Metabolic syndrome predicts postoperative complications after gastrectomy in gastric cancer patients: Development of an individualized usable nomogram and rating model

Abstract Background Metabolic syndrome (MetS), a public health problem, is reportedly related to an increased risk of postoperative complications after surgery. However, whether MetS have an effect on complications after gastric cancer (GC) surgery are unknown. This study aimed to investigate the effects of preoperative MetS on complications after gastrectomy. Methods Altogether, 718 gastric cancer patients who planned to receive radical gastrectomy between June 2014 and December 2016 were enrolled, demographic and clinicopathological characteristics were analyzed. Univariate and multivariate analyses were performed to identify potential risk factors for postoperative complications. A predictive model for postoperative complications was constructed in the form of a nomogram, and its clinical usefulness was assessed. Results Of the 628 patients ultimately included in the study (mean age 62.92 years, 450 men and 178 women), 84 were diagnosed with MetS preoperatively. Severe postoperative complications (Clavien‐Dindo grade ≥II) were significantly more common in patients with MetS (41.7% versus 23.7%, P < .001). Predictors of postoperative complications included MetS (odds ratio [OR] = 1.800, P = .023), age (OR = 1.418, P = .050), Charlson score (OR = 1.787, P = .004 for 1‐2 points) and anastomosis type (OR = 1.746, P = .007 for Billroth II reconstruction). The high‐risk rating had a high AUC (ROC I = 0.503, ROC Ib = 0.544, ROC IIa = 0.601, ROC IIb = 0.612, ROC IIc = 0.638, ROC III = 0.735), indicating that the risk‐rating model has good discriminative capacity and clinical usefulness. Conclusions MetS was an independent risk factor for complications after gastrectomy. The nomogram and rating model incorporating MetS, Billroth II anastomosis, age, and Charlson score was useful for individualized prediction of postoperative complications.


| INTRODUCTION
Metabolic syndrome (MetS) is characterized by a series of metabolic disturbances including obesity, hyperglycemia, dyslipidemia, and high blood pressure. [1][2][3] Recently, studies have indicated the carcinogenic function of MetS in variety cancer including liver, 4 pancreatic, 5 colorectal, 6 and gastric 7 cancers, and associations between MetS and poor prognoses have been reported in patients with these types of cancers. 2,8 A number of studies have yielded discordant results however, 9,10 thus, it remains controversial whether preoperative MetS is related to postoperative complications.
Worldwide, the incidence and lethality of the gastric cancer (GC) has increased during the past decades, GC has been the 4th common cancer and the 3rd cause of cancer-related-mortality. 11,12 To date surgical resection remains the preferred therapeutic approach for gastric cancer. Postoperative complications after gastrectomy affect life quality, delay subsequent therapies, and even overall survival, and despite the development of minimally invasive surgery and the enhancement of postoperative recovery program, their reported prevalence still ranges from 10% to 46%. [13][14][15] Risk assessment preoperatively is therefore significant, to identify postoperative complication prone patients.
Considering that the aforementioned previous ambiguous results are probably because of retrospective research designs as well as limited patient cohorts, we conducted the present prospective multicenter study incorporating a large sample size with the aim of better evaluating the potential effects of preoperative MetS on postoperative complications after gastrectomy.

| Patient
GC patients who received radical gastrectomy with D2 lymphadenectomy at the Department of Gastrointestinal Surgery of the Second Affiliated Hospital of Wenzhou Medical University and Yuying children's Hospital between 01 June 2014 and 31 December 2016 were screened for inclusion in the study. The inclusion criteria were as followed: (a) postoperative pathology confirmed gastric adenocarcinoma; (b) aged ≥ eighteen years old; (c) Sign the participation informed consent. The following exclusion criteria were applied: (a) Underwent palliative surgery or emergency surgery; (b) received neoadjuvant chemotherapy or radiotherapy; (c) had a severe immune disease, blood disease, endocrine disease, or thyroid dysfunction; (d) had a history of other malignant tumors; (e) had a history of previous surgery; (f) there was a substantial absence of clinical data. Further details and a flow chart depicting the study procedure are shown in Figure 1.

