What about Alkaline Phosphatase in Endocarditis?

1. Introduction

Infective endocarditis (IE) is a rare but severe manifestation of valvular disease associated with significant morbidity and mortality. [1] The Duke criteria are used for diagnostic purposes. [2] Besides extended antimicrobial therapy, surgery should be performed in case of heart failure, uncontrolled infection and in order to prevent embolism. If indicated, surgery should be performed as early as possible. [3,4]

Transient bacteremia, as it may occur after dental, gynecologic, gastrointestinal, or urologic procedures or even after daily activities such as brushing teeth, may lead to bacterial colonization of heart valves resulting in sepsis, causing a spectrum of end-organ damage. [5] With up to two-thirds developing acute kidney injury (AKI) in the setting of IE, end-organ damage of the kidney is one of the most common. [6] Renal inflammation and resulting hypoxia are considered to be the causative. [7] The binding of pathogen-associated molecular patterns (PAMPs) such as lipopolysaccharide (LPS) to toll-like receptor 4 leads to the further release of inflammatory mediators. [8] The pervasive alkaline phosphatase (AP) neutralizes LPS and converts proinflammatory molecules to anti-inflammatory ones (such as adenosine). [7,9,10] (Severely) elevated AP levels have been considered a surrogate parameter for infection or bacteremia. [11,12,13] Consequently, AP has been considered a treatment target for sepsis patients in improving AKI or survival (in animal models). [7,14,15]

Additionally, to the inflammatory process caused by endocarditis, cardiac surgery induces some systemic inflammation due to the heart-lung machine (HLM). In up to 30%, a systemic inflammatory response syndrome (SIRS) occurs, associated with an elevated risk of end-organ injury. [16,17,18] No sole mechanism is responsible, but various factors are considered contributing [7,19,20,21,22] Increased consumption of AP during cardiac surgery or extracorporeal membrane oxygenation is associated with a higher morbidity and mortality rate. [23,24,25,26]

This study aims to evaluate the AP metabolism in the setting of an inflammatory state in IE patients undergoing surgery. To our knowledge, no study evaluating the course of AP following cardiac surgery for IE has been published previously.

2. Results

Patient Population and Cut-off Values

After reviewing the exclusion criteria 314 patients were included in the retrospective analysis. According to an initial drop in AP of ≥ 30% (with an area under curve (AUC) 0.638; sensitivity 71%, specificity 53% for 1-year mortality) two groups were created– with 179 patients with an AP drop ≥ 30% and 135 patients with an AP drop < 30%. The selection of cut-off values was guided by previous research findings. [25,26] In the following sections, the first value always refers to the lower initial AP drop cohort.

Preoperative Characteristics

The higher initial AP drop cohort was at significantly higher surgical risk (EuroScore II 9.1 (15.9;3.5) vs. 16.1 (34.3;6.4) %; p=0.000), had a significantly higher rate of preoperative dialysis (5.9 vs. 13.4%; p=0.030) and prosthetic valve endocarditis (PVE) (20.0 vs. 40.2%; p=0.000). All other patient characteristics were similar among the cohorts (Table 1).

Procedural Data

A significantly higher rate of re-sternotomy (21.5 vs. 40.8%; p=0.000) and double valve replacement (8.9 vs. 22.9%; p=0.001) were observed after an initial higher AP drop. Surgical times were prolonged in the initial higher AP cohort: total surgery time (240 (300;195) vs. 348 (455;255) min; p=0.000), total cardiopulmonary bypass time (110 (155;90) vs. 162 (237;117) min; p=0.000) and total aortic cross clamp time (81 (113;62) vs. 111 (159;79) min; p=0.000). Procedural data is illustrated in Table 2.

Laboratory data

While AP values at baseline were significantly higher in the higher initial AP drop cohort (74 (96;62) vs. 93 (120;76) U/L; p=0.000), there was no significant difference at 30 days (116 (141;89) vs. 128 (172;99) U/L; p=0.089). Patients with the higher initial AP drop required a longer time to reach baseline values (days till baseline value surpassed: 4 (6;3) vs. 6 (7;4) days; p=0.000; baseline within 3 days: 43.7 vs. 33.0%; p=0.052; baseline within 5 days: 74.1 vs 60.3 %; p=0.011). The postoperative course of AP dependent on the initial AP drop is depicted in Figure 1. On postoperative day (POD) 30 AP levels were significantly higher compared to baseline in both cohorts (p=0.000 for both). Figure 2 illustrates the differences of AP levels at baseline to POD 1 and POD 30.

