Abstract
Background:
Endoscopic biliary stenting (EBS) and percutaneous transhepatic biliary drainage (PTBD) are two techniques used for preoperative biliary drainage prior to hepatobiliary resection. The objectives of this study were to determine predictors of the drainage technique selection and to evaluate the association between drainage technique and postoperative outcomes.
Methods:
Using ACS NSQIP data (2014–2017), patients who underwent preoperative biliary drainage prior to hepatobiliary resection for malignancy were identified. Separate multivariable-adjusted, propensity score (PS) adjusted, and PS matched logistic regression models were constructed to evaluate the association between drainage technique and postoperative outcomes.
Results:
Of 527 patients identified, 431 (81.8%) received EBS and 96 (18.2%) received PTBD. Patients who underwent PTBD had more preoperative co-morbidities, including higher ASA class, recent weight loss, and hypoalbuminemia (all p<0.05). After multivariable adjustment, PTBD was significantly associated with 30-day DSM (OR 1.92, 95% CI 1.24–2.97, p=0.004), overall SSI (OR 1.74, 95% CI 1.10–2.76, p=0.019), and superficial SSI (OR 2.08, 95% CI 1.20–3.60, p=0.010). These findings remained significant for both PS-adjusted and PS-matched models.
Conclusion:
Patients undergoing hepatobiliary resection selected for PTBD had significantly more preoperative co-morbidities and nutritional deficits. Compared to EBS, PTBD was associated with significantly higher odds of postoperative morbidity and mortality.
Introduction
Surgical resection remains the only potentially curative therapy for patients with hepatobiliary malignancies. These patients often present with or develop biliary obstruction and cholestasis, which is associated with biliary sepsis, increased hepatic toxicity, and impaired postoperative liver regeneration.(1) Hepatobiliary resection in the presence of cholestasis is known to increase the risk for postoperative morbidity and mortality.(2, 3) Preoperative biliary drainage is thought to reduce postoperative mortality by resolving cholestasis, improving nutritional status and liver function, reducing bacterial translocation, and preserving postoperative liver function.(4)
There are two commonly used preoperative biliary drainage techniques: endoscopic biliary stenting (EBS) and percutaneous transhepatic biliary drainage (PTBD). However, the optimal technique has not been established. EBS is advantageous as it approximates normal physiologic drainage, but is associated with endoscopic complications such as cholangitis from biliary contamination and pancreatitis.(4) PTBD, on the other hand, has higher technical success rates and lower risk for biliary contamination, but is less desirable for patients due to external catheters and catheter-associated complications.(5)
Little is known about the association of preoperative biliary drainage technique with postoperative outcomes. Several meta-analyses have found that PTBD is superior to EBS with regards to technical success rates, conversion rates, and post-procedural complications.(4, 6–8) A handful of single institution retrospectives series have reported mixed results regarding postoperative outcomes with both techniques.(9–11) No prior study has investigated differences in patient selection and postoperative outcomes in a multi-institutional cohort of patients with resectable hepatobiliary malignancy. The objectives of this study were 1) to determine predictors of drainage technique selection and 2) to evaluate the association between preoperative biliary drainage technique with postoperative outcomes.
Methods
Data Source and Patient Population
Patients who underwent preoperative biliary drainage prior to resection for hepatobiliary malignancy from January 1, 2014 to December 31, 2017 were identified using the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) Targeted Hepatectomy database. The ACS NSQIP database includes patient characteristics, operative details, and 30-day postoperative outcomes, which are collected from the medical record by dedicated, trained data abstractors.(12) The Targeted Hepatectomy database includes all hepatic resections defined by the following CPT codes: 47120, 47122, 47125, and 47130. Hepatobiliary malignancy was defined by final histologic diagnosis of either primary (i.e., hepatocellular carcinoma, cholangiocarcinoma, gallbladder cancer, and other unspecified primary hepatobiliary cancer) or secondary (i.e., metastatic) hepatobiliary cancer. Because this study used preexisting, deidentified data, the Chesapeake Institutional Review Board deemed it exempt from oversight.
Preoperative Drainage Technique
Preoperative drainage was defined as placement of an EBS or PTBD catheter prior to surgery. Patients with no or unknown drainage catheter or stent were excluded from analysis. ACS NSQIP does not record whether more than one drainage technique was used in a single patient (i.e., whether PTBD was used following the failure of EBS).
