Nonheart failure-associated elevation of amino terminal pro-brain natriuretic peptide in the setting of sepsis

Abstract

In addition to its importance in clinical assessment, N-terminal pro-brain natriuretic peptide (NT pro-BNP) is a valuable marker for evaluation of treatment and prognosis of heart failure. However, there are situations where NT pro-BNP is not related to myocardial dysfunction. Two cases of sepsis with markedly elevated NT pro-BNP levels that are not indicative of depressed myocardial function are described.

One-half of a century has been spent debating the etiology of depressed myocardial function in septic shock (1). Brain natriuretic peptide (BNP) is a new biomarker for cardiac dysfunction, and most data in the literature associate its increased levels with depressed myocardial function (2–5). However, increased levels of BNP have also been found in critical illnesses associated with shock (eg, sepsis), apparently not related to myocardial dysfunction. In human myocardium, BNP is secreted in response to myocardial stretch. Its release promotes natriuresis and diuresis, and it has a vasodilatory effect on systemic circulation (6). As such, it reduces preload, venous return and filling pressures with a subsequent reduction in cardiac output (7). The measured BNP increases in response to increases in left ventricular (LV) end-diastolic pressure, pulmonary artery wedge pressure, atrial pressure and LV hypertrophy (8). BNP level is also associated with echocardiographic dysfunction and clinical outcomes (9). The degree of cardiac dysfunction determines the magnitude of increase in BNP level (10).

The BNP gene contains an encoded protein consisting of 108 peptides, called pro-BNP (11). The cleavage of pro-BNP within the heart results in two components produced in equal amounts: BNP 32 (the active hormone) and N-terminal (NT) pro-BNP (the inactive fragment) (12). The plasma half-life of NT pro-BNP is approximately 2 h in comparison with that of BNP 32 (20 min) (13), which explains the higher level of NT pro-BNP in myocardial dysfunction (14).

In addition to the documented relationship between BNP and myocardial dysfunction mentioned above, there are other conditions associated with increased BNP level that include primary pulmonary hypertension, myocarditis, cardiac allograft rejection, arrhythmogenic right ventricle with decreased LV ejection fraction (EF), renal failure, Kawasaki disease, ascitic cirrhosis and endocrine diseases (Cushing’s syndrome, primary hyperaldosteronism) (8). Other patient characteristics that may influence the BNP level, independent of disease, are advancing age, probably reflecting LV subclinical abnormalities (15) and female sex, both of which result in increased BNP level (15,16). Two additional studies have confirmed the correlation of NT pro-BNP with age, but not with sex (17,18). In contrast, obesity is associated with decreased levels of BNP, probably as a result of increased clearance through a receptor contained in adipose tissue (19).

Parker et al (20) demonstrated that myocardial depression is also present in human septic shock and is associated with ventricular dilation and decreased LVEF (20). We present two cases of septic shock with markedly increased levels of NT pro-BNP with no evidence of myocardial dysfunction based on evaluation using echocardiography and catheterization. Also described are possible explanations of elevated NT pro-BNP in a setting not related to myocardial dysfunction. Thus, there are situations where the predictive value of BNP is not correlated with the magnitude of heart failure. Consequently, we emphasized the need for a comprehensive evaluation using clinical assessment in addition to measurement of NT pro-BNP.

CASE PRESENTATIONS

Patient 1

A 79-year-old man with documented hypertension and mitral regurgitation was admitted for mitral valve repair and pulmonary vein ablation therapy for chronic atrial fibrillation. Normal coronaries were documented angiographically. His LVEF before surgery was 65%. Mitral valve annuloplasty and pulmonary vein ablation were performed, and a dual-mode, dual-pacing, dual-sensing pacemaker was subsequently inserted following a Mobitz II AV block. During the follow-up period, the patient developed progressive dyspnea with minimal exertion. NT pro-BNP concentration was 3431 pg/mL (laboratory normal less than 855 pg/mL). Echocardiography revealed severe mitral regurgitation caused by a dehiscent mitral valve ring. He returned to the operating room to have a redo tissue mitral valve replacement. On the third postoperative day, his temperature increased to 37.9°C, and he developed acute renal failure with a creatinine concentration of 371 μmol/L (normal range 40 μmol/L to 120 μmol/L). A second NT pro-BNP concentration was 17,470 pg/mL. The postoperative course was complicated by aspiration pneumonia with Gram-negative Serratia that were revealed in blood cultures together with coagulase-negative Staphylococcus. The patient was intubated and treated with broad-spectrum antibiotics (vancomycin and ceftazidime). A new echocardiographic study showed an LVEF of 60%, moderate tricuspid regurgitation and trivial mitral regurgitation. There was no echocardiographic evidence of endocarditis. Pulmonary artery (PA) systolic pressure was 62 mmHg. The patient developed markedly increased work of breathing. Pulmonary capillary wedge pressure (PCWP) was 14 mmHg, cardiac index (CI) was 4.8 L/min/m², and systemic vascular resistance (SVR) was 339 dynes/s × cm⁻⁵.

