Hepatectomy is a complex surgical procedure and remains the most effective treatment in cases of both primary and secondary liver tumors.
Morbidity from liver resection remains high, despite improvements in patient selection, the application of parenchymal preservation techniques, and the evolution of surgical technology in the effective treatment of postoperative complications [1-3] .
A significant proportion of these postoperative occurrences originate in the chest, and have a marked impact on patient outcomes and mortality [4].
Pulmonary morbidity after liver resection is varied and includes: acute respiratory failure, pulmonary edema with hypoxia, nosocomial pneumonia, and large pleural effusion and/or biliopleural fistula requiring invasive intervention.
The etiology of these complications is related to the patient’s natural physiological environment, as well as the surgical risk profile of the hepatectomy procedure, itself. Pleural effusions are reported as the most common deviation from the normal physiological postoperative course, and are often an incidental finding that is frequently considered innocuous, without clinical significance, estimated to occur in 35% of hepatectomies [5-7].
This work examines the predictable factors for post-hepatectomy morbidity from these subclinical findings detected early; Additionally, it examines the range of postoperative pulmonary complications (PPCs) and their severity, and reviews management strategies.
Early recognition and management of these complications are key to improving surgical outcomes after hepatectomy.
| Material and methods |
> Patient population
A retrospective review of the prospectively maintained institutional database of the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) was performed to identify patients undergoing hepatectomy at Indiana University (IU) Health University Hospital , among 2013 and 2018.
The inclusion criteria were: patients over 18 years of age, and those undergoing hepatectomy for any diagnosed disease. Patients were excluded if they had a thoracic operation concomitantly, or an operative return during the initial admission.
Through an extensive review of medical records, patients were categorized according to the presence or absence of a CPP. The Institutional Review Board approved the conduct of this study, and the data stored complied with the Health Insurance Portability and Accountability Act .
> Clinical data
All variables monitored in the ACS-NSQIP were evaluated, and the clinical database was augmented through individual review of patients’ medical records. Preoperative diuretic dependence, pattern for end-stage liver disease, diagnosis of mental illness, and biliary stent placement were recorded along with standardized ACS-NSQIP variables, when available, within 30 days. of the operation.
Additional preoperative and postoperative variables included: additional anesthesia (epidural or transabdominal plane block), intraoperative colloid and crystalloid volume, transfusion requirement, total resuscitation volume within 12 and 24 hours postoperatively, intraoperative drain placement, hepatectomy major versus minor, Postoperative diuresis, day of initiation of postoperative diuresis, status of Foley and nasogastric tubes (NGT), reinsertion of the NGT, and total parenteral requirement.
The primary outcome was CPP, categorized as mild versus severe, as described below. Secondary endpoints included the need for intervention for the thoracic complication, length of stay, morbidity, and mortality.
> Definitions
The Strasbourg system was used to divide liver resections based on the complexity index. Minor liver resections included: peripheral wedge resection, left lateral segmentectomy, left hepatectomy without caudate, right hepatectomy, and right posterior segmentectomy.
On the other hand, major liver resections included: left hepatectomy with caudate, caudate alone, right anterior segmentectomy, middle anterior hepatectomy, and left trisegmentectomy with caudate.
Bile leak was defined according to the guidelines of the International Study Group of Liver Surgery [2]. The degree of liver failure post hepatectomy was also defined according to these guidelines[8].
The Clavien-Dindo classification was used to group PPCs based on identifiable radiological changes (chest x-ray, computed tomography [CT]), and/or symptomatic respiratory exacerbation.
- Mild pulmonary complication (Clavien-Dindo 2 or 3): simple or medium effusion with or without symptoms, diagnosed on a radiological image.
- Severe pulmonary complication (Clavien-Dindo 4 or 5): acute respiratory failure, pneumothorax, severe pneumonia, biliopleural fistula, pulmonary embolus, complex pleural effusion, reintubation secondary to respiratory failure, prolonged ventilation (> 48 hours), or those requiring intervention invasive (chest tube, thoracentesis, surgery, intubation, bronchoscopy).
