Acute kidney injury (AKI) is a common postoperative complication for children undergoing congenital heart surgery with cardiopulmonary bypass (CPB), accounting for 30% to 50% of them [1,2]. AKI is independently associated with elevated mortality and morbidity in pediatric patients following cardiac surgery, including prolonged duration of mechanical ventilation and longer ICU stay [3,4]. Unfortunately, the specific mechanism of AKI currently remains unclear, and its occurrence may be a result of a combination of multiple risk factors.

Fluctuation of blood glucose levels occurs frequently in cardiac surgery, especially with CPB [5]. Some studies demonstrated that AKI occurrence was associated with intra-, postoperative hyperglycemia and dramatic blood glucose variability in adults undergoing cardiac surgery [6,7]. Additionally, strict perioperative blood glucose control could significantly reduce the incidence of postoperative renal impairment and failure in adult cardiac surgery patients [8]. Concerning pediatric patients having cardiac surgery, especially with CPB, they are prone to varieties of blood glucose levels intraoperatively. Unexpectedly, studies as to the harm from intraoperative blood glucose levels or glucose variability are inconclusive [9–12]. Meanwhile, tight blood glucose control might not be beneficial to a decreased incidence of AKI [13]. So we conducted this study to further explore the effect of intraoperative mean blood glucose levels and variability on AKI in children following their congenital cardiac surgery.

As shown in Fig. 1, from April 01, 2022, and July 30, 2022, we screened 688 eligible patients from 791 patients, including 196 cases and 492 controls. After propensity score matching by age, sex, and baseline SCr, 140 cases and 140 controls were paired.

Table 1 demonstrated the univariate analyses of demographic data and potential risk factors by unmatched and 1:1 matched method. In the entire study population, the median age was 25.7 months and 49.6% were male patients. We selected the patients undergoing all kinds of congenital heart surgeries, and 48.3% (332/688) of the patients experienced high-risk surgery (RACHS-1 ≥ 3) [18]. The overall incidence of AKI was 28.5% (196/688), and patients who were categorized into AKI stages 1, 2, and 3 were 19.9% (137/688), 6.1% (42/688), and 2.5% (17/688) separately (data are given in Appendix A).

Before matching, there were no significant differences between patients with and without AKI in sex, gestation weeks, or preoperative LVEF (all p > 0.05), but those who developed AKI were more likely to be younger and have cyanotic heart disease, undergo more complex surgeries (RACHS-1 ≥ 4), experience lower core temperature during CPB, more frequent use of modified ultrafiltration and intraoperative platelet transfusions (all p < 0.05). Fig. 2A showed the distributions of intraoperative glycemic metrics between AKI and Non-AKI groups. The median [IQR] pre-CPB blood glucose levels of AKI and Non-AKI patients were 4.8 (4.3 to 5.6) and 5.1 (4.5 to 5.6) mmol/L respectively, the intra-CPB blood glucose levels were 6.3 (5.2 to 7.4) and 6.6 (5.6 to 7.7) mmol/L, and the post-CPB blood glucose levels were 7.7 (6.2 to 9.2) and 7.9 (6.7 to 9.4) mmol/L. Intra-GFs between the two groups were similar at 1.6 (0.9 to 2.3) and 1.6 (1.0 to 2.2) mmol/L. By univariate analysis, higher intra-CPB (OR 0.896; 95% CI, 0.812 to 0.989; p = 0.029) and post-CPB blood glucose levels (OR 0.903; 95% CI, 0.831 to 0.982; p = 0.017) during the perioperative time were significantly associated with postoperative AKI, while pre-CPB and intra-GF were not. After accounting for known and potential confounding factors (p < 0.2), the multivariate regression analysis showed a consistent result, that higher intra-CPB (adjusted OR 0.802; 95% CI, 0.706 to 0.912; p = 0.001) and post-CPB blood glucose levels (adjusted OR 0.830; 95% CI, 0.744 to 0.925; p = 0.001) both significantly associated with lower risk of postoperative AKI. However, the pre-CPB and intra-GF were not associated with AKI in this study. Results are shown in Table 2.