| Baseline data collection
Demographic details including age and sex were collated for each patient, as were details pertaining to the operation such as tumor size, tumor location, and tumor stage. Severe postoperative complications were classified as Clavien-Dindo 16 grade II or above in this study.

| Diagnosis of MetS
MetS is defined in accordance with the recommendations of the Chinese Diabetes Society 17 preoperatively, with three or more of the following five items: (a) Body mass index (BMI) determined obesity (BMI≥ 25 kg/cm 2 ); (b) with fasting blood glucose above 6.1 mmol/L or diabetes history; (c) systolic blood pressure above 140 mm Hg or diastolic blood pressure ≥90 mm Hg or hypertension history; (d) triglycerides above 1.70 mmol/L; (e) low high-density lipoprotein cholesterol (<0.9 mmol/L for men, <1.0 mmol/L for women).

| Statistical analysis
Parameters Distribution equality was first assessed by Kolmogorov-Smirnov test, means/standard deviation (SD) and medians/interquartile ranges (IQR) were chosed to present normally and non-normally distributed data, respectively. Chi-squared test or Fisher's exact test were further performed to compare categorical data. Univariate logistic regression analysis was used and all the variables yielding p values < .05 were then selected into multivariable logistic regression (forward stepwise variable selection) to establish the regression model. A nomogram that could quantitatively predict the incidence of secondary or higher complications was constructed, subsequently. Decision curve analysis (DCA) and Receiver operating characteristic (ROC) curve analysis were finally performed to evaluated the clinical usefulness of the nomogram as well as rating model. Statistically significant was determined as P < .05. All the statistical analyses were performed using SPSS version 23.0 and R version 4.0.0.

| Baseline characteristics
Among the 718 patients initially included, 90 were subsequently excluded, the reasons for exclusion are detailed in Figure 1. Finally, 628 patients were included in the analysis. Table 1 exhibited the overview of patient population. The mean age of the patients was 62.92 ± 11.45 years, and they included 450 men (71.7%) and 178 women (28.3%).
With regard to tumor clinicopathological characteristics, the tumors were located in the antrum in 550/628 patients (87.6%), and 439/628 patients (69.9%) had a tumor that was >2.1 cm along its longest axis. Differentiated tumor histopathology was detected in 549 of the patients (87.4%). Based on the above-described Chinese Diabetes Society 17 criteria 84 patients (13.4%) were diagnosed with MetS. The incidence of MetS increased with age, as shown in Figure 2

| Nomogram development and usefulness
A model incorporating the independent predictors that yielded P < .05 in multivariate analysis was developed and structured as a nomogram (Figure 4). Further DCA demonstrated that a use the nomogram adds more benefit than either the treat all or treat none scheme while the threshold probability of a patient is 13% to 54% ( Figure 5). For example, if the personal threshold probability of a patient is 30% (ie, the patient would opt for treatment if their probability of postoperative complications was 30%), then the net benefit when using the nomogram to make a decision as to whether to undergo treatment is 0.246, constituting added benefit compared with the treat-all or treat-none scheme.