Risk Adjustment Variables
Patient demographics (e.g., sex, race), preoperative comorbidities (e.g., diabetes mellitus), American Society of Anesthesiologists (ASA) class, functional status, laboratory values (e.g., albumin level), and operative details (e.g., surgical procedure) were considered for risk adjustment. Additional hepatectomy-specific details (e.g., neoadjuvant chemotherapy, concurrent intraoperative ablation, concurrent intraoperative partial hepatic resections, biliary reconstruction) were included.
Thirty-Day Outcomes
The primary outcome studied was a composite measure of 30-day death or serious morbidity (DSM). Per ACS NSQIP definitions, a serious morbidity event included one or more the following complications: cardiac arrest, myocardial infarction, pneumonia, renal failure, reoperation, deep or organ-space surgical site infection (SSI), systemic sepsis, unplanned intubation, urinary tract infection, or wound disruption. Seven other standard ACS NSQIP outcomes (i.e., overall SSI, superficial SSI, deep or organ-space SSI, systemic sepsis, reoperation, prolonged length of stay [LOS], and readmission) and three hepatectomy-specific outcomes (i.e., post-hepatectomy liver failure [PHLF], bile leakage, and need for invasive postoperative intervention) were studied as secondary outcomes. PHLF was defined as a postoperatively acquired deterioration in liver function resulting in a concomitant rise in international normalize ratio and total bilirubin level after postoperative day 5, in accordance with the International Study Group of Liver Surgery (ISGLS) definition.(13) Bile leakage was defined as a persistent drainage of bilious fluid after postoperative day 3. Invasive postoperative interventions included any invasive procedure except for reoperation (e.g., aspiration or drain placement for bile leakage or abscess, biliary stent for biliary obstruction or leakage, angiography and embolization for hemorrhage, etc.) following surgery.
Statistical Analyses
Baseline patient, oncologic, and operative characteristics and 30-day postoperative outcomes were compared between patients who received EBS versus PTBD using student’s ttest, Pearson’s chi-squared test for association, or Fisher’s exact test, where appropriate. Standard NSQIP variables missing <5% of data were imputed using maximum likelihood estimation.(14) All tests of statistical significance were 2-sided with α = 0.05.
Multivariable logistic regression models with robust standard errors were constructed to determine independent predictors of PTBD use while accounting for patient clustering within hospitals. Risk adjustment variables were included in the model if they were clinically relevant or found to be significant with α = 0.10 in bivariate analysis.
Next, several risk adjustment modeling approaches were used to evaluate the association of preoperative biliary drainage technique with postoperative outcomes. The first approach was a standard multivariable logistic regression model accounting for differences in baseline characteristics. The second approach was a propensity score (PS) adjusted multivariable logistic regression model accounting for potential selection bias in the surgeon’s choice of drainage technique. PS was calculated from a separate logistic regression model predicting the probability of PTBD use and added to the standard multivariable logistic regression models as a five-level categorical variable in this approach. Finally, the third approach was a multivariable logistic regression model using a 1:1 PS matched cohort. The PS match was performed using a greedy 1:1 algorithm with a 0.2 caliper width based on the same variables included in the calculation of the PS in the second approach. Balance was assessed graphically comparing standardized differences in the means of variables before and after matching using a threshold of 0.1 to define excellent balance.(15, 16)
Risk-adjusted associations are presented as odds ratios (ORs) with 95% confidence intervals (CIs) comparing those who received PTBD with those who received EBS, where ORs greater than 1 represented greater odds of adverse outcome in those who received PTBD. All statistical analyses were performed in SAS, version 9.4 (SAS Institute, Cary, NC).
Sensitivity Analysis
For the PS adjusted analysis, the distribution of the estimated probabilities was assessed graphically to evaluate the extent to which procedure type was influenced by potential selection bias. Sensitivity analysis was conducted using the subset of patients with PS in the region of common support. An additional sensitivity analysis was performed in patients with the most complex tumor and procedure characteristics. In this analysis, only the subset of patients with hilar cholangiocarcinoma or gallbladder cancer who underwent either a total lobectomy or trisegmentectomy with biliary reconstruction were included.
Results
Patient Characteristics
Of the 15,748 patients identified, 4,494 (28.5%) had primary hepatobiliary cancers and 7,228 (45.9%) had metastatic cancers. Of these 11,722 patients with either a primary or metastatic hepatobiliary malignancy, 527 (4.5%) patients from 84 hospitals underwent preoperative biliary drainage: 431 (81.8%) received EBS and 96 (18.2%) received PTBD (Table 1). One hundred and fifty-four (29.2%) patients underwent partial lobectomy, 98 (18.6%) patients underwent left total lobectomy, 111 (21.1%) patients underwent right total lobectomy, and 164 (31.3%) patients underwent trisegmentectomy (Table 2).