After matching, the differences between age, sex, and baseline SCr were well-balanced (p < 0.05). The median [IQR] age of the cases and controls were 11.2 (3.9 to 30.4) and 10.6 (6.0 to 31.8) months, respectively. Male patients made up about 50% of both two groups. The distributions of intraoperative glycemic metrics were presented in Fig. 2B. Multivariate conditional logistic regression showed that intra-CPB (adjusted OR 0.784; 95% CI, 0.649 to 0.948; p = 0.012) and post-CPB glucose levels (adjusted OR 0.803; 95% CI, 0.691 to 0.933; p = 0.004) were significantly associated with AKI after adjusting for confounders (p < 0.2), including preoperative LVEF, CPB time, the lowest core temperature during CPB and intraoperative platelet transfusion, RACHS-1 score ≥4, cyanosis, neonate. Results were consistent in the study population before and after matching. The univariate and multivariate analyses of the associations between intraoperative glycemic metrics and postoperative AKI after matching are presented in Table 3, and the results of multivariate analyses are shown in Fig. 3.

This prospective nested-case control study showed that higher intraoperative blood glucose levels during and after CPB were independent protective factors against AKI in pediatric patients following cardiac surgery. However, there was no significant difference in blood glucose levels before CPB and variability of intraoperative blood glucose levels between AKI and control groups.

Our study assessed the potential effects of intraoperative blood glucose and glycemic fluctuations on postoperative AKI in pediatric patients undergoing congenital heart surgery. Based on previous studies, hyperglycemia, hypoglycemia, and glucose variability were significantly associated with poor clinical outcomes [19]. Some experiments reported that hyperglycemia-induced reactive oxygen species could inhibit Na⁺/glucose cotransporter activity in renal proximal tubule cells and result in renal oxidative stress and kidney injury [20,21]. Certainly, hypoglycemia could decrease glucose intake and compromise cellular biosynthetic abilities, and also result in acute organ impairment [19,22]. Patients with a minimum glucose level less than 75 mg/dL both intraoperatively and postoperatively had a 3-fold greater adjusted odds ratio of developing the composite morbidity-mortality endpoints including renal failure requiring dialysis [18]. Besides, acute glycemic fluctuations were independently associated with postoperative AKI induced by endothelial apoptosis, greater endothelial dysfunction, and activation of oxidative stress in non-cardiac surgery patients and children who underwent congenital heart surgery [11,23–25]. Overall, these findings above display a potential biological association between intraoperative blood glucose metrics and postoperative AKI in children undergoing congenital heart surgery though the underlying mechanism is unclear.

In this study, we selected peri-CPB as the points of investigation mainly because CPB was strikingly associated with intraoperative elevated blood glucose levels and variability. The process of CPB induces stress response and systemic inflammatory reaction which are associated with hyperinsulinemia and insulin resistance, posing a dominating contribution to perioperative hyperglycemia [14,26]. During stress, increased secretion of hormones like glucagon, catecholamines, growth hormone and cortisol promotes gluconeogenesis in the liver and impairs insulin action in peripheral tissues, resulting in elevated serum glucose [27,28]. At the same time, prophylactic glucocorticoids have been routinely administered to infants undergoing heart surgery with CPB in the perioperative period to attenuate the systemic inflammatory response. Recent studies have demonstrated a significant association between glucocorticoids administration and hyperglycemia in pediatric patients undergoing cardiac surgery [29,30].

Several studies have examined the effect of hyperglycemia and adverse outcomes in the pediatric cardiac population [31,32]. Average postoperative glucose >143 mg/dL and peak glucose level ≥250 mg/dL were associated with greater adjusted odds of reaching a composite morbidity-mortality end point after complex congenital heart surgery [18]. In O’Brien’s study, intraoperative hyperglycemia which was defined more than 175 mg/dL had a significant association with postoperative bacteremia in a pediatric cardiac population aged less than 18 years old [33]. In contrast to previous studies involving postoperative hyperglycemia in the intensive care unit, our study showed no harmful effects of higher mean blood glucose on postoperative AKI during surgery. The underlying factors affecting intraoperative glucose levels, including surgical insults and the process of CPB, differ from the causes of postoperative hyperglycemia. Duncan and colleagues suggested that intraoperative mild hyperglycemia (141 to 170 mg/dL/7.8 to 9.4 mmol/L) was well tolerated in cardiac surgical patients and may be associated with benefits on postoperative outcomes [12]. In Lou’s study, perioperative hyperglycemia (>150 mg/dL) was not detrimental to infants undergoing cardiac surgery with CPB [9]. What’s more, a blood glucose level of no more than 8.1 mmol/L (146 mg/dL) during CPB is associated with a greatly lower risk probability of severe systemic inflammatory response syndrome in pediatric patients [34].