Although there was no statistic significant difference at baseline CRP values (3.9 (8.3;1.7) vs. 4.5 (11.7;1.5) mg/dL; p=0.456), a significantly higher value after 30 days in the higher initial AP drop cohort was observed (2.1 (5.7;0.7) vs. 4.6 (10.6;1.8) mg/dL; p=0.001). Laboratory values are provided in Table 3.

Adverse Events and mortality

An overview of adverse events and mortality is available in Table 4. A higher rate of revision due to bleeding (7.4 vs. 15.6%; p=0.027), postoperative renal replacement therapy with or without previous one (7.4 vs 21.8 %; p=0.001 and 6.7 vs. 15.6%; p=0.015), extracorporeal membrane oxygenation (2.2 vs. 19.0%; p=0.000) was seen after higher initial AP drop. While there was a strong trend towards prolonged hospitalization (p=0.054), a prolonged ICU stay was significantly more likely in the higher initial AP drop cohort (31.9 vs. 48.0%; p=0.004).

Short-term mortality (30day: 3.0 vs. 10.6; p=0.010; in-hospital: 5.9 vs. 17.3%; p=0.002; 1-year: 14.1 vs. 25.7%; p=0.012) and long-term mortality (p=0.008, see Figure 3) were significantly higher in the higher initial AP drop cohort.

ECMO – extracorporeal membrane oxygenation; prolonged hospital stay longer than 30days; prolonged ICU (intensive care unit) stay defined as more than 7 days; prolonged intubation defined as reintubated, longer as 48h, tracheostoma;

3. Discussion

In line with previous studies by Schaefer et al. (mitral valve patients with impaired left ventricular function) and Poschner et al. (postcardiotomy VA-ECMO patients), we were able to demonstrate a relation between higher initial AP drop and morbidity and mortality in patients undergoing cardiac surgery for left-sided IE. [25,26] Not only short-term (i.e., hospital mortality) but also long-term survival (over ten years) was significantly impaired following a higher initial AP drop (p=0.008).

Prosthetic valve endocarditis occurs in up to 20% of all endocarditis cases and is associated with a worse prognosis. [27,28] This finding could also be highlighted in our study cohort. Indeed, PVE was approximately twice as common in the higher initial AP drop cohort (40.2% vs. 20.0%; p=0.000). Even though preoperative AP levels were within the normal range in both cohorts, AP levels were significantly higher in the higher initial AP drop cohort (see Table 3). Interestingly, baseline AP values, as well as baseline CRP values, showed no significant difference in PVE versus native IE (baseline AP in PVE 87 (109; 71) vs. native 84 (113; 68) U/L; p=0.465; CRP in PVE 4.3 (12.6; 1.3) vs. native 4.2 (8.7; 1.8); p=0.617). Hence, we may conclude that PVE seems to be not related to a more pertinent systemic inflammation and that the significant impaired outcome was instead primarily due to the longer surgical times (total surgical time: PVE 389 (485;320) vs. native 245 (319;205) min; total cardiopulmonary bypass time PVE 192 (266;147) vs. native 118 (172;94) min and total aortic cross-clamp time PVE 121 (173;95) vs. native 82 (120;62) min all with a p=0.000). As demonstrated by Doenst T et al., Nissinen J et al., and others, longer operation times are an independent predictor of worse outcomes. [29,30]. Increased permeability in the gastrointestinal tract induced by HLM leads to a wash-in of endotoxins into the bloodstream, inducing the release of proinflammatory mediators. [20,22] AP-mediated dephosphorylation may convert a portion into anti-inflammatory substances. [31] This process leads to the consumption of systemic alkaline phosphatase. [23,32] Thus, prolonged duration of surgery is also likely to lead to an increased decrease in AP due to prolonged exposure to endotoxins and associated mediators.

Iso-forms of AP are present in the proximal tubule cells of the kidney. Without a complete understanding of the role of AP in the kidney, animal studies have demonstrated increased levels of AP in the urine after ischemia-reperfusion injury and administration of LPS, thus, suggesting injury to the proximal tubule brush border in these situations. [33,34] In the receiver operating characteristics (ROC) analysis in our cohort, the drop in alkaline phosphatase was associated with the need for postoperative dialysis with an AUC of 0.668. After an initial more significant drop, the need for postoperative dialysis was significantly more frequent (p=0.001). Interestingly, the rate of postoperative dialysis was significantly higher with and without preoperative dialysis in the cohort with a higher initial AP drop. This suggests that patients with and without injured renal parenchyma are significantly more likely to have AKI due to SIRS (represented by the higher AP drop). Patients with AKI have significantly worse outcomes after cardiac surgery. [35]

Zhang et al. highlighted the alkaline phosphatase on the surface membrane of neutrophils as a good tool for differentiating between SIRS patients with or without bacteremia. [36] Similarly, Kerner A. et al. and Tung C.B. et al. showed (severely) elevated AP levels as a surrogate parameter for bacteremia. [11,12] However, these findings could not be confirmed in this study. In 82.5% of the total cohort, positive blood cultures were present, which were associated with a significantly elevated CRP level (CRP positive BC 4.9 (11.9; 1.9) vs. negative BC 2.7 (5.7;0.8) mg/dL; p=0.001); however, without any differences in AP (AP positive BC 86 (114;70) vs. negative BC 79 (106;64) U/L; p=0.116).