Table 1.
Baseline characteristics of patients undergoing hepatobiliary resection for hepatobiliary malignancy
| Overall N=527 |
EBS N=431 (81.8%) |
PTBD N=96 (18.2%) |
p-value | |
|---|---|---|---|---|
| Age (mean ± standard deviation) | 63.6 ± 11.3 | 64.1 ± 11.2 | 61.3 ± 11.6 | 0.029* |
| Female sex | 224 (42.5%) | 186 (43.2%) | 38 (39.6%) | 0.522 |
| Race class | 0.482 | |||
| White | 367 (69.6%) | 305 (70.8%) | 62 (64.6%) | |
| African American | 30 (5.7%) | 24 (5.6%) | 6 (6.3%) | |
| Other/Unknown | 130 (24.7%) | 102 (23.7%) | 28 (29.2%) | |
| Hispanic | 21 (4.0%) | 19 (4.4%) | 2 (2.1%) | 0.292 |
| ASA class | 0.003* | |||
| I or II | 111 (21.1%) | 100 (23.2%) | 11 (11.5%) | |
| III | 374 (71.0%) | 303 (70.3%) | 71 (74.0%) | |
| IV or V | 42 (8.0%) | 28 (6.5%) | 14 (14.6%) | |
| Dependent functional status | 1 (0.2%) | 1 (0.3%) | 0 (0%) | 0.637 |
| Admitted not from home | 27 (5.1%) | 22 (5.1%) | 5 (5.2%) | 0.967 |
| Weight loss | 104 (19.7%) | 75 (17.4%) | 29 (30.2)% | 0.004* |
| Smoking | 98 (18.6%) | 81 (18.8%) | 17 (17.7%) | 0.805 |
| Serum albumin level <3.5 g/dL | 222 (42.1%) | 164 (38.1%) | 58 (60.4%) | <0.001* |
| BMI | 0.771 | |||
| Underweight | 11 (2.1%) | 10 (2.3%) | 1 (1.0%) | |
| Normal | 196 (37.2%) | 160 (37.1%) | 36 (37.5%) | |
| Overweight | 198 (37.6%) | 162 (37.6%) | 36 (37.5%) | |
| Class 1 obesity | 96 (18.2%) | 76 (17.6%) | 20 (20.8%) | |
| Class 2 obesity | 20 (3.8%) | 17 (3.9%) | 3 (3.1%) | |
| Class 3 obesity | 6 (1.1%) | 6 (1.4%) | 0 (0%) | |
| Hypertension requiring medication | 238 (45.2%) | 194 (45.0%) | 44 (45.8%) | 0.884 |
| Congestive heart failure | 1 (0.2%) | 1 (0.2%) | 0 (0%) | 0.637 |
| Preoperative dyspnea | 0.585 | |||
| At rest | 1 (0.2%) | 1 (0.2%) | 0 (0%) | |
| Moderate exertion | 33 (6.3%) | 25 (5.8%) | 8 (8.3%) | |
| Chronic obstructive pulmonary disease | 21 (4.0%) | 19 (4.4%) | 2 (2.1%) | 0.292 |
| Ventilator dependence | 0 (0%) | 0 (0%) | 0 (0%) | 1.0 |
| Renal insufficiency | 0 (0%) | 0 (0%) | 0 (0%) | 1.0 |
| Dialysis requirement | 1 (0.2%) | 0 (0%) | 1 (1.0%) | 0.034* |
| Preoperative SIRS or sepsis | 14 (2.7%) | 10 (2.3%) | 4 (4.1%) | 0.309 |
| Diabetes mellitus requiring medication | 0.999 | |||
| Oral | 61 (11.6%) | 50 (11.6%) | 11 (11.5%) | |
| Insulin | 33 (6.3%) | 27 (6.3%) | 6 (6.3%) | |
| Ascites | 3 (0.6%) | 0 (0%) | 3 (3.1%) | <0.001* |
| Disseminated cancer | 86 (16.3%) | 68 (15.8%) | 18 (18.8%) | 0.476 |
| Chronic steroid use | 11 (2.1%) | 10 (2.3%) | 1 (1.0%) | 0.428 |
| Bleeding disorder | 14 (2.7%) | 10 (2.3%) | 4 (4.2%) | 0.309 |
| Preoperative viral hepatitis | 25 (4.7%) | 22 (5.1%) | 3 (3.1%) | 0.409 |
| Neoadjuvant chemotherapy | 118 (22.4%) | 97 (22.5%) | 21 (21.9%) | 0.893 |
: statistically significant at α = 0.05
EBS: endoscopic biliary stenting; PTBD: percutaneous transhepatic biliary drainage; ASA: American Society of Anesthesiologists; BMI: body mass index; SIRS: systemic inflammatory response syndrome
Table 2.
