FormalPara Key Summary Points

Why carry out the study?

This study establishes an association between blood urea nitrogen (BUN) and prognosis of patients with infective endocarditis (IE).

It is one of the first studies to demonstrate the prognostic value of BUN in evaluating short-term mortality of patients with IE.

What was learned from the study?

BUN is suitable for independently predicting short-term mortality in patients with infective endocarditis, which could be a medical marker assisting risk stratification in patients with IE at the first clinical encounter.

Introduction

Infective endocarditis (IE) is a rare but life-threatening disease that involves the cardiac endothelium as the initial site of infection, extending to native valves, prosthetic valves, and even indwelling cardiac devices [1,2,3]. Over the past decades, advanced diagnostic technologies and therapeutic strategies have been explored for the management of IE. However, the prognosis of IE is poor with high mortality and severe complications [4]. Data have indicated that the annual incidence of IE has increased since 2000 with an average of 2–8 cases per 100,000 person-years [5,6,7,8]. In 2011, there were even 15 cases per a population of 100,000 [7]. Even with timely antimicrobial medication and diagnostic tools, the incidence of mortality remained high [5,6,7,8]. About 15–30% of patients with IE died in hospital, and the 1-year mortality of IE reached 30% [5, 9, 10]. Therefore, it is of great importance to explore the hematological and biochemical biomarkers with high specificity and sensitivity to facilitate the diagnosis and risk stratification of IE.

Blood urea nitrogen (BUN) is a metabolic product in the urea cycle after detoxifying the ammonia in the liver [11]. It circulates in the blood and is finally excreted through the kidneys. The formation of this non-toxic product avoids the toxic effect of ammonia from damaging the central nervous system. Moreover, BUN is also involved in the formation of concentrated urine and maintenance of the osmotic gradient in the nephrons. Therefore, level of BUN can somehow reflect kidney function. Studies have shown that BUN was an independent predictor in patients with acute heart failure, acute myocardial infarction, and acute gastrointestinal bleeding [11,12,13].

A similar conclusion was also drawn with respect to infectious state. It was reported that BUN can help predict the survival of patients infected with Escherichia coli [14] or coronavirus disease 2019 (COVID-19) [15]. As an infectious disease, IE is commonly complicated with heart dysfunction, which can also affect the normal renal function and increase the level of urea in the blood. However, rarely have there been studies further probing the prognostic value of BUN in patients with IE. We investigated the association of BUN and short-term mortality in patients with IE, which will help promote faster triage of patients with IE.

Methods

Study Population

This was a retrospective study conducted in Guangdong Provincial People’s Hospital, Guangzhou, China. The need for informed consent was waived given the retrospective nature of the study, but an oral informed consent would be provided during follow-up (GDREC2020098). A total of 2634 consecutive patients with diagnosis of IE were selected from January 2009 to February 2020. Selected patients were diagnosed with IE according to the pathologic or clinical criteria based on the modified Duke criteria [3]. Patients with (i) age < 18 years; (ii) noninfective vegetation; (iii) nosocomial IE; (iv) possible IE; (v) missing BUN data and (vi) readmission records were excluded. Finally, 1371 patients were enrolled (Fig. 1).

Fig. 1
figure 1

The flow chart for patients’ screening

Full size image

Data Extraction

Data were collected from the hospital’s electronic system by trained study coordinators. Clinical information of patients including demographic features (age, sex, symptoms for duration of longer than 1 month before admission), past medical history (history of dialysis, injection drug usage), comorbidities (hypertension, diabetes mellitus, heart diseases), laboratory results [complete blood count, C-reactive protein (CRP), serum creatinine, estimated glomerular filtration ratio (eGFR), blood microorganism culture (Staphylococcus, Streptococcus), and blood urea nitrogen (BUN)], New York Heart Association (NYHA) classification, ultrasonographic results [left ventricular ejection fraction (LVEF), heart valve deformities/vegetation (aortic valve, mitral valve)], and type of treatment (surgery, conservative therapy) were extracted and evaluated. The normal range of BUN in the hospital laboratory was 3.60–9.50 mmol/L. Prognostic events including in-hospital mortality and events were also collected and recorded by two independent researchers. The residual clinical data was recorded by one researcher and randomly checked by another.

