.jamcollsurg.2013.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mise Y, Vauthey JN, Zimmitti G, Parker NH, Conrad C, Aloia TA, et al. Ninety-day postoperative mortality is a legitimate measure of hepatopancreatobiliary surgical quality. Ann Surg. 2015 doi: 10.1097/SLA.0000000000001048. PMID: 25590497 [Epub ahead of print] [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hall BL, Hamilton BH, Richards K, Bilimoria KY, Cohen ME, Ko CY. Does surgical quality improve in the American College of Surgeons National Surgical Quality Improvement Program: an evaluation of all participating hospitals. Ann Surg. 2009;250:363–376. doi: 10.1097/SLA.0b013e3181b4148f. [DOI] [PubMed] [Google Scholar]
- Loehrer AP, House MG, Nakeeb A, Kilbane EM, Pitt HA. Cholangiocarcinoma: are North American surgical outcomes optimal? J Am Coll Surg. 2013;216:192–200. doi: 10.1016/j.jamcollsurg.2012.11.002. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1. List of Current Procedural Terminology (CPT) and International Classification of Diseases, Ninth Revision (ICD-9) codes used for patient study selection.