Operative characteristics of patients undergoing hepatobiliary resection for hepatobiliary malignancy
| Overall N=527 | EBS N=431 (81.8%) | PTBD N=96 (18.2%) | p-value | |
|---|---|---|---|---|
| Surgical procedure | 0.219 | |||
| Partial lobectomy | 154 (29.2%) | 132 (30.6%) | 22 (22.9%) | |
| Left total lobectomy | 98 (18.6%) | 83 (19.3%) | 15 (15.6%) | |
| Right total lobectomy | 111 (21.1%) | 89 (20.7%) | 22 (22.9%) | |
| Trisegmentectomy | 164 (31.1%) | 127 (29.5%) | 37 (38.5%) | |
| Emergency | 7 (1.3%) | 6 (1.4%) | 1 (1.0%) | 0.786 |
| Laparoscopic approach | 21 (4.0%) | 17 (3.2%) | 4 (4.2%) | 0.920 |
| Preoperative blood transfusion | 4 (0.8%) | 2 (0.5%) | 2 (2.1%) | 0.098 |
| Wound class | 0.478 | |||
| Clean or clean/contaminated | 403 (76.5%) | 334 (77.5%) | 69 (71.9%) | |
| Contaminated | 81 (15.4%) | 64 (14.9%) | 17 (17.7%) | |
| Dirty | 43 (8.2%) | 33 (7.7%) | 10 (10.4%) | |
| Liver Texture | 0.405 | |||
| Normal | 151 (28.7%) | 125 (29.0%) | 26 (27.1%) | |
| Fatty | 42 (8.0%) | 38 (8.8%) | 4 (4.2%) | |
| Congested | 19 (3.6%) | 14 (3.3%) | 5 (5.2%) | |
| Cirrhotic | 52 (9.9%) | 44 (10.2%) | 8 (8.3%) | |
| Not documented | 263 (49.9%) | 210 (48.7%) | 53 (55.2%) | |
| Concurrent intraoperative ablation | 20 (3.8%) | 19 (4.4%) | 1 (1.0%) | 0.119 |
| Concurrent partial liver resection | 239 (45.4%) | 195 (45.2%) | 44 (45.8%) | 0.916 |
| Intraoperative Pringle maneuver | 148 (28.1%) | 121 (28.1%) | 27 (28.1%) | 0.992 |
| Biliary reconstruction | 391 (74.2%) | 314 (72.9%) | 77 (80.2%) | 0.137 |
| Surgical drain placement | 454 (86.2%) | 367 (85.2%) | 87 (90.6%) | 0.160 |
| Final histology | 0.127 | |||
| Hepatocellular carcinoma | 35 (6.6%) | 29 (6.7%) | 6 (6.3%) | |
| Intrahepatic cholangiocarcinoma | 162 (30.75) | 131 (30.4%) | 31 (32.3%) | |
| Hilar cholangiocarcinoma/gallbladder cancer | 228 (43.3%) | 189 (43.9%) | 39 (40.6%) | |
| Other primary hepatobiliary cancer | 24 (4.6%) | 15 (3.5%) | 9 (9.4%) | |
| Metastatic cancer | 78 (14.8%) | 67 (15.6%) | 11 (11.5%) |
Predictors of Percutaneous Transhepatic Biliary Drainage Use
Compared to EBS, patients who were selected for PTBD had significantly more preoperative comorbidities, including higher ASA class, recent weight loss, hypoalbuminemia, ascites, and dialysis requirement (all p<0.05; Table 1). There were no differences between the two groups with regards to surgical procedure or final cancer histology (Table 2). On multivariable analysis, younger age, higher ASA class, and preoperative hypoalbuminemia were independently associated with PTBD use (Table 3). Patients with a primary hepatobiliary cancer other than hepatocellular carcinoma, cholangiocarcinoma, or gallbladder cancer were more likely to receive PTBD, whereas patients with metastatic cancer were more likely to receive EBS.
Table 3.
