Kaplan-Meier curves for overall survival after resection of colorectal liver metastases. A, Relationship between postoperative complications and overall survival (P = .007). B, Relationship between postoperative complications and overall survival when subclassified into infectious and noninfectious complications. No complication vs noninfectious complication, P = .23; no complication vs infectious complication, P = .007 (log-rank test).
Kaplan-Meier curves for disease-free survival after resection of colorectal liver metastases. A, Relationship between postoperative complications and disease-free survival (P = .04). B, Relationship between postoperative complications and disease-free survival when subclassified into infectious and noninfectious complications. No complication vs noninfectious complication, P = .82; no complication vs infectious complication, P = .009 (log-rank test).
Neal CP, Mann CD, Garcea G, Briggs CD, Dennison AR, Berry DP. Preoperative Systemic Inflammation and Infectious Complications After Resection of Colorectal Liver Metastases. Arch Surg. 2011;146(4):471-478. doi:10.1001/archsurg.2011.50
Copyright 2011 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2011
Postoperative complications are associated with a poor long-term prognosis after resection of colorectal liver metastases via an undetermined mechanism. The preoperative systemic inflammatory response, itself a predictor of poor survival, was recently shown to independently predict postoperative infectious complications after primary colorectal cancer resection.
To examine the association of postoperative infectious complications with preoperative systemic inflammation and survival in patients undergoing resection of colorectal liver metastases.
Retrospective study based on a prospectively updated database.
A United Kingdom tertiary referral hepatobiliary unit.
A total of 202 consecutive patients with colorectal liver metastases undergoing hepatectomy between January 1, 2000, and April 30, 2006.
Main Outcome Measures
Multivariable analyses were performed to correlate preoperative and operative variables with postoperative complications and to correlate complications with long-term survival after metastasectomy.
Ninety-day mortality and morbidity were 2.0% and 25.7%, respectively. The preoperative systemic inflammatory response independently predicted the development of infectious complications (P = .009) and major infectious complications (P = .005) after hepatectomy, along with performance of trisectionectomy. Infectious complications were associated with poor long-term survival after metastasectomy but lost independent significance when systemic inflammatory variables were included in multivariable analyses.
The preoperative systemic inflammatory response independently predicts the development of infectious complications after colorectal liver metastases resection. Although infectious complications are associated with adverse long-term prognosis after hepatectomy, they lacked independent prognostic value when systemic inflammatory variables were also considered, suggesting that much of their prognostic value arises from their association with the preoperative systemic inflammatory response.
Hepatectomy is the only potentially curative therapy for patients with colorectal liver metastases, with reported 5-year survival of 30% to 50%.1- 3 Postoperative mortality after resection of colorectal liver metastases is less than 5% in most specialist centers, but morbidity remains high, ranging from 15% to 35%.2- 6 Postoperative morbidity is associated with increased length of hospitalization and greater economic cost.7 Recent data suggest that postoperative complications are also associated with shortened overall and cancer-specific survival after resection of a range of malignant neoplasms,8,9 including primary colorectal cancer10,11 and colorectal liver metastases.4,12- 14
The mechanism linking postoperative complications with poor long-term prognosis after cancer resection remains to be determined. It has been postulated that enhancement of the systemic inflammatory response may be involved.11 The local immune response to malignancy involving tumor environment infiltration by myeloid cells and lymphocytes is accompanied by a systemic response, driven by the release of proinflammatory cytokines by tumor cells and recruited immune cells.15 The resultant systemic inflammatory response involves hematopoietic changes (including alterations in the relative numbers of leukocytes), changes in acute-phase reactants (including an elevation in C-reactive protein [CRP] levels and a decline in serum albumin levels), alterations in the neuroendocrine system, and changes in protein and energy metabolism.16 The preoperative systemic inflammatory response, evidenced by an elevated neutrophil count, neutrophil to lymphocyte ratio, and CRP level and a low serum albumin level, has been demonstrated to independently predict poor prognosis after resection of a range of malignant neoplasms,17,18 including primary colorectal cancer19 and colorectal liver metastases.2,20- 22 Evidence suggests that circulating leukocytes and CRP have direct roles in promoting aggressive tumor phenotypes and disease recurrence.23,24 It has been hypothesized that postoperative complications may predispose to tumor recurrence by increasing the magnitude of the systemic inflammatory response11
An alternative link between systemic inflammation and postoperative morbidity was, however, recently identified by Moyes et al.25 In their study, the presence of the preoperative systemic inflammatory response itself independently predicted postoperative infectious complications in patients undergoing curative resection for colorectal cancer. The authors concluded that preoperative systemic inflammation is an underlying host characteristic that predisposes to postoperative infection, suggesting that the prognostic value of complications may relate, at least in part, to this association.
