Analysis of factors contributing to postoperative body weight change in patients with gastric cancer: based on generalized estimation equation

Study design

Prospective longitudinal experimental design was used to identify factors contributing to body weight change during three months after gastrectomy in GC patients. Ethics approval of this study was obtained from Ruijin Hospital Ethics Committee, Shanghai Jiao Tong University School of Medicine (Shanghai, China).

Participant recruitment

GC Patients undergoing gastrectomy between January to March 2019 were identified in gastrointestinal surgery department prospectively. Volunteers meeting the following criteria were recruited: (1) diagnosed with GC; (2) scheduled for open gastrectomy; (3) aged 18 or above; (4) informed and agreed to participate in this study. Participates were excluded when they met condition below: (1) combined with other malignancies besides GC except distant metastasis; (2) palliative treatment; (3) recurrence within residual stomach; (4) patients with tube feeding after discharge. A written informed consent was obtained from all patients enrolled in the study. The sample size was estimated by G*power software (version 3.1) with three repeated measures, within factors design, a power of 0.80, significant level of 0.05, and effect size of 0.30.A sample size of 43 was required for analyzing body weight change during three months after gastrectomy by considering the attrition rate of 10%.

Data collection

Volunteers were asked to complete a demographic questionnaire and a Digestive Cancer Patients Nutrition Knowledge, Attitude and Practice Questionnaire (DCNKAPQ). DCNKAPQ was designed to assess digestive cancer patients’ nutrition knowledge, nutrition attitude, and nutrition practice with Cronbach’s alpha 0.822 (Jing, Wei-li & Xin-qiong, 2016). Specifically, nutrition knowledge dimension contains 17 questions, one point would be taken when the question was given the right answer, otherwise 0. Nutrition attitude dimension includes 5 items, which are scored on a 0-4 response scale, with answers ranging from “strongly disagree” to “strongly agree”. Nutrition practice dimension covers 8 items, which are scored on a 0-4 response scale, with answers ranging from “never” to “always”. Total scores of DCNKAPQ range from 0 to 69, with score 0-17 from nutrition knowledge, score 0-20 from nutrition attitude, and 0-32 from nutrition practice. The demographic information included age, height, body weight, gender, marriage status, educational level, and religious faith.

Data related with disease and treatment variables were collected from electronic patient record, including cancer stage, cancer treatment, postoperative complication, nutrition support at home, and comorbidity. The tumor staging was evaluated using the 7th edition tumor-node-metastasis (TNM). The Charlson comorbidity index (CCI) was applied to represent patient’s comorbidity severity. Categories of preoperative BMI were based on Asian standards proposed by WHO obesity expert advisory group in 2002, where BMI < 18.5 kg/m², BMI between 18.5 kg/m² to 22.99 kg/m², BMI between 23 kg/m² to 24.99 kg/m², and BMI ≥ 25 kg/m² indicate underweight, normal weight, overweight, and obesity, respectively. Body weight were collected at admission (T0), one month (T1), two month (T2), and three month (T3) after surgery. The body weight at four time points were reported by the patient self. The patients were required to wear the same outfit without accessories every time they measured the body weight for technical consistency. The percentage of body weight loss at each point was defined as follows: (preoperative body weight –body weight at each point after surgery) *100% / preoperative body weight.

Statistical analysis

Data analyses were performed by IBM SPSS 20.0. Measurement data meeting normal distribution were expressed as mean ± standard deviation (SD), otherwise median combined with 25% quartiles and 75% quartiles. Categorical data were showed in the form of frequency or percentage. Generalized estimating equations (GEEs) were used to analyze the univariate and multivariable influence of demographic variables and clinical variables on body weight change in GC patients over time. Variables with at least a significant of P value less than 0.1 on univariate analysis (all characteristic, disease-, treatment-, nutritional variables) went on to be included in multivariate analysis. Quasi-likelihood under the independence model criterion (QIC) value is an index to evaluate whether working correlation matrix is appropriate for GEE. The lower the QIC value is, the fitter the GEE parameter estimation is. All statistics were based on two-sided test with level of significance set at 0.05.

