Factors associated with postoperative efficacy evaluation in patients with morbid obesity

The global obesity problem is becoming increasingly serious, with eight of the top ten causes of death in Taiwan in 2020 being related to obesity. Morbid obesity poses a significant threat to one’s health and well-being. In recent years, bariatric surgery has emerged as a more effective treatment option for patients with morbid obesity. However, the procedure is not without risks. This study aims to examine the factors that impact the postoperative efficacy evaluation of patients with morbid obesity. This study uses a retrospective cross-sectional design, with medical records being collected retrospectively. The data was collected from patients who underwent bariatric surgery between July 1, 2017 and June 30, 2020 at a hospital in southern Taiwan. A total of 663 patients were included in the study and were observed for 1 year after the surgery. The independent variables included demographic variables, perceived symptoms variables, perceived lifestyle variables, and surgery-related variables, while the dependent variables included weight loss outcomes and complications. The prognostic factors affecting the postoperative efficacy evaluation of patients with pathological obesity were determined using multiple regression analysis and binary regression analysis. The study found that 65.6% of the participants were female, with an average age of 36.8 years. The results of the multiple regression and binary logistic regression showed that gender, age, BMI, diabetes, and smoking habit were the predictors of postoperative weight loss. Hypertension, diabetes, liver disease, kidney disease, smoking habit, drinking habit, and operation time were the predictors of postoperative complications. The study found that the presence of the aforementioned 12 significant factors can affect the success of weight loss after surgery and the incidence of postoperative complications. This information can serve as a reference for clinical care institutions and patients to improve the postoperative efficacy evaluation.


