Analysis of the Relationship Between Preoperative Arterial Oxygen Partial Pressure and Acute Kidney Injury after Surgery for Tetralogy of Fallot and Explore the Related Risk Factors

Background: Acute kidney injury (AKI) is a severe complication of pediatric cardiothoracic surgery (CTS). It is debatable whether patients with the low preoperative arterial partial pressure of oxygen (PaO2) are more likely to develop AKI after surgery. The study aims to investigate the incidence and possible inuencing factors of AKI in patients undergoning the radical operation of tetralogy of Fallot (TOF) with different preoperative oxygen partial pressure. Methods: In this retrospective clinical study, 36 pediatric patients who underwent CTS were enrolled in this study. The patients were divided into 4 groups according to preoperative PaO2. We examined the baseline data and outcomes of the study population among groups. Results: Of the 36 patients, 17 developed AKI. Compared with the high preoperative PaO2 group, the low preoperative PaO2 group mostly had severe AKI and persistent, but there was no signicant difference in AKI among groups (P (cid:0) 0.05). In the 48-hour continuous monitoring after surgery, the oxygen metabolism indexes (Pv-aCO2/Ca-vO2) were correlated with AKI and there were signicant differences among the groups. Conclusions: Low preoperative PaO2 does not signicantly increase the incidence of AKI and Pv-aCO2/Ca-vO2 is associated with postoperative AKI and persistent. Shapiro-Wilk normality of continuous variables with normal distribution expressed in terms of mean and standard deviation. Group comparisons were analyzed by independent sample t-tests or ANOVA test. The continuous variables with the non-normal distribution were expressed as median (P25 and P75) and were compared using non-parametric factorial Kruskal-Wallis sum-rank test. Frequencies and proportions were estimated for categorical variables and were compared using the chi-squared test or Fisher's test. Odds ratios (ORs) with 95% condence intervals (CIs) for the development of AKI in each group were calculated using a logistic regression model. Persistent AKI was dened as a continuance of AKI according to the KDIGO criteria beyond 48 h according to the consensus report of the ADQI 16 workgroup (21). Transient AKI was dened as AKI of less than 48 h duration. To investigate the relationship between preoperative PaO2 and postoperative AKI, Univariate analysis was performed on the general preoperative data, hemodynamic indexes and oxygen metabolism indexes of the 4 groups at 24 and 48 hours after operation and the incidence, severity, and duration of AKI in the 4 groups were compared. Subgroup analysis was performed to explore the relationship between oxygen metabolism index (Sa-vO2, Pa-vO2, Pv-aCO2, Ca-vO2, Pv-aCO2/Ca-vO2, O2ER) and persistent AKI. Single-factor regression analysis was performed to explore the risk factors with certain signicance in ) to arterial-to-central venous O 2 content difference (Ca-vO 2 ) ; O 2 ER = Oxygen Extraction Ratio; PAP = Pulmonary Artery Pressure; LVEF = Left Ventricular Ejection Fraction; LVFS = Left Ventricular Fraction Shortening; VIS Vasoactive

