Abstract
Background
Post-hepatectomy liver failure (PHLF) remains a significant complication after hepatic resection. This study aims to determine the rate of PHLF in patients undergoing resection of 3 or fewer segments and analyze the association of PHLF with perioperative characteristics and postoperative complications.
Methods
The ACS hepatectomy-targeted NSQIP database was queried for patients undergoing left hemi-hepatectomy or partial resection from 2014-2018. The primary outcome was PHLF, defined by ISGLS. Multivariable logistic regression models assessed the association between PHLF, preoperative and operative variables and postoperative complications.
Results
Among 7029 patients, 187 (2.7%) experienced PHLF, with clinically significant (grade B/C) PHLF in 1.4%. PHLF was associated with older age, male gender, higher ASA classification, ascites, and elevated SGOT. Preoperative ascites (OR 4.94, 95%CI:2.45-9.94, p<0.001) had the strongest association with PHLF. There was no association between PHLF and concurrent colorectal resection, neoadjuvant therapy, or concurrent ablation. Surgical site infection (OR 3.64, 95%CI:2.40-5.54, p<0.001), sepsis (OR 3.78, 95%CI:2.16-6.61, p<0.001), postoperative invasive procedure (OR 6.92, 95%CI:4.91-9.76, p<0.001), and bile leak (OR 4.65, 95%CI:3.04-7.12, p<0.001) were associated with PHLF.
Discussion
PHLF after minor hepatectomy is rare and associated with signs of preoperative liver dysfunction. The association with infectious complications suggests a multifactorial etiology and provides targets for quality improvement.
Introduction
Partial hepatectomy (PH) is a potentially curative treatment option for primary and secondary hepatic malignancies. While perioperative morbidity and mortality associated with PH have significantly decreased due to improved patient selection and perioperative management1-4, post-hepatectomy liver failure (PHLF) remains a significant cause of postoperative mortality after PH.5,6 Reported incidence of PHLF is widely variable, ranging from 0.7 to 32%, likely attributable to patient selection, extent of hepatic resection, baseline liver function, and perioperative management.7-9 Given the gravity of this complication, multiple risk-calculators and models have been developed to predict PHLF by incorporating clinical predictors, laboratory values, and imaging features.5,7,10-14
Continued efforts to adopt parenchymal-sparing techniques in liver surgery have reduced morbidity and altered the composition of postoperative complications. A recent multi-institutional series evaluating non-cirrhotic patients undergoing open PH observed lower 90-day mortality and severe morbidity for minor hepatectomy (MiH) compared to right hepatectomy (RH). In this series, PHLF was observed in 2.4% of MiH and 11.6% of RH.15 Previously published data by Kingham et al. from Memorial Sloan Kettering Cancer Center further illustrated a decreased rate of PHLF over time, likely in part due to increased use of parenchymal-sparing resections, with a reported rate of 2%.1 Notably, in this study, only 2 of the 31 mortalities after MiH were due to PHLF, a total of 0.1% of all MiH. Based on this single institution experience, the current study aims to test the hypothesis that national rates of PHLF after MiH are extremely low. Such a finding would support further adoption of parenchymal-sparing techniques, whenever possible.
The introduction of a hepatectomy-targeted dataset in 2014 from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) offers the unique opportunity to both quantify the national rates of PHLF after MiH and to determine associated clinical variables with this complication. While the number of patients and hospitals included in the ACS-NSQIP vary per year, the hepatectomy-targeted dataset includes up to 133 hospitals per year from which data on between 3,854 and 4,773 patients are submitted annually.
Understanding the factors associated with PHLF in the perioperative period can help guide therapeutic decision-making and prognosticate for patients requiring PH. With increased adoption of parenchymal-sparing techniques, studies that focus specifically on outcomes after MiH are indicated, as these may differ from more extensive resections. The aims of this project were to 1) determine the rate of PHLF in patients undergoing MiH, defined as resection of 3 or fewer segments; 2) understand the high-risk patient and perioperative characteristics associated with PHLF in patients undergoing MiH and 3) characterize the association between PHLF and other postoperative complications that may serve as other targets for quality improvement.
Methods
National Surgical Quality Improvement Program
ACS NSQIP is a national prospective clinical data registry, which includes over 250 variables for a variety of surgical procedures. These variables include patient demographics, comorbidities, operative variables, and postoperative complications. Full details of the NSQIP database have been previously described elsewhere.16 Since 2014, NSQIP has maintained a targeted database for patients who undergo hepatectomy. This dataset includes procedure-specific variables such as neoadjuvant therapy; viral hepatitis; liver texture; intraoperative Pringle maneuver and biliary reconstruction, as well as hepatectomy-specific complication data including PHLF and bile leakage. Data collected by NSQIP are de-identified and Health Insurance Portability and Accountability Act (HIPAA)-compliant. Institutional Review Board approval at Memorial Sloan Kettering Cancer Center was obtained for this study.