Results
The results showed that among the 663 bariatric surgery patients, 435 (65.6%) were mostly female.The mean age was 36.88 ± 10.25 years, and the body mass index (BMI) ranged from 32.5 to 40, with a total of 450 (67.9%).The number of perceived symptom variables were greater in those without hypertension (51.0%) than in those with hypertension (49.0%), and in those without heart disease (94.9%) than in those with heart disease (5.1%).Those without diabetes (82.8%) were greater than those with diabetes (17.2%), those with liver disease (68.8%) were greater than those without liver disease (31.2%), and those without kidney disease (78.7%) were greater than those with kidney disease (21.3%).Among the perceived lifestyle variables, those without smoking habits (76.8%) were greater than those with smoking habits, and those without drinking habits (87.5%) were greater than those with drinking habits (12.5%).Surgery-related variables were more common among sleeve gastrectomy patients (98.3%), and the average surgery time was 81.22 ± 18.05 min.The data were compiled in Table 1.
The independent sample T-test results of weight loss results for demographic variables (gender, body mass index), perceived symptom variables, perceived lifestyle variables, surgery-related variables (surgical procedure) and postoperative efficacy evaluation variables are shown in Table 2.In terms of gender (t = 3.351, p < 0.001), the results of weight loss after surgery were statistically significant, indicating that the effect of female was higher than that of male.There were statistically significant differences between body mass index (t = 9.101, p < 0.001) and postoperative weight loss, indicating that patients with body mass index between 32.5 and 40 kg/m 2 had higher weight loss than those with body mass index greater than 40.In terms of perceived symptom variables, diabetes (t = 3.866, p < 0.001) and weight loss results showed a statistically significant difference, indicating that the weight loss effect of patients without diabetes was higher than that of patients with diabetes.The data are shown in Table 2.
In terms of demographic variables, women were used as the reference group.standardized regression coefficient of postoperative weight loss for men was − 0.102, which was statistically significant (p = 0.011), indicating that men had lower weight loss efficiency than women.Age was a continuous variable, and the standardized regression coefficient was − 0.175, which was statistically significant (p < 0.001), indicating that the lower the age, the higher the weight loss effect.For body mass index (BMI), the standardized regression coefficient of 32.5 < body mass index (BMI) < 40 was − 0.313, which was statistically significant (p < 0.001).This suggests that patients with morbid obesity with a body mass index (BMI) of 32.5 to 40 will lose more weight after weight-loss surgery than those with a BMI greater than 40.
The multivariate regression analysis of perceived symptom variables on postoperative weight loss results showed that patients without hypertension were taken as the reference group, and the standardized regression coefficient of postoperative weight loss results in patients with hypertension was 0.005, which was not statistically significant (p = 0.920).In terms of cardiac disease variables, the standardized regression coefficient of postoperative weight loss in patients with cardiac disease was 0.047, which was not statistically significant (p = 0.200).In terms of diabetes variables, patients without diabetes were taken as the reference group.The standardized regression coefficient of postoperative weight loss in patients with diabetes was − 0.126, which was statistically significant (p = 0.001), indicating that patients with diabetes had lower weight loss results than those without diabetes.In terms of liver disease variables, patients without liver disease were used as the reference group.The standardized regression coefficient of postoperative weight loss in patients with liver disease was − 0.017, www.nature.com/scientificreports/which was not statistically significant (p = 0.726).In terms of kidney disease variables, patients without kidney disease were used as the reference group.The standardized regression coefficient of postoperative weight loss in patients with kidney disease was 0.073, which was not statistically significant (p = 0.064).The multivariate regression analysis of the effect of perceived lifestyle variables on postoperative weight loss results showed that patients without smoking habits were taken as the reference group in terms of smoking habits variables.The standardized regression coefficient of postoperative weight loss results in patients with smoking habits was 0.109, which was statistically significant (p = 0.006).These results indicated that smokers had higher weight loss results after weight loss surgery than those without smoking habit.Patients without drinking habits were taken as the reference