Interleukin-8 is superior to CRP for the prediction of severe complications in a prospective cohort of patients undergoing major liver resection

Continued advances in diagnostics and treatments have expanded the indication and extent of liver resections. Currently, it has become an established curative treatment option, particularly for hepatic metastases from colorectal cancer. The success in terms of 5- and 10-year survival rates in patients undergoing curative R0 resections for metastatic colorectal cancer has dramatically increased the frequency and extent of hepatic resections [1]. These extensive resections in increasingly multimorbid patients have resulted in postoperative complication rates of up to 50% [2], which may impact not only quality of life but also overall survival [3].

Early diagnosis of complications could mitigate their negative impact on patients’ overall outcome of patients by reducing morbidity, hospital stay and cost, and improving quality of life and overall survival. Several cellular and acellular blood components belong to the standard repertoire for the detection of postoperative complications, as these markers are subject to significant changes during acute or chronic inflammatory responses.

Serum C-reactive protein (CRP) is synthesized by the liver upon secretion of IL-6 by T cells and macrophages. CRP is a robust and well-described marker in elective colorectal cancer resection [4, 5], but a retrospective study has shown the limited predictive power of CRP in patients undergoing major liver resection [6]. The neutrophil-lymphocyte ratio (NLR) and thrombocyte-lymphocyte ratio (TLR) are primarily liver independent and have shown some predictive power for complications in elective gastric surgery and in intensive care patients [7‐9], as well as predictive power for poor survival in several cancers, including colorectal cancer, hepatocellular carcinoma, and hepatopancreaticobiliary malignancies [10‐12]. Proinflammatory cytokines, such as IL-6 and IL-8, are predominantly produced at the site of infection and can therefore be considered early response markers of the acute phase response (APR). They are closely associated with inflammation, liver disease, and liver regeneration [13, 14]. IL-6 is secreted by T lymphocytes, endothelial cells, and macrophages. It reaches the liver via the bloodstream and induces APR, namely, the production of CRP in the liver. IL-8 is produced by macrophages, epithelial cells, endothelial cells, and fibroblasts and initiates the recruitment of neutrophils; thus, its release is upstream of the NLR. Both proinflammatory cytokines have shown promising results in critically ill patients and in acute pancreatitis [15, 16], but have not yet sufficiently been described in liver resected patients.

The aim of the present study was to identify predictive markers for the development of severe postoperative complications (Clavien-Dindo ≥ 3) in patients undergoing various degrees of liver resection and to demonstrate the feasibility of IL-8 analysis in a clinical setting.

A total of 139 patients undergoing hepatic resection were prospectively enrolled in the study. The group consisted of 56 female (40%) and 83 male (60%) patients with a mean age of 61 years (27–84 years). Only 4% of patients had an ASA (American Society of Anesthesiologists) score of 1, 30% had an ASA score of 2, and 67% had an ASA score of 3. Almost all patients underwent liver resection for malignant disease (n = 128; 92%), with most resections performed for colorectal liver metastases (n = 78), followed by cholangiocarcinoma (n = 21), HCC (n = 17), and other metastases (n = 12). Of the 139 patients, 45 patients (32%) underwent wedge resections, 49 (35%) underwent minor resections, and 45 (32%) underwent major resections (≥ 3 liver segments). Further characteristics of the cohort in terms of clinical data, complications (CDC), and the operations performed are shown in Table 1 and Supplementary Table 1.

A total of 75 complications occurred in 64 out of 139 patients (46%). Minor complications (CDC I–II) were observed in 26 patients (19%), while 33 patients (24%) had at least one severe complication (CDC III–IV), and 5 patients (4%) died during hospitalization. Major liver resections were associated with more complications and longer intensive care unit (ICU) stay than wedge or minor resections (Tables 1 and 2).

