Perioperative Diabetes Insipidus in Patients Undergoing Off-Pump Coronary Bypass Grafting Surgery: A Study on Incidence, Related Factors and Clinical Signicance

This was a retrospective study of 199 patients who underwent OPCABG surgery in Anzhen hospital, Beijing, China, between January 2019 to December 2019. Patients were divided into a DI(+) group and a DI(-) group according to perioperative urine condition. The incidence of perioperative DI in patients undergoing OPCABG surgery was calculated as the main outcome. Multivariable binary logistic regression analysis was used to identify independent prognostic factors of DI.


Abstract Background
Perioperative diabetes insipidus (DI) is a serious complication occurring in patients undergoing off-pump coronary artery bypass grafting (OPCABG).

Methods
This was a retrospective study of 199 patients who underwent OPCABG surgery in Anzhen hospital, Beijing, China, between January 2019 to December 2019. Patients were divided into a DI(+) group and a DI(-) group according to perioperative urine condition. The incidence of perioperative DI in patients undergoing OPCABG surgery was calculated as the main outcome. Multivariable binary logistic regression analysis was used to identify independent prognostic factors of DI.

Conclusion
Coronary artery disease(CAD) patients underwent OPCABG surgery were easy to have perioperative DI, which had a higher ICU stay and mechanical ventilation time than those without DI. BMI, Crystal quantity, perioperative creatinine was associated with the presence of DI. Prospective studies are needed to validate these ndings.

Background
Off-pump coronary bypass grafting (OPCABG) surgery is an important therapy to treat CAD, the rst pioneer of OPCABG surgery was Kolessov in the 1960s, the proportion of this treatment to treat CAD patients has been increasing in recent years 1 . In OPCABG surgery, the operation may cause signi cant uctuations in the patients' hemodynamics. Maintaining blood pressure stability is essential in the perioperative of OPCABG surgery, therefore, it is very important to maintain a stable internal environment and uid balance during the perioperative period 2 . Diabetes insipidus (DI) often occurs in OPCABG surgery, excessive urine production can exacerbate circulatory uctuations. However, we often ignore the emergence of DI and risk factors for poor outcomes in patients undergoing OPCABG surgery. The role of perioperative DI in determining post-operative condition has not been evaluated, while there are seldom researches focus on the related factor of DI in patients undergoing OPCABG. In this research we detect the risk factors of DI in patients undergoing OPCABG surgery and the outcomes of these patients.

Patients
All patients diagnosed with CAD were included in this study. Inclusion criteria were (i) age 18-80 years and (ii) surgical treatment of OPCABG (iii) BMI 18-30 Kg/m 2 . Exclusion criteria were (i) Combined with heart valve disease and requires valve replacement or repair (ii) Combined with heart failure, renal failure, uremia, primary aldosteronism (iii) surgical method was changed intraoperative.

Grouping
We de ned intraoperative DI(+) as the speed of urine was greater than 2 ml/kg/h intraoperatively, postoperative DI(+) as the total amount of urine was greater than 3000 ml postoperatively within 24 hours. The DI (+) group of patients had both intraoperative DI(+) and postoperative DI(+), The DI (-) group of patients had both intraoperative DI(-) and postoperative DI(-), intraoperative DI(+) and postoperative DI(-) or intraoperative DI(-) and postoperative DI(+).

Data collection
Information regarding patient gender, age, history of hypertension (yes/no), type 2 diabetes (yes/no), body mass index (BMI), Blood routine (including hemachrome, white blood cells, platelet), Serum biochemistry (including AST, Creatinine, Na + , K + ) Intraoperative condition of the lowest temperature, maximal blood pressure, total infusion quantity, crystal quantity, colloidal quantity, blood loss, operating time, perioperative total urine volume were collected from the database. The speed of urine intraoperatively was calculated by total urine(ml)/body weight(Kg)/operating time(h).
Recovery condition of mechanical ventilation time, ICU stay, hospital stay was collected from the database Serologic examination. Routine blood tests and biochemical, total volume of urine were monitored till 24 hours after surgery.

Anesthetic management
Patient vital signs were regularly monitored after hospital admission. All patients received standard anesthesia including 10 mg morphine i.v. before entering the operating room, and 5-lead ECG, pulse oximetry (SpO 2 ), blood pressure, central venous pressure (CVP) and bispectral index (BIS) monitoring.
Primary outcome and secondary outcome Our primary outcome was the incidence of perioperative DI in patients undergoing OPCABG surgery to treat CAD. Our secondary outcomes including measurement of factors related to perioperative DI, extubation time, ICU duration, and activity of daily living scale score.
Statistical analysis SPSS Statistics Desktop (version 21.0.0 for Mac OS, IBM, Armonk, NY, USA) was used for statistical analysis. Mean ± standard deviation was used to express continuous data, and frequencies were used to express categorical data. Normally distributed continuous variables were compared using a two-tailed Student's t-test. Wilcoxon rank sum testing was used for inter-group comparisons when parametric data were not normally distributed. χ 2 testing was performed to compare categorical variables. Multivariable binary logistic regression analysis was used to identify independent prognostic factors. A P value less than 0.05 indicated a signi cant difference. The sample size was calculated using a 1-sample, 1-sided test with a power of 0.99 and α < 0.05. Formulas used are shown in the appendix.