Baseline characteristics of the study population

Eighty-two potentially eligible patients were available, but 27 patients did not meet the eligibility criteria, and 6 patients did not participate in follow up from T1 to T3. A total of 49 patients were involved in the study, including 1 patient lost at T1, but recovered at T2 and T3. There was no case of relapse observed during the follow-up period. The median age of them was 57 years old (range 29-75 years old), the median preoperative body weight of male was 69.25 (62.50, 72.50) kg in male GC patients and 57.00 (51.00, 61.00) kg in female GC patients. The majority of them were married (91.80%), without specific religious faith (85.71%), and with a stage III or IV cancer (95.92%). There were 13 cases with postoperative complications, including 6 cases with infection, 3 cases with delayed gastric empty, 1 case with diarrhea, 1 with bleeding, 1 with duodenal stump fistula, and 1 with deep vein thrombosis. The median value of CCI, nutrition knowledge, attitude, and practice score were 5.00 (3.00, 6.00) points, 12.00 (8.00, 13.00) points,15.00 (14.00, 18.00) points, and 21.00 (18.00, 23.00) points, respectively. Most of them were treated with neoadjuvant chemotherapy combined with gastrectomy (81.63%), where the chemotherapy was mainly based on paclitaxel (75.00%). Around half of the subjects underwent subtotal gastrectomy (51.02%). After discharge, 57.14% patients took oral nutrition supplements (ONS) during the follow up (Tables 1 and 2)

Body weight change before and during three months after gastrectomy

The median body weight loss in GC patient at T1, T2, and T3 was 9.85% (7.41%, 11.76%), 12.14% (8.70%, 12.14%), and 11.63% (8.51%, 14.40%), respectively, for the whole cohort (Fig. 1). The GEE results showed a significant effect of time (Wald χ² = 333.917, P < 0.001), indicating a decrease in body weight over time. According to the GEE parameter estimation, patient’s body weight decreased by 9.2% at T1 (Wald χ² = 271.173, P < 0.001), 11.0% at T2 (Wald χ² = 277.267, P < 0.001), and 11.4% at T3 compared to baseline at T0 (Wald χ² = 284.076, P < 0.001) (Table 3).

Factors contributed to body weight change during three months after gastrectomy

Each study variable entered GEE separately for univariate analysis. The results of the GEEs for univariate analysis showed that total gastrectomy patients had greater body weight loss compared with distal gastrectomy patients (β = 3.1%, P = 0.003), patients with postoperative complication was associated with greater body weight loss than patients without postoperative complication (β = 3.0%, P = 0.010), BMI ≥ 25 kg/m² was associated with greater body weight loss compared with BMI < 18.5 kg/m² (β = 3.9%, P = 0.080), and higher nutrition practice score was associated with greater weight loss (β = 0.3%, P = 0.070) (Table 4). Therefore, those variables with P value less than 0.1 (operation type, preoperative BMI, post-operative complication, and nutrition practice score) were entered into GEE to do multivariable analysis. The results of GEE multivariable analysis indicated that surgery type and preoperative BMI were contributing factors of body weight change, while other variables, including postoperative complication and nutrition practice were not relevant. Compared with distal gastrectomy patients, total gastrectomy patients experienced greater body weight loss (β = 2.8%, P = 0.014). Compared with patient with BMI<18.5 kg/m², patients with BMI ≥ 25 kg/m² experienced greater body weight loss (β = 4.5% P = 0.026) (Table 5).

Body weight change after gastrectomy

GC Patients often experience body weight loss after gastrectomy, which involves complicated mechanism and a lot of factors, including remnant stomach, malabsorption, hormonal changes, and life style (Aoyama, 2019). Some scholars deemed that subtotal or total excision of gastric tissue negatively influence digestion and absorption of food inevitably. Also, operation itself may result in anorexia, thereby leading to body weight loss (Kim et al., 2019). Eom et al.( 2018) found that postoperative body weight loss and anemia might originate from altered absorptive function and metabolic change after gastrectomy rather than decreased nutrient intake. Aoyama et al. (2016) revealed that mean body weight loss at 1 month reached 3.4 kg (5.90% of body weight) compared with preoperative mean body weight in Japanese subjects . The team of Aoyama et al. also found that mean body weight loss value at 1 month and 3 months after surgery were 6.5% and 9.0% respectively in the elderly, while 6.0% and 8.1% respectively in the non-elderly from the preoperative value, but the difference of weight loss over time between elderly and non-elderly was not significant (Aoyama et al., 2019). In China, researchers found that mean body weight loss up to 5.8–8.8 kg at 1 month (Shuai et al., 2019), and 6.33 kg at 3 months (Tang et al., 2018) after gastrectomy compared with preoperative body weight. Our results show that the percentage of body weight loss up reached 9.2% (5.67 kg) at 1 month, 11.0% (6.83 kg) at 2 months, and 11.4% (7.14 kg) at 3 months after surgery from preoperative mean body weight. The higher body weight loss at three months in our study is possibly related the fact that most of the enrolled patients were advanced staged and underwent extensive chemotherapy.