One-quarter of the complications occurred before day 4 (n = 21; 25%), 14 (17%) occurred between days 4 and 6, 34 (41%) between days 7 and 14, 9 (11%) between days 15 and 30, and 5 (6%) between days 30 and 90. Fifteen patients (11%) had major complications by day 6, 2 after wedge resections, 4 after segmental resections, and 9 after major liver surgery. Of these, 11 patients (8%) had major complications before day 4. Complications before day 4 consisted mostly of early biliary leaks, sometimes leading to vascular erosion with hemorrhage, which were detected by drainage fluid inspection, as well as some medical complications such as pulmonary embolism or one case of a brief episode of postoperative asystole of unclear origin that requiring brief CPR and the administration of atropine (Supplementary Fig. 1).

Our study is the first to prospectively evaluate a wide range of serum markers in patients undergoing liver resection. The large number of patients included (n = 139), as well as the wide spectrum ranging from wedge resections to minor liver resections and major liver resections, provides a representative picture of the clinical reality in a maximum care hospital. The subgroup analyses provide further insight into the clinical issue of detecting major complications in patients undergoing major liver resection. A weakness of the study is the relatively small sample size in the subgroups. In particular, the subgroups of patients who underwent wedge resections and minor resections have fewer major complications (18%), which increases the risk of statistical variance in these groups. Furthermore, the novel serum markers IL-6 and IL-8 are not part of the standard repertoire of every hospital. However, the total number of resected patients is very large (n = 139), and the subgroup of patients who underwent major liver resection represents a high-risk cohort with a high complication rate (46% severe complications), making the results reliable. In addition, major hepatic resections are technically complex procedures that require a significant amount of resources and should be reserved for specialized centers. We do not expect any problems with the implementation of IL-8 into their repertoire should these results be confirmed in further studies.

Our prospective study provides a comprehensive summary of the predictive values of a panel of widely used and novel serum markers for the detection of clinically relevant complications in patients undergoing liver surgery. After comparing the predictive values of all markers, leukocytes, CRP, IL-6, and IL-8 showed the most consistent results, with acceptable predictive values (AUC ≥ 0.7) and significant differences (p ≤ 0.05) in more than one of our defined subgroups. As commonly used in daily clinical practice, CRP works best when its dynamic course is taken into account (∆CRP). Then, CRP is a robust predictive marker of postoperative outcome for patients undergoing wedge or minor liver resection. However, in patients undergoing major liver surgery, the predictive correlation with postoperative complications is rather poor. These findings are consistent with previous publications of retrospective data [6, 20] and can be explained by the synthesis and release of CRP by the liver, which limits the response after extensive liver tissue resection. In contrast, proinflammatory cytokines, particularly IL-8, showed a stronger correlation with the occurrence of postoperative complications, especially in major liver surgery. IL-6 and IL-8 are both released at the site of inflammation, which should make them independent of liver function and therefore of the extent of liver resection [21, 22]. Nevertheless, IL-6 could be used to predict postoperative complications with relative accuracy only in the subgroup with wedge and minor resections, similar to CRP. This could be explained by the fact that Kupffer cells (resident macrophages of the liver) are an important source of IL-6 [23]. The residual liver tissue after major liver resection may not be sufficient to synthesize IL-6, thus limiting the IL-6 response after extensive liver tissue resection, similar to the CRP response.

In contrast, the release of IL-8 is more independent of functional liver tissue. In addition to macrophages and Kupffer cells, fibroblasts and endothelial cells at the site of infection also secrete IL-8 [22, 24, 25]. The release of IL-8 leads to an increased recruitment of neutrophils, which is less affected by liver function and its residual volume. This may contribute to the fact that IL-8, in particular, is equivalent to CRP after minor surgery and far superior to CRP as a predictor of postoperative severe complications after major liver surgery. The weaker performance in wedge and minor resections is unclear, but might be partly explained by a lower number of major complications in these groups (wedge resections n = 8, minor resections n = 9), making these groups prone to higher statistical variability. This could also explain why IL-8 performs much better in patients undergoing major liver surgery, as well as in all patients. However, this needs to be confirmed in future studies.