These data suggest that complex links exist among systemic inflammation, postoperative morbidity, and long-term survival in patients with primary colorectal cancer. The present study aimed to further define these links in patients with resectable colorectal liver metastases. First, we examined the association of the preoperative systemic inflammatory response with postoperative complications in patients undergoing colorectal liver metastases resection. Subsequently, we set out to further define the prognostic value of postoperative complications in this context using multivariable analyses that both included and omitted systemic inflammatory variables.
Patients undergoing resection of colorectal liver metastases between January 1, 2000, and April 30, 2006, at a single tertiary referral institution (Leicester General Hospital) were identified using a prospectively maintained database. These data were analyzed retrospectively, with additional information obtained from hospital records. Preoperative staging comprised computed tomography of the chest, abdomen, and pelvis; magnetic resonance imaging of the liver; and staging laparoscopy with intraoperative ultrasound. Clinicopathologic variables, including the clinical risk score devised by the Memorial Sloan-Kettering Cancer Center,1 were recorded for each patient. This score is composed of the following components (1 point is assigned to each): a positive nodal status of the primary tumor, a disease-free interval of less than 12 months, more than 1 metastasis, diameter of the largest metastasis greater than 5 cm, and a preoperative carcinoembryonic antigen level higher than 200 ng/mL (to convert to micrograms per liter, multiply by 1.0). Hematologic variables and serum albumin levels were obtained from blood tests routinely performed before surgery. No patient had clinical evidence of infection or other inflammatory conditions at the time of sampling.
The liver parenchyma was dissected using an ultrasonic dissector (Sonoca; Söring Medical, Quickborn, Germany), with hemostasis achieved using diathermy, argon beam coagulation, and suturing. If necessary, intermittent clamping of the portal triad (the Pringle maneuver) was performed during parenchymal dissection, with periods of occlusion of up to 10 minutes alternating with 5-minute release intervals. After hepatic resection, patients were transferred to the intensive care unit or the high-dependency unit, where an established clinical care protocol was followed. All the patients received intravenous albumin and intravenous vitamin K for 3 to 5 days. Early enteral feeding was encouraged, with parenteral nutrition introduced in cases in which this was delayed.
The main short-term postoperative outcome measures were mortality, morbidity (graded according to the Clavien classification26 [Table 1]), major morbidity (Clavien grade III-V complications, ie, requiring surgical, endoscopic, or radiologic intervention; involving high-dependency unit or intensive care unit management; or resulting in death), infectious morbidity, major infectious morbidity (Clavien grade III-V infectious complications), noninfectious morbidity, and major noninfectious morbidity. Morbidity and mortality were defined as occurring within 90 days of surgery. Criteria used to define infectious complications were as described previously.27 Infectious complications recorded included wound infection (superficial or deep infection that required treatment with antibiotic agents or wound drainage), intra-abdominal abscess (intra-abdominal collection associated with fever or leukocytosis that discharged spontaneously or required surgical or radiologically guided drainage, with positive blood or fluid culture), and lower respiratory tract infection (respiratory symptoms and signs and infiltrate on chest radiography associated with fever or leukocytosis requiring antibiotic drug treatment). Sepsis was defined as the presence of 2 or more systemic inflammatory response syndrome criteria28 coupled with positive findings on blood culture. Hepatic dysfunction was defined as a serum bilirubin level higher than 7.0 mg/dL (to convert to micromoles per liter, multiply by 17.104) unrelated to biliary obstruction or leak or an international normalized ratio greater than 2 more than 2 days postoperatively or clinically significant ascites or encephalopathy.29
The follow-up protocol included clinical review 6 weeks after hospital discharge, followed by clinical evaluation, serum investigations (including liver function tests and carcinoembryonic antigen measurement), and abdominal ultrasonography, all performed every 3 months for the first year, every 6 months for the second year, and then annually until 5 years after hepatectomy. Abnormal results during surveillance triggered further investigation. Long-term outcome measures assessed were overall and disease-free survival (as of December 2008). Patients who died within 90 days of surgery (Clavien grade V complications) were excluded from long-term survival analyses.