Patient and Variable Selection
The hepatectomy-targeted participant use file was queried for Current Procedural Terminology (CPT) codes 47125 (left hemi-hepatectomy) and 47120 (partial hepatectomy) to identify all patients undergoing resection of three or fewer liver segments from 2014 through 2018. Patients with concurrent partial resections were excluded to ensure that the study cohort was limited to those with three or fewer resected segments. Selected variables included demographic factors, such as age, race/ethnicity, gender, and body mass index (BMI); comorbidities including diabetes mellitus (DM), hepatitis, smoking, preoperative ascites, preoperative bleeding disorders and American Society of Anesthesiologists (ASA) physical status classification; preoperative laboratory variables including hematocrit, international normalized ratio (INR) and liver function tests; operative variables including surgical approach, wound class, liver texture, drain placement, Pringle maneuver and biliary reconstruction; and postoperative complications. Bleeding disorders are defined by NSQIP as any chronic or active condition that places a patient at an increased risk for excessive bleeding; these include vitamin K deficiency, hemophilia, thrombocytopenia, or chronic anticoagulation therapy that has not been discontinued prior to surgery and exclude chronic aspirin or NSAID use, remote history of heparin-induced thrombocytopenia, and use of prophylactic anticoagulants. Liver texture, defined as cirrhotic, fibrotic, congested, fatty, or normal, is also captured in the hepatectomy-targeted dataset. Cirrhotic, fibrotic, congested, and fatty liver textures were defined as abnormal for this study. Thirty-day mortality was also captured.
Outcomes
The primary outcome of the study was PHLF after MiH, defined by the International Study Group of Liver Surgery (ISGLS) as elevated INR and hyperbilirubinemia on or after postoperative day 5 (POD 5) or hyperbilirubinemia and the need for fresh frozen plasma (FFP) on or after POD 5. Patients with PHLF are further categorized in the NSQIP dataset based on ISGLS grade, which characterizes the severity of PHLF based on the required intervention. In this classification system, Grade A PHLF represents deterioration of liver function in the postoperative period requiring no change in clinical management; Grade B PHLF is defined as a deterioration requiring deviation from regular postoperative care without the need for invasive intervention; and Grade C PHLF requires an invasive procedure. While clinically relevant PHLF may be limited to grade B and C in practice, all PHLF patients were included in this study to appropriately characterize the severity of this outcome after MiH.
Statistical Analysis
The association between PHLF and patient preoperative and operative variables was tested using Fischer's exact test for binary variables, Chi-square testing for non-binary categorical variables, t-tests for continuous parametric variables, and Wilcoxon rank-sum for continuous non-parametric variables. To estimate adjusted associations, a multivariable logistic regression model was constructed including all covariates with p < 0.1 on univariable analysis. A backward stepwise selection method was used to incrementally remove noncontributory predictors. Next, the relationship between PHLF and other postoperative complications including surgical site infection, sepsis, postoperative invasive intervention, unplanned reoperation, bile leak, pneumonia, failure to wean from ventilator > 48 hours, pulmonary embolism and myocardial infarction was evaluated using Fisher’s exact test. Individual multivariable logistic regression models were estimated to evaluate the association between PHLF and each complication, including all covariates from the prior model. Estimation of variance in models was adjusted to account for clustering at the hospital level.17 Tests of significance were two-sided with a threshold of p < 0.05. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).
Results
A total of 7029 patients were identified between 2014-2018. Of these, 187 (2.7%) patients experienced PHLF. Eighty-seven patients experienced Grade A PHLF (46.5% of PHLF, 1.2% of overall sample); 49 patients Grade B PHLF (26.2% of PHLF, 0.7% of overall sample); and 51 patients Grade C PHLF (27.3% of PHLF, 0.7% of overall sample). The most common indication for resection was secondary hepatic malignancy (2864/7029, 40.8%) and most patients underwent open resection (4928/7029, 70.1%). Receipt of neoadjuvant chemotherapy or concurrent ablation was heterogenous in this sample, with rates of 19.3% (1357/7029) and 9.5% (669/7029), respectively. Concurrent colorectal resection was performed in 5.9% of patients (415/7029) and hepatic arterial infusion pumps (HAIP) were concurrently placed in 62 patients (0.9%). Clinicopathologic characteristics associated with PHLF are listed in Table 1 and 2.
Table 1.
