Prognostic significance of the Controlling Nutritional Status (CONUT) score in predicting postoperative complications in patients with Crohn’s disease

Preoperative immune-nutritional status is correlated with postoperative outcomes. The Controlling Nutritional Status (CONUT) score is a useful tool for predicting the postoperative outcomes of cancer surgery. This study aimed to evaluate whether the CONUT score could predict postoperative complications in Crohn’s disease (CD) patients. In total, 202 CD patients were eligible. Univariate and multivariate analyses were performed to identify risk factors for postoperative complications. Receiver operating characteristic (ROC) curves were generated to examine the cutoff value for predictors of postoperative complications. Among all the patients, 66 developed postoperative complications. The cut-off value of the CONUT score was 3.5 for complications. Eighty-one patients had a low CONUT score (< 3.5), and 121 patients had a high CONUT score (> 3.5). There was a significant difference in postoperative complications between the groups with low and high CONUT score (17.3% vs. 43.0%, p < 0.001). Patients with high CONUT score had low body mass index (BMI), more mild postoperative complications (p = 0.001) and a longer postoperative stay (p = 0.002). Postoperative complications were correlated with BMI, preoperative albumin, the preoperative CONUT score, and preoperative infliximab use. Then, the preoperative CONUT score was an independent risk factor for complications (OR 3.507, 95% CI 1.522–8.079, p = 0.003). ROC analysis showed that the CONUT score was a better predictor of postoperative complications in CD patients than albumin and the prognostic nutritional index. Thus, a preoperative CONUT score cut-off value of more than 3.5 could help to identify patients with a high possibility of malnutrition and postoperative complications.

www.nature.com/scientificreports/ hazardous surgical complications, and postoperative complications in patients with low albumin can be minimized by nutritional support. Grass et al. 12 reported that malnutrition was a major risk factor for postoperative complications, and postoperative morbidity can be decreased efficiently by enteral and parenteral routes. Additionally, a high preoperative CDAI score with severe inflammatory response could predict negative postoperative outcomes 13 . Therefore, it is necessary to find a scoring system to assess both nutritional status and immune status.
The Controlling Nutritional Status (CONUT) score is calculated based on the serum albumin level, total blood cholesterol concentration, and total peripheral lymphocyte count, and this objective tool was first reported for the early detection of hospital malnutrition in 2005 14 . The CONUT system is considered to be associated with the host's nutritional and immune status based on these three laboratory parameters 15 . The CONUT is widely used as a prognostic indicator in cancer patients. Kuroda et al. 16 reported that the CONUT score tended to estimate nutritional status and predict long-term overall survival in gastric cancer patients after surgery. The early postoperative CONUT score was found to be an independent risk factor for postoperative complications in patients with hepatocellular carcinoma after liver resection 17 . Recently, a meta-analysis showed that the CONUT score was helpful for predicting which patients with gastrointestinal and hepatopancreatobiliary cancers are at an increased risk of mortality and postoperative complications 18 . However, the clinical significance of the CONUT system for predicting postoperative outcomes in CD remains unknown.
This study analysed a large cohort of CD patients in the Inflammatory Bowel Disease Center of our hospital to demonstrate the role of the CONUT score in CD patients. The aim of the current study was to evaluate the relationship between the CONUT score and postoperative complications and to find a reliable and simple indicator to reflect immune-nutritional status in CD.

