AIM: To evaluate the impact of postoperative infectious complications on hepatocellular carcinoma following curative hepatectomy.

METHODS: We performed a retrospective analysis of 200 hepatocellular carcinoma patients who underwent hepatectomy at our institution between September 2003 and June 2011. The patients’ demographics, clinicopathological characteristics and postoperative infectious complications were analyzed. The Clavien-Dindo classification was adopted to assess the severity of complications. The dynamic change in the neutrophil-to-lymphocyte ratio, defined as the absolute neutrophil count divided by the absolute lymphocyte count, after surgery was also investigated. The observation endpoints for this study were recurrence-free survival and overall survival of the patients. Statistical analysis of the survival curves was performed using the Kaplan-Meier method and the log-rank test. The prognostic value of each variable for predicting prognosis was assessed via multivariate Cox proportional hazards regression analysis. The cutoff score for each variable was selected based on receiver operating characteristic curve analysis. All statistical tests were two-sided, and significance was set at P < 0.05.

RESULTS: The median age of the patients was 49 years, and the majority of patients were male (86%) and had been infected with hepatitis B virus (86%). The 30-d postoperative infectious complication rate was 34.0% (n = 68). Kaplan-Meier survival analysis revealed that postoperative infection was significantly correlated with tumor recurrence (P < 0.001). The postoperative intra-abdominal infection group exhibited a worse prognosis than the non-intra-abdominal infection group (P < 0.001). A significantly increased incidence of postoperative intra-abdominal infection was observed in the patients with hepatic cirrhosis (P = 0.028), concomitant splenectomy (P = 0.007) or vascular invasion (P = 0.026). The patients who had an elevated postoperative neutrophil-to-lymphocyte ratio change (> 1.643) clearly exhibited poorer recurrence-free survival than those who did not (P = 0.009), although no significant correlation was observed between overall survival and the change in the postoperative neutrophil-to-lymphocyte ratio. Based on multivariate analysis, hepatitis B surface antigen positivity, Child-Turcotte-Pugh class B, an elevated postoperative neutrophil-to-lymphocyte ratio change and intra-abdominal infection were significant predictors of poor recurrence-free survival. Hepatic cirrhosis, the maximal tumor diameter and intra-abdominal infection were significant predictors of overall survival.

CONCLUSION: Postoperative intra-abdominal infection adversely affected oncologic outcomes, and the change in postoperative neutrophil-to-lymphocyte ratio was a good indicator of tumor recurrence in hepatocellular carcinoma patients after curative hepatectomy.

Hepatocellular carcinoma (HCC) is one of the most common cancers in the world, leading to 695900 deaths annually. Globally, nearly half of new HCC cases occur in China due to the prevalence of hepatitis B virus infection[1]. Surgical resection is considered a curable treatment for HCC, but the recurrence and the long-term prognosis of HCC remain unsatisfactory[2]. The 5-year recurrence-free survival rate has been reported to be approximately 30%[3-5]. Due to improvements in surgical techniques and perioperative management, the mortality rate of hepatectomy has significantly decreased in recent decades. However, the high incidence of postoperative infectious complications (ranging from 15.6% to 25.0% based on previous reports) remains a major problem, especially in patients with underlying hepatic disease, such as chronic hepatitis and liver cirrhosis[6,7].

Recently, the negative impact of infectious complications on prognosis has been demonstrated in various cancers. The study by Farid et al[8] suggested that postoperative sepsis was an independent predictor of both disease-free survival and overall survival in colorectal cancer patients with liver metastasis who underwent hepatic resection. Amin Andalib et al[9] analyzed a health care database of lung cancer patients after curative-intent surgery in Quebec, and discovered a strong negative correlation between major infectious complications and the long-term survival rate. However, the prognostic impact of postoperative infectious complications on patients with HCC remains unclear.

Inflammation and its associated immunomodulation have been suggested as causes of cancer progression. Systemic inflammation augments the adhesion of circulating tumor cells, thereby accelerating tumor metastasis[10], and bacterial antigens directly increase the metastatic ability of cancer cells[11,12]. Additional evidence has demonstrated the correlation between inflammation and poor oncologic outcomes of cancer patients. Markers of systemic inflammatory responses, such as C-reactive protein (CRP) and the preoperative monocyte count have been investigated as prognostic factors[13,14]. The neutrophil-to-lymphocyte ratio (NLR) is a novel marker of the systemic inflammatory response. An elevated NLR has been experimentally demonstrated to serve as a predictor of poor prognosis in cancer patients[15-18].

Long-term survival was reported to be poorer in HCC patients with operative complications after hepatectomy[19,20]. However, to date, few studies have focused on the potential association between postoperative infectious complications and oncologic outcome of HCC. This study aimed to evaluate the impact of postoperative infectious complications on recurrence-free survival (RFS) and overall survival (OS) among patients receiving curative hepatectomy for HCC.