Patient Characteristics associated with Post-Hepatectomy Liver Failure after Minor Hepatectomy
| Total, n (%) | PHLF, n (%) | No PHLF, n (%) | P value | |
|---|---|---|---|---|
| Overall | 7029 | 187 (2.7) | 6842 (97.3) | |
| Age, mean (SD) | 59.8 (14.0) | 64.9 (11.2) | 59.7 (14.0) | < 0.001 |
| Sex | ||||
| Male | 3399 (48.4) | 132 (3.9) | 3267 (96.1) | < 0.001 |
| Female | 3630 (51.6) | 55 (1.5) | 3575 (98.5) | |
| BMI, mean (SD) | 28.6 (6.4) | 28.3 (5.8) | 28.6 (6.4) | 0.650 |
| Race/Ethnicity | 0.349 | |||
| Non-Hispanic white | 4383 (62.4) | 107 (2.4) | 4276 (97.6) | |
| Non-Hispanic black | 529 (7.5) | 13 (2.5) | 516 (97.5) | |
| Hispanic | 332 (4.7) | 9 (2.7) | 323 (97.3) | |
| Other/Unknown | 1785 (25.4) | 58 (3.3) | 1727 (96.8) | |
| ASA Class | < 0.001 | |||
| I/II | 1972 (28.1) | 21 (1.1) | 1951 (98.9) | |
| III/IV/V | 5057 (71.9) | 166 (3.3) | 4891 (96.7) | |
| Diabetes | 0.098 | |||
| Yes | 1247 (17.7) | 42 (3.4) | 1205 (96.6) | |
| No | 5782 (82.3) | 145 (2.5) | 5637 (97.5) | |
| Smoking | 0.303 | |||
| Yes | 1079 (15.4) | 34 (3.1) | 1045 (96.9) | |
| No | 5950 (84.7) | 153 (2.6) | 5797 (97.4) | |
| Preoperative ascites, n (%) | < 0.001 | |||
| Yes | 48 (0.7) | 9 (18.8) | 39 (81.3) | |
| No | 6981 (99.3) | 178 (2.6) | 6803 (97.5) | |
| Preoperative sepsis, n (%) | 0.095 | |||
| Yes | 91 (1.3) | 5 (5.5) | 86 (94.5) | |
| No | 6938 (98.7) | 182 (2.6) | 6756 (97.4) | |
| Bleeding disorder, n (%) | < 0.001 | |||
| Yes | 250 (3.6) | 20 (8.0) | 230 (92.0) | |
| No | 6779 (96.4) | 167 (2.5) | 6612 (97.5) | |
| Preoperative transfusion, n (%) | 0.095 | |||
| Yes | 41 (0.6) | 3 (7.3) | 38 (92.7) | |
| No | 6988 (99.4) | 184 (2.6) | 6804 (97.4) | |
| Emergency surgery, n (%) | 0.052 | |||
| Yes | 53 (0.8) | 4 (7.6) | 49 (92.5) | |
| No | 6976 (99.3) | 183 (2.6) | 6793 (97.4) | |
| Bilirubin >1.2 | < 0.001 | |||
| Yes | 421 (6.0) | 26 (6.2) | 395 (93.8) | |
| No | 6608 (94.0) | 161 (2.4) | 6447 (97.6) | |
| SGOT > 40 | < 0.001 | |||
| Yes | 1216 (17.3) | 82 (6.7) | 1134 (93.3) | |
| No | 5813 (82.7) | 105 (1.8) | 5708 (98.2) | |
| INR > 1.5 | 0.148 | |||
| Yes | 187 (2.7) | 4 (5.2) | 73 (94.8) | |
| No | 6842 (97.3) | 183 (2.6) | 6769 (97.4) | |
| Preoperative alkaline phosphatase | 106.1 (80.6) | 134.1 (106.0) | 105.3 (79.7) | < 0.001 |
| Preoperative hematocrit | 39.6 (4.9) | 39.4 (5.9) | 39.6 (4.9) | 0.625 |
| Neoadjuvant Therapy, n (%) | 0.511 | |||
| Yes | 1357 (19.3) | 32 (2.4) | 1325 (97.6) | |
| No/Unknown | 5672 (80.7) | 155 (2.7) | 5517 (97.3) | |
| Viral Hepatitis, n (%) | < 0.001 | |||
| B and/or C | 888 (12.6) | 46 (5.2) | 842 (94.8) | |
| None/Other/Unknown | 6141 (87.4) | 141 (2.3) | 6000 (97.7) | |
| Biliary Stent | 0.024 | |||
| Yes | 248 (3.5) | 13 (5.2) | 235 (94.8) | |
| No | 6781 (96.5) | 174 (2.6) | 6607 (97.4) |
Abbreviations: SD, standard deviation; BMI, body mass index; ASA, American Society of Anesthesiologists Classification; SGOT, aspartate transaminase; INR, international normalized ratio
Table 2.
Treatment, Pathologic and Operative Characteristics associated with Post-Hepatectomy Liver Failure after Minor Hepatectomy
| Total, n (%) | PHLF, n (%) | No PHLF, n (%) | P value | |
|---|---|---|---|---|
| Overall | 7029 | 187 (2.7) | 6842 (97.3) | |
| Wound Class | 0.239 | |||
| Clean, n (%) | 1117 (15.9) | 23 (2.1) | 1094 (97.9) | |
| Clean-Contaminated, n (%) | 5455 (77.6) | 148 (2.7) | 5307 (97.3) | |
| Contaminated/Dirty, n (%) | 457 (6.5) | 16 (3.5) | 441 (96.5) | |
| Surgical Pathology | < 0.001 | |||
| Benign, n (%) | 1964 (27.9) | 19 (1.0) | 1945 (99.0) | |
| Primary Hepatobiliary, n (%) | 2201 (31.3) | 120 (5.5) | 2081 (94.6) | |