Inclusion criteria for patients were: (1) age from 1-month to 12-month; (2) TOF was con rmed by preoperative echocardiography and intraoperative ndings. TOF consists of a tetrad, or a group of 4 defects, which are ventricular septal defect, pulmonary stenosis, overriding aorta, and right ventricular hypertrophy (12); (3) the rst time to receive CTS treatment and undergo complete surgical repair. Complete surgical repair includes closing the ventricular septal defect, resecting muscle bundles within the right ventricular out ow tract with or without patch augmentation (13); (4) complete clinical information such as history and laboratory examination.
The exclusion criteria were: (1) preexisting renal dysfunction (RD) or requirement of renal replacement therapy before surgery. RD was de ned as kidney injury denoted by pathological changes, or other indicators such as abnormal blood, urine, or imaging ndings or an estimated glomerular ltration rate (eGFR) of less than 60 ml/min*1.73 m2 for more than 3 months (14); (2) a history of nephrotoxic drug use within 7 days before surgery; (3)a lack of postoperative renal data; (4) delayed sternal closure or reoperation was required due to bleeding and other reasons.
Finally, 36 patients were included in the study, of which 17 had AKI. Diagnosis and staging of AKI were performed according to the Kidney Disease: Improving Global Outcomes (KDIGO) 2012 clinical practice guidelines (15).
Patients scheduled for CTS should have preoperative tests including the complete history, transthoracic echocardiography, and blood test. Venous blood used for the preoperative blood test was collected via scalp vein needle from scalp vein or femoral vein. The radial artery catheter was needed to obtain a sample of arterial blood for gas analysis and invasive monitoring. A central venous catheter (CVC) was placed through the femoral venous or jugular vein and central venous pressure (CVP) was measured hourly.
All postoperative children were admitted to the pediatric intensive care unit (PICU) and had real-time monitoring, which included measurement of vital signs (temperature, blood pressure, pulse, and respiration rate), quanti cation of all uid intake and output. Blood used for routine postoperative test were collected from indwelling arterial lines after CTS, and simultaneous blood samples were obtained from a central venous catheter and an arterial catheter at 48 hours after surgery. Echocardiographic reexamination was performed on the 7th day after the operation. Postoperative management, systematic monitor and programmatic therapy were performed based on Handbook of Pediatric Cardiac Surgical Intensive Care (16).
Baseline data including demographics, clinical manifestation, and laboratory data were extracted and analyzed.
Preoperative data included weight, height, age, sex, serum creatinine (SCr), eGFR, arterial oxygen pressure (PaO2), Hct, hemoglobin (Hb), McGoon ratio, Nakata index, pulmonary arterial pressure (PAP), left ventricular ejection fraction (LVEF) and fractional shortening (LVFS); surgical characteristics were CTS time, cardiopulmonary bypass (CPB) time and American Society of Anesthesiologists grade (ASA). Postoperative data including temperature, heart rate (HR), mean arterial pressure (MAP), central venous pressure (CVP), lactic acid, arterial oxygen saturation (SaO2), PaO2 and urine output were recorded up to the rst and second 24 hours after admission to PICU. Simultaneous monitoring of venous blood gas at 48 hours after CTS. The highest vasoactive inotrope score (VIS) within postoperative 48 hours was used to re ect the application of vasoactive drugs after CTS (17). Lengths of PICU and hospital stay were also recorded. To better show the oxygen metabolism in tissues, we calculated central venous-to-arterial carbon dioxide difference (Pv-aCO2), central arterial-to-venous oxygen saturation difference (Sa-vO2), central arterial-to-venous oxygen pressure difference (Pa-vO2), arterial-tovenous oxygen content difference (Ca-vO2), the rate of Pv-aCO2/Ca-vO2, and oxygen extraction ratio (O2 ER) according to the following formulas (18)(19)(20): · CaO2 = (1.34 × SaO2 × Hb) + (0.003 × PaO2) · CcvO2 = (1.34 × ScvO2 × Hb) + (0.003 × PcvO2) Shapiro-Wilk test was used to check the normality of the data; continuous variables with normal distribution were expressed in terms of mean and standard deviation. Group comparisons were analyzed by independent sample ttests or ANOVA test. The continuous variables with the non-normal distribution were expressed as median (P25 and P75) and were compared using non-parametric factorial Kruskal-Wallis sum-rank test. Frequencies and proportions were estimated for categorical variables and were compared using the chi-squared test or Fisher's test. Odds ratios (ORs) with 95% con dence intervals (CIs) for the development of AKI in each group were calculated using a logistic regression model. Persistent AKI was de ned as a continuance of AKI according to the KDIGO criteria beyond 48 h according to the consensus report of the ADQI 16 workgroup (21). Transient AKI was de ned as AKI of less than 48 h duration. To investigate the relationship between preoperative PaO2 and postoperative AKI, Univariate analysis was performed on the general preoperative data, hemodynamic indexes and oxygen metabolism indexes of the 4 groups at 24 and 48 hours after operation and the incidence, severity, and duration of AKI in the 4 groups were compared. Subgroup analysis was performed to explore the relationship between oxygen metabolism index (Sa-vO2, Pa-vO2, Pv-aCO2, Ca-vO2, Pv-aCO2/Ca-vO2, O2ER) and persistent AKI. Single-factor regression analysis was performed to explore the risk factors with certain signi cance in univariate analysis (P < 0.05) of postoperative AKI and persistent AKI. All statistical analyses were performed with SPSS version 23 (IBM Corp. Released (2015) IBM SPSS Statistics for Windows. IBM Corp., Armonk, NY). The difference was considered signi cant when the two-tailed P-value was less than 0.05.

