Association between mineral and bone disorder in patients with acute kidney injury following cardiac surgery and adverse outcomes

Background Numerous studies have evaluated the prevalence and importance of mineral and bone disorders among patients with chronic kidney disease (CKD) and end-stage renal disease (ESRD). However, little is known about dysregulated mineral and bone metabolism in acute kidney injury (AKI). Methods We evaluated the association between mineral and bone metabolites and clinical outcomes in 158 patients who underwent cardiac surgery and developed AKI between June 2014 and January 2016. The baseline characteristics of the patients were recorded, and the levels of mineral and bone metabolites, including calcium, phosphate, intact parathyroid hormone (iPTH), 25-hydroxyvitamin D (25D), bone-specific alkaline phosphatase (BAP), tartrate-resistant acid phosphatase 5b (TRACP-5b) and C-terminal fibroblast growth factor 23 (cFGF23) were measured within 12 h after establishing the clinical diagnosis. Results The serum phosphate, iPTH and cFGF23 levels were significantly associated with the 28-day mortality (phosphate: Hazard Ratio [HR] =2.620, 95% CI: 1.083 to 6.338, p = 0.035; iPTH: HR = 1.044, 95% CI: 1.001 to 1.090, p = 0.046; cFGF23: HR = 1.367, 95% CI: 1.168 to 1.599, p < 0.001). Moreover, higher serum cFGF23 and BAP levels were independently associated with an increased risk of adverse outcomes. Additionally, we found that the serum cFGF23 levels rose most significantly and were associated with the severity of AKI (P < 0.001). Conclusions Mineral and bone metabolites are dysregulated and are associated with adverse clinical outcomes among patients with AKI. Trial registration www.clinicaltrials.gov NCT 00953992. Registered 6 August 2009.


Background
Acute kidney injury (AKI) is a common problem in seriously ill patients and is associated with several poor clinical outcomes [1][2][3][4][5][6]. Numerous studies have proven that chronic kidney disease-mineral and bone disorder (CKD-MBD) is associated with adverse clinical outcomes [7][8][9]. In contrast, little is known about mineral and bone disorder and its association with outcomes among patients with AKI.
Reduced vitamin D levels, elevated parathyroid hormone (PTH) and fibroblast growth factor 23 (FGF23) levels and decreased or increased calcium and phosphate levels have been reported in patients with AKI in small studies [10][11][12][13][14], but their mechanisms are defined insufficiently. In the case of chronic kidney disease (CKD), the FGF23 level rises early in the course of the disease, stimulates urinary phosphate excretion and inhibits the activation of 25-hydroxyvitamin D (25D), and contributes to the development of secondary hyperparathyroidism [15].
In recent studies, the role of markers of bone metabolism in the prediction of mortality in patients with CKD has been prospectively evaluated. These studies have indicated a link between bone metabolism and cardiovascular events in patients with CKD. Markers of bone formation (bone-specific alkaline phosphatase [BAP]) and bone resorption (tartrate-resistant acid phosphatase 5b [TRACP-5b]) can serve as predictors of cardiovascular morbidity and mortality in patients with CKD [16].
Based on these observations, we hypothesized that the serum markers of mineral and bone metabolism can be used to predict the mortality of patients with AKI. To systematically evaluate the predictive value of the markers of mineral and bone metabolism for poor outcomes in patients with AKI, we designed a prospective cohort study to evaluate the predictive value of the markers of mineral and bone disorder for the mortality of patients who underwent cardiac surgery and developed AKI.

Study design
We conducted a single-center, prospective cohort study and enrolled patients who underwent cardiac surgery and developed incident AKI at a comprehensive hospital between June 2014 and January 2016. This study was approved by the hospital ethics committee of Shanghai Ninth People's Hospital, School of Medicine, Shanghai Jiaotong University (approval number: [2014]45). All of the experimental protocols were implemented in accordance with the relevant guidelines and regulations. In accordance with the Declaration of Helsinki, patients or their legal guardians had obtained written informed consent prior to participation,. This study was registered at www.clinicaltrials.gov with the identification number NCT00953992.

Study patients
The inclusion criteria were an age > 18 years and the development of incident AKI following cardiac surgery. AKI was defined as an increase in serum creatinine of ≥0.3 mg/ dL within 48 h or that of ≥50% within 7 days, which conforme with the criteria established by the Kidney Disease Improving Global Outcomes (KDIGO) work group [17].