> Statistics
Categorical differences in preoperative variables between patients with CPP and patients without CPP were evaluated using Fisher’s exact test as well as c2. For continuous variables, the Mann-Whitney test was used. Thereafter, all relevant perioperative variables with significant group differences, with a level of P < 0.1, were included in a multivariable logistic regression model.
Finally, a predictive model for the development of PPC was developed by running a multiple logistic regression containing only significant perioperative variables, with a level of P < 0.1 in the previous stage. The biostatistical analysis was performed by the Indiana University Center for Outcomes Research in Surgery , using the STATA program.
| Results |
> Study population
A total of 702 patients underwent hepatectomy during the 5-year period. Of them, 119 were performed laparoscopically (17%), and 29 were performed laparoscopically with manual assistance (4%). CPP rates were reported at 20% (n = 137). The overall pneumonia rate was around 10%, and all cases were categorized as CPP, based on the study design.
Baseline comparison of demographic data and comorbidities between the groups revealed that the CPP group, compared to the control group, was an average of 5 years older (63 vs 58; P < 0.01), was more likely to have hypertension (55% vs 41%; P < 0.01), and chronic lung disease (10% vs 5%; P = 0.02), and twice as likely to be malnourished (10 vs 5%; P = 0.03), and having been admitted with a non-home status (19% vs 5%; P < 0.01).
The remaining demographics and comorbidities were statistically similar between the groups. Perioperative characteristics identified more frequent placement of biliary stenting (20% vs 9%; P < 0.001), longer duration of operation (212 vs 185 minutes; P < 0.01), greater blood loss (500 vs 300 ; P < 0.001), transfusion requirements (23% vs 10%; P < 0.01), as well as a greater number of intraoperative drains in the CPP cohort (65% vs 54%; P < 0.01).
Major hepatectomy and biliary reconstructions were more likely in the CPP cohort (60% vs 37%; P < 0.01, and 18% vs 7%; P < 0.01, respectively). Intraoperative liver texture was predominantly normal in both groups; However, fatty, cirrhotic, and congestive textures were observed more frequently in the CPP group. Operative pathology confirming malignancy was higher among patients with CPP.
Perioperative management varied significantly between the 2 groups, in relation to fluid administration, diuresis requirements, and use of NGT and Foley. Patients who developed CPP received a larger intravenous volume 24 hours postoperatively. About 70% of patients with CPP required the administration of a diuretic. Indwelling catheters were more likely to be placed in the CPP group.
The CPP group had generally worse outcomes and delayed recovery. CPP was associated with a higher rate of biliary fistulas (15% vs 5%; P < 0.01), and postoperative liver failure (PHF) (22% vs 6%; P < 0.01); This was associated with prolonged hospital stay (8 vs 5 days; P < 0.01), higher rate of non-home discharge (18% vs 5.5%; P < 0.01), and higher rates of readmission (20% vs 6.6&; P < 0.01).
When early detection of pleural effusion on a chest radiograph was used as a predictor, the occurrence of the radiographic phenomenon was associated with 3-fold higher rates of biliary fistula and FHP (OR 3.5 and 3, respectively; P < 0.001 ) ; length of hospital stay 3 days longer ( P < 0.001), and readmission rates 3 times higher (OR 3, P < 0.001).
> Mild versus severe complications
A total of 138 patients developed PPC (severe: n = 51 [37%]; mild: n = 87 ¨63%]). Baseline characteristics did not differ by the degree of complication. However, univariate analysis identified the duration of surgery (hours) (4.3 ± 1.7 vs 3.7 ± 1.3; P = 0.018), the volume of crystalloid used intraoperatively (3.1 ± 1. 2 vs 2.5 ± 0.5; P = 0.021), and biliary reconstruction (14 [28%] vs 11 [13%]; P = 0.038), as risk factors for severe CPP.
In multivariate analysis, predictors for severe CPP after any liver resection were postoperative biliary fistula (OR 5; P < 0.01), FHP (OR 2.5; P < 0.01), and major hepatectomy (OR 3; P < 0.01).