The changes of hemodynamic indexes and oxygen metabolism indexes of the 4 groups at PICU, 24 hours and 48
hours after CTS were shown in Table 3  The total incidence of acute kidney injury was 47.2% (n = 36) and the incidence of AKI was signi cantly lower in group 1 (n = 1, 2.8%). Most patients developed postoperative AKI in the early period after CTS (n = 11, 64%), and patients with low PaO2 (group 3 and 4) tended to have persistent AKI (n = 8, 80%), however it was not statistically signi cant (P > 0.05).
Subgroup analysis was performed based on whether the patients were persistent or short AKI which was shown in Table 4. Patients with persistent AKI had signi cantly higher Pv-aCO2, Ca-vO2, Pv-aCO2/Ca-vO2 and O2ER than those with short AKI (P > 0.05). Even after grouping by PaO2, this phenomenon persisted and those values of group 3 and 4 were higher than group 1 and 2. Predictors of AKI and persistent AKI in single-factor logistic regression were shown in Fig. 2 and Fig. 3, respectively. Single-factor logistic regression analysis revealed a signi cant correlation between the occurrence of an AKI with age, height, Hb, Hct, PICU stay days, total hospitalization expenses, VIS, Pv-aCO2 and Pv-aCO2/Ca-vO2, while only Pv-aCO2/Ca-vO2 showed a signi cant correlation with persistent AKI.

Discussion
AKI is one of the common and serious complications after CPB operation, which has a negative impact on the prognosis of patients and increases the fatality rate (21). Most studies believe that chronic hypoxia has a great effect on the organism, especially for the kidney and even put forward the concept of cyanotic nephropathy (CN) (22). Studies have indicated that CN occurs in about 30-50% of the patients with C-CHD. The incidence of AKI after CTS was as high as 47.2% in this study and 10 patients had persistent AKI according to the de nition.
Hence it is important to identify patients who are at risk of developing AKI after CTS. Preoperative symptoms, surgical interventions, and postoperative management are related to the postoperative AKI. Previous studies have clearly demonstrated that prolonged CTS time and CPB time were associated with the development of AKI (23)(24), however, it was not the case in our study. We considered that this is because the advances in surgical techniques and perioperative management that the CTS time (range, 99 to 282 minutes) and CPB time (range, 40 to 127 minutes) were controlled within a relatively safe range and had led to dramatic improvements in survival outcomes for children with TOF. We also analyzed the hemodynamic indicators and oxygen metabolism indicators at the time of PICU admission, postoperative 24 h and 48 h. The arterial blood and venous blood gas analysis were measured simultaneously after the 48 hours CTS, and oxygen transport data (Sa − vO2, Pa − vO2, Ca − vO2, Pv-aCO2, O2 ER and Pv-aCO2 /Ca − vO2) were calculated. Due to the lack of echocardiographic data in the study population at 48 hours after surgery, we were unable to directly calculate oxygen delivery and oxygen consumption, which was a major aw of this study.
MAP, CVP and SO2 are commonly used as the indices of adequate tissue perfusion and oxygenation in patients, but tissue hypoxia may still exist even when these indices are within the normal range (25). And in this study, normalization of SO2 and PO2 does not exclude the persistence of tissue hypoperfusion and tissue hypoxia.
Some studies have found out that Sa-vO2 is an excellent indicator of adequacy of cellular oxygenation, and it has shown a good correlation with oxygen delivery (26); Ca − vO2 can to some extent re ect poor tissue perfusion (27); In this study, Pv-aCO2 /Ca − vO2 and Pv-aCO2 were independent risk factors for acute kidney injury, and Pv-aCO2 /Ca − vO2 was associated with persistent AKI. These results are consistent with previous studies. Mekontso-Dessap A et al. found that the Pv-aCO2 /Ca − vO2 ratio could detect global anaerobic metabolism in critically ill patients (28). As an index of the presence of anaerobic metabolism, its basic theory is: during anaerobic metabolism carbon dioxide production decreases less than oxygen consumption. The present study showed that a high Pv-aCO2/Ca-vO2 ratio was associated with low survival rates at day 28 (P < 0.005) (29). Mukai A et al. found that Pv-aCO2 and Pv-aCO2 /Ca-vO2 at the end of surgery had a superior ability for predicting postoperative complications (30) The patients in this study were divided into 4 groups according to preoperative PaO2 and patients with low PaO2 had higher Hb and Hct due to chronic hypoxia. Multiple studies have shown that high Hb and Hct are associated with low cardiac output, increased blood viscosity, increased platelet aggregation, and impaired blood ow (31).
Although there was no signi cant difference between the Hb and Hct among groups in this study, patients with higher Hb and Hct had longer CTS time and CPB times, longer postoperative PICU stays, higher VIS scores, and higher expenditures. However, there was no signi cant difference in the incidence of postoperative AKI among the 4 groups, several hypotheses might be used to justify this observed discrepancy between lung tissue and systemic responses: (1) following CTS, all patients were admitted to the PICU and received routine mechanical ventilation, vasoactive drugs, moderate rehydration, therefore, the mature surgical technique and rational postoperative management strategies were closely associated with the decreased of the incidence of AKI.
(2) the short follow-up time and the small number of patients involved, it was necessary to extend the follow-up time and increase the number of patients included. Overall, patients in group 1 had less kidney damage among the groups, and patients in group 3 and 4 were more likely to have severe AKI and persistent AKI.
During chronic hypoxia, increases in red blood cell number or formation of new vasculature occur, adapting an organism to decreased oxygen conditions. In this study, we found that there was not a signi cant increase in postoperative AKI of patients with low PaO2 (group 3 and 4) during the early period after CTS. We hypothesized that it might be related to these adapt changes. The animal study suggested that hypoxic preconditioning of animals in vivo increased hypoxic tolerance (32). Alexander Zarbock et al. found that among high-risk patients undergoing cardiac surgery, remote ischemic preconditioning compared with no ischemic preconditioning signi cantly reduced the rate of AKI and use of renal replacement therapy (33).
The limitation of this study includes the following aspects. First, the relatively small sample may have lacked the power to detect signi cant interactions. Second, Hemodynamic indexes which the study collected cannot directly represent renal perfusion. Third, the evaluation criteria of microcirculation and oxygen metabolism are still controversial at present, and the indexes we adopted in this paper need to be further veri ed.