group, and the standardized regression coefficient of postoperative weight loss effect of patients with drinking habits was − 0.059, which was not statistically significant (p = 0.142).Multiple regression analysis of surgery-related variables on postoperative weight loss results.In terms of surgical variables, patients undergoing sleeve gastrectomy were selected as the reference group.The standardized regression coefficient of postoperative weight loss results in patients undergoing other surgical procedures was − 0.008, which was not statistically significant (p = 0.834).The operation time was a continuous variable, and the standardized regression coefficient was 0.006, which was not statistically significant (p = 0.876).The above results were shown in Table 4.
In terms of the statistical results of binary logistic regression analysis on complications of postoperative efficacy evaluation, Omnibus was first used to verify whether the model was representative, and the fitness of the model of the self-variable terms of this study was tested by binary logistic regression, and the results showed significant (p < 0.001).It can be expressed that at least one of the selected self-variable terms can predict the probability of complications in the outcome variable.The next is the difference between the number of observations and the expected number, including all the predicted variables.Hosmer and Lemeshow was used to detect the fit statistics of the model, and the test result was no statistically significant difference (chi-square statistic was 8.747, p = 0.364), indicating that there was no difference between the number of observations and the expected number.It means that the model has a good fit.Among the demographic variables, females were used as the reference group, and Adjusted odds ratio (AOR) for males was 1.066 times of that for females, with no significant difference (p = 0.767).In terms of age, the incidence of postoperative complications increased by 0.994 times if the age www.nature.com/scientificreports/increased by one year, with no significant difference (p = 0.541).In terms of body mass index (BMI), patients with 32.5 < body mass index (BMI) < 40 as the reference group, the AOR of 32.5 < body mass index (BMI) < 40 were 0.831 times higher than those with ≧ 40, with no significant difference (p = 0.358).Patients without hypertension were taken as the reference group.The AOR adjusted for postoperative complications was 2.617 times higher in patients with hypertension than in patients without hypertension (p < 0.001, 95% CI 1.648-4.154)as a predictor of postoperative complications in patients with morbid obesity; In terms of cardiac disease variables, the AOR of patients with cardiac disease was 1.647 times higher than that of patients without cardiac disease (p = 0.259).
In terms of diabetes variables, patients without diabetes were taken as the reference group.The AOR of patients with diabetes was 2.062 times that of patients without diabetes, showing a significant difference (p = 0.008, 95% CI 1.204-3.533),which was a predictor of postoperative complications in patients with morbid obesity.In terms of liver disease variables, patients with liver disease were taken as the reference group, and the AOR for postoperative complications was 1.968 times that of patients without liver disease, with a significant difference (p = 0.016, 95% CI 1.135-3.413), is a predictor of postoperative complications in patients with morbid obesity; In terms of kidney disease variables, patients with kidney disease were used as the reference group, and the AOR for postoperative complications was 2.811 times that of patients without kidney disease, with a significant difference (p < 0.001, 95% CI 1.702-4.644) is a predictor of postoperative complications in patients with morbid obesity.In terms of the smoking habit variable, patients with smoking habit were used as the reference group, and the AOR of patients with smoking habit was 3.489 times higher than those without smoking habit, showing a significant difference (p < 0.001, 95% CI 2.136-5.697), is a predictor of postoperative complications in patients with morbid obesity; In terms of the variables of drinking habits, patients without drinking habits were taken as the reference group, and the AOR of patients with drinking habits was 2.770 times that of patients without The results of binary logistic regression analysis of postoperative complications for surgery-related variables showed that the AOR for patients undergoing sleeve gastrectomy was 0.864 times higher than that for patients undergoing sleeve gastrectomy, with no significant difference (p = 0.864).In terms of operation time, if the operation time was increased by one minute, the incidence of postoperative complications would increase by 1.014 times, with a significant difference (p = 0.021, 95% CI 1.002-1.026),which is a predictor of postoperative complications in patients with pathological obesity.The data are shown in Table 5.