Results
Study population and baseline characteristics. A total of 202 CD patients were enrolled in the study; 135 (66.8%) were male, and the mean age of the patients was 36.5 ± 0.9 years. In total, 202 patients received an anastomosis after bowel resection, and 60 patients received a protective ostomy. The mean disease duration before surgery was 48.7 ± 3.4 months. Ileocolonic involvement was the most common disease pattern. Preoperatively, 39 (19.3%) patients had a medical history of azathioprine, 24 (11.9%) of infliximab, 55 (27.2%) of 5-ASA, 10 (5.0%) of corticosteroids, and 70 (36.1%) of enteral nutrition. The Montreal classification and the history of medical treatment are shown in Table 1. The mean CONUT score was 4.3 ± 0.2, and the mean PNI score was 41.7 ± 0.5. The ROC curve analysis showed that the CONUT cut-off value was 3.5. Thus, 81 (40.1%) patients were in the low CONUT group, with a score less than 3.5, and 121 (59.9%) were in the high CONUT group, with a score greater than 3.5.
Operative data and postoperative complications. For 131 (64.9%) patients, this was their first surgery. A total of 120 (59.4%) patients underwent laparoscopic surgery successfully; however, 40 (19.8%) experienced conversion to laparotomy. The mean operative time was 191.1 ± 4.2 min. The low CONUT group had a higher rate of laparoscopic surgery than the high CONUT group (74.1% vs. 49.6%, p = 0.001), which might be associated with nutrition status, disease activity, and surgical history. After surgery, a total of 136 (67.3%) CD patients recovered uneventfully, and 66 (32.7%) patients had postoperative complications: 41 (20.3%) had mild complications, and 34 (16.8%) had major complications. The common complications in our hospital were wound infection (10.4%), early postoperative bowel obstruction (6.4%), and gastrointestinal bleeding (5.9%). The incidence of intra-abdominal abscess and anastomotic leakage was only 2.0% and 3.0% in our centre, and there was no significant difference between the low and high CONUT groups. The mean postoperative stay was 10.1 ± 0.4 days. In the low CONUT group, 14 (17.3%) patients had postoperative complications, which was significantly lower than the number in the high CONUT group (17.3% vs. 43.0%, p < 0.001). Therefore, the mean postoperative stay was also longer in the high CONUT group (8.6 ± 0.6 vs. 11.1 ± 0.5, p = 0.002). More details are shown in Table 2.
Factors associated with postoperative complications. Univariate and multivariate analyses were performed to identify risk factors for complications in CD patients. The univariate analysis showed that BMI, preoperative albumin, the CONUT score, and preoperative use of infliximab were associated with postoperative complications. Moreover, there was no significant difference in PNI between the group with and without complications (42.3 ± 0.5 vs. 40.4 ± 0.9, p = 0.079). Then, the multivariate logistic regression analysis identified the preoperative CONUT score and infliximab as independent risk factors associated with postoperative complications (OR = 3.507, 95% CI 1.522-8.079, P = 0.003; OR = 2.619, 95% CI 1.050-6.531, P = 0.039) (Tables 3 and 4).
Predictive accuracy of the CONUT score vs. other scoring systems for postoperative complications. It has been reported that ALB and PNI constitute a practical predictive index for postoperative complications in patients with CD or other digestive diseases 8,10 . ROC curve analysis was performed to examine the predictive accuracy of the CONUT score, ALB, and PNI. The areas under the curve (AUC) for the CONUT score, ALB, and PNI were 0.611, 0.399, and 0.418, respectively, which indicated that the CONUT score might be a better predictor of postoperative complications (Fig. 1). Additionally, the sensitivity of the CONUT score was 78.8%, and the specificity was 50.7%. The Youden index of the CONUT score was 0.281.