Evaluation of outcomes among patients who underwent major liver resection showed an association between the development of severe CPP and the occurrence of intra-abdominal complications: deep organ infection (OR 5.5; P < 0.001), fistula biliary (OR 3.5; P < 0.001); and FHP (OR 2.8; P < 0.001), causing a significant delay in recovery, length of stay (OR 4.3; P < 0.001), discharge not home (OR 2.2; P < 0.001 ), 0.001) and readmission rates (OR 3.1; P < 0.001)
Regarding perioperative variables only among patients undergoing major hepatectomy, patients with CPP were older, had more diabetes and hypertension, as well as also having a stent placed preoperatively. Intraoperatively, those patients had higher crystalloid resuscitation and more surgical drains left in place, in the setting of bilioenteric reconstruction.
Postoperatively, the CPP group required more resuscitation and subsequent diuresis. Septic shock, organ/space infection, biliary fistula, and postoperative liver failure were more predictive of PPC after major hepatectomy.
> Risk factors for mild and severe pulmonary complications
When the risk of developing severe PPC was investigated only among major hepatectomies, diabetics who were insulin-dependent had more than 4 times the probability of developing severe complications: 4.16 (1.22-14.79; P = or,022); the occurrence of a postoperative biliary fistula increased the risk of severe complications by more than 8 times: 8.5 (1.57-46.30; P = 0.013).
| Discussion |
In this study, CPPs were frequent and comparable to those reported in previous series. The high frequency of these complications has been the objective of previous works in the literature, and it was reported that they impact up to 30% of patients with hepatectomy, similar to the present experience [9,10]. Similarly, Lo et al., reported only 3 cases of pleural effusions in their multicenter analysis of more than 1500 patients [11].
The most frequently encountered CPP by the authors was pleural effusion (PE), most of which was identified incidentally on a chest x-ray.
When effusion occurred, in the authors’ experience, it was almost exclusively identified on the 3rd postoperative day (90%); Therefore, the recognition of the radiographic finding on the 3rd postoperative day was selected as the primary effusion, for the purpose of this study.
Although Uchiyama et al. have interestingly reported an association between PD and worse outcomes after hepatectomy for hepatocellular carcinoma, their recommendation of a chest CT on the 7th postoperative day may not be applicable, because the majority of hepatectomy patients , in the care center of the authors of the present study, could have been discharged from the hospital at that time [12].
The association observed in the data set between “pleural effusion without clinical significance” and worse outcomes was an early and possibly valuable sign of poor outcomes in that patient population, when compared to the control group. Therefore, it could be used as an early tool in the clinical evaluation of possible pulmonary morbidity in the postoperative period.
Haines et al. identified baseline variables that increase the risk of developing these CPPs, including advanced age, preexisting comorbidity, particularly hypertension or chronic obstructive pulmonary disease, poor preoperative nutritional status (i.e., low albumin), blood loss, significant weight gain before surgery, or pre-admission originating in a nursing care facility (probably a surrogate marker for the presence of other comorbidities) [13].
In fact, meta-analysis studies were consistent with the present findings, and have identified in previous work age as a risk factor for complications after abdominal surgery [9,14].
Tzeng et al. investigated age in more detail and defined elderly status as 75 years or older; These authors successfully associated elderly status with greater postoperative morbidities, as well as a double effect on mortality [14,15]. Furthermore, low albumin level was considered a marker for worse postoperative outcomes, and facilitated the formation of pleural effusion, secondary to decreased oncotic pressure [13].
Other notable findings in this series imply that underlying malignancy, major hepatectomy, and excess peri- and postoperative fluid resuscitation were all independent risk factors for developing CPP. All of these factors have a single theme in common: excessive fluid administration.
Major hepatectomy, especially for malignant pathology, was typically associated with greater blood loss, and required the administration of greater quantities of fluids and blood products [5]. Major hepatectomy, such as an extended right hepatectomy, has been reported as an independent variable for worse outcomes [12,16,17].
In this work, the authors hypothesized that an extended right hepatectomy could increase the risk for severe PPCs, particularly the risk of biliopleural fistulization. However, the analysis did not establish statistical significance in the clinical difference observed between the groups (OR 4.2 [0.3-0.6; P = 0.26]).
