Analysis On The Association of Intraoperative Fluid Balance and Short-Term Outcomes After Radical Gastrectomy in Aged Patients

Background: To observe the relationship between uid balance and the short-term outcomes of aged patients after gastrectomy for gastric cancer in Nanjing Drum Tower Hospital. Methods: The clinical data of patients with gastrectomy for gastric cancer from January 2016 to December 2018 in Nanjing Drum Tower Hospital were retrospectively analyzed. According to the criteria of inclusion and exclusion, 691 patients who met the study conditions were analyzed according to intraoperative uid balance recorded on patients who has undergone radical gastrectomy. Patients were classied into three uid administration groups representing incremental quartiles of the primary exposure variable. Preoperative characteristics used for statistical adjustment included gender, age, weight, admission type, ASA degree. Operative factors included procedure duration, estimated blood loss, urine output, and so on. The primary outcomes included acute kidney injury (AKI).and postoperative respiratory complications (PRCs) Exploratory outcomes included length of stay, postoperative length of stay and total cost of hospitalization. The association between perioperative factors and AKI/PRCs in hospital was tested with multivariable logistic regression analyses. Results: 16 cases were diagnosed as AKI and 23 cases were diagnosed as PRCs. The association between intraoperative uid balance and the incidence of acute kidney injury (AKI)/postoperative respiratory complications (PRCs) remained U-shaped but the difference was not statistically signicant (P>0.05). After adjustment for potential confounders, lower urine output (P=0.017, OR=0.997,95%CI=0.994-0.999) and coronary heart disease (P=0.032, OR=4.867,95%CI=1.142-20.75) were independent predictor of AKI in aged patients after radical gastrectomy. Besides, coronary heart disease(OR=3.371,95%CI=1.021-11.129,P=0.049) and intestinal obstruction (OR=12.501,95%CI=3.058-51.107,P (cid:0) 0.0005) were independent predictor of PRCs in aged patients after radical gastrectomy.

hospital was tested with multivariable logistic regression analyses.
Conclusion: There were no signi cant association between the incidence of AKI or any other complications and intraoperative uid balance during radical gastrectomy in aged patients. Lower urine output and coronary heart disease were independent predictors of AKI in aged patients after radical gastrectomy. Coronary heart disease and intestinal obstruction were independent predictors of PRCs in aged patients after radical gastrectomy.

Background
At least 310 million operations took place each year and this number is still rising nowadays [1]. Perioperative uid management, one of the key components of enhanced recovery pathways, has been widely used in surgical procedures. Optimal intraoperative uid management is important because both under and over uid balance (FB) may associated with harm [2][3][4][5]. Prior studies found that exposure to positive or negative FB was associated with long-term mortality compared with even FB [6], while another trial recently found that even FB was associated with a higher rate of acute kidney injury [7]. Therefore, the evidence for uid therapy during and immediately after surgery is still inconclusive.
Fluid therapy should only be given in well-de ned protocols according to individual needs [8,9], especially the aged [10]. Markedly increased cases of operations in fast-elderly population have posed a signi cant challenge to medical health eld. Elderly subjects have reduced ability of regulating homeostasis, decreased myocardial function and impaired pulmonary reserve function, and these may signi cantly increase susceptibility for multiple diseases after surgery. Thus, identifying an optimal strategy of uid resuscitation for elderly patients during perioperative will improve outcomes.
Therefore, aged patients undergone gastric cancer radical surgery were chosen, a common surgical operation. We conducted this retrospective study to explore the optimal strategy of uid therapy on postoperative clinical outcomes like acute renal injury (AKI), postoperative respiratory complications (PRCs) and so on. We also tried to identify the independent predictors of such complications.