Repeat NT pro-BNP was 9980 pg/mL. White blood cell count was 12.0 g/L (normal range 4 g/L to 11 g/L), and creatinine rose to 460 μmol/L (baseline 111 μmol/L). Chest x-ray showed extensive consolidation on the right lower lobe and left middle lung. The patient was diagnosed as having ventilator-associated pneumonia with Staphylococcus aureus (cultured in sputum) and, subsequently, septic shock. Therapy with antibiotics and inotropes were instituted. The patient was transferred to the intensive care unit with hypoxemic respiratory failure and hemodynamic compromise, and measures of ventilatory and inotropic support were reinstituted. Despite maximal therapy, progression of septic shock was irreversible, and multiple organ system failure followed.

Patient 2

An 83-year-old woman with a previous myocardial infarction and aortic stenosis was admitted for coronary artery bypass graft surgery and aortic valve replacement. A previous heart catheterization showed normal right heart pressures, with a PA systolic pressure of 31/12 mmHg, PCWP of 11 mmHg, and a CI of 2 L/min/m². Coronary angiography showed a right coronary artery stenosis of 90% and a proximal left anterior descending artery stenosis of 80%. She also had chronic renal impairment with a baseline creatinine concentration of 115 μmol/L and a glomerular filtration rate of 40 mL/min. Preoperative echocardiography documented an LVEF of 65%, mild mitral regurgitation, aortic valve area of 0.8 cm² with a mean gradient of 30 mmHg and LV hypertrophy. Coronary artery bypass graft surgery was performed together with a bovine pericardial bio-prosthetic aortic valve replacement with no postoperative complications. An intraoperative transesophageal echocardiogram showed normal left and right ventricular function with an LVEF of 70% and moderate mitral regurgitation. During the following days, she experienced episodes of atrial fibrillation with rapid ventricular response and sinus pauses of up to 8 s with subsequent insertion of a ventricle-paced, ventricle-sensed, inhibited, rate-responsive pacemaker. Escherichia coli was found in the urine culture and treated. During the postoperative period, she became progressively dyspneic with no symptomatic improvement despite escalating doses of diuretics (furosemide up to 280 mg/day intravenously and metolazone). Laboratory investigations included NT pro-BNP, extremely elevated at 28,000 pg/mL, that suggested significant myocardial depression. Objective examination revealed cool skin, a temperature of 35°C and blood pressure of 75/30 mmHg on milrinone; the heart was paced at 60 beats/min and the respiratory rate was 30 breaths/min. PA catheter measurement revealed a CI of 8.3 L/min/m², PA systolic pressure of 40/18 mmHg, PCWP of 18 mmHg and an SVR of 105 dynes/s × cm⁻⁵. The diagnosis of septic shock was apparent. An echocardiographic study showed a well seated aortic valve prosthesis, LVEF of 70%, aortic valve area of 1.2 cm², gradient of 17 mmHg, severe mitral regurgitation, moderate tricuspid regurgitation, PA systolic pressure of 50 mmHg and no evidence of endocarditis. Vasopressor therapy (noradrenaline) was initiated, and antibiotic therapy was started empirically (piperacillin-tazobactam/vancomycin) for treatment of presumed endocarditis. Despite maximal therapy with inotropes and antibiotics, the progression of septic shock was irreversible. A subsequent/postmortem blood culture revealed coagulase-negative Staphylococcus sensitive to vancomycin.

DISCUSSION

We described two patients with septic shock, markedly elevated levels of NT pro-BNP and normal myocardial function documented on complementary investigations (heart catheterization and echocardiography). We have also described possible explanations for elevated NT pro-BNP in this setting unrelated to myocardial dysfunction. Even though BNP has been considered an optimal marker for cardiac dysfunction, the lack of data to provide a rationale for its changes during sepsis suggests that the clinical assessment should be incorporated when evaluating patients with presumed myocardial depression.