Discussion
In patients with CD, malnutrition and systemic inflammation are the main risk factors for postoperative complications. In the current study, our results linked high CONUT scores in CD patients undergoing bowel resection to poor postoperative outcomes. Furthermore, the CONUT score was found to be an independent risk factor Scientific Reports | (2020) 10:19040 | https://doi.org/10.1038/s41598-020-76115-0 www.nature.com/scientificreports/ for postoperative complications in CD. To the best of our knowledge, this was the first study to determine the correlation between the preoperative CONUT score and postoperative outcomes in CD patients. In addition, this study found that the CONUT score was better than other clinical predictors for evaluating nutritional status, such as ALB and PNI. The CONUT score is determined by the serum albumin concentration, total cholesterol concentration, and total lymphocyte count. The concentration of serum albumin will be affected by the nutritional status and other factors, including infection, inflammatory response, and fluid retention status 19,20 . Malnutrition is common in CD, especially in active CD. Studies have reported that hypoalbuminemia is a risk factor for complications following gastrointestinal surgery 21 . Galata et al. 22 reported that a preoperative albumin level greater than 32.6 g/L was associated with a reduced risk of complications, and hypoalbuminemia was the only independent risk factor for major postoperative complications in CD patients undergoing colorectal surgery. Preoperative nutritional optimization in CD was recommended in patients with low albumin levels to minimize postoperative complications 11 . Our results were consistent with previous findings. Our study revealed that CD patients with postoperative complications had significantly lower serum albumin concentrations.
The total peripheral lymphocyte count, one necessary component of the CONUT score, indicates the immunological status of the patient 23 . Various studies have indicated that T lymphocytes affected by the systemic www.nature.com/scientificreports/ inflammatory response play an important role in the depression of innate cellular immunity in cancer patients 24 . A poor prognosis in cancer patients was correlated with decreased T lymphocytes due to an inadequate immune response 20 . Neubauer et al. 25 found the role of apoptosis of peripheral lymphocytes in intestinal inflammation in CD patients. Additionally, peripheral B1a lymphocytes were decreased in association with the severity of disease activity in CD, and B1a lymphocytes played a vital role in immune protection 26 . Thus, the total peripheral lymphocyte count might be associated with postoperative complications in CD. PNI, which includes the total peripheral lymphocyte count, has already been demonstrated to predict complications in CD; the cut-off value is less than 40 10 . Similarly, the CONUT score in our study also included the total peripheral lymphocyte count and was found to be an independent risk factor for complications in CD. The compositional difference between the PNI and CONUT score lies in the total cholesterol concentration. It is suggested that a lower cholesterol concentration has a detrimental effect on postoperative outcomes by affecting antioxidant reserve and inflammatory response 27 . Takagi et al. reported that a low cholesterol level was associated with postoperative complications in patients who underwent gastrointestinal and hepatopancreatobiliary surgery 18 . The serum cholesterol level also plays an important role in the poor prognosis of cancer patients, because cell membrane fluidity is influenced by hypocholesterolemia, which is related to the mobility of cell surface receptors and the ability to transmit transmembrane signals 28 . Therefore, a higher CONUT score including a low cholesterol level predicted poor outcomes in CD patients undergoing surgical therapy.
Although there are some common factors between the CONUT score, PNI, and ALB, the prediction values of these three indicators are different. The ROC analyses revealed that the predictive accuracy of the CONUT score was better than that of PNI and ALB. This might be because of the higher emphasis placed on the serum albumin concentration, total cholesterol concentration, and total lymphocyte count. The combinations of these three factors enhance the utility of the CONUT score for evaluating patients' general condition, which is consistent with previous results in cancer patients 16,20 .
The current study has several limitations to acknowledge. First, this was a retrospective observational analysis, and some residual confounding factors remained. First is the selection bias created due to the exclusive recruitment of CD patients. Second, the cutoff value needs to be evaluated in other cohorts to verify the conclusion of the current study. Last but not least, a multi-centre prospective observational study is warranted, because the outcomes might be influenced by our local experience.
In conclusion, the current study confirmed that the preoperative CONUT score predicted postoperative complications in CD patients undergoing bowel resection. The CONUT system was better than ALB and PNI at predicting postoperative complications in CD. CD patients with a CONUT score over 3.5 should be intensively monitored so that postoperative complications can be detected early.

Materials and methods
Patients. This study was a retrospective review of 202 CD patients undergoing surgical resection from June 2016 to June 2019 at the IBD Center, a teaching hospital of Zhejiang University. All the CD patient data were collected from patients' medical charts in the IBD database. All CD patients were diagnosed using the accepted criteria 29 . This study was approved by the Ethics Committee of our hospital and conformed to the ethical guidelines of the 1975 Declaration of Helsinki. Written informed consent was obtained from all patients.    29 . Patients with undetermined IBD or UC, those without a surgical bowel resection, those younger than 18 years old, those with malignant disease, and those with incomplete laboratory data were excluded. Patients with emergency surgery were also excluded. Thus, 148 patients were excluded in total, 80 due to incomplete laboratory data.
Data collection. All the baseline characteristics data, perioperative data, and laboratory data were collected from the IBD database. Baseline characteristics included age, body mass index (BMI), sex, comorbidity, smoking history, medication, and Montreal classification. CD patients exposed to preoperative infliximab were defined as having a documented dose of infliximab more than 4 weeks before surgery. For patients on preoperative corticosteroids, attempts were made to wean them to a daily dose of 5 mg prednisolone or 4 mg methylprednisolone 4 weeks before surgical intervention. Intraoperative data included operation time and surgical approach (open vs laparoscopy). Laboratory data included white blood count (WBC), red blood count (RBC), haemoglobin (Hb), total blood cholesterol, albumin (ALB), C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and lymphocyte count. The prognostic nutritional index (PNI) was calculated from the serum ALB level and total peripheral lymphocyte count (TLC), and the formula was PNI = 10 × ALB (g/dL) + 0.005 TLC (per mL) 30 . The CONUT score was calculated based on serum albumin level, peripheral lymphocyte counts, and total cholesterol concentrations as in a previous study 15 . The details of the CONUT scoring system are shown in Table 5.

Definition of outcomes.
The primary outcome of this study was to evaluate the predictive value of the CONUT score for postoperative complications. Postoperative complications were defined as those occurring within 30 days from the date of operation or discharge, and complications were documented using the Clavien-