As previously,19,30 grouping of systemic inflammatory response variables was performed using standard thresholds. Univariable and multivariable logistic regression analyses were used to examine the effect of variables on the development of postoperative complications. Univariable associations with long-term survival were determined using univariable Cox regression analysis, Kaplan-Meier analysis, and the log-rank test. Multivariable analyses were performed by Cox proportional hazards regression analysis (using a stepwise backward procedure), incorporating all variables with P < .10 on univariable analysis. Statistical significance was defined as P < .05. These analyses were conducted using a software program (SPSS version 16.0; SPSS, Inc, Chicago, Illinois).
A total of 202 consecutive patients underwent hepatic metastasectomy during the study, with all resections performed with curative intent. There were 126 men (62.4%) and 76 women (37.6%). The median patient age at the time of surgery was 61.5 years (mean age, 61.8 years; age range, 32-79 years). The median number of tumors was 2 (range, 1-12). Tumor size ranged from 5 to 140 mm, with a median size of 40 mm. A poor preoperative clinical risk score (score, 3-5) was recorded in 96 patients (47.5%). Eighty-four patients (41.6%) had systemic chemotherapy in the 6 months before liver resection. Use of preoperative chemotherapy showed no correlation with preoperative systemic inflammatory response variables. No patient had biliary obstruction or cholangitis at any stage during his or her preoperative course, and no patient underwent preoperative biliary manipulation. One hundred twenty-five patients (61.9%) had a major resection (>3 Couinaud segments), and 72 patients (35.6%) underwent hepatic trisectionectomy (resection of ≥5 Couinaud segments).31
Postoperative morbidity occurred in 52 patients (25.7%) after hepatectomy, with a total of 67 events recorded (Table 2). Infectious complications were most common, developing in 36 patients (17.8%). Major infectious complications occurred in 12 patients (5.9%): 8 with perihepatic abscesses requiring percutaneous or surgical drainage and 4 with lower respiratory tract infections requiring invasive ventilatory assistance. Noninfectious complications occurred in 24 patients (11.9%), with major noninfectious complications occurring in 6 patients (3.0%). Multiple complications developed in 9 patients (4.5%), with 8 patients (4.0%) developing both infectious and noninfectious complications. Ninety-day mortality after hepatectomy was 2.0% (n = 4), with 3 of these patients developing major infectious complications.
The results of univariable and multivariable analyses for factors associated with infectious complications and major infectious complications after hepatectomy are given in Table 3 and Table 4, respectively. On univariable analysis (Table 3), performance of trisectionectomy, operative duration of 5 hours or greater, high neutrophil count, and high neutrophil to lymphocyte ratio were significantly associated with postoperative infectious morbidity. Performance of trisectionectomy, high leukocyte count, high neutrophil count, high neutrophil to lymphocyte ratio, and high monocyte count were significantly associated with the development of major infectious complications after hepatectomy.
On multivariable analysis (Table 4), performance of trisectionectomy and high neutrophil count were the only variables independently associated with the development of infectious complications. These factors were also the only variables independently associated with major infectious complications after hepatectomy. Inflammatory variables showed no correlation with the development of noninfectious complications or major noninfectious complications after hepatectomy (data not shown).
Overall median follow-up was 40 months (range, 6-70 months), with no patients lost to follow-up. Three- and 5-year overall survival after metastasectomy were 47.5% and 36.9%, respectively (median, 33 months; 95% confidence internal [CI], 26.1-39.9 months). Three- and 5-year disease-free survival were 28.8% and 23.1%, respectively (median, 14 months; 95% CI, 11.6-16.4 months).