| Secondary/Unknown, n (%) | 2864 (40.8) | 48 (1.7) | 2816 (98.3) | |
| Liver Texture | < 0.001 | |||
| Normal/Not documented, n (%) | 5320 (75.7) | 97 (1.8) | 1619 (94.7) | |
| Cirrhotic/Congested/Fatty/Fibrotic, n (%) | 1709 (24.3) | 90 (5.3) | 5223 (98.2) | |
| Number of tumors, median (IQR) | 1 (1 - 2) | 1 (1 - 1) | 1 (1 - 2) | < 0.001 |
| Operative Time (min), mean (SD) | 207.1 (107.1) | 262.5 (121.9) | 205.6 (106.4) | < 0.001 |
| Surgical Approach | < 0.001 | |||
| Minimally Invasive, n (%) | 2101 (29.9) | 25 (1.2) | 2076 (98.8) | |
| Open/Other, n (%) | 4928 (70.1) | 162 (3.3) | 4766 (96.7) | |
| Procedure Type | 0.111 | |||
| Partial Lobectomy, n (%) | 6047 (86.0) | 149 (2.5) | 5898 (97.5) | |
| Left hemi-hepatectomy, n (%) | 982 (14.0) | 38 (3.9) | 944 (96.1) | |
| Concurrent colorectal resection | 0.521 | |||
| Yes, n (%) | 415 (5.9) | 9 (2.2) | 406 (97.8) | |
| No, n (%) | 6614 (94.1) | 178 (2.7) | 6436 (97.3) | |
| Concurrent ablation, n (%) | 0.379 | |||
| Yes | 669 (9.5) | 14 (2.1) | 655 (97.9) | |
| No | 6360 (90.5) | 173 (2.7) | 6187 (97.3) | |
| Pringle Maneuver, n (%) | < 0.001 | |||
| Yes | 1291 (18.4) | 59 (4.6) | 1232 (95.4) | |
| No | 5738 (81.6) | 128 (2.2) | 5610 (97.8) | |
| Biliary Reconstruction, n (%) | < 0.001 | |||
| Yes | 274 (3.9) | 21 (7.7) | 253 (92.3) | |
| No | 6755 (96.1) | 166 (2.5) | 6589 (97.5) | |
| Surgical Drain, n (%) | < 0.001 | |||
| Yes | 2866 (40.8) | 113 (3.9) | 2753 (96.1) | |
| No | 4163 (59.2) | 74 (1.8) | 4089 (98.2) |
Abbreviations: SD, standard deviation; IQR, interquartile range
Multivariable analysis of variables predicting PHLF are shown in Table 3. The multivariable model was subsequently performed evaluating only patients with grade B or C PHLF. While preoperative ascites, elevated SGOT, primary hepatobiliary pathology, and abnormal liver texture remained associated with PHLF in this subgroup, the associations with patient age, gender, and ASA classification were no longer statistically significant (p = 0.057-0.739). (Table 4)
Table 3.
Multivariable Analysis for All Patients with PHLF
| PHLF n (%) |
Adjusted Odds Ratio (95% CI) |
P value | |
|---|---|---|---|
| Age, mean (SD)* | 64.9 (11.2) | 1.21 (1.05 - 1.38) | 0.008 |
| Sex | |||
| Male | 132 (3.9) | 1.00 (REF) | - |
| Female | 55 (1.5) | 0.62 (0.45 - 0.85) | 0.003 |
| ASA Class | |||
| I/II | 21 (1.1) | 1.00 (REF) | - |
| III/IV/V | 166 (3.3) | 1.85 (1.10 - 3.11) | 0.021 |
| Ascites | 9 (18.8) | 4.94 (2.45 - 9.94) | < 0.001 |
| Bleeding disorder | 20 (8.0) | 2.10 (1.18 - 3.44) | 0.011 |
| Elevated SGOT > 40 | 82 (6.7) | 2.11 (1.53 - 2.92) | < 0.001 |
| Surgical Pathology | |||
| Benign | 19 (1.0) | 1.00 (REF) | - |
| Primary Hepatobiliary | 120 (5.5) | 2.14 (1.29 - 3.54) | 0.003 |
| Secondary/Unknown | 48 (1.7) | 1.09 (0.74 - 1.59) | 0.677 |
| Operative Time (hours), mean (SD)** | 4.4 (2.0) | 1.17 (1.09 - 1.26) | < 0.001 |
| Surgical Approach | |||
| Minimally Invasive | 25 (1.2) | 1.00 (REF) | - |
| Open/Other | 162 (3.3) | 1.98 (1.33 - 2.94) | < 0.001 |
| Liver Texture | |||
| Normal/Not documented | 97 (1.8) | 1.00 (REF) | - |
| Cirrhotic/Congested/Fatty/Fibrotic | 90 (5.3) | 1.92 (1.36 - 2.72) | < 0.001 |
| Pringle Maneuver | 59 (4.6) | 1.60 (0.91 - 2.82) | 0.100 |
| Surgical Drain | 113 (3.9) | 1.57 (0.95 - 2.60) | 0.078 |
OR reported for each additional 10 years
OR reported for each additional hour
Abbreviations: PHLF, post-hepatectomy liver failure; SD, standard deviation; ASA, American Society of Anesthesiologists Classification; SGOT, aspartate transaminase; MIS, minimally invasive surgery
Table 4.
