Prediction of postoperative inflammatory complications after esophageal cancer surgery based on early changes in the C-reactive protein level in patients who received perioperative steroid therapy and enhanced recovery after surgery care: a retrospective analysis

The patients were selected from the medical records of consecutive patients who underwent esophagectomy for esophageal cancer at Kanagawa Cancer Center from January 2011 to September 2015. The patients met the following inclusion criteria: (1) histologically proven primary esophageal squamous cell carcinoma located at thoracic esophagus, (2) clinical stage I to III (excluding T4) disease as evaluated using the 7th edition of the tumor-node-metastasis classification established by the Union for International Cancer Control [16], and (3) neoadjuvant chemotherapy followed by curative resection with radical lymph node dissection.

The patients received two courses of cisplatin plus 5-fluorouracil. Cisplatin was administered at a dose of 80 mg/m² by intravenous drip infusion on day 1, and 5-fluorouracil was administered at a dose of 800 mg/m² by continuous infusion on days 1-5 [1].

Surgical resection was generally performed 4-6 weeks after the completion of chemotherapy. Our standard procedures consisted of open subtotal esophagectomy via right anterolateral thoracotomy, reconstruction with a gastric tube through the posterior mediastinal route or retrosternal route, and anastomosis in the cervical incision. In principle, two-field lymph node dissection is indicated when tumors are located at the middle thoracic to lower thoracic esophagus, while three-field is applied for upper thoracic tumors. Multiple drains were placed; one to the posterior side of the thoracic cavity and the others on either side of the neck. A feeding tube was routinely placed at the stomach or duodenum.

All of the patients received perioperative management by the clinical path based on the ERAS program, which routinely included antibiotic prophylaxis and steroid therapy. Cefazolin (1 g) was administered 30 min before surgical incision and then again every 3 hours during surgery and at 2 g on postoperative day (POD) 1. Methylprednisolone was administered at a dose of 500 mg on the day of surgery, 250 mg on POD 1, and 125 mg on POD 2 [13, 14]. Our ERAS program satisfied the 15 items proposed by Fearon et al. [17]. Briefly, the patients were allowed to eat 30% rice porridge until midnight the day before the surgery and were required to drink the contents of two 500-ml plastic bottles containing oral rehydration solution by 3 h before surgery. Intraoperatively, we conducted epidural anesthesia with morphine for pain control during surgery. Previous study showed the use of the epidural anesthesia with morphine has clinical benefits such as, a selective analgesia with no motor or sympathetic blockade and a long analgesia at low use of rescue medication [18, 19]. However, the use of the epidural anesthesia with morphine could cause delayed respiratory depression and apnea as late as 12 hours after administration [20]. Therefore, the patients remained on ventilation for 12 hours after surgery. After 12 hours, we carefully observe respiratory condition and extubate. Ambulation and enteral nutrition was started on POD 1. Oral intake was initiated on POD 6, beginning with water and gelatinous foods. The patients began to eat solid food on POD 9, starting with rice gruel and soft food and progressing in three steps to regular food intake.

All data were retrospectively retrieved from the patients’ records. ICs were defined as complications of anastomotic leakage, pneumonia, abdominal abscess, and/or pyothorax according to the Clavien-Dindo classification [21] occurring during hospitalization within 30 days after surgery. Of these, ICs ≥ grade IIIa were defined as serious ICs (SICs). The complications were assessed based on the clinical symptoms, blood tests, and X-ray imaging at POD 1, 2, 4, 6, 8, and thereafter. If ICs were suspected, precise examinations, such as computed tomography, esophagography, and esophagoduodenoscopy, were performed.

A two-sided P value < 0.05 was considered significant. Continuous data are presented as the median with the range. The Mann-Whitney U test and Fisher's exact test were employed to evaluate the differences in continuous and categorical variables, respectively. The patients were classified as those with SICs (SICs group) and those without SICs (NSICs group). The diagnostic accuracy was determined based on the area under the receiver operating characteristic (ROC) curve (AUC) [22]. The optimal cut-off value of CRP was determined by maximizing Youden’s index. The optimum value of CRP was then determined based on the AUC and the earliest prediction of SICs. The predictive value of CRP, categorized as high or low by the cut-off value at the optimum point, was examined using univariate and multivariate logistic regression analyses. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Australia). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics [23].

A total of 208 patients underwent esophagectomy for esophageal squamous cell carcinoma between January 2011 and September 2015. Excluding 3 patients with no survival information available, 7 patients who were not diagnosed with squamous cell carcinoma, 74 patients who did not receive neoadjuvant chemotherapy, and 7 patients who did not receive curative resection, one hundred and seventeen of these patients were eligible for the present study (56.3%). The patient characteristics are summarized in Table 1. The SICs group received three-field lymph node dissection more frequently (p = 0.023) and had greater blood loss (p = 0.018) than the NSICs group.

SICs were observed in 20 patients (17.1%). The details of the complications and duration from surgery to their diagnosis are shown in Table 2. The median duration until the diagnosis of any SICs was 7 days (range: 4-14).

The changes in the CRP level after esophagectomy are shown in Fig. 1. The preoperative CRP level was not markedly different between the SICs and NSICs groups. After surgery, the CRP level reached its first peak on POD 1 and 2, with no significant differences between the two groups, and then decreased to its lowest value on POD 4. However, the subsequent CRP levels on POD 4, 6, and 8 were significantly higher in the SICs group than in the NSICs group. The AUC for prediction by CRP was 0.778 (95% CI, 0.673-0.884) on POD 4 (Fig. 2a), 0.875 (95% CI, 0.799-0.952) on POD 6 (Fig. 2b), and 0.883 (95% CI, 0.813-0.953) on POD 8 (Fig. 2c). Considering the AUC and earliest prediction of SICs, the optimum cut-off value of CRP was determined to be 4.0 mg/dl on POD 4. By this cut-off, 40 patients had high CRP with median of 6.95 mg/dl (range: 4.01-28.51), while 77 patients had CRP with median of 1.50 mg/dl (range: 0.13-3.99). Among 40 patients with high CRP levels, 16 developed SICs; anastomotic leakage in 12 patients, pneumonia in 2, abdominal abscess in 1, and pyothorax in 3. The sensitivity and specificity were 80.0% and 75.3%, respectively, and the negative and positive predictive values (NPV and PPV) were 94.8% and 40.1%, respectively.

Table 3 shows the results of univariate and multivariate analyses (Table 3 ). Among these, CRP ≥ 4.0 mg/dl on POD 4 (odds ratio, 18.600; 95% CI, 4.610–75.200) and three-field lymph node dissection (odds ratio, 7.950; 95% CI, 1.900–33.400) were identified as significant independent predictive factors for SICs.