Lepere et al. identified intraoperative hemodynamic status, hepatic ischemia, and increased blood loss requiring transfusion as independent risk factors for PPCs within 7 days of liver resection [18]. These findings reflect the present ones, and suggest a common mechanism for the development of these CPPs: hypovolemia. Over-resuscitation increases hydrostatic pressure, altering fluid balance and distribution, resulting in pleural effusions, pulmonary edema, and secondary pneumonia. Furthermore, the present analysis provided several interesting observations.
Multivariate logistic analysis found an association between poor nutrition (evidenced by significant preoperative weight loss), hypoalbuminemia, and the presence of cirrhosis, with a higher incidence of postoperative biliary fistula. That was likely attributable to the impaired healing seen in those with poor nutrition (including those with significantly low levels of serum albumin, seen particularly in patients with some degree of liver cirrhosis).
In a systematic review of perihepatectomy nutritional optimization, postoperative morbidities were reduced by a quarter when nutritional supplements were used [19]. In the present study, an association between postoperative liver failure and intraoperative liver texture (hepatic steatosis) or preoperative cirrhosis was also reported. This is related to the reduced amount of functional liver tissue present in these patients, making FHP inevitable in some cases.
Steatohepatitis secondary to neoadjuvant chemotherapy has been well established as increasing morbidity and mortality after hepatectomy [20,21], although this did not differ between groups in the present cohort. These associations need further investigation with large multicenter studies. However, this observation will be important in the development of preoperative nutritional regimens, and appropriate selection of patients for different management strategies, to optimize outcomes.
In this study, postoperative pulmonary morbidity was also found to strongly impact surgical outcomes. PPC has greater postoperative morbidity, including increased number of unscheduled intubations, length of stay, discharge to a nursing facility, and unplanned readmission. Perhaps the most important finding of this study was the identification of a correlation between postoperative surgical pathology and PPC, which is contrary to what has been previously reported by Dondero et al. [17].
Biliary fistula was a common complication after hepatectomy, and has been reported as clinically important in 8% of cases [3,22]. The authors of this work found that there was a significant correlation between the development of a pleural effusion and the presence of an underlying biliary fistula. Given that pleural effusions occur early in the postoperative course, it was safe to use that finding as a predictor for hepato-specific outcome variables. Surprisingly, when a pleural effusion occurred, the risk of biliary fistula or HPF increased at least 3-fold.
The consequences of those CPPs had a devastating effect on surgical outcomes and substantially impacted the healthcare system as a whole, increasing hospital stays and unplanned readmissions, with all the associated financial burden. By considering the results of this study and then implementing strategies to address those various risk factors, significant progress can be made in reducing those complications and overall healthcare costs [23].
Goal-directed fluid resuscitation would significantly reduce the incidence of such complications as a possible strategy. Furthermore, instituting a restrictive strategy for blood transfusion could also limit that, and the intermittent transfusion-related morbidities that may result in CPP. Reducing the incidence of those complications, in turn, could reduce patients’ length of hospital stay and any additional adjunctive procedures needed to evaluate pulmonary complications.
In conclusion , this study is one of the first to examine CPP after major hepatectomy for several surgical pathologies. Most studies with low CPP have examined the occurrence of pulmonary issues after living donor hepatectomies, or certain indications such as hepatocellular carcinoma [11,12]. In those studies, pulmonary complications were markedly suppressed among highly selective patient populations. Although the present study was able to replicate and validate several findings from other studies, it was limited by the fact that it was the experience of a single healthcare center.
Surgeon patient selection bias, as well as specific intraoperative decisions and postoperative management, were not uniform among surgeons. Furthermore, not all hepatectomy cases in the present series had a radiographic image in the chest, resulting in potential selection bias. A larger multicenter study might be necessary to further affirm those findings in hepatectomy patients. Likewise, a subsequent larger trial could possibly show a causal relationship between several independent risk factors and pulmonary complications.
Considering the current financial state of healthcare focused on quality outcomes and efficiency, this study provided information on what modifiable risk factors could be addressed (e.g., poor nutrition status), and signs of impending serious pulmonary morbidity after hepatectomy, to help improve postoperative outcomes and reduce overall morbidity.