Participants
All methods were carried out in accordance with relevant guidelines and regulations of the A liated Drum Tower Hospital of Nanjing University. Since it is not harmful to patients with data collected, a statement on consent waiver for this study was approved by Ethics Committee of the A liated Drum Tower Hospital of Nanjing University (Registration number: 2018-162-01). Subjects in this research were consecutively recruited from January 2016 to December 2018 and undergoing gastric cancer radical surgery. The inclusion criteria included: 1) age ≥ 65, 2) American Society of Anesthesiologists (ASA) ≤ . The exclusion criteria were as follows: 1) no renal function test was performed during the perioperative period. 2) hepatic and or kidney dysfunction. 3) vital cardiovascular dysfunction 4) preoperative coagulation dysfunction 5) severe surgery within the past year 6) incomplete medical records. A total of 691 subjects were enrolled. The subjects were divided into 3 groups according to the intraoperative liquid intake (input minus output) by percentile method: low intake group (< 25%), medium intake group(26%~75%),and high intake group (> 75%). The subjects were further divided into AKI group and non-AKI group according to whether acute renal injury (AKI) developed or not. The subjects were divided into PRCs groups and non-PRCs according to whether postoperative respiratory complications (PRCs) appeared.

Anesthesia treatment
All participants received general anesthesia according to a standardized protocol. All patients received standard monitoring, heart rate, arterial pressure, central venous pressure, PETCO2, SpO2 and body temperature were recorded continuously. All aspects of clinical care were documented in each patient's medical record.
Short-term prognosis after operation All patients were transferred to PACU or ICU after operation and experienced the process of recovery. They were transferred to the general ward when condition stable. The time of defecation was recorded to judge the recovery of intestinal function. Urine volume and serum creatinine were recorded to judge whether acute renal injury (AKI) developed or not after surgery [11]. In this study, PRCs were recorded with the de nition of pneumonia or pulmonary edema with imaging evidence, respiratory failure, and / or intubation within 3 days after surgery.

Statistical analysis
The measurement data which presented as the mean ± SD or Median (Quartiles) were statistically analyzed with One-way ANOVA, Kruakal Wallis rank sum test or t test. The numeration data were statistically analyzed with Fisher test or χ 2 test. All statistics were performed with SPSS v24.0, Empowerstats software and R software. P < 0.05 was considered to be statistically signi cant.
After adjusting for confounding variables (age, albumin level preoperative, anemia, use of Dizosin and liquid balance), Preoperative coronary disease (OR 4.867, 95% CI: 1.142-20.750, P = 0.032)and decreased urine output (OR 0.997, 95% CI: 0.994-0.999, P = 0.017) were respectively independent risk factors for the incidence of AKI in elderly patients after radical gastrectomy ( Table 2). After put the intraoperative net uid intake and intraoperative uid balance group as exposure variables and adjusted for confounding variables like age, albumin level preoperative, preoperative coronary disease and so on, no correlation has been found between the net uid input and the incidence of AKI in elderly patients after radical gastrectomy (Table 3).  (Table 4) After adjusting for confounding variables (age, hypertension and so on), the logistic regression analysis identi ed that the occurrence of coronary disease (OR 3.371, 95% CI: 1.021-11.129, P = 0.049) and digestive tract obstruction before surgery(OR 12.501, 95% CI: 3.058-51.107, P < 0.0005) remained independent predictors of the incidence of PRCs in elderly patients after radical gastrectomy. (Table 5A). Put the intraoperative net liquid intake and intraoperative uid balance group as exposure variables and adjusted for confounding variables like age, hypertension, preoperative coronary disease and so on, no links has been shown between the net uid input and the incidence of PRCs in elderly patients after radical gastrectomy (Table 5B).

Prediction of defecating time
The average defecation time was 5.79 ± 2.25 days in these 691 patients included. the univariate analysis showed that colloidal infusion volume, crystal and colloid ratio, baseline C-reactive protein level, hypoalbuminemia and use of vasoactive drugs during operation were suspected factors of in uencing postoperative defecation time (Table 6A). After adjusting for these suspected variables, we found no relationship between the net uid input and defecating time in elderly patients after radical gastrectomy (Table 6B).