Study procedures
The serum creatinine data of the participants were monitored daily in the hospital's information system (HIS) to rapidly identify patients with incident AKI. We collected and stored the serum samples within 12 h after establishing the clinical diagnosis. The samples were centrifuged, aliquoted, and stored at − 80°C within 2 h of collection.

Clinical outcomes
We adjudicated all of the outcomes by reviewing the electronic medical records from the HIS. The pre-specified primary endpoint was the 28-day all-cause mortality.
The secondary endpoints were Renal Replacement Therapy or in-hospital mortality (RRT/death). Additional endpoints included the ventilator-free days and hospital-free days in 28 days. The patients who died before 28 days were graded a score of zero.

Laboratory assessment
The bone-specific alkaline phosphatase and tartrateresistant acid phosphatase 5b levels were measured using a standard ELISA kit (Immunodiagnostic Systems Limited, UK). The C-terminal FGF-23 levels were measured using a standard ELISA kit (Immutopics Inc., USA). The serum calcium and phosphate levels were measured on an automated chemistry analyzer (AU5800, Beckman Coulter, USA). The iPTH level was measured on an automated immunoassay analyzer (UniCel DxI800, Beckman Coulter, USA). The 25-hydroxyvitamin D level was measured on an automated immunoassay analyzer (LI-AISON, DiaSorin, Italy).

Statistical analyses
The data are reported as the median and interquartile range (IQR: 25th percentile-75th percentile) for continuous variables and as the n (%) for categorical variables. The baseline and operative characteristics were compared in patients with different stages of AKI using the Kruskal-Wallis test for continuous variables with a non-normal distribution and using the chi-square test for category variables. The correlation between the enrollment mineral metabolite levels and the other variables was analyzed using the Spearman rank correlation coefficient. The comparison of the mineral metabolite levels among the patients who died and those of the patients who did not die was assessed using the Mann-Whitney U test.
Univariate Cox regression was used to identify potential confounding variables, including demographics (sex and age), preoperative renal function (serum creatintine and eGFR), comorbidities (hypertension, chronic heart failure and diabetes mellitus), operative characteristics (operation type, whether cardiopulmonary bypass or not), postoperative characteristics (low blood pressure, applying diuretics), Acute Physiology and Chronic Health Evaluation (APA-CHE) II score and biochemical parameters. After that, we used multivariate Cox regression to adjust for covariates using two different models: model 1 was adjusted for age, sex, baseline eGFR, hypertension, chronic heart failure (New York Heart Association functional class [NYHA] ≥ 2) and diabetes mellitus; model 2 was further adjusted for the operation type and APACHE II score on enrollment, according to P value in the univariable regression (P < 0.1) or clinical importance. Logistic and linear regression models were used to assess the association between the mineral metabolite levels and secondary endpoints, adjusting for the same covariates as above. Receiver-operating characteristics (ROC) curves and areas under the curves (AUCs) were calculated using the survival and time ROC packages to compare the predictability of the mineral and bone metabolite levels for the 28-day mortality, Furthermore, the C-index was calculated using the survival package. The evaluation criterion of the C-index was similar to that of the AUC.
We used linear mixed models for the repeated measures to test for significant differences in the mineral and bone metabolite levels over time between the patients with different severities of AKI. The comparison of the mineral and bone metabolite levels at individual time points was assessed using the Spearman rank correlation coefficient (for the correlation between the severity of AKI and the mineral and bone metabolite levels).
All of the comparisons were two-tailed, with P < 0.05 considered as being statistically significant. The statistical analysis was performed using the software package R version 3.4.0 (www.r-project.org).

Patient characteristics
We enrolled 158 patients who developed incident AKI following cardiac surgery but had no evidence of AKI at baseline into a prospective cohort study.
Among these three groups, the patients with aortic dissection and those with low blood pressure after the operation presented with higher ratios of severe kidney injury (P = 0.002 and P < 0.001, respectively). Meanwhile, the ratios of AKI stage 3 were lowest in the patients with coronary artery bypass alone (P = 0.009). Moreover, patients with more severe kidney injury had higher scores of the APACHE II (P = 0.007). However, there were no significant differences among three groups in the demographics, preoperative renal function, comorbidities, cardiopulmonary bypass during the operation, and applying diuretics after the operation (P > 0.05).