Independent predictors of percutaneous transhepatic biliary drainage technique
| Characteristic | OR (95% CI) | P-value |
|---|---|---|
| Age | 0.97 (0.95–0.99) | 0.002* |
| ASA class | 0.005* | |
| I-II | 1.00 | |
| III | 2.49 (1.25–4.96) | 0.010* |
| IV or V | 5.02 (1.96–12.85) | 0.001* |
| Serum albumin level <3.5 g/dL | 2.64 (1.51–4.62) | 0.001* |
| Histology | 0.029* | |
| Hilar cholangiocarcinoma/gallbladder cancer | 1.00 | |
| Hepatocellular carcinoma | 1.55 (0.54–4.49) | 0.413 |
| Intrahepatic cholangiocarcinoma | 1.23 (0.60–2.49) | 0.571 |
| Other primary hepatobiliary cancer | 4.04 (1.42–11.51) | 0.010* |
| Metastatic cancer | 1.00 (0.34–2.93) | 0.999 |
| Female sex | 1.00 (0.66–1.54) | 0.988 |
| Race | 0.877 | |
| White | 1.00 | |
| African American | 1.09 (0.47–2.52) | 0.833 |
| Other/Unknown | 1.18 (0.60–2.32) | 0.621 |
| Hispanic ethnicity | 0.25 (0.04–1.42) | 0.116 |
| Weight loss | 1.53 (0.84–2.80) | 0.163 |
| Emergency | 0.35 (0.04–3.56) | 0.373 |
| Neoadjuvant chemotherapy | 1.07 (0.54–2.11) | 0.843 |
| Procedure | 0.549 | |
| Partial lobectomy | 1.00 | |
| Left total lobectomy | 0.98 (0.41–2.35) | 0.956 |
| Right total lobectomy | 1.25 (0.58–2.66) | 0.567 |
| Trisegmentectomy | 1.54 (0.79–2.99) | 0.204 |
| Concurrent intraoperative ablation | 0.36 (0.04–3.51) | 0.374 |
| Concurrent partial resection | 1.06 (0.59–1.91) | 0.838 |
| Biliary reconstruction | 1.24 (0.60–2.58) | 0.558 |
: statistically significant at α = 0.05
ASA: American Society of Anesthesiologists; OR: odds ratio; CI: confidence interval
Association Between Biliary Drainage Technique and Postoperative Outcomes
A DSM occurred in 251 of 527 (47.6%) patients within 30 days of surgery. In unadjusted analysis, DSM was significantly more frequent in the PTBD group compared to the EBS group (61.5% vs 44.6%, p=0.003), as were overall SSI, superficial SSI, PHLF, bile leakage, and prolonged LOS (Table 4).
Table 4.
Postoperative complications for patients who received percutaneous transhepatic biliary drainage (PTBD) versus endoscopic biliary stenting (EBS)
| Unadjusted OR (95% CI) | p-value | Adjusted OR (95% CI) | p-value | PS Adjusted OR (95% CI) | p-value | PS Matched OR (95% CI) | p-value | |
|---|---|---|---|---|---|---|---|---|
| Death or serious morbidity | 1.99 (1.26–3.13) | 0.003* | 1.92 (1.24–2.97) | 0.004* | 1.97 (1.27–3.07) | 0.003* | 1.96 (1.11–3.49) | 0.022* |
| Overall SSI | 1.78 (1.14–2.79) | 0.012 | 1.74 (1.10–2.76) | 0.019* | 1.77 (1.11–2.81) | 0.016* | 2.38 (1.21–4.70) | 0.013* |
| Superficial SSI | 2.01 (1.07–3.78) | 0.031 | 2.08 (1.20–3.60) | 0.010* | 2.01 (1.16–3.48) | 0.013* | 3.19 (1.02–9.97) | 0.046* |
| Deep/Organ space SSI | 1.34 (0.85–2.14) | 0.210 | 1.35 (0.83–2.22) | 0.228 | 1.40 (0.85–2.29) | 0.183 | 1.72 (0.86–3.43) | 0.122 |
| Sepsis or septic shock | 1.41 (0.84–2.37) | 0.189 | 1.38 (0.86–2.20) | 0.180 | 1.43 (0.90–2.28) | 0.131 | 1.57 (0.68–3.61) | 0.284 |
| Reoperation | 1.23 (0.62–2.42) | 0.558 | 0.92 (0.47–1.83) | 0.819 | 1.10 (0.52–2.32) | 0.806 | 1.17 (0.44–3.10) | 0.755 |
| Posthepatectomy liver failure | 1.88 (1.12–3.17) | 0.017* | 1.51 (0.78–2.91) | 0.217 | 1.53 (0.76–3.07) | 0.235 | 1.61 (0.68–3.84) | 0.274 |
| Bile leakage | 1.73 (1.09–2.75) | 0.019* | 1.50 (0.92–2.45) | 0.102 | 1.76 (1.05–2.97) | 0.033* | 1.19 (0.58–2.43) | 0.631 |
| Invasive intervention | 1.39 (0.88–2.20) | 0.161 | 1.32 (0.75–2.34) | 0.333 | 1.35 (0.75–2.45) | 0.316 | 1.25 (0.61–2.60) | 0.536 |
| Prolonged length of stay | 1.81 (1.05–3.12) | 0.033 | 1.66 (0.80–3.45) | 0.174 | 1.74 (0.84–3.61) | 0.133 | 2.18 (0.88–5.40) | 0.092 |
| Readmission | 0.90 (0.53–1.52) | 0.695 | 0.94 (0.53–1.69) | 0.844 | 0.97 (0.53–1.76) | 0.907 | 0.78 (0.35–1.71) | 0.523 |
: statistically significant at α = 0.05