Multivariable Analysis for Patients with Grade B/C PHLF
| Grade B/C PHLF n (%) |
Adjusted Odds Ratio (95% CI) |
P value | |
|---|---|---|---|
| Age, mean (SD)* | 65.0 (10.5) | 1.19 (0.99 - 1.42) | 0.069 |
| Sex | |||
| Male | 65 (1.9) | 1.00 (REF) | - |
| Female | 35 (1.0) | 0.93 (0.59 - 1.46) | 0.739 |
| ASA Class | |||
| I/II | 10 (0.5) | 1.00 (REF) | - |
| III/IV/V | 90 (1.8) | 1.99 (0.98 - 4.05) | 0.057 |
| Ascites | 6 (12.5) | 4.63 (1.56 - 13.78) | 0.006 |
| Bleeding disorder | 14 (5.6) | 2.62 (1.09 - 6.32) | 0.032 |
| Elevated SGOT > 40 | 50 (4.1) | 2.51 (1.72 - 3.64) | < 0.001 |
| Surgical Pathology | |||
| Benign | 6 (0.3) | 1.00 (REF) | - |
| Primary Hepatobiliary | 70 (3.2) | 4.03 (2.12 - 7.64) | < 0.001 |
| Secondary/Unknown | 24 (0.8) | 1.79 (0.94 - 3.40) | 0.074 |
| Operative Time (hours), mean (SD)** | 259.1 (130.8) | 1.16 (1.06 - 1.27) | 0.002 |
| Surgical Approach | |||
| Minimally Invasive | 12 (0.6) | 1.00 (REF) | - |
| Open/Other | 88 (1.8) | 2.42 (1.44 - 4.06) | < 0.001 |
| Liver Texture | |||
| Normal/Not documented | 47 (0.9) | 1.00 (REF) | - |
| Cirrhotic/Congested/Fatty/Fibrotic | 53 (3.1) | 2.18 (1.40 - 3.39) | < 0.001 |
| Pringle Maneuver | 29 (2.3) | 1.35 (0.75 - 2.41) | 0.315 |
| Surgical Drain | 56 (2.0) | 1.21 (0.78 - 1.88) | 0.401 |
OR reported for each additional 10 years
OR reported for each additional hour
Abbreviations: PHLF, post-hepatectomy liver failure; SD, standard deviation; ASA, American Society of Anesthesiologists Classification; SGOT, aspartate transaminase; MIS, minimally invasive surgery
As patients who developed PHLF appeared to have evidence of preoperative liver dysfunction, a subgroup analysis was performed evaluating only patients without evidence of liver dysfunction, by excluding patients with preoperative ascites, abnormal preoperative bilirubin, SGOT, or alkaline phosphatase, viral hepatitis, abnormal liver texture, or elevated INR. A total of 3353 patients met these criteria, of which 41 (1.2%) developed PHLF. The distribution of PHLF by grade in this cohort was as follows: 22 (0.7%) grade A; 12 (0.4%) grade B; and 7 (0.2%) grade C. In these patients, older age (mean 67.2 years, OR 1.59, 95%CI: 1.29-1.97, p < 0.001), longer operative time (mean 4.3 hours, OR 1.24, 95%CI: 1.07-1.44, p = 0.005), and open approach (1.7%, OR 6.22, 95%CI: 1.57-24.63, p = 0.010) were independently associated with PHLF, while female gender was protective (0.6%, OR 0.43, 95%CI: 0.23-0.81, p = 0.009).
An additional subgroup analysis assessed outcomes for patients with primary hepatobiliary malignancy including hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC), as these patients are presumably at higher risk for PHLF, given potential associated underlying liver disease. We identified 1894 patients with HCC or CC, with a PHLF rate of 5.9% (n = 111). Older age (mean 67.1, OR 1.28, 95%CI: 1.05-1.55, p = 0.014), preoperative ascites (36.8%, OR 6.53, 95%CI: 2.55-16.71, p < 0.001) or bleeding disorders (13.7%, OR 2.46, 95%CI: 1.18-5.14, p = 0.017), SGOT > 40 (9.7%, OR 2.10, 95%CI: 1.30-3.38, p = 0.003), longer operative time (mean 4.3 hours, OR 1.20, 95%CI: 1.11-1.30, p < 0.001), open approach (7.0%, OR 2.31, 95%CI: 1.34-4.00, p = 0.003), and abnormal liver texture (7.6%, OR 1.85, 95%CI: 1.21-2.82, p = 0.005) were all independently associated with PHLF in patients with primary hepatobiliary malignancy.
Postoperative complications including surgical site infection, sepsis, postoperative invasive procedures, bile leak, and unplanned reoperation were associated with PHLF on univariable analysis. Notably, 16.6% of patients with PHLF experienced death within 30 days after surgery compared to 0.4% in the cohort without PHLF (p < 0.001). A multivariable model confirmed an independent association between surgical site infection, sepsis, unplanned return to the operating room, postoperative invasive procedures, bile leak, pneumonia, failure to wean from the ventilator > 48 hours, myocardial infarction and PHLF. Notably, pulmonary embolism was not associated with PHLF. (Table 5) Similarly, for the subgroup of patients without evidence of liver dysfunction and for the subgroup of patients with primary hepatobiliary malignancy, PHLF remained independently associated with all complications, except pulmonary embolism.
Table 5.