This also provided the rationale to additionally investigate the NLR and TLR as surrogate markers for postoperative complications after liver surgery. Despite being fundamentally liver-independent, thrombocytopenia is a common finding after liver resection [26] and in liver disease [27]. Similarly, neutrophil function is severely impaired in patients with chronic liver dysfunction [28]. Therefore, it was reasonable to hypothesize that NLR or TLR could be a robust surrogate marker for liver dysfunction as well as for posthepatectomy-associated complications. However, in our prospective cohort, we identified NLR solely as a diagnostic marker for postoperative complications in patients undergoing wedge resection, whereas TLR showed no predictive value. Moreover, our investigation revealed that NLR and TLR exhibited only marginal predictive capabilities as markers for postoperative complications in major hepatic surgery. These findings are in line with previous scientific studies, such as the research conducted by McCluney et al., which also reported limited discriminative capacity of NLR in the context of postoperative complications in hepatic surgery [20].

Furthermore, our study findings revealed the occurrence of significant complications in 15 patients (11%) within the initial 6 days following the procedure. Notably, 10 patients (7%) experienced such complications before the fourth postoperative day. The early complications primarily consisted of early biliary leaks which were detected prior to the manifestation of a substantial infectious component, or postoperative hemorrhage, both readily detectable through drainage fluid inspection. Additionally, noninfectious medical complications, such as pulmonary embolism or temporary postoperative asystole of uncertain origin, posed diagnostic challenges when utilizing infectious serum markers.

The apparent limitations of IL-8 as a predictive marker before the fourth day after surgery can be attributed to a combination of factors. Firstly, our analysis considered the aforementioned noninfectious complications, which constituted a significant portion of early adverse events. Secondly, the vast majority (92%) of the patients underwent liver resection for malignant indications, a well-known factor associated with elevated IL-8 levels [29, 30]. Notably, patients with larger tumors undergoing major liver resection exhibited elevated preoperative IL-8 values, which further increased after resection and subsequently normalized around day four in patients without postoperative complications (Fig. 1). Moreover, the act of liver resection itself led to increased postoperative IL-8 values, which normalized by the fourth day. Considering these factors, selecting the fourth day post-surgery appears to strike a reasonable balance between sensitivity and specificity for analyzing IL-8. This is corroborated by the observation of good prognostic differentiation on the fourth day in patients who underwent major liver resection. However, the reasons for IL-8’s diminished prognostic performance in patients undergoing minor liver surgery remain elusive, and potential explanations may be linked to the lower incidence of major complications within this subgroup. Additionally, it is plausible that IL-8’s involvement in cancer metabolism, inflammation, and progression is more pronounced in larger cancers, contributing to the observed differences in prognostic efficacy. These observations may also shed light on the association between preoperative IL-8 levels and the occurrence of major postoperative complications. The presence of an inflamed microenvironment in patients with malignant liver lesions might influence the development of postoperative complications. However, due to the heterogeneity of malignant tumors included in this study, a more comprehensive analysis is not currently feasible. Nevertheless, these results are highly intriguing and warrant confirmation in further studies.

In summary, our study demonstrates that established markers commonly used for clinical monitoring after surgery fail to predict severe complications following major liver surgery. Our data suggest that leukocyte count and dynamic changes in CRP remain reliable markers in patients undergoing wedge or minor liver surgery, but are not suitable for predicting postoperative complications after major liver surgery. Additionally, it should be noted that inflammation markers specific to the liver, such as CRP, are influenced by the remaining liver volume, particularly after major liver resection.

While IL-8 is not a routine marker that is easily measured in every laboratory, our study showcases the feasibility of routine IL-8 assessment in serum using standardized testing equipment, which can easily be reproduced. Although the cost of determining IL-8 in the serum is significant (approximately 10–15 Euros per test), it may be worthwhile to consider its assessment in specific scenarios, such as major liver surgery.

To effectively predict and diagnose complications in these patients during the early postoperative stage, liver-independent markers like IL-8 should be validated in future trials with the aim of incorporating them into the standard repertoire of blood tests in the future.