EBS: endoscopic biliary stenting; PTBD: percutaneous transhepatic biliary drainage; PS: propensity score; OR: odds ratio; CI: confidence interval; SSI: surgical site infection;
After multivariable adjustment, preoperative biliary drainage with PTBD remained significantly associated with DSM (OR 1.92, 95% CI 1.24–2.97, p=0.004), overall SSI (OR 1.74, 95% CI 1.10–2.76, p=0.019), and superficial SSI (OR 2.08, 95% CI 1.20–3.60, p=0.010) (Table 4). When the estimated probability of PTBD use was included in the PS adjusted model, DSM, overall SSI, and superficial SSI remained significantly associated with PTBD use (Table 4). Finally, a 1:1 PS matched cohort was constructed, where 94 patients in the PTBD group was matched with 94 patients in the EBS group. There were no differences in the baseline characteristics in the matched cohort with absolute standardized differences less than 0.1, except for preoperative hypoalbuminemia (standardized difference = 0.17). On multivariable analysis of this balanced cohort, DSM, overall SSI, and superficial SSI also remained significantly associated with PTBD use (Table 4).
Sensitivity Analysis
Patients who received PTBD had PS between 0.028 and 0.654, representing an estimated 2.8% to 65.4% probability of receiving PTBD based on their baseline characteristics (Figure 1). Conversely, patients who received EBS had PS between 0.013 and 0.667, representing an estimated 1.3% to 66.7% probability of receiving PTBD. Five hundred and fourteen (97.5%) patients had PS in the region of common support (i.e., PS between 0.028 and 0.654, representing 2.8%−65.4% likelihood of receiving PTBD), while 13 (2.5%) patients had PS outside of this region indicating an estimated zero probability of receiving the alternative procedure based on baseline characteristics. When only the subset of patients in the region of common support was analyzed, the association between PTBD and DSM remained significant (OR 1.97, 95% CI 1.28–3.02, p=0.002).
Figure 1.
Estimated probability to receive percutaneous transhepatic biliary drainage
EBS: endoscopic biliary stenting; PTBD: percutaneous transhepatic biliary drainage
An additional sensitivity analysis of the subset of 168 patients with hilar cholangiocarcinoma or gallbladder cancer who underwent either a total lobectomy or trisegmentectomy with biliary reconstruction was performed. Among this subset of patients with the most complex tumor and procedure characteristics, the primary outcome, DSM, remained significantly associated with PTBD use (OR 2.69, 95% CI 1.09–6.64, p=0.033).
Discussion
Preoperative biliary drainage has been shown to improve outcomes following hepatobiliary resection, but considerable debate over the best drainage technique still exists.(17) This study aimed to use a large national clinical registry to evaluate postoperative outcomes in patients who underwent preoperative biliary drainage with PTBD versus EBS prior to resection for hepatobiliary malignancy. The results showed that patients selected for PTBD had more preoperative co-morbidities with higher ASA scores and poorer nutritional status and had higher odds of 30-day morbidity and mortality.
Few other studies have specifically examined the association between patient characteristics and biliary drainage type. Kim et al. found that patients receiving PTBD were significantly more likely to have a preoperative diagnosis of hypertension.(11) Jo et al. found that the PTBD group had more Bismuth type IV tumors, while the EBS group had more type I and II tumors.(10) The results of this study similarly found that patients who received PTBD were significantly more likely to have preoperative co-morbidities, including hypoalbuminemia and higher ASA class.