Our data showed that the median intra-CPB blood glucose levels were 6.3 and 6.6 mmol/L, and the post-CPB blood glucose were 7.7 and 7.9 mmol/L respectively in AKI and Non-AKI groups before matching, which suggested that most of our pediatric patients were at a relatively normal and safe glycemic threshold. Higher intraoperative glucose levels may ensure increased glucose uptake and confer protection from hypoxic-ischemic injury by promoting anti-apoptotic pathways and favoring angiogenesis [35], yielding the roles of organ protection in infants undergoing deep hypothermic circulatory arrest [36,37]. In our study, the average serum blood glucose in the Non-AKI group was higher than that in the AKI group and the difference was statistically significant. Besides, there was no significant difference in intraoperative steroids administration between AKI and Non-AKI groups, which meant that the difference of serum glucose between AKI and Non-AKI groups may be not associated with steroids administration in this study. So we reported that higher blood glucose levels during and after CPB may be nephroprotective. Since no optimal perioperative serum glucose range has been made in pediatric patients with a large age range undergoing cardiac surgery, we did not define a cutoff value of hyperglycemia which would be associated with increased incidence of adverse outcomes. But we believe that there is a certain upper limit which the intraoperative serum glucose exceeds would cause a harmful effect.

This study did not show an association between pre-CPB blood glucose and postoperative AKI. Our findings are somehow consistent with another observational study in pediatric patients undergoing congenital heart surgery, which found that blood glucose before CPB was not a predictor of mortality, infection, or increased duration of ventilation time [38]. Unlike adults who are commonly associated with obesity, dyslipidemia, insulin resistance, and hypertension, the children in our study have no diabetic complications or other metabolic syndromes. In addition, the pre-CPB blood glucose defined in our study did not represent a long-term but acute glycometabolic status. Pre-CPB blood glucose may be not an independent predictor of postoperative AKI in pediatric patients following cardiac surgery, and the subsequently elevated CPB blood glucose levels may pose a more important role in AKI.

Previous observational studies in adults and children have demonstrated that intraoperative glycemic fluctuation is an independent predictor of postoperative AKI [11,39]. However, this association was not observed in this study. Similarly high levels of intraoperative glycemic fluctuations were seen in AKI and Non-AKI patients in our study. This could be partly explained by the different study design that we included all kinds of congenital heart surgeries in the final analysis and almost half of them (51.7%, 356/688) are not complex congenital heart surgeries (RACHS-1 < 3) [18]. The whole blood glucose variability during surgery is not so great. In addition, we included patients aged less than 18 years which represents a wide pediatric age range. Though uniform normoglycemia thresholds and definitions of diabetes have been adopted in children, there are still age-related differences in response to dysglycemia [40,41]. For example, cortisoland growth hormone responses to spontaneous hypoglycemia are highly age-dependent, and impaired fasting glucose risk is positively correlated with increasing age [42,43]. Previous studies also showed a age influence on glucose levels or insulin requirement to achieve target glycemic control [44,45]. Considering the age-related effects on glucose metabolism, we did not provide specific cutoff limits in this study. Furthermore, it remains unknown whether age-related differences affect the biological sequelae of intraoperative hyperglycemia in children undergoing cardiac surgery.

Our study has some limitations that merit the next explorations. Firstly, we collected glucose levels intermittently intraoperatively and might miss the peak values of blood glucose. However, we acquired the blood glucose data at the same time points for all enrolled patients, thereby generating relatively good comparability between the two groups. Secondly, although we obtained that higher blood glucose levels may be a protective factor against AKI, we did not provide the specific cutoff value of optimal blood glucose level. We will enlarge our sample size to further explore the optimal blood glucose thresholds for postoperative renal function according to different age gaps. Thirdly, though the incidence of AKI in our study is consistent with that in a previous study diagnosing AKI solely according to serum creatinine [46], it is probable that some additional AKI would have been identified when combined with the urine output KDIGO criteria. Lastly, this was a single-center observational study, some confounding factors were unavoidable, which might affect the final results. However, we prospectively collect the data, thereby guaranteeing the integrity and reliability of the data as far as possible. Besides, we performed strictly univariate and multivariate analyses under both unmatched and matched methods and obtained consistent results.

This nested-case control study demonstrated that higher intraoperative glucose levels during and after CPB may be protective factors against AKI in the pediatric population having congenital heart surgery. The intraoperative glucose levels before CPB and the glycemic fluctuations during the whole surgery were not significantly associated with postoperative AKI. These results suggested that intraoperative blood glucose levels after CPB initiation could produce a great influence on AKI occurrence in pediatric cardiac surgery patients. Unluckily, we did not identify the specific cutoff value of optimal blood glucose level due to the large age range of enrolled patients. Another prospective study is in progress in our center to investigate the optimal cutoff values of intraoperative blood glucose associated with reduced AKI incidence in children following cardiac surgery based on different age gaps.