The characteristics of the studied patients are summarized in Table 1. The median follow-up duration was 32 mo (range: 4-103 mo). At the time of the most recent follow-up, 101 (50.5%) patients had developed tumor recurrence, 43 (21.5%) patients had died of HCC progression, and two (1.0%) patients had died of non-cancerous disease. The 1-, 3-, and 5-year OS rates for all the patients in this study were 96.0%, 83.8% and 81.3%, respectively, and the 1-, 3-, and 5-year RFS rates were 74.5%, 56.0% and 51.0%, respectively. Overall, the 30-d postoperative infectious complication rate was 34.0% (n = 68), and the peak time of infectious complications occurred within one to two weeks after surgery. Table 1 also details the specific number and categories of infectious complications. Five patients exhibiting C-D grade I wound complications were excluded from the analysis of infectious complications. No urinary tract infections or venous catheter related-infections were observed in this study.

Kaplan-Meier survival analysis showed that postoperative infection was significantly correlated with tumor recurrence (P < 0.001), although no significant difference was detected between the different C-D grades (P = 0.798). Both intra-abdominal infection and infection at multiple sites were significantly associated with decreased RFS (P = 0.001 and P < 0.001), although the respiratory infection group was not significantly different from the non-infection group (P = 0.338). The source of infection in the patients with infection at multiple sites was the abdominal cavity, thus, we considered intra-abdominal infection as the primary investigated variable in this study. As shown in Figure 1, the difference between the groups with and without intra-abdominal infection was significant (log-rank test, P < 0.001). The 3-year RFS rate in the intra-abdominal infection group and the non-intra-abdominal infection group was 24.0% and 60.6%, respectively. Intra-abdominal infection was also associated with a poorer 3-year OS rate than non-intra-abdominal infection (52.0% vs 88.6%, P < 0.001; Figure 2). Based on further analysis, a significantly increased incidence of postoperative intra-abdominal infection was observed in patients with hepatic cirrhosis (P = 0.028), concomitant splenectomy (P = 0.007) or vascular invasion (P = 0.026) in univariate analysis, and these variables were all significant in multivariate analysis. However, the results showed no significant correlation between delta-NLR and any of the clinicopathological factors (Table 2). Figure 1 shows the RFS curve for HCC patients with and without post-operative intra-abdominal infection. The 3-year RFS rate in the intra-abdominal infection group and the non-intra-abdominal infection group was 24.0% (n = 6) and 60.6% (n = 106), respectively. The difference between these two groups was statistically significant (log-rank test, P < 0.001).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Graphs displaying the recurrence-free survival curves for hepatocellular carcinoma patients with and without post-operative intra-abdominal infection.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Graphs displaying the overall survival curves for hepatocellular carcinoma patients with and without post-operative intra-abdominal infection.

Figure 2 shows the OS curve for HCC patients with and without post-operative intra-abdominal infection. The 3-year OS rate in the intra-abdominal infection group and the non-intra-abdominal infection group was 52.0% (n = 13) and 88.6% (n = 155), respectively. The difference between these two groups was significant (log-rank test, P < 0.001).

Postoperative NLR change, which was used as a measurement of inflammatory response, was another factor investigated in this study. The optimal cutoff value of delta-NLR for RFS was determined via receiver operating characteristic (ROC) curve. The area under the curve (AUC) was recorded as 0.513 (95%CI: 0.441-0.585). The delta-NLR value of 1.643 was used as the optimal cutoff value. This value resulted in the most appropriate sensitivity of 77.23% and specificity of 36.46%. The patients were stratified into the elevated (> 1.643) and non-elevated (≤ 1.643) delta-NLR groups. As shown in Figure 3, the 3-year RFS rate in the patients with an elevated delta-NLR was significantly poorer than that in the patients with a non-elevated delta-NLR (50.0% vs 68.4%, P = 0.009). Nevertheless, no significant correlation was detected between OS and postoperative NLR change.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Graphs displaying the recurrence-free survival curves for hepatocellular carcinoma patients according to the postoperative neutrophil-to-lymphocyte ratio change (> 1. 643 or ≤ 1.643).

Figure 3 shows the RFS curves of the HCC patients according to the postoperative NLR change (> 1.643 or ≤ 1.643). The 3-year RFS rate in the elevated group and the non-elevated group was 50.0% (n = 70) and 68.4% (n = 39), respectively. The elevated postoperative NLR change group exhibited significantly poorer RFS (log-rank test, P = 0.009).