Several studies have compared PTBD to EBS with respect to procedural success rates and post-procedure complications. One meta-analysis of 8 studies, including 3 randomized controlled trials (RCTs) and 5 observational studies, found that patients receiving PTBD were less likely to experience post-procedural cholangitis, but had similar rates therapeutic success and other complications.(8) However, it focused on palliative relief of obstructive jaundice rather than preoperative optimization. Several other systematic reviews and meta-analyses found that preoperative PTBD is slightly superior to EBS with regards to technical success rates, conversion rates, and post-procedure complications.(4, 6, 7) However, these published studies did not specifically evaluate postoperative outcomes.
One recent multi-center RCT randomized patients with resectable perihilar cholangiocarcinoma requiring major liver resection to receiving either endoscopic or percutaneous biliary drainage.(18) Although the primary outcome of the trial was severe complications between randomization and surgery, secondary outcomes included postoperative mortality and morbidity. The trial was stopped prematurely due to higher all-cause mortality in the PTBD group vs EBS group (11 [41%] vs 3 [11%], p=0.03). Three of the 11 deaths in the PTBD group occurred preoperatively, but the majority occurred following resection, similar to the findings of the this study. Several single-institution, observational studies have attempted to evaluate postoperative outcomes, but detected no significant differences between PTBD and EBS with respect to mortality and postoperative complications, except one study that found significantly lower long-term survival rates following PTBD.(9–11) However, these studies were all limited in sample size (fewer than 150 patients) and only evaluated patients with perihilar cholangiocarcinoma.
The present study used a large national dataset with rigorously collected 30-day postoperative outcomes in 527 patients from 84 hospitals. The patient population was not limited to only perihilar cholangiocarcinoma and instead included all hepatobiliary malignancies with the intent of improving generalizability. Further, several modeling approaches were used to account for differences in baseline characteristics and potential selection bias. Finally, this study is the first multi-institutional study that specifically evaluated the relationship between preoperative biliary drainage technique and postoperative outcomes following hepatobiliary resections.
Limitations
There were several important limitations to this study. First, the database only contained patients who underwent resection at a hospital participating in the ACS NSQIP Targeted Hepatectomy program, which excludes patients with resectable hepatobiliary malignancies who did not ultimately proceed to surgery, possibly secondary to complications from a drainage procedure. Second, the database did not include information on multiple drainage attempts. A subset of patients who underwent PTBD might have underwent prior failed EBS attempts and this more complex subset of patients might have biased the group to more complications. However, the PS adjustment and matching should have accounted for at least some of these differences as the propensity score represents the probability of receiving PTBD based on measured patient characteristics. The third limitation is related to the retrospective nature of the study. Patients selected for PTBD were found to have more preoperative co-morbidities, suggesting possible selection bias for drainage procedure. PTBD may have been preferentially used in patients too ill to tolerate anesthesia or those with tumors causing complex proximal obstruction. Although PS adjustment and matching can account for differences in known confounders contributing to the bias, it is unclear the extent to which these findings were related to differences in unmeasured factors. In addition to the variety of statistical methods used for adjustment, two sensitivity analyses were performed to address the potential effect of selection bias on the primary outcome. One sensitivity analysis included only the subset of patients with PS in the region of common support. Another sensitivity analysis included only the subset of patients with the most complex tumor and procedure characteristics (e.g., biliary reconstruction). Both sensitivity analyses yielded similar results to the primary analysis, suggesting a true association between PTBD and postoperative DSM.
In conclusion, PTBD was more likely to be used in patients with more preoperative comorbidities and malnutrition. Compared to EBS, PTBD was associated with significantly increased odds of DSM and SSI after adjustment for patient and operative characteristics. Although several statistical methodologies were used to account for patient selection, it is unclear the extent to which these findings are related to unmeasured factors. Particularly, PTBD is frequently used to relieve biliary obstruction in patients in which ERCP had failed or could not be performed due to prior surgically altered anatomy, potentially biasing the PTBD group to more complex cases. Further investigations are still needed to conclusively determine the independent effect of preoperative drainage technique on postoperative outcomes. The optimal approach should continue to be evaluated on a case-by-case basis according to the individual patient characteristics, clinical scenario, and available expertise.
Acknowledgments
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Drs. Q.L. Hu and J.Y. Liu receives salary support through a contract with Agency for Healthcare Research & Quality (HHSP233201500020I). Dr. R.J. Ellis is supported by a postdoctoral research fellowship (AHRQ 5T32HS000078). Dr. R.P. Merkow is supported by the Agency for Healthcare Research and Quality (K12HS023011) and an Institutional Research Grant from the American Cancer Society (IRG-18–163-24).