Study Endpoint

All enrolled patients were followed up for 6 months either telephonically, as outpatients, or via hospital admission. The primary endpoint was all-cause mortality during hospitalization. The secondary endpoint was all-cause mortality after the 6-month follow-up, which was determined by the date of death from date of hospital admission.

Statistical Analysis

All data were processed using R-software (version 3.6.2; http://www.R-project.org) and SPSS 22.0 (IBM Corporation, Armonk, NY, USA). Analysis of variance (ANOVA) was used to compare normally distributed continuous data that were presented as mean ± standard deviation. H-test was used for comparison of non-normally distributed continuous data that were presented as median and interquartile range. The chi-squared test was used to compare categorical data that were presented as percentages. After analyzing the BUN’s odds ratio (OR) curves for mortality, we identified the threshold of BUN at the points where risk of mortality ceased to decline or started to rise steeply, if any. For convenient clinical application, the nearest integer was selected. The association between variables and the 6-month mortality were assessed by Cox proportional hazard analyses. Significant variables in the univariate regression analysis were inputted into the multivariate regression analysis. Receiver operating characteristics curve (ROC) analysis and the Youden index were applied for prediction of the in-hospital mortality with the greatest sensitivity and specificity. The OR and 95% confidence interval (CI) were calculated. The cumulative 6-month mortality in patients with different levels of BUN was evaluated through a Kaplan–Meier curve and compared using the log-rank test.

Results

Baseline Characteristics

A total of 1371 patients (mean age: 45 ± 16 years old, 30.2% female) were finally enrolled in the present study and the baseline characteristics were listed in Table 1. Overall, about 59.2% patients presented with symptoms for a duration longer than 1 month before admission to the hospital, and 70.7% patients received surgical therapy on basis of antimicrobial therapy. The studied population was classified into four groups with BUN: < 3.5 mmol/L (n = 343), 3.5–4.8 mmol/L (n = 343), 4.8–6.8 mmol/L (n = 341), and ≥ 6.8 mmol/L (n = 344). According to the demographic features of the population, groups with higher BUN level were found with more elderly ages and with a greater proportion being male. Patients with history of previous dialysis were only found in the group of BUN level over 6.8 mmol/L. Patients with higher BUN level were more commonly found with history of hypertension and diabetes than patients with lower BUN level. Among the 1371 patients, native-valve IE was the most dominant type of infection, in which percentages of patients with congenital heart disease were obviously higher than those with rheumatic heart disease. However, patients with either rheumatic heart disease or congenital heart disease showed no significant difference in groups with different level of BUN. Patients with higher BUN level were more commonly found with decreased left ventricular ejection fraction (LVEF) (65.6% versus 65.7% versus 65.0% versus 62.8%, P < 0.001). Moreover, patients with increased BUN level were commonly found with exercise intolerance or the presence of vegetation, especially in either the aortic or the mitral valve. Patients with BUN ≥ 6.8 mmol/L had the highest serum creatinine (62.0 versus 71.8 versus 81.0 versus 119.8 μmol/L, P < 0.001). However, patients with lower BUN had a higher percentage of Streptococcus infection compared with patients with higher BUN (44.9% versus 43.7% versus 40.8% versus 19.8%, P < 0.001).

Table 1 Baseline characteristics of included patients
Full size table

BUN and In-Hospital Events

During hospitalization, 101 (7.4%) patients died. The observed in-hospital mortality rate went higher as BUN level ascended. The incidence of acute heart failure and renal replacement treatment during hospitalization were significantly higher in patients with higher BUN level. An approximately linear correlation could be observed between the level of BUN and risk of in-hospital death (Fig. 2). After adjustment, the risk of in-hospital mortality increased significantly when the level of BUN exceeded 5 mmol/L. ROC curve analysis revealed that BUN showed good discrimination for prediction of in-hospital mortality [area under the curve (AUC) = 0.705], and 6.8 was the optimal cut-off value for BUN to predict in-hospital mortality with a sensitivity and specificity of 56.4% and 77.4%, respectively (Fig. 3). In multivariate logistic regression analysis, BUN higher than 6.8 mmol/L was analyzed as an independent predictor of higher risk of in-hospital mortality in patients with IE infection, after adjustment for age, symptom duration for longer than 1 month before admission, diabetes, prosthetic valve, NYHA classification, elevated CRP, lower eGFR, LVEF, aortic and mitral valve vegetation, and surgical treatment (aOR = 2.365, 95% CI 1.292–4.328, P = 0.005; Table 2).