An association of elevated BNP with sepsis in critically ill patients has been reported (21), but the cause of BNP increase is not well understood. Sepsis is a systemic response to a localized injury, and a very common cause of mortality in intensive care units (22). One of the most important consequences of septic shock is cardiac dysfunction. Both systolic (23) and diastolic impairment (24) are present during sepsis. Recent studies have demonstrated a correlation of the severity of cardiac dysfunction with mortality (9). In survivors of the septic shock, cardiac dysfunction is transient with EF and ventricular dilation returning to normal in seven to 10 days (23). BNP increase is also transient. In contrast, in patients who do not survive, the lack of ventricular dilation and preservation of EF indicates a poor compensatory mechanism (9). Our cases are unusual in that NT pro-BNP levels were markedly elevated, suggesting significant myocardial depression. Nevertheless, normal LV systolic function and filling pressures were identified. The recent ProBNP Investigation of Dyspnea in the Emergency Department (PRIDE) (15) study has reported three patients with elevated NT pro-BNP in the setting of septic shock. A level of NT pro-BNP greater than 10,000 pg/mL has confirmed the diagnosis of heart failure in more than 99% of newly diagnosed cases and in 94% of heart failure cases. Additionally, age-stratified cutoff points have been included to increase diagnostic accuracy of heart failure (Table 1) (15). The association of NT pro-BNP with New York Heart Association (15) symptom severity described in the PRIDE study is presented in Table 2. NT pro-BNP assessment seems to surpass clinical evaluation by direct comparison; however, we submit that a higher diagnostic yield will be achieved by using both clinical and laboratory evaluation (15).

TABLE 1.

Correlation of age with concentration of N-terminal pro-brain natriuretic peptide (NT pro-BNP)

	Younger than50 years of age	Between 50 and 75 years of age	Older than75 years of age
NT pro-BNP concentration (pg/mL)	<450	<900	<1800

Adapted from reference 15

TABLE 2.

Correlation of New York Heart Association (NYHA) class with N-terminal pro-brain natriuretic peptide (NT pro-BNP) concentration

NYHA class	II	III	IV
NT pro-BNP concentration (pg/mL)	1591	3438	5564

Adapted from reference 15

Both of our patients progressed to septic shock. An increased level of BNP and vasodilation in septic shock is a consequence of the cytokine production (9,25,26). Tumour necrosis factor-alpha and interleukin-1beta are important cytokines released during sepsis, which are associated with myocardial depression (25). It has been demonstrated that these proinflammatory cytokines upregulate the BNP gene expression and secretion (27). In addition, cardiotrophin-1, a member of the interleukin-6 family, increases the BNP messenger RNA expression in vivo (28). One of the mechanisms by which BNP exerts its vasodilatory effect is nitric oxide (NO) production (29). The basal NO production is correlated with the severity of cardiac dysfunction and is reflected in the levels of plasma BNP (30). An increased NO production in sepsis can have both beneficial and deleterious effects. Beneficial effects manifest by vasodilation in systemic and pulmonary territories and an increase in myocardial oxygen supply, as well as a preservation of the CI. The deleterious effects manifest as a decrease in blood pressure and SVR, concomitant with poorer tissue oxygenation (31). Additionally, BNP exerts a cardiac antihypertrophic action independent of the endothelial NO production in diabetic rats (32). In septic shock, Gram-negative organisms mediate myocardial depression by a mechanism that involves NO release (33). By a similar mechanism, S aureus alpha toxin produces coronary vasoconstriction and myocardial depression (34). An additional study (35) reveals that endotoxin, which also modulates BNP expression and secretion, synergizes with a component of pathogen and nonpathogen Gram-positive bacteria to cause the release of NO (36). NO has the ability to autoregulate its own production, and by negative feedback (37) may counteract myocardial depression (31) and vasoconstriction in sepsis (38). Therefore, small quantities of NO are necessary to maintain myocardial contractility in sepsis (31), and NO generation may represent a compensatory mechanism in sepsis-induced coronary vasoconstriction and myocardial dysfunction.

Our patients had Gram-positive and Gram-negative bacteria in cultures at different times during their illness and they did not survive, which is consistent with a prior report that demonstrated a lack of LV dilation in patients who did not survive (9). It has been suggested that critically ill patients who do not present LV dilation and reduction of EF in response to sepsis are more likely to die and less likely to have an elevated BNP. In contrast, patients with septic shock who survive have elevated BNP levels and reduced EF which normalize quickly once medically stable (9). Interestingly, in our two cases, NT pro-BNP was markedly elevated. However, no LV dilation or systolic dysfunction resulted. This may illustrate a potential mechanism for increased mortality in septic patients who do not exhibit expected compensatory hemodynamic changes.

There are limitations to our study in terms of the measurement of the NO and cytokine production. In addition, both of our patients developed renal failure during their sepsis, a condition known to modify the NT pro-BNP level through decreased clearance or associated subclinical structural heart disease (15).

Heart failure is a chronic medical condition with a very poor prognosis (39). Hence, there is an imperative need for a greater diagnostic accuracy. The literature is expanding suggesting that NT pro-BNP is not a stand-alone test for predicting heart failure and likely requires inclusion in a scoring system where clinical assessment remains important in the diagnosis and management of heart failure.