Footnotes
Disclosures: None
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Iacono C, Ruzzenente A, Campagnaro T, Bortolasi L, Valdegamberi A, Guglielmi A. Role of preoperative biliary drainage in jaundiced patients who are candidates for pancreatoduodenectomy or hepatic resection: highlights and drawbacks. Annals of surgery. 2013;257(2):191–204. [DOI] [PubMed] [Google Scholar]
- 2.Dixon JM, Armstrong CP, Duffy SW, Davies GC. Factors affecting morbidity and mortality after surgery for obstructive jaundice: a review of 373 patients. Gut. 1983;24(9):845–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Farges O, Regimbeau JM, Fuks D, et al. Multicentre European study of preoperative biliary drainage for hilar cholangiocarcinoma. The British journal of surgery. 2013;100(2):274–283. [DOI] [PubMed] [Google Scholar]
- 4.Tang Z, Yang Y, Meng W, Li X. Best option for preoperative biliary drainage in Klatskin tumor: A systematic review and meta-analysis. Medicine. 2017;96(43):e8372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hwang S, Song GW, Ha TY, et al. Reappraisal of percutaneous transhepatic biliary drainage tract recurrence after resection of perihilar bile duct cancer. World journal of surgery. 2012;36(2):379–385. [DOI] [PubMed] [Google Scholar]
- 6.Al Mahjoub A, Menahem B, Fohlen A, et al. Preoperative Biliary Drainage in Patients with Resectable Perihilar Cholangiocarcinoma: Is Percutaneous Transhepatic Biliary Drainage Safer and More Effective than Endoscopic Biliary Drainage? A Meta-Analysis. Journal of vascular and interventional radiology : JVIR. 2017;28(4):576–582. [DOI] [PubMed] [Google Scholar]
- 7.Hameed A, Pang T, Chiou J, et al. Percutaneous vs. endoscopic pre-operative biliary drainage in hilar cholangiocarcinoma - a systematic review and meta-analysis. HPB : the official journal of the International Hepato Pancreato Biliary Association. 2016;18(5):400–410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhao XQ, Dong JH, Jiang K, Huang XQ, Zhang WZ. Comparison of percutaneous transhepatic biliary drainage and endoscopic biliary drainage in the management of malignant biliary tract obstruction: a meta-analysis. Digestive endoscopy : official journal of the Japan Gastroenterological Endoscopy Society. 2015;27(1):137–145. [DOI] [PubMed] [Google Scholar]
- 9.Hirano S, Tanaka E, Tsuchikawa T, et al. Oncological benefit of preoperative endoscopic biliary drainage in patients with hilar cholangiocarcinoma. Journal of hepato-biliary-pancreatic sciences. 2014;21(8):533–540. [DOI] [PubMed] [Google Scholar]
- 10.Jo JH, Chung MJ, Han DH, et al. Best options for preoperative biliary drainage in patients with Klatskin tumors. Surgical endoscopy. 2017;31(1):422–429. [DOI] [PubMed] [Google Scholar]
- 11.Kim KM, Park JW, Lee JK, Lee KH, Lee KT, Shim SG. A Comparison of Preoperative Biliary Drainage Methods for Perihilar Cholangiocarcinoma: Endoscopic versus Percutaneous Transhepatic Biliary Drainage. Gut and liver. 2015;9(6):791–799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ko CY, Hall BL, Hart AJ, Cohen ME, Hoyt DB. The American College of Surgeons National Surgical Quality Improvement Program: achieving better and safer surgery. Joint Commission journal on quality and patient safety. 2015;41(5):199–204. [DOI] [PubMed] [Google Scholar]
- 13.Rahbari NN, Garden OJ, Padbury R, et al. Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS). Surgery. 2011;149(5):713–724. [DOI] [PubMed] [Google Scholar]
- 14.Allison PD. Handling Missing Data by Maximum Likelihood. In: SAS Global Forum 2012: Statistics and Data Analysis. .2012. [Google Scholar]
- 15.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Statistics in medicine. 2009;28(25):3083–3107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Stuart EA. Matching methods for causal inference: A review and a look forward. Statistical science : a review journal of the Institute of Mathematical Statistics. 2010;25(1):1–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kennedy TJ, Yopp A, Qin Y, et al. Role of preoperative biliary drainage of liver remnant prior to extended liver resection for hilar cholangiocarcinoma. HPB : the official journal of the International Hepato Pancreato Biliary Association. 2009;11(5):445–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Coelen RJS, Roos E, Wiggers JK, et al. Endoscopic versus percutaneous biliary drainage in patients with resectable perihilar cholangiocarcinoma: a multicentre, randomised controlled trial. The lancet Gastroenterology & hepatology. 2018;3(10):681–690. [DOI] [PubMed] [Google Scholar]