The variation of postoperative body weight loss in different studies indicated that postoperative body weight was affected by a lot of factors. Aoyama et al. (2016) found that body weight loss during the first week after surgery was significantly greater than that during subsequent 3 weeks, and loss of lean body mass accounted for a significant part of body weight loss during the first week due to surgical stress . Abdiev et al. (2011) discovered that both fat and muscle mass reduction accounted for body weight loss during the first month after gastrectomy, and fat mass loss followed was responsible for weight loss after that period . Therefore, since energy deficiency and fat mass consumption rather than protein consumption are more prominent after discharge, continuous oral intake and exercise intervention should be given to GC patients after discharge to ensure the sufficient energy intake. In addition, body weight loss at 2 months was not significantly different from at 3 months after gastrectomy in this study, indicating that clinician should pay attention to the underlying causes of weight loss and direct nutrition intervention before 2 months after gastrectomy as soon as possible.

Factors contributed to body weight change after gastrectomy

The results of GEE showed significant main effect of preoperative BMI categories on body weight change after gastrectomy. Same as the findings in previous study (Aoyama et al., 2019; Shuai et al., 2019; (Tang et al., 2018)), we found that the percentage of body weight loss showed a downward trend as a whole within 3 months after gastrectomy, and the percentage of body weight loss within the first month postoperative period was significantly more than that in the third month postoperative. Similar to the findings in the study of Tanabe et al. who found that higher preoperative BMI, was independent predictors of greater body weight loss at the first one year mark after gastrectomy (Tanabe et al., 2017). We found that BMI ≥ 25 kg/m² was risk factors to body weight loss during three months after gastrectomy in GC patients. However, the result was contract with the study by Park et al. (2018) who found that preoperative BMI < 23 kg/m ²were associated with severe weight loss after gastrectomy . Similar to previous studies (Davis et al., 2016; Nishigori et al., 2017; Segami et al., 2018), we found that total gastrectomy was a risk factor for postoperative weight loss. Previous study observed that decreased calorie intake after total gastrectomy compared with distal gastrectomy (Noguchi et al., 1992), which may explain total gastrectomy patients experienced more body weight loss.

Similar to the findings in the study of Davis et al. (2016), we found that postoperative complication was not found to be independent factor on postoperative body weight loss in this study. While Segami et al. found that surgical complication was a significant risk factor for severe body weight loss the first month after gastrectomy (Segami et al., 2018). The conflict results prompt the complexity of weight loss in GC patients after gastrectomy. As for higher nutrition knowledge, attitude, and practice score were positively related with higher self-efficacy (Jing, Wei-li & Xin-qiong, 2016), which is a vital theory to implement nutrition education, less weight loss was expected to appear in patients with higher nutrition knowledge, attitude, and practice score in this study. However, there was no significant effect of nutrition knowledge, attitude, and practice on postoperative body weight loss. Consistent with some scholars’ findings (Aoyama et al., 2019; Davis et al., 2016), our study verified that age was not a risk factor for postoperative body weight loss during 3 months after gastrectomy. Recently, Aoyama et al. found that postoperative lean body mass decreased significantly in elderly compared with non-elderly rather than body weight (Aoyama et al., 2019).

There were some limitations in this study. Firstly, this was a single center study, 95.92% patients involved in our study were with cancer stage III or IV GC. So, the results that total gastrectomy and preoperative BMI were factors contributing to body weight change may not generalizable to patients with stage I or II. Secondly, we did not collect data about patient nutritional intake amount during 3 months after gastrectomy. So, the nutrition therapy factors on body weight change may be ignored in our study. Thirdly, the body weight at each point was reported by the patient self.