The results of univariate and multivariate analysis are shown in Table 3. Indicators of underlying liver disease, such as the hepatitis B surface antigen (HBsAg) status and the Child-Turcotte-Pugh (CTP) class were identified as significant predictors of poor RFS. The postoperative NLR change and intra-abdominal infection were also identified as predictors of RFS. Multivariate analysis revealed HBsAg positivity, CTP class B, elevated delta-NLR and intra-abdominal infection as independent predictors of poor RFS. The tumor diameter displayed borderline significance. For OS, multivariate analysis identified hepatic cirrhosis, the maximal tumor diameter (> 5 cm) and intra-abdominal infection as independent predictors of poor OS.

The impact of postoperative infection on prognosis has been evaluated in various malignancies[8,22,23]. The presence of intra-abdominal and respiratory infections was reported to be associated with poorer long-term outcomes in patients who underwent hepatic resection for colorectal liver metastasis[8,14]. In the present study, we analyzed the data from a cohort of 200 HCC patients who underwent curative liver resection and demonstrated that postoperative intra-abdominal infection was an independent risk factor affecting RFS and OS.

Our results showed a relatively higher rate of postoperative infectious complications (34.0%) compared to the results of previous studies[3-5]. This difference may be due to distinct patient population characteristics, particularly the status of liver parenchyma. Most of our patients suffered from hepatic cirrhosis (74.0%), which highly correlated with postoperative intra-abdominal infection (Table 2). In addition to cirrhosis, vascular invasion and concomitant splenectomy correlated with an increased risk of postoperative intra-abdominal infection. However, the precise mechanism underlying these correlations remains unclear. The postoperative infectious complications were stratified according to severity using the C-D classification system in our cohort. No significant differences were detected in the outcomes of the patients according to the C-D grades, suggesting the homogeneity of the pro-neoplastic effects of inflammation on HCC carcinogenesis, despite differences in the severity of the postoperative complication.

The association between postoperative intra-abdominal infection and poor prognosis is complex and remains to be elucidated. Several possible mechanisms may explain the prognostic value of postoperative intra-abdominal infection. One possible mechanism is that the deregulated host immunological response during infection may promote tumorigenesis[10,24]. It has been well documented that infectious complications correlated with the excessive synthesis and release of proinflammatory cytokines, such as interleukin IL-1, IL-6 and TNFα, and these cytokines can affect the function and regulation of cytotoxic T lymphocytes, natural killer cells, and other immunocompetent circulating cells such as dendritic antigen-presenting cells and induce immunosuppression[25,26]. Cell-mediated immune function was also suppressed by the post-surgical stress response. As a result, the residual malignant cells may progress rapidly during the periods of relative immunosuppression caused by postoperative infection and post-surgical stress. In addition, infectious complications may stimulate the release of angiogenesis-regulating chemokines, growth factors and proteases, which are major contributors to tumor-related angiogenesis[27]. Furthermore, it has recently been demonstrated that bacterial antigens, which are frequently involved in postoperative infectious complications, exert a direct effect on the metastatic capacity of cancer cells[11].

The presence of a systemic inflammatory response can be evaluated based on both the elevation of the CRP level and the NLR[14,28]. Because the CRP level was not routinely measured at our center, we assessed the prognostic significance of the NLR as a potential objective biomarker representing the perioperative change in the inflammatory response. The patients with an elevated NLR more frequently exhibited lymphocytopenia and neutrophilic leukocytosis, shifting the balance to a pro-tumor inflammatory response. Various studies have demonstrated that an elevation in the preoperative NLR correlated with a poor clinical outcome in HCC patients treated by different methods[29-32]. The postoperative NLR was also considered to be a useful marker that displayed remarkable prognostic value[33]. However, neither the preoperative nor postoperative NLR reflects the dynamic change in the balance between the host inflammatory response and the immune response after curative surgery. Recently, the postoperative NLR change was found to be an independent prognostic factor for patients with small HCCs receiving RFA. The patients with a decreased NLR exhibited a higher survival rate than those with an increased NLR[30]. Our investigation showed a similar result among the HCC patients who underwent liver resection. The postoperative NLR change was an independent factor for RFS. Neither preoperative nor postoperative NLR was significantly associated with RFS or OS.

Our results should be interpreted with caution as this retrospective study contains some limitations. First, this study was based on retrospective data. Second, these data were collected from a single institution. Thus, further validation in prospective studies enrolling patients from additional institutions is needed. Third, we chose to stratify the patients according to a delta-NLR of > 1.643 or ≤ 1.643 based on ROC curve analysis. However, it is uncertain whether a different cutoff value would serve as a better predictor of disease recurrence for these HCC patients or whether NLR would be better categorized into “high”, “intermediate”, and “low” groups rather than two groups.

In conclusion, the results of our study suggest that postoperative intra-abdominal infection has an adverse effect on the oncologic outcomes (RFS and OS) of HCC patients after hepatectomy. The postoperative NLR change (delta NLR), a simple and objective inflammation index, is an independent predictor of tumor recurrence in HCC patients.

The authors thank Heng Zhang and Fan Zhang for their contribution in this study.