Univariable and multivariable survival analyses were performed to determine factors associated with overall survival (Table 5) and disease-free survival (Table 6) of patients who did not die within 90 days of surgery (n = 198). The development of postoperative complications was associated with shortened overall (Figure 1A) and disease-free (Figure 2A) survival after metastasectomy. Further analysis revealed that only infectious complications correlated with shortened overall and disease-free survival after hepatectomy, with noninfectious complications demonstrating no correlation with either outcome (Figures 1B and 2B). The prognostic value of the major subtypes of infectious complication was also examined. Wound infections demonstrated no prognostic significance after hepatectomy to either overall survival (P = .26) or disease-free survival (P = .65). In contrast, nonwound infectious complications were significantly associated with shortened overall survival (P = .02) and disease-free survival (P = .005) after hepatectomy. The effect of severity of complications was also examined. Although patients with Clavien grade I/II and grade III/IV infectious complications had shorter overall survival than did patients without infectious complications (P = .048 and P = .04, respectively), there was no significant difference in survival when grade I/II and grade III/IV infectious complications were compared (P = .58). A similar result was noted for disease-free survival. Noninfectious complications did not achieve prognostic significance after subdivision according to severity.
When significant variables from univariable analysis were entered into multivariable analyses incorporating inflammatory variables (Tables 5 and 6), postoperative complications were not independently associated with either overall or disease-free survival after hepatectomy. Similarly, neither infectious complications nor specific types of infectious complication exerted any independent prognostic value in relation to either outcome in multivariable analyses. Poor clinical risk score, high neutrophil count, and low serum albumin level were the only variables independently associated with shortened overall and disease-free survival after metastasectomy.
To allow comparison of the present data with those from previous studies4,12,13 examining the prognostic value of complications after colorectal liver metastases resection, multivariable analyses were repeated excluding inflammatory biomarkers. In these analyses, postoperative complications demonstrated an independent association with overall survival (hazard ratio, 1.58; 95% CI, 1.01-2.50; P = .048) but not with disease-free survival after hepatectomy. However, infectious complications maintained independent prognostic significance in relation to overall survival (hazard ratio, 1.66; 95% CI, 1.04-2.66; P = .03) and disease-free survival (1.65; 1.09-2.49; P = .02) in analyses that did not incorporate inflammatory variables.
The results of this study demonstrate that the preoperative systemic inflammatory response is independently associated with the development of infectious complications after resection of colorectal liver metastases. In the only other study to date to correlate systemic inflammation with complications after oncologic surgery,25 the Glasgow Prognostic Score (a composite score derived from a serum CRP level >10.0 mg/L [to convert to nanomoles per liter, multiply by 9.524] and an albumin level <35 g/L [to convert to grams per liter, multiply by 10.0]) and the total leukocyte count independently predicted infectious but not noninfectious complications after primary colorectal cancer resection. In the present study, several inflammatory variables, including leukocyte count and neutrophil to lymphocyte ratio, were associated with the development of infectious or major infectious complications. Neutrophil count showed the most significant association, being independently associated with the development of infectious and major infectious complications. These correlations were independent of the extent of hepatectomy, the only other independent predictor of infectious complications and a factor that is well established to be associated with adverse outcome after metastasectomy.32 The CRP, the other major biomarker of the systemic inflammatory response, was not evaluated in the present study, and further studies are required to determine whether CRP is also independently associated with the development of complications after metastasectomy.
The basis of the independent correlation between the systemic inflammatory response and postoperative infection in patients undergoing resection of primary colorectal cancer and colorectal liver metastases is not clear. Moyes et al25 suggested that the independent association of CRP with infectious complications after colorectal cancer resection arose from its association with impaired T-lymphocyte response and complement activation.33,34 In relation to myeloid cells, it is well established that the ability of a cell to synthesize pro-inflammatory and anti-inflammatory mediators is affected by its previous state.35 Chronically activated phagocytes may, therefore, demonstrate an inadequate response to infectious stressors in the postoperative period, resulting in an increased risk of perioperative infectious events. Although the exact mechanism remains to be delineated, the results of this study suggest that a simple preoperative inflammatory biomarker may identify patients at increased risk for infectious complications after hepatectomy for colorectal liver metastases.