Results
Baseline characteristics 247 of these patients planned to receive CABG and 234 patients planned to receive OPCABG. 35 additional patients were excluded (21 patients changed the operation style and received CABG which was planned to receive OPCABG, 5 patients was diagnosed chronic renal failure, 9 patients missing data in the case database). One hundred and ninety-nine cases were evaluable for this study(Graph 1.).
The result of primary outcome and secondary outcome Eighty-six (43.2%) patients were DI (+). In all these patients, one hundred and twenty-six (63.3%) were male, their average age was 62.5 years. Their average speed of urine was 5.6 ± 3.3 ml/kg/h. intraoperatively, the average total amount of urine was 3373 ± 1695 ml postoperatively within 24 hours in the ICU. In the DI(+) group, the average speed of urine was 7.0 ± 3.4 intraoperatively, the average total amount of urine was 4708 ± 1298 ml postoperatively. In the DI(-) group, the average speed of urine was 4.5 ± 2.7 ml/kg/h intraoperatively, the average total amount of urine was 2348 ± 1176 ml postoperatively. The percentage of imbalance postoperative PH condition (P < 0.05), duration of post-operative mechanical ventilation (P < 0.01), ICU stay (P = 0.04) was longer in DI (+) patients, compared to that of DI (-) patients (Table 1).  The occurrence of DI intraoperatively and postoperatively The occurrence of DI was very high in the patients undergoing OPCABG surgery, during the period of operation, 175 patients' speed of urine more than 2 ml/kg/h (87.9%, B + C in Graph 2.), during the period of ICU, the number of postoperative amount of urine more than 3000 ml was 125 (62.8, C + D in Graph 2.). The number of patients without DI was 14 (7.0%, A in Graph 2.)

Discussion
OPCABG surgery is an important therapy to treat CAD, as OPCABG surgery may improve long-term outcomes by reducing the rates of perioperative myocardial injury, stroke, and cardiac-related mortality, more and more patients receive this therapy 3 . However, OPCABG surgery is also known to cause a number of complications, including uid, electrolyte, and acid-based imbalances 3 . The changes of these internal environment could cause DI which is a condition that the kidney are unable to conserve water 4 . DI is either due to de cient secretion of arginine vasopressin (AVP/ADH) (central) or to tubular unresponsiveness (nephrogenic). It has a high mortality and carries severe morbidity. DI after OPCABG surgery can lead to increased medication requirements. Predicting which patients are at high risks for developing DI can help direct services to ensure adequate care and follow-up.
The objective of this study was to retrospectively review our institution's data on patients undergoing OPCABG surgery and determine which clinical/laboratory variables are associated with DI in this patient population.
In our study, the occurrence of DI in the patients underwent OPCABG was very common, the percentage of patients suffered from DI perioperatively was 43.2%, the percentage of DI intraoperatively and postoperatively was 87.9% and 62.8%.
DI morbidity and mortality is mostly dependent on the electrolyte imbalance it produces, osmotic disturbances, acid base changes and the effect that has on end organ function. As it has shown in our results, the DI(+) group had a higher imbalanced postoperative PH condition than DI(-) group. We highly considered that the occurrence of DI was central DI due to the touch to aortic arch and left atrium of the heart and stimulating to the osmotic pressure receptor. Although the creatinine concentration was higher in the DI (-) group, both the concentration of these two group were in normal condition. After the occurrence of DI, uid imbalance was further aggravated in patients with a low BMI compared to those with a high BMI 5  DI is caused by insu cient production and secretion of antidiuretic hormone (ADH), or the inability of the kidney tubules to respond to ADH. ADH is a key factor to adjust kidney to conserve water. Secretion of ADH is primarily regulated by plasma osmolarity, but other factors such as left atrial distention, circulating blood volume, exercise, and certain emotional states can also alter ADH releasing 6 .
The release of ADH directly acts on the kidney. Non-osmotic stimulus of ADH releasing is mediated through volume receptors located in the left atrium of the heart, aortic arch, and carotid artery. Osmotic stimulus of ADH release is mediated through osmoreceptors in the hypothalamus 7 . What's more, ADH were found to de ciency developed in patients undergoing aortocoronary bypass operations in some studies 8,9 The reason of DI in patients undergoing OPCABG may associated with both non-osmotic and osmotic stimulus of ADH. For non-osmotic of ADH, the operative procedures of OPCABG including changing the position of the heart always lead to touch to aortic arch and left atrium of the heart. For osmotic stimulus, Cerebral hypoperfusion can occur from intraoperative hypotension and declined cardiac output, Embolic strokes associated with the operation are predominantly attributable to thromboembolism and aeroembolism, these factors may restrict the function of hypothalamus 10 .
For CAD patients underwent OPCABG surgery, the use of propofol and fentanyl can in uence blood pressure, and the operator's operation on the heart also aggravates the di culty of hemodynamic maintenance. The use of noradrenaline is very common in OPCABG surgery to maintain blood pressure. Adrenal hormone inhibits the synthesis and secretion of AVP, thus adrenal insu ciency can result in increased levels of AVP, which are then lowered following initiation of glucocorticoid replacement.
There were several limitations to this study. This was a retrospective study that needs independent validation. Patients evaluated here underwent OPCABG surgery by different operator and anesthetist, it is hard to avoid the in uence caused by the operator and anesthetist. The use of drugs during perioperative period was not recorded very accuracy, there may be some factors of drugs associated with DI.
The occurrence of DI in patients undergoing OPCABG surgery is very common, prospective studies are needed to validate these ndings and we need to focus more on this point.

Conclusion
About 43.2% of the CAD patients underwent OPCABG surgery had DI. Their recovery from surgery was slower than that of patients without DI. BMI, Crystal quantity, perioperative creatinine was associated with the presence of perioperative DI. Prospective studies are needed to validate these ndings Abbreviations Figure 1 Patients selection of perioperative diabetes insipidus in patients undergoing OPCABG surgery.

Figure 2
The occurence of DI intraoperatively and postoperatively.