Discussion
In recent years, patients with morbid obesity are gradually using bariatric surgery as a treatment method.The weight loss effect and complications will affect the postoperative efficacy evaluation and patient safety.Multiple regression analysis showed that gender (β = − 0.012, p = 0.011), age (β = − 0.175, p < 0.001) and body mass index (BMI) (p < 0.001) were all predictors of pathological obesity.In terms of gender, the weight loss effect of females was better than that of males, with a statistically significant difference (p < 0.001), which was different from the research of domestic scholars, which showed that the weight loss effect of males was better than that of females.The reason for the difference may be that the number of cases in this study was more female than male.Age was consistent with previous studies, which found a statistically significant difference between age and postoperative weight loss (p < 0.001) 11 .In terms of body mass index (BMI), morbidly obese patients with body mass index (BMI) between 32.5 and 40 will have better weight loss results than those with body mass index (BMI) greater than 40, which is similar to the results of foreign studies that show that body mass index (BMI) less than 40 6,12 .In terms of perceived symptom variables, diabetes (p = 0.005) showed that patients with diabetes had lower weight loss after weight loss surgery than those without diabetes, with a statistically significant difference (p = 0.005).There was no significant relationship between the percentage of postoperative excess weight loss (% EWL) of patients with diabetes symptoms (p = 0.848) 13 .Among the perceived lifestyle habit variables, smoking habit (p = 0.002), the study results showed that patients with smoking habit had higher weight loss results after weight loss surgery than those without smoking habit.Patients with smoking habits had a significant relationship (p = 0.005) with the percentage of excess weight lost after surgery (% EWL) 13 .
The results of binary logistic regression analysis of complications showed that in hypertensive patients, the AOR of postoperative complications were 2.617 times higher in morbidly obese patients with hypertension than in patients without hypertension (p < 0.001, 95% CI 1.648-4.154),indicating that hypertension is a predictor of postoperative complications, which is similar to the results of Dutch scholars' study, showing a significant difference between hypertension and postoperative complications in patients undergoing weight loss surgery (p = 0.012, 95% CI 1.076-1.812),and similar to the results of this study 14 .The past study has shown that blood vessel walls in hypertensive patients may become more rigid and fragile, increasing the risk of cardiovascular events after surgery, and fluctuations in blood pressure may impose an additional burden on cardiac and renal function, affecting postoperative recovery 15 .In terms of diabetes, the AOR for postoperative complications in morbidly obese patients with diabetes symptoms were 2.062 times higher than those without diabetes (p = 0.008, 95% CI 1.204-3.533),indicating that diabetes is a predictor of postoperative complications.Studies by American scholars showed that patients with diabetes after weight loss surgery had 0-1 months after surgery (OR 1.3, p = 0.017, 95% CI 1.04-1.50)and 4-6 months after surgery (OR 1.8, p = 0.001, 95% CI 1.31-2.47)had a higher probability of infection, which was similar to this study 16 .The past study has shown that people with diabetes are more prone to infections and have slower wound healing due to poor blood sugar control.High blood sugar levels may also cause damage to blood vessels and affect blood circulation, which can affect recovery after surgery 17 .In terms of liver disease, the AOR for postoperative complications in morbidly obese patients with liver disease symptoms were 1.968 times higher than those without liver disease (p = 0.016, 95% CI 1.135-3.413),indicating that liver disease was a predictor of postoperative complications, which was significantly different from the results of Dutch scholars' studies, showing that there was a significant difference between the presence of liver disease and postoperative complications in weight-loss surgery patients (OR 1.396, p = 0.042, 95% CI 1.012-1.928).The previous study has shown that poor liver function affects drug metabolism and blood clotting, increasing the risk of postoperative bleeding and infection.The nutritional status of people with liver disease may also be poor, further affecting the recovery process 18 .Moreover, the Kidney disease of AOR for postoperative complications in morbidly obese patients with liver disease symptoms were 2.811 times higher than those without liver disease (p ≤ 0.001, 95% CI 1.702-4.644).The previous study has also shown that renal dysfunction can affect fluid and electrolyte balance in the body, increase the risk of cardiovascular disease and infection after surgery, and may require adjustment of medication dosages to avoid drug accumulation and toxicity 19 .Similar to the results of this study 14 , since both this study and foreign literature have found that patients suffering from hypertension, diabetes, liver disease or kidney disease may increase the incidence of postoperative complications, nursing units can be provided as the basis for preoperative evaluation.
For the perceived lifestyle variables, the AOR for postoperative complications were 3.489 times higher in patients with smoking habit symptoms than in patients without smoking habit (p < 0.001, 95% CI 2.136-5.697),indicating that smoking habit is a predictor of postoperative complications, which is statistically significant different from American scholars' studies showing that smoking habit and postoperative complications in weight-loss surgery patients (OR 1.2, p = 0.03), which is similar to the results of this study 20 .The past study indicated that smoking affects lung function and oxygenation, increasing the risk of postoperative respiratory complications such as lung infections and respiratory failure.Smoking also slows wound healing and increases the risk of infection 21 .The AOR for postoperative complications in morbidly obese patients with drinking habits were 2.770 times higher than those without drinking habits (p = 0.009, 95% CI 1.437-5.337),indicating that drinking habit is a predictor of postoperative complications, which is statistically significantly different from that of American scholars (OR 1.55, p = 0.009), which is similar to the results of this study 22 .These results found that smoking or drinking habits may increase the risk of postoperative complications.The past study has suggested that chronic alcohol use may lead to liver insufficiency and malnutrition, increase the risk of bleeding after surgery, and affect drug metabolism 23 .
Among surgery-related variables, the AOR for postoperative complications in patients with morbid obesity in terms of surgical time spent performing bariatric surgery was 1.014 (p = 0.021, 95% CI 1.002-1.026),indicating that operation time was a predictor of postoperative complications, which was statistically significant different from that of American scholars (OR 1.27 p < 0.001, 95% CI (1.19, 1.35)).Similar to the results of this study 24 , since both foreign studies and the results of this study found that the duration of surgery may increase the risk of complications, the study results can be provided to weight loss centers and clinical care related units as relevant evidence before surgery.The past study suggested that longer surgery times may mean that the procedure is more complex, more physically demanding and tolerable to the patient, and that prolonged anesthesia and surgical procedures may increase the risk of postoperative complications, including thrombosis, infection, and prolonged recovery time 25 .

Conclusions
This study found that gender, age, BMI, diabetes and smoking habits were predictors of postoperative weight loss.Hypertension, diabetes, liver disease, kidney disease, smoking, alcohol consumption, and operation time were predictors of postoperative complications.If the patient has symptoms related to these 12 significant factors, it will affect the outcome of postoperative weight loss and the occurrence of complications.To provide reference

Table 1 .
Descriptive statistical results in patients with morbid obesity (n = 663).

Table 2 .
Independent sample T-test of variables and postoperative weight loss in patients with morbid obesity (n = 663).

Table 3 .
Chi-square test of variables and postoperative complications in morbidly obese patients (n = 663).