Adjusted odds of PHLF among patients with and without other postoperative complications
| PHLF among patients with complication n (%) |
PHLF among patients without complication n (%) |
Adjusted Odds Ratio (95% CI) |
P value | |
|---|---|---|---|---|
| Surgical Site Infection | 61/604 (10.1) | 126/6425 (2.0) | 3.64 (2.40 - 5.54) | < 0.001 |
| Sepsis | 24/190 (12.6) | 163/6839 (2.4) | 3.78 (2.16 - 6.61) | < 0.001 |
| Unplanned return to OR | 31/138 (22.5) | 156/6891 (2.3) | 7.35 (4.21 - 12.83) | < 0.001 |
| Invasive Intervention Post-op | 77/511 (15.1) | 110/6518 (1.7) | 6.92 (4.91 - 9.76) | < 0.001 |
| Bile Leak | 47/357 (13.2) | 140/6672 (2.1) | 4.65 (3.04 - 7.12) | < 0.001 |
| Pneumonia | 35/180 (19.4) | 152/6849 (2.22) | 6.92 (4.63-10.34) | < 0.001 |
| Failure to wean from ventilator >48h | 38/97 (39.2) | 149/6932 (2.2) | 15.83 (9.02-27.76) | < 0.001 |
| Pulmonary embolism | 6/71 (8.5) | 181/6958 (2.6) | 2.37 (0.73-7.76) | 0.152 |
| Myocardial infarction | 9/54 (16.7) | 178/6975 (2.6) | 5.15 (2.32-11.43) | < 0.001 |
Discussion
PHLF is a significant, potentially fatal complication after PH. With increased utilization of parenchymal-sparing surgical techniques and better perioperative management, rates of PHLF have decreased.18-20 While prior studies have focused on developing diagnostic criteria and predictive models for PHLF, few have attempted to characterize patients with PHLF after MiH and understand the relationship between PHLF and other complications. In this large, national database, a very low rate of PHLF was observed in patients who underwent resection of 3 or fewer segments, with clinically significant PHLF in only 1.4% of patients. High risk patients included those with preoperative ascites, primary hepatobiliary disease, bleeding disorders, elevated liver function tests, and abnormal liver texture. Notably, other complications, including infection, bile leak, and sepsis, were associated with PHLF, signifying a potential multifactorial etiology of PHLF in these patients. These associated complications may serve as additional targets for quality improvement initiatives.
Rates of PHLF after PH are heterogenous in the literature.5,8 MiH has traditionally been associated with lower rates of PHLF compared to major resection.21 A recently published study by Vigano et al. evaluating MiH reported PHLF rates ranging from 1.4% in limited resections to 9.7% for right posterior sectionectomy.15 In our series, the rate of PHLF after MiH was 2.7%. Notably, only 1.4% of these patients had clinically significant (grade B or C) liver dysfunction, suggesting that this complication is rare and likely mild for most affected patients. As NSQIP does not report prior PH, this rate of PHLF may overestimate PHLF for patients with no prior surgical history.
Predictive factors associated with PHLF have been previously described.22-27 Patient-related risk factors include diabetes, obesity, malnutrition, renal insufficiency, lung disease, and advanced age. Similarly, patients with PHLF in our series were more likely to be older and have higher ASA classification, a proxy for patient fitness. Interestingly, other co-morbidities including diabetes, obesity, renal insufficiency and pulmonary disease were not found to be associated with PHLF for patients undergoing MiH. In this study, the primary risk of PHLF appeared to be related to preoperative evidence of liver dysfunction, given the association with preoperative ascites, abnormal liver texture and elevated liver function tests. These associations remained present when evaluating only patients with grade B and C PHLF, whereas other patient factors such as age and ASA classification were no longer statistically significant in this select cohort. When excluding patients with evidence of liver disease, rates of clinically significant PHLF were less than 1%. Taken together, these results, which are specific to MiH, are clinically informative and demonstrate that patients undergoing MiH may require a different risk stratification strategy than those undergoing major hepatectomy.
In addition to those with evidence of preoperative liver dysfunction, patients with primary hepatobiliary malignancies were also more likely to develop PHLF, with a rate of 5.9% in this subgroup. Risk factors for PHLF were similar, including older age, ascites, bleeding disorders, open surgery, longer operative time, and abnormal liver texture. Ascites remained the most significant risk factor for PHLF, suggesting that even MiH in patients with HCC or CC and ascites may carry substantial risk. Given the high risk of cirrhosis and potential chronic liver dysfunction in these patients, careful, upfront evaluation of liver function and portal hypertension is essential for appropriate patient selection prior to MiH for these indications.
Operative factors previously associated with PHLF include blood loss > 1200 milliliters, transfusion requirements, operative time, major hepatectomy, and postoperative complications.23,26 In this series, operative approach and total operative time were associated with PHLF, whereas use of Pringle maneuver and drain placement were not. One anecdotal hypothesis for this finding is that patients with grossly diseased livers are less likely to be offered minimally invasive surgery, given the technical difficulty associated with this approach in these patients. As such, while not directly related to underlying liver disease, operative approach may serve as a surrogate for liver dysfunction. Longer operative times may similarly be a proxy for worse underlying liver disease or increased case complexity.
In this study, preoperative chemotherapy and ablation were not associated with PHLF. For patients undergoing PH for malignancy, prior associations between chemotherapy use, steatohepatitis, and mortality have been demonstrated.28,29 A recent ACS-NSQIP based risk calculator assessing all PH patients similarly found an association between PHLF and chemotherapy.10 This discordant finding suggests that the effects of hepatic toxicity secondary to chemotherapy may not be as clinically significant with resection of less parenchyma. Such data may advocate for greater use of cytotoxic chemotherapy, HAIP, or concurrent use of ablation, as techniques to decrease tumor burden to minimize parenchymal resection. Further studies to confirm this association are indicated.