Serum mineral metabolite levels and their relationship with in-hospital mortality at 28 days
Among the 158 patients enrolled in this study, 18 (18.8%) had died within 28 days, and the median time from the clinical diagnosis to death was 7 days (IQR: 2-17 days). The causes of death included multiple organ dysfunction syndromes (n = 7), heart failure (n = 4), adult respiratory distress syndrome (n = 3), acute cerebral infarction (n = 2), cute gangrenous cholecystitis (n = 1) and circulatory failure of unknown disease (n = 1). The serum mineral metabolite levels in patients who died and in those who did not die are shown in Table 2. The patients who died had higher levels of serum cFGF23 (P = 0.007), as well as lower levels of 25D (slightly insignificant, P = 0.072), compared with the patients who were alive. The levels of the other mineral metabolites were similar between the two groups (Table 2).
Then, we analyzed the association between each mineral metabolite biomarker and the risk of death after AKI. Table 3 shows the unadjusted and adjusted hazard ratios for the 28-day mortality risk according to the levels of the mineral metabolites. In the unadjusted analyses, the serum iPTH and cFGF23 levels were both significantly associated with the 28-day mortality (iPTH: HR = 1.045, 95% CI: 1.010 to 1.080, p = 0.009; cFGF23: HR = 1.258, 95% CI: 1.149 to 1.378, p < 0.001). These associations were slightly attenuated but remained significant when the model was adjusted for age, sex, preoperative eGFR, hypertension, congestive heart failure, and diabetes mellitus (Model 1), as well as when it was further adjusted for operation type and APACHE II score (Model 2; iPTH: HR = 1.044, 95% CI: 1.001 to 1.090, p = 0.046; cFGF23: HR = 1.367, 95% CI: 1.168 to 1.599, p < 0.001). The serum phosphate level was also associated with the 28-day mortality in the adjusted analyses (HR = 2.620, 95% CI: 1.083 to 6.338, p = 0.035). The other biomarkers, including the corrected calcium, 25D, BAP, and TRACP-5b, were not associated with the 28day mortality. Additionally, Table 4 shows the area under the receiver operating characteristic for the 28day mortality risk according to the levels of the mineral metabolites. The strengths of the AUC appeared to be greater for cFGF23 than for the other mineral biomarkers that were tested (P = 0.007).
In an exploratory analysis, we evaluated whether the phosphate, iPTH, and cFGF23 levels remained associated with 28-day mortality after the adjustment for each. As shown in Additional file 1: Table S2, only the cFGF23 level, and not the iPTH or phosphate levels, remained as being significantly associated with the 28-day mortality even after the adjustment for each (HR = 1.380, 95% CI: 1.162 to 1.639, p < 0.001).
Serum mineral metabolite levels and other adverse outcomes Table 5 shows the adjusted logistic and linear regression models of the secondary endpoints according to the

Kinetics of the serum mineral metabolite levels in AKI after cardiac surgery
To compare the time course of changes in the mineral and bone metabolite biomarkers in patients with AKI after cardiosurgery, we also measured the levels in a subgroup of 71 patients, and these results are shown in Fig. 1. We found that there was a hierarchical association between the AKI severity and the levels of cFGF23 over time (P < 0.001). Among the AKI stage 3 patients, we found rapid and significant increases in the cFGF23 levels like that the median levels at enrollment were1 8-fold higher than the baseline (preoperative) levels (P < 0.001) and the Day 5 levels were~14-fold higher (P < 0.001) than the baseline levels. Among the AKI stage 1 patients, the median cFGF23 levels increased bỹ 4.5-fold at the time of enrollment and by~1.6-fold on Day 5. The AKI stage 2 group of patientshad medium rises in the cFGF23 levels that were between those of the AKI stage 1 group and those of the AKI stage 3 group. The serum creatinine and iPTH levels showed a similar hierarchical association, which higher levels were associated with a greater severity of AKI. Additionally, when the phosphate and 25D levels were evaluated at individual time points, only the Day 5 levels were significantly different between the groups. The longitudinal calcium levels, as well as the BAP and TRACP-5b levels, were similar in these three groups.