An impaired barrier function of the gastrointestinal tract (GIT) leads to a wash-in of LPS. Hamarneh et al. demonstrated an impaired barrier function of the GIT in patients deprived of enteral feeding due to a loss of AP expression reversible to AP supplementation. [37] Given the critical state of IE patients, they often lack adequate enteral feeding. Hence, considering additionally the inflammatory state of endocarditis patients with the high rate of need for postoperative dialysis (and the likely even higher rate of any AKI) and the further AP consumption during cardiac surgery, supplementation of AP may be especially promising in IE patients. [8] Currently, clinical trials (e.g., APPIRED III – ClinicalTrial.gov NCT03050476) are evaluating the external administration of AP during (elective) cardiac surgery and are expected to reduce morbidity after cardiac surgery.

Our study found a significantly higher rate of ECMO implantation after a higher initial AP drop (2.2% vs. 19.0%; p=0.000), presumably due to the longer operation times and the higher preoperative risk found in this cohort. In-hospital mortality after postcardiotomy ECMO support is reported in up to two-thirds of patients. [38,39] The significantly worse outcome after a higher initial AP drop may thus be attributed to or at least influenced by the higher rate of ECMO implantation (30-day mortality 3.0% vs. 10.6%; p=0.010; in-hospital mortality 5.9% vs. 17.3%; p=0.002; 1-year mortality 14.1% vs. 25.7%; p=0.012). However, as already shown in previous work, ECMO support did not increase AP consumption, evident in reaching baseline values within five days in both studies. [25]

4. Materials and Methods

Patients

All patients who required valve surgery for isolated left-sided IE between January 2009 and October 2022 at the Department of Cardiac Surgery (Vienna General Hospital, Austria) were included into this retrospective data analysis. The ethics committee of the Medical University of Vienna approved this retrospective study (2101/2022). Modified Duke criteria were used for IE diagnosis. [40] Exclusion criteria included IE involvement of the right side, surgical valve repair, homograft implantation and missing laboratory values at baseline or on POD 1. After reviewing the exclusion criteria 314 patients were analyzed.

Laboratory Data

Blood samples were drawn at baseline and on each of the first seven POD as well as on POD 14 (±3) and 30 (±5). Baseline was defined as a preoperative blood draw within a maximum of 7 days prior surgery. The normal range of alkaline phosphatase is between 40 and 130 U/L and is determined as a routine parameter at our department. To determine the plasma concentrations of alkaline phosphatase, enzyme kinetic measurements are performed on native or heparinized blood samples.

Follow-Up

To evaluate the individual postoperative outcome and adverse events, the local patient documentation system AKIM was used. Additionally, federal statistics (Statistics Austria, Vienna, Austria) was used to determine the survival rates.

Statistical Analysis

Categorical variables are presented by numbers and percentages. For continuous variables, the mean ± standard deviation (SD) is reported if the data follows a normal distribution; otherwise, the median and interquartile range (IQR) are provided. The normal distribution assumption is tested using the Kolmogorov-Smirnov test. The patients were grouped according to the initial AP drop.

AP drop = 1 – AP POD 1 / AP Baseline

Cut-offs were determined using a ROC analysis for 1-year mortality. Furthermore, the first day when alkaline phosphatase returned or exceeded baseline level was determined.

The Mann-Whitney-U-Test was used to assess differences between the continuous variables with non-normal distributions. For the comparison of categorical variables, the chi-square test was applied. Differences between AP at Baseline and AP POD 1 and AP POD 30, respectively, were determined using the Wilcoxon signed-rank test. The findings are visually presented using boxplots. Survival was visualized with Kaplan Meier Curves and the differences between the groups were tested using the log rank test. All analyses were performed using SPSS version 27.0 (IBM Corp, Armonk, NY, USA). A two-sided p-value of less than 0.05 was considered to indicate statistical significance.

5. Conclusions

A significant higher morbidity and an impaired outcome after a higher initial drop of alkaline phosphatase at 30-days, 1-year and in the long-term survival was evident. Future risk assessment scores for cardiac surgery should consider alkaline phosphatase.