Several other postoperative complications, including surgical site infection, sepsis, pneumonia, myocardial infarction and bile leak were associated with PHLF in this study. Although the direction of causality is unknown, patients with cardiopulmonary or infectious complications postoperatively may be more likely to subsequently develop PHLF. An association between sepsis and PHLF has been previously described in a series of 19 patients with intraperitoneal sepsis after hepatectomy, where reported PHLF-related death occurred in 13 patients.30 Endotoxin production has also been associated with PHLF, through a mechanism of impaired cytokine release and reduced availability of growth factors necessary for hepatic regeneration.31,32 Based on these findings, it is possible that postoperative complications in patients with MiH contribute to a multifactorial process that drives the uncommon scenario of PHLF. Notably, this association is present in patients with and without evidence of preoperative liver dysfunction and regardless of indication for resection. While further study must be performed to establish this causal association in MiH patients, complications such as postoperative infection may serve as a reasonable target for hepatectomy-specific quality improvement initiatives.
This study has several limitations. First, data from ACS-NSQIP is retrospective, may not be fully generalizable, and does not allow exploration of specific process measures that may be targets for quality improvement. Future studies evaluating rates of PHLF based on hospital volume and practice patterns may help identify specific processes and areas for improvement. Second, thorough evaluation of preoperative risk factors such as duration and type of neoadjuvant therapy, etiology of preoperative ascites or indication for biliary stent placement is limited in the NSQIP dataset. Clear delineation of the clinical decision making to perform resection versus ablation is also not readily available. Third, as NSQIP diagnoses are categorized by CPT code, more granular evaluation of specific minor anatomic resections is not possible. Previous published data by Vigano et al. suggest that minor resections have wide variation in outcomes based on anatomy and complexity of resection.15 Those findings cannot be reproduced by these data. Fourth, there is no variable in NSQIP to capture prior liver resection or liver volume prior to PH, limiting any conclusions about volumetrics and PHLF. The dataset also does not provide information regarding different modalities used clinically to risk-stratify patients with primary hepatic malignancy. Fifth, PHLF is characterized by NSQIP using the ISGLS classification; while this classification scheme is widely accepted for PHLF, validation using other classification methods is not possible. Finally, morbidity and mortality data in NSQIP is limited to 30 days postoperatively.
In this large national hepatectomy-specific dataset, PHLF was found to be rare and mild in presentation after MiH. High-risk patients, such as those with evidence of preoperative liver dysfunction, remain at elevated risk of PHLF, despite utilization of parenchymal-sparing surgical techniques. These data suggest that PHLF, while devastating, may not be a predominant driver of morbidity for patients undergoing MiH. Efforts to increase adoption of hepatic parenchymal-sparing surgery and to focus attention on the prevention of associated complications, such as postoperative infections, after MiH should be pursued.
Funding:
This work was supported by the NIH/NCI Cancer Center Support Grant P30CA008748 at Memorial Sloan Kettering Cancer Center.
Footnotes
Meetings: This work has not been presented at an academic or society meeting.
Declarations of interest: 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.Kingham TP, Correa-Gallego C, D'Angelica MI, et al. Hepatic parenchymal preservation surgery: decreasing morbidity and mortality rates in 4,152 resections for malignancy. J Am Coll Surg. 2015;220(4):471–479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Edwards WH Jr., Blumgart LH. Liver resection in malignant disease. Semin Surg Oncol. 1987;3(1):1–11. [DOI] [PubMed] [Google Scholar]
- 3.Fortner JG, Kim DK, Maclean BJ, et al. Major hepatic resection for neoplasia: personal experience in 108 patients. Ann Surg. 1978;188(3):363–371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dokmak S, Fteriche FS, Borscheid R, Cauchy F, Farges O, Belghiti J. 2012 Liver resections in the 21st century: we are far from zero mortality. HPB (Oxford). 2013;15(11):908–915. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mullen JT, Ribero D, Reddy SK, et al. Hepatic insufficiency and mortality in 1,059 noncirrhotic patients undergoing major hepatectomy. J Am Coll Surg. 2007;204(5):854–862; discussion 862-854. [DOI] [PubMed] [Google Scholar]
- 6.Gilg S, Sandstrom P, Rizell M, et al. The impact of post-hepatectomy liver failure on mortality: a population-based study. Scand J Gastroenterol. 2018;53(10-11):1335–1339. [DOI] [PubMed] [Google Scholar]
- 7.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]