Discussion
In this prospective cohort study, we confirmed that higher serum cFGF23 levels obtained within 12 h of a clinical diagnosis for AKI after cardiac surgery are independently associated with a greater 28-day mortality risk, as well as with several other adverse outcomes, including in-hospital mortality and the need for RRT, and fewer ventilator-free and hospital-free days. Additionally, we also observed that other mineral and bone metabolites, such as phosphate, iPTH, and BAP, are marginally associated with poor outcomes. Moreover, we found that the serum cFGF23 levels rose most significantly and were associated with the severity of AKI. These findings suggest that there is a complex interplay between AKI, mineral and bone metabolism, and adverse clinical outcomes after cardiac surgery.  Multiple studies have demonstrated that mineral and bone disorders are prevalent and associated with mortality in patients with CKD and ESRD [7][8][9]. However, studies on minerals, particularly those on bone metabolite levels in AKI, are limited. In contrast to our prior study, which investigated the levels of serum vitamin D (including 1,25dihydroxyvitamin D and 25-hydroxyvitamin D) in patients with hospital-acquired AKI only [14], here, we evaluated both mineral and bone metabolite levels in patients with AKI after cardiac surgery. Thus, the findings from this study not only demonstrate an independent and prospective association between the mineral metabolite levels and poor outcomes of AKI but also represent the comprehensive study of bone metabolites in clinical setting, which is a rare finding in similar AKI cohort studies.
BAP and TRACP-5b are sensitive biomarkers of osteoblastic bone formation and osteoclastic bone resorption [8,19,20]. The few prior studies that measured the serum BAP and TRACP-5b levels were conducted almost exclusively among patients with CKD or ESRD. Astrid et al. measured the levels of bone markers in serum samples from 627 patients with CKD and found that the markers of bone formation (BAP: OR: 1.01, 95% CI: 1.01-1.02) and bone resorption (TRACP-5b: OR: 0.86, 95% CI: 0.75-0.99) can serve as predictors of cardiovascular morbidity and mortality in patients with CKD [16]. Further contrasting with the results of prior studies, our data indicated that a high BAP serum level, which implies that there is high bone formative activity, is potentially associated with an increased risk of death in patients with AKI.
In addition to the 28-day mortality, we also evaluated the association between the mineral and bone metabolite levels and several secondary endpoints. The serum cFGF23 levels were markedly and independently associated with RRT/Death and longer durations of mechanical ventilation and hospital length of stay. Coincidentally, higher serum BAP levels were also associated with longer durations of mechanical ventilation and hospital length of stay.
Whether mineral and bone metabolite levels are merely an illness severity marker or directly contribute to poor outcomes is a critical question that could not be answered by this study. Higher FGF23 levels are related to pathogenesis of left ventricular hypertrophy [21][22][23] and inflammation [24][25][26] in CKD. In fact, a large number of epidemiologic data suggest that decreased vitamin D metabolite levels are associated with adverse outcomes in those with critical illness [27,28]. Maybe increased FGF23 levels could lead to adverse outcomes via inhibiting the activation of vitamin D metabolite [15,26]. However, we did not find positive or inverse correlations between the serum cFGF23 and 25D levels, and we also did not find a correlation between the 25D level and the 28mortality risk, which is consistent with the results of our previous study. Thus, the associations between the FGF23 and vitamin D metabolite levels remain unclear.
We acknowledge several limitations of this study. All samples used in our study were from one single centre. Therefore, selection bias in the enrollment was inevitable. 18 patients had died within 28 days, due to the limitation of the sample size in this cohort study, no further subgroup analysis has been done for the cause of death, and so it was difficult to determine the link between mineral and bone disorder and the cause of death. Additionally, urine output was an important criterion, which was not accessed in this study. Thus, the incidence of AKI may be underestimated and the estimation of AKI stages may be affected. Of course, this result was only uncovered in a small sample size and a larger clinical sample is needed to confirm this result. AKI acute kidney injury, RRT/death renal replacement therapy or in-hospital mortality, iPTH intact parathyroid hormone, 25D 25-hydroxyvitamin D, BAP bonespecific alkaline phosphatase, TRACP-5b tartrate-resistant acid phosphatase 5b, cFGF23 C-terminal fibroblast growth factor 23 . P values are from repeated-measures linear mixed models comparing the 3 groups. *P < 0.05, **P < 0.01, and ***P < 0.001; a comparison of the biomarker levels cross AKI categories at individual time points. The data are shown as the median (interquartile range). Scr: serum creatinine; iPTH: intact parathyroid hormone; 25D: 25-hydroxyvitamin D; BAP: bone-specific alkaline phosphatase; TRACP-5b: tartrate-resistant acid phosphatase 5b; cFGF23: C-terminal fibroblast growth factor 23