In agreement with recent series of colorectal liver metastases resection,4,12,13 postoperative complications were associated with shortened overall and disease-free survival after metastasectomy. In line with the only other study14 to assess the prognostic value of specific postoperative complications after colorectal liver metastases resection, only infectious complications correlated with long-term outcome, with noninfectious complications showing no association with survival. The prognostic value of postoperative infection was restricted to nonwound infections, with wound infections having no prognostic relevance. These data agree with those of a recent study10 of 1657 patients with resectable colorectal cancer that found that of all complications, non–wound-related septic events had, by far, the most significant association with overall survival and recurrence, with wound complications exerting no prognostic significance. Although the mechanism linking infectious complications to poor prognosis remains to be determined, these data suggest that events that are triggered by the occurrence of specific infectious events or that are involved in their development are implicated in disease recurrence and poor long-term survival.
Major surgery elicits a degree of systemic inflammation and immunosuppression in patients.36 Postoperative complications may increase the magnitude and duration of this systemic inflammatory response, thereby predisposing the growth of metastatic tumor cells and tumor recurrence.8 In particular, postoperative sepsis leads to an extended period of inflammation and immunosuppression,37 which may contribute to metastatic proliferation.38 Noninfectious complications also affect the postoperative systemic inflammatory response, however, questioning the hypothesis that the specific association of infectious complications with prognosis is solely due to an effect on postoperative systemic inflammation. It remains possible, however, that the specific molecular events involved or the magnitude of the systemic inflammation elicited differ between complication types. Of note, however, it was previously reported that only the preoperative, and not the postoperative, systemic inflammatory response was associated with shortened cancer-specific survival after primary colorectal cancer resection.39
An alternative explanation for the link between infectious complications and cancer recurrence is that postoperative infection is a manifestation of preexisting immune alteration that itself predisposes recurrence. This is supported by the observation that the preoperative systemic inflammatory response, a known independent predictor of adverse outcome,2,20- 22 was independently associated with the development of postoperative infectious complications and by the results of the multivariable analyses reported previously herein. Although several studies4,12,13 have found postoperative complications to exert independent prognostic significance after colorectal liver metastases resection, all but one have not included inflammatory variables in multivariable survival analyses. In the present study, although infectious complications independently predicted shortened overall and disease-free survival after metastasectomy in analyses that omitted inflammatory biomarkers, prognostic value was not maintained when inflammatory variables were incorporated. These data suggest that the prognostic value of infectious complications is derived, to a significant degree, from their independent association with the preoperative systemic inflammatory response. In the only other study to date to simultaneously examine the prognostic value of infectious complications and systemic inflammation,14 both factors were found to maintain independent prognostic significance after colorectal liver metastases resection, although the prognostic value of infectious complications was inferior to that of the systemic inflammatory response. In addition, systemic inflammatory response variables were unavailable in 19.0% of patients, limiting the conclusions that can be drawn regarding the true independent prognostic value of complications in this study.
In summary, in agreement with recently published data for primary colorectal cancer,25 these results suggest that the preoperative systemic inflammatory response, easily measurable before surgery, may be used clinically to identify patients at increased risk for postoperative infectious complications and to predict long-term survival after hepatectomy. In addition to potentially guiding the allocation of existing treatment modalities, inflammatory variables themselves represent attractive therapeutic targets. Although it remains to be determined whether the preoperative systemic inflammatory response may be moderated, the development of novel therapeutic agents targeting varied aspects of the systemic inflammatory response, some of which are in clinical trial,40 raises the possibility that moderation of systemic inflammation may be used as a means of improving perioperative outcomes and long-term prognosis after resection for cancer. Further prospective studies are required to confirm the present findings and to explore these hypotheses.
Correspondence: Christopher P. Neal, MRCS, PhD, Department of Hepatobiliary and Pancreatic Surgery, Leicester General Hospital, Gwendolen Road, Leicester LE5 4PW, United Kingdom (email@example.com).
Accepted for Publication: April 9, 2010.
Author Contributions: Dr Neal had full access to all the data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. Study concept and design: Neal, Dennison, and Berry. Acquisition of data: Neal, Mann, and Briggs. Analysis and interpretation of data: Neal, Mann, Garcea, and Briggs. Drafting of the manuscript: Neal and Mann. Critical revision of the manuscript for important intellectual content: Garcea, Briggs, Dennison, and Berry. Statistical analysis: Neal, Mann, Garcea, and Briggs. Administrative, technical, and material support: Dennison and Berry. Study supervision: Dennison and Berry.
Previous Presentation: This study was presented at the annual meeting of the Association of Upper Gastrointestinal Surgeons; September 3, 2009; Nottingham, United Kingdom.