Fig. 2
figure 2

The association of BUN with in-hospital death. a Unadjusted, b Adjusted

Full size image
Fig. 3
figure 3

The receiver operating characteristics curve (ROC) analysis of BUN for predicting in-hospital death

Full size image
Table 2 Univariate and multivariate logistic regression analysis for in-hospital mortality
Full size table
Table 3 Univariate and multivariate Cox regression analysis for 6-month mortality
Full size table

BUN and 6-Month Mortality

In total, 1273 (92.9%) patients finally completed the 6-month follow-up, with a 6-month mortality rate of 9.8%. The Kaplan–Meier curve indicated that the cumulative 6-month mortality was significantly higher in patients with BUN > 6.8 mmol/L than others (log rank = 97.9, P < 0.001, Fig. 4). Univariate Cox regression analysis revealed that compared with BUN in other groups, patients with BUN > 6.8 mmol/L may have higher risk of 6-month death. With adjustment of potential confounders, it was demonstrated that BUN > 6.8 mmol/L could independently predict the 6-month mortality in patients with IE (aHR = 2.171, 95% CI 1.355–3.479, P = 0.001, Table 3).

Fig. 4
figure 4

Cumulative rate of 6-month mortality according to different BUN levels

Full size image

BUN and Potential Risk Factors

To identify the risk factors associated with high BUN level, both univariate and multivariate logistic regression analysis were further carried out between the potential factors and BUN higher than 6.8 mmol/L (Table 4). The results showed that age (aOR = 0.976, 95% CI 0.964–0.990, P < 0.001), female (aOR = 0.452, 95% CI 0.295–0.692, P < 0.001), NYHA class III–IV (aOR = 2.124, 95% CI 1.501–3.007, P < 0.001), LVEF (aOR = 0.977, 95% CI 0.956–0.999, P = 0.038), and Streptococcus infection (aOR = 0.470, 95% CI 0.317–0.697, P < 0.001) were independently associated with high BUN level.

Table 4 Univariate and multivariate logistic regression analysis for high BUN level (> 6.8 mmol/L)
Full size table

Discussion

To the best of our knowledge, the present study is the first exploration of the association between BUN and short-term mortality in patients with IE. Our results showed that BUN had a good discrimination for in-hospital death. We also discovered that higher BUN level is related to higher possibilities of suffering from cardiac and renal insufficiency, even death. Moreover, BUN > 6.8 mmol/L was found to be an independent predictor of in-hospital and 6-month mortality in patients with IE. Therefore, BUN can be a prognostic indicator for patients diagnosed with IE.

IE is a lethal cardiovascular disorder initiated by the injury of valvular endothelium or endocardium that subsequently forms infective thrombi with bacterial colonization [2]. With the expansion of bacterial colonization, IE can be further complicated with valvular destruction or even heart failure, vascular embolization, metastatic infection, and immunological dysfunction [2]. In our study, we observed more patients with native-valve infection than those with prosthetic-valve infection. Streptococcus was the main causative pathogen accounting for 37.2% of the IE population, followed by Staphylococcus infection accounting for 12.3% of the IE population, which were similar to the results of epidemiological studies of patients with IE in China from 2001 to 2018 [16, 17]. However, according to a systematic review aiming to analyze the global situation of IE from 1990 to 2010 [18], native-valve infection was not so common in that approximately 2–10 cases per 100,000 person-years were diagnosed [2, 18]. Staphylococcus was still the predominant causative pathogen of IE in countries including France and the USA [2, 10, 18]. It was speculated that the difference of incidence may be attributed to the epidemiological features shifting towards more advanced ages, more utility of prosthetic valves, implantable cardiac devices, indwelling catheters, and hemodialysis in developed countries, but less utility of implantable devices and late intervention of cardiac deformities in China [2, 17, 19,20,21]. Besides, our study found that patients with Streptococcus infection had low BUN level (P < 0.001), which might be explained by some type of Streptococcus-containing urease to further catabolize urea [22].