| Development of postoperative complications risk-rating model
Based on the univariate analyses, a postoperative complications risk-rating model was further developed (Table 3). ROC analyses showed that higher risk-rating predicted the occurrence of complications more accurately with a higher area under the curve (AUC) (ROC I = 0.503, ROC Ib = 0.544, ROC IIa = 0.601, ROC IIb = 0.612, ROC IIc = 0.638, ROC III = 0.735) ( Figure 6). 13.4%, which is comparable to that previously reported. 7 MetS was further turned out to be an independent risk factor for severe postoperative complications after gastrectomy as expected. Given that MetS is consist of a bunch of metabolic abnormalities such as obesity, dyslipidemia, hypertension, and hyperglycemia, insulin resistance may be the most probable pathophysiological cause. Mounting evidence suggests that insulin resistance influences abnormal metabolism in adipocytes, with succeeding increasing of pro-inflammatory cytokines and decreaing of adiponectin (protective adipokine), 20 thus leading to infective complications including infection of surgical site, abdominal infection, and pneumonia. Abdominal obesity is also reportedly correlated with insulin resistance-and thus reduced oxygen tension within surgical wounds-which can give rise to infectious complications. 21 Furthermore, the massive and fragile fat tissue in abdominally obese patients is easily torn and greatly reduces surgical precision, 22 which could prolong the surgery duration and increase intraoperative hemorrhage, resulting in an increased risk of wound infection. In another study hypertension and dyslipidemia were both related to poor microvascular circulation, 23 which impaired tissue repair and increased the risk of wound complications such as anastomotic leakage. Although each individual component may only have a small influence, the risk of postoperative complications was substantially increased when these minor forms of damage occurred together. As has been reported previously, 2,24 we also found the incidence rate of MetS was higher in the elderly. Age was another independent risk factor for postoperative complications in the present study. Considering the difference in physical condition between younger patients and older patients, we investigated whether MetS correlated with outcome separately in the younger and older groups. Interestingly MetS could only independently predicted postoperative complications in the young. It may be that the value of MetS predicting postoperative complications in the elderly is reduced by age and age-related comparatively poorer physical condition. Thus under the same diagnostic criteria, more attention should be paid to young patients.

| DISCUSSION
In the present study, Billroth II anastomosis was also an independent risk factor for postoperative complications in our study, consistent with Kang KC demonstrated that the postoperative complication was much less in a Billroth I group than in a Billroth II group. 25 Notably however, in another study there was no statistical differences in morbidity derived from postoperative complications in a Billroth I group and a Billroth II group. Despite of differences in operation technique, although Billroth II reconstruction was tension-free with the application of anastomosis of gastric-jejunum, it has the disadvantages of altering the digestive tract and taking longer time to empty, which potentially leading to duodenal stump leakage. 26,27 Consistent with previous reports, 28,29 Charlson score was independently relevant to postoperative complications in the current study.
The nomogram incorporating MetS, anastomotic type, age, and Charlson score to predict individual risk of postoperative complications developed in the current study has potential clinical applications. Patients at a high risk of developing postoperative complications could be monitored more closely. With this aim, DCA, which offers insight into clinical consequences based on net benefit derived threshold probability, was applied in this study. The net benefit was considered as the difference between proportion of true positives and false positives and weighted by the relative harm of false positive and false negative. DCA demonstrated that while a patient's threshold probability determined by the nomogram developed in the present study is > 13% and < 54%, offering prophylactic measures adds more benefit than either the treat all or treat none scheme.
A postoperative complications risk rating model was finally developed in the present study. Further ROC analysis showed that a higher risk rating corresponded with a higher AUC indicating that the model could be used effectively to screen susceptible patients of postoperative complications. The current study is the first to investigate relationships between MetS and postoperative complications, and also first constructed the nomogram to quantify the possibility of postoperative complications.
The present study had several limitations. Although it can be considered representative because the patients in the study were prospectively enrolled in two large centers, the conclusions of the study is needed to confirm via a randomized controlled multicenter study. Another limitation is that the study only included short term outcomes (within 30 days postoperatively). Long term follow-up data are also needed.
In conclusion, it is the first report of relationship between preoperative MetS and short-term prognosis. MetS, as well as Billroth II anastomosis, age, and Charlson score affected postoperative complications after gastrectomy. A nomogram incorporating these independent risk factors was developed. Preoperative use of this prediction tool could provide useful information to surgeons with respect to better preoperative preparation to decrease the risks associated with the surgery economically and conveniently.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE
All participants provided their written informed consent, and the protocol for this study was approved by the ethics committee of The Second Affiliated Hospital of Wenzhou Medical University and Yuying children's Hospital.