Discussion
Fluid therapy serves to achieve homeostasis by restoring and maintaining body water, electrolytes and tissue perfusion [12,13]. ERAS programs recommended to avoid too much intravenous uid [14,15]. Some studies determined restrictive uid therapy to be optimal which reduced the complication rate compared with liberal uid management policy [16]. However, inappropriate uid-balance approaches may be harmful [2,17]. Our ndings showed that the incidence of AKI and PRCs in low intake group and high intake group increased when compared with medium intake group, no signi cant difference although. That is, liquid balance is not associated with postoperative clinical outcomes in aged patients undergone gastric cancer radical surgery. Myles' research found that restrictive uid regimen was not associated with increased rate of disability-free survival and was associated with increased incidence of AKI compared with a liberal uid therapy [7]. We suggest that the differences in the observed results may be due to our small sample size relatively and a single-center study.
AKI occurrence after operation could induce increased mortality, hospitalization and medical expenses [18][19]. In our trial the mortality of AKI group was 6.25% while non-AKI group was 0.3% in hospital after operation(P = 0.068). No signi cant difference was found which may due to short follow-up time.
However, noted that one third of the dead cases have been suffered from AKI. In addition, AKI group has longer hospital stays. Some studies found that body mass index, hyperlipidemia, preoperative use of ACE-I or ARB, COPD and diabetes could be the independent risk factors for AKI after postoperative [20][21][22][23]. Our ndings showed that aged, anemia, preoperative coronary disease incidence low albumin level preoperative, decreased urine output and the use of Dizosin were associated with increased frequency of AKI. By multivariable analyses, the independent risk factors for AKI were preoperative coronary disease and decreased urine output. The decrease of urine volume during operation is the external sign of renal perfusion insu ciency, which is related to the occurrence of AKI. While in this study, we did not nd that COPD, diabetes and so on were AKI Independent risk factors. We thought that the event rate of AKI was relatively low, which may have limited our ability to test for a large number of risk factors in a multivariable model.
Our study analyzed the risk for PRCs that is linked with commonly prescribed medications for comorbid conditions in subjects who undergo radical gastrectomy. The data indicate the occurrence of coronary disease and digestive tract obstruction preoperative were associated with increased odds for development of PRCs. Interestingly, Coronary heart disease preoperative turned out to be an independent risk factor both of AKI and PRCs in elderly patients after radical gastrectomy. The possible mechanism is that patients with a long history of coronary heart disease may result in ischemic cardiomyopathy, decreased cardiac contractility, and reduced cardiac output. The decrease of renal blood ow, di cult of pulmonary venous return, increased pressure in pulmonary capillaries and pulmonary interstitial edema, and this eventually developed into AKI and PRCs. Patients with digestive tract obstruction before operation are of in ammation and water electrolyte disorder in degree, which may be a potential mechanism of PRCs.

Conclusion
In conclusion, in patients at aged for complications while undergoing radical gastrectomy, uid therapy was not associated with a higher rate of AKI and PRCs. Preoperative coronary disease and decreased urine output are independently associated with increased risk for postoperative AKI. Coronary disease and digestive tract obstruction preoperative were associated with increased incidence of PRCs. Limited by its retrospective design and relatively small sample size in our study, we suggest a larger-scale prospective study be conducted in future. Availability of data and materials: We declared that materials described in the manuscript, including all relevant raw data, will be freely available to any scientist wishing to use them for non-commercial purposes, without breaching participant con dentiality. If anyone wants to request the data from this study, please contact with corresponding author.
Competing interests: The authors declare that they have no competing interests. Authors' contributions: Yin Cui has seen the original study data, reviewed the analysis of the data, and was a major contributor in writing the manuscript. Yuhui Wu designed the study, conduct the study, analyze the data. Xue Han conducted the study. Beibei Zhu analyze the data. Zhengliang Ma helped design the study and write the manuscript. Xiaoping Gu helped design the study and write the manuscript. All authors read and approved the nal manuscript. Table 5B Relationship between intraoperative net uid input and the incidence of PRCs postoperative in different models Table 6A Single factor analysis of postoperative defecation time Table 6B The relationship between the net uid input and defecating time after radical gastrectomy