Conclusions
In conclusion, low preoperative PaO2 did not signi cantly increase the incidence of AKI, but patients with low PaO2 were more likely to have severe AKI and persistent AKI. Pv-aCO2/Ca-vO2 and Pv-aCO2 are independent risk factors for postoperative AKI and Pv-aCO2/Ca-vO2 is an independent risk factor for persistent AKI. Monitoring from their parent or guardian. This permits the researchers to engage in research, the use of identi able biospecimens and identi able data during the peri-operative period and future follow-up without the requirement to obtain additional consent for the future storage, maintenance, or research uses, so long as the future activities are within the scope of the broad consent. The study protocols as well as the application form were fully reviewed, and we certify that this study did not raise any issues of patient risk or cause any harm to patients. We also certify that the study was strictly in accordance with the Declaration of Helsinki and International Ethical Guidelines for Health-related Research Involving Humans. This study was approved by the Ethics Committee of Xin Hua Hospital A liated to Shanghai Jiao Tong University School of Medicine (Approval No. XHEC-D-2020-016).

Consent for publication
Written informed consent for publication was obtained from all participants

Availability of data and materials
The data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare that they have no con icts of interest.

Funding
This research did not receive any speci c grant from funding agencies in the public, commercial, or not-for-pro t sectors.
Authors' Contributions YX contributed towards the study design, data collection and writing of the manuscript; YZ contributed towards the study design and literature search; ZL contributed towards the data collection and statistical analyses; LX contributed towards the statistical analyses; LC contributed towards data interpretation; WX contributed towards data interpretation and XZ contributed towards data interpretation and writing of the manuscript. All authors approved the nal version of the manuscript. Figure 1 Flow diagram of the study participants TOF= Tetralogy of Fallot; CTS=cardiothoracic surgery; AKI=acute kidney injury; KDIGO=Kidney Disease: Improving Global Outcomes