In the present study, with increased BUN level, patients with IE showed significantly increased incidence of acute heart failure and renal replacement treatment, supported by gradually decreased LVEF and increased creatinine, respectively. ROC analysis and Youden index revealed that BUN at the level of 6.8 mmol/L showed good discrimination for prediction of in-hospital mortality. With increased previous studies focusing on BUN, it has been validated as a prognostic indicator in many situations, including acute myocardial infarction [11, 23], acute and chronic heart failure [12, 24], gastrointestinal bleeding [13], acute stroke [25], and acute pulmonary embolism [26]. Our study showed that the in-hospital mortality rate of patients with IE reached 7.3%, which is similar to that of an epidemiological study carried out in China from 2007 to 2016 (7.3% versus 11.3%) [16]. BUN was an independent predictor in patients with IE for both in-hospital and 6-month mortality, even after adjustment for age, medical history, and cardiac function. The pathophysiological mechanism underlying the association remains to be elucidated. With bacterial colonization in the cardiac configuration, the cardiac electronic system and structure could be affected, thereby leading to cardiac arrhythmia and abnormal pumping, eventually affecting the normal blood supply for essential organs, including the kidneys [2, 11]. With decreased cardiac output, reduced blood flow through the renal vasculature can activate the renin–angiotensin–aldosterone system (RAAS), thus releasing renin and causing renal vasoconstriction. The vasoconstriction can decrease eGFR, thus reducing the excretion of BUN [11, 27]. Furthermore, vasoconstriction increases the renal tubular reabsorption, by which urea is transferred back to the circulation through proximal convoluted tubules [11, 13, 27, 28]. Meanwhile, the RAAS further stimulates the sympathetic nervous system (SNS), causing expansion of the same effect. With volume restriction, aldosterone is secreted in the RAAS axis and creates volume expansion by increasing sodium and water reabsorption, causing passive reabsorption of BUN in the distal tubules [11, 13]. Additionally, neurohormonal axis is also involved in the urea reabsorption, mediated by arginine vasopressin (AVP) activation of urea receptors distributed in the collecting ducts, thus leading to increased serum BUN overall [27, 28].

Traditionally, a lot of previous studies have demonstrated the prognostic value of CRP in evaluating the severity of infectious diseases, but the biomarker was gradually questioned as it was easily affected by some non-infectious inflammation or neoplasms, thus calling for more biomarker exploration [29]. Recently, many reports have approved the use of the prognostic value of BUN in some infectious states, including infection with Escherichia coli [14], COVID-19 [15], and septic shock [25, 30]. Li et al.’s study even indicated that BUN was correlated with the infection marker including procalcitonin (PCT), high-sensitive C-reactive protein (hsCRP), and neonatal Sequential Organ Failure Assessment (qSOFA), which was demonstrated as an independent predictor of the presence and severity of neonatal sepsis [30]. It was hypothesized that patients with sepsis are prone to have higher basal rate of BUN production and increased BUN was related with systemic inflammation [30]. In our study, patients with increased BUN level were prone to have elevated CRP. We could also find a prognostic value of elevated CRP for either in-hospital or 6-month mortality. However, it was clearly found that only about 55.9% patients with IE infection presented with an elevated CRP. Elevated CRP showed no association with higher BUN. We contributed this phenomenon to the reason that most of the patients in our study were transferred from community hospitals, who may have presented with symptoms for a duration longer than 1 month before admission. Therefore, compared with CRP, it was speculated that the variation of BUN may not be affected easily by the duration of disease progression as well as previous therapy, and thus may be a better biomarker for IE infection evaluation.

Limitations

Limitations exist in our study. First, the present study was a retrospective analysis and inherent bias was unavoidable, although efforts were made to balance the cofounding factors through multivariate analysis. Second, the study was a single-centered study with limited sample size. Third, we only concentrated on the prognostic value of BUN in patients with IE in short-term evaluation and only the BUN values during hospitalization admission were analyzed. However, as IE is essentially a lethal cardiovascular disease, long-term follow-up should be carried out with experienced clinical professionals to extend patient survival and promote their health status and overall quality of life.

Conclusion

BUN is validated as an independent predictor of short-term mortality in patients with IE. The validation could help fasten the initial triage and risk stratification in the clinical situation and assist clinicians with quick diagnosis and management, thereby improving survival.