- 8.van den Broek MA, Olde Damink SW, Dejong CH, et al. Liver failure after partial hepatic resection: definition, pathophysiology, risk factors and treatment. Liver Int. 2008;28(6):767–780. [DOI] [PubMed] [Google Scholar]
- 9.Lafaro K, Buettner S, Maqsood H, et al. Defining Post Hepatectomy Liver Insufficiency: Where do We stand? J Gastrointest Surg. 2015;19(11):2079–2092. [DOI] [PubMed] [Google Scholar]
- 10.Liu JY, Ellis RJ, Hu QL, et al. Post Hepatectomy Liver Failure Risk Calculator for Preoperative and Early Postoperative Period Following Major Hepatectomy. Ann Surg Oncol. 2020. [DOI] [PubMed] [Google Scholar]
- 11.Bilimoria KY, Liu Y, Paruch JL, et al. Development and evaluation of the universal ACS NSQIP surgical risk calculator: a decision aid and informed consent tool for patients and surgeons. J Am Coll Surg. 2013;217(5):833–842 e831-833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Balzan S, Belghiti J, Farges O, et al. The "50-50 criteria" on postoperative day 5: an accurate predictor of liver failure and death after hepatectomy. Ann Surg. 2005;242(6):824–828, discussion 828-829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rahbari NN, Reissfelder C, Koch M, et al. The predictive value of postoperative clinical risk scores for outcome after hepatic resection: a validation analysis in 807 patients. Ann Surg Oncol. 2011;18(13):3640–3649. [DOI] [PubMed] [Google Scholar]
- 14.Dasari BVM, Hodson J, Roberts KJ, et al. Developing and validating a pre-operative risk score to predict post-hepatectomy liver failure. HPB (Oxford). 2019;21(5):539–546. [DOI] [PubMed] [Google Scholar]
- 15.Vigano L, Torzilli G, Troisi R, et al. Minor Hepatectomies: Focusing a Blurred Picture: Analysis of the Outcome of 4471 Open Resections in Patients Without Cirrhosis. Ann Surg. 2019;270(5):842–851. [DOI] [PubMed] [Google Scholar]
- 16.Fink AS, Campbell DA Jr., Mentzer RM Jr., et al. The National Surgical Quality Improvement Program in non-veterans administration hospitals: initial demonstration of feasibility. Ann Surg. 2002;236(3):344–353; discussion 353-344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.The SURVEYLOGISTIC Procedure. 2016. [Google Scholar]
- 18.Billingsley KG, Jarnagin WR, Fong Y, Blumgart LH. Segment-oriented hepatic resection in the management of malignant neoplasms of the liver. J Am Coll Surg. 1998;187(5):471–481. [DOI] [PubMed] [Google Scholar]
- 19.Bismuth H Surgical anatomy of the liver. Recent Results Cancer Res. 1986;100:179–184. [DOI] [PubMed] [Google Scholar]
- 20.Bismuth H Surgical anatomy and anatomical surgery of the liver. World J Surg. 1982;6(1):3–9. [DOI] [PubMed] [Google Scholar]
- 21.Donadon M, Cescon M, Cucchetti A, et al. Parenchymal-Sparing Surgery for the Surgical Treatment of Multiple Colorectal Liver Metastases Is a Safer Approach than Major Hepatectomy Not Impairing Patients' Prognosis: A Bi-Institutional Propensity Score-Matched Analysis. Dig Surg. 2018;35(4):342–349. [DOI] [PubMed] [Google Scholar]
- 22.Aloia TA, Fahy BN, Fischer CP, et al. Predicting poor outcome following hepatectomy: analysis of 2313 hepatectomies in the NSQIP database. HPB (Oxford). 2009;11(6):510–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Kauffmann R, Fong Y. Post-hepatectomy liver failure. Hepatobiliary Surg Nutr. 2014;3(5):238–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Eguchi H, Umeshita K, Sakon M, et al. Presence of active hepatitis associated with liver cirrhosis is a risk factor for mortality caused by posthepatectomy liver failure. Dig Dis Sci. 2000;45(7):1383–1388. [DOI] [PubMed] [Google Scholar]
- 25.Golse N, Bucur PO, Adam R, Castaing D, Sa Cunha A, Vibert E. New paradigms in post-hepatectomy liver failure. J Gastrointest Surg. 2013;17(3):593–605. [DOI] [PubMed] [Google Scholar]
- 26.Tzeng CW, Cooper AB, Vauthey JN, Curley SA, Aloia TA. Predictors of morbidity and mortality after hepatectomy in elderly patients: analysis of 7621 NSQIP patients. HPB (Oxford). 2014;16(5):459–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Little SA, Jarnagin WR, DeMatteo RP, Blumgart LH, Fong Y. Diabetes is associated with increased perioperative mortality but equivalent long-term outcome after hepatic resection for colorectal cancer. J Gastrointest Surg. 2002;6(1):88–94. [DOI] [PubMed] [Google Scholar]
- 28.Vauthey JN, Pawlik TM, Ribero D, et al. Chemotherapy regimen predicts steatohepatitis and an increase in 90-day mortality after surgery for hepatic colorectal metastases. J Clin Oncol. 2006;24(13):2065–2072. [DOI] [PubMed] [Google Scholar]
- 29.Rubbia-Brandt L, Audard V, Sartoretti P, et al. Severe hepatic sinusoidal obstruction associated with oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer. Ann Oncol. 2004;15(3):460–466. [DOI] [PubMed] [Google Scholar]
- 30.Yanaga K, Kanematsu T, Takenaka K, Sugimachi K. Intraperitoneal septic complications after hepatectomy. Ann Surg. 1986;203(2):148–152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Panis Y, McMullan DM, Emond JC. Progressive necrosis after hepatectomy and the pathophysiology of liver failure after massive resection. Surgery. 1997;121(2):142–149. [DOI] [PubMed] [Google Scholar]
- 32.Helling TS. Liver failure following partial hepatectomy. HPB (Oxford). 2006;8(3):165–174. [DOI] [PMC free article] [PubMed] [Google Scholar]