B-Type Natriuretic Peptide in Patients with Sepsis and Preserved Left Ventricular Ejection Fraction

B-type natriuretic peptide (BNP) as a marker of left ventricular end-diastolic pressure 1 has proven to be a useful diagnostic tool to differentiate dyspnea caused by congestive heart failure (CHF) from non-cardiac dyspnea in patients presenting at the emergency room 2,3. It has been demonstrated that a normal BNP level (<80 to 100 pg/mL) has a high negative predictive value to exclude CHF 2,3, whereas a markedly elevated BNP has a high positive predictive value for CHF 4. Therefore, BNP testing seems to be very useful not only for emergency room settings, but also for the intensive care unit (ICU), where a rapid assessment of the cause of a patient's worsening oxygenation is required. However, while introducing BNP testing in our ICU we have observed surprisingly high BNP levels in critically ill patients with echocardiographically preserved left ventricular ejection fraction (LVEF). From a prospectively conducted registry, patients with sepsis admitted to our medical ICU with LVEF>50% in absence of regional wall abnormalities were assessed over a period of 10 months. Patients were included in the present analysis if they had sepsis according to the current definitions 5, and BNP measurements were provided within 36 h of echocardiographically assessed LVEF. Patients with any history of heart disease were excluded from the sepsis group. They were compared to patients hospitalised due to CHF and echocardiographically assessed severely impaired LVEF (LVEF<35%, systolic CHF group, S-CHF group) and BNP measurement within 36 h before or after echocardiography. Patients with acute coronary syndrome, valvular or congenital heart disease, patients after mechanical resuscitation, cardiac surgery, percutaneous coronary intervention, and those with severe renal failure (estimated creatinine clearance <30 mL/min at any time during hospitalisation) were excluded. For BNP testing 3 to 5 ml of blood from a peripheral vein or a central line were drawn into an ethylene-diamin-tetra-acetate containing tube. B-type natriuretic peptide was measured on a daily basis using the Triage B-Type Natriuretic Peptide test (Biosite Diagnostics Inc., San Diego, California). For each patient clinical and laboratory variables for exactly the time of BNP measurement were extracted. Eight patients with sepsis (5 women) and 8 S-CHF patients (1 woman) were included to the study (age 50.9+17.4 vs. 71.5+9 years; p=0.01) (Table 1). In the S-CHF group the underlying heart disease was coronary artery disease (n=4), hypertensive heart disease (n=3), or dilated cardiomyopathy (n=1), whereas no patient of the sepsis group had a history of either heart disease or hypertension. According to the inclusion criteria LVEF was higher in the sepsis group (63.1+4.6 vs. 23.6+5.3%; p<0.0001). B-type natriuretic peptide values were similar in both groups (fig. Fig. 1). All of the S-CHF patients were treated with daily angiotensin-converting-enzyme inhibitors or angiotensin-receptor blockers and loop diuretics, but none of them was intubated or received positive inotropic drugs. Seven patients with sepsis fulfilled criteria for septic shock 5 and were given norepinephrine (NE) in dosages up to 100 mcg/min during their ICU stay. While BNP measurements were done NE was administered in 5 patients. Mean arterial pressure and heart rate were similar in both groups. Central venous pressure measurements were performed only in the sepsis group. Six patients of the sepsis group were mechanically ventilated due to acute respiratory distress syndrome (ARDS, n=4), pneumonia (n=1), or impaired consciousness (n=1) while BNP measurements were done. Patients of the sepsis group had a cumulative positive fluid balance of 3675+5577 mL between admission to the ICU and BNP measurement, whereas S-CHF patients had lost 1362+1792 mL. Four sepsis patients suffered from acute pathologies of the central nervous system (2 meningitis, 1 ischemic stroke, 1 intracerebral hemorrhage) compared to no patient of the S-CHF group. In-hospital mortality showed no significant difference between patients with sepsis (2/8 died) as compared to patients with S-CHF (0/8 died; p=0.46). In the present study we have for the first time demonstrated that ICU patients with sepsis and echocardiographically preserved LVEF may have BNP levels as high as patients admitted due to severe systolic pump failure. The source of massive BNP release in our patients with sepsis is not evident, and several possibilities have to be considered. Previously published data suggest that a BNP level >400 pg/mL is highly predictive for CHF, meaning left heart failure 4. In patients with sepsis the heart often undergoes substantial functional changes due to the cytokine-mediated “sepsis-induced myocardial depression” resulting in biventricular dilatation and depression of ventricular function, which is detectable by echocardiography or radionuclide ventriculography 6. At the time of echocardiography our patients obviously did not suffer from overt sepsis-induced cardiac depression. However, minor changes in ventricular contractility may have escaped due to lowered systemic vascular resistance and left ventricular unloading respectively 6. Witthaut and co-workers recently reported that elevated BNP levels in patients with septic shock were inversely correlated to cardiac index 7. However, they did not assess LVEF, and a comparison with our data is therefore not possible. In contrast to the lowered systemic vascular resistance, pulmonary vascular resistance in patients with sepsis often increases due to associated acute lung injury (ALI) or ARDS leading to right ventricular (RV) overload, possibly causing BNP release. Acute cor pulmonale develops in 25% of patients with ARDS even in those submitted to protective ventilatory support 8. Six of eight sepsis patients were mechanically ventilated, which might have caused at least transient RV overload inducing BNP release, as has been described in pulmonary hypertension 9. Moderately elevated BNP levels in patients with acute lung injury have previously been reported 10. Furthermore, transient increase in left ventricular and RV afterload due to systemic and pulmonary vasoconstriction following high-dose NE administration 11 may account for BNP release. On the other hand, NE has been shown even to enhance cardiac index in patients with septic shock due to its β-adrenergic properties 11 possibly leading to overestimation of LVEF in the sepsis group. In addition, fluid resuscitation in the sepsis group may have influenced BNP levels by rising filling pressures, whereas ongoing diuretic treatment may have attenuated very high BNP levels in the S-CHF group. But it could also be speculated that marked hypalbuminemia in the sepsis group may have caused a fluid shift into the interstitium rather than a substantial increase of left ventricular preload and myocyte stretch. A recent study reported a weak correlation between creatinine and BNP levels in critically ill patients 12. Due to the fact that there was a trend to higher creatinine clearance values in the sepsis group renal dysfunction is unlikely to be a confounder for BNP elevation in the sepsis group. Interestingly, in the sepsis group we found 4 patients with acute disease of the central nervous system. B-type natriuretic peptide, initially called “brain natriuretic peptide” after its detection in the porcine brain 13, was previously reported to be elevated in patients with subarachnoid hemorrhage 14. Recent data suggest that brain diseases other than subarachnoid hemorrhage are not associated with BNP elevation 12, and high BNP levels in our 4 sepsis patients can probably not be attributed to the acute disease of the central nervous system. Although it is not possible to precisely define the stimulus and the source of BNP elevations in the presented septic patients, this phenomenon has to be taken in account while interpreting BNP levels in critically ill patients. BNP has a half-life of about 20 min and has been shown to vary up to 50% within 24 h following diuretic or vasodilator therapy for decompensated CHF 15. Since the maximal time interval between BNP assessment and echocardiography was allowed to be 36 h significant changes of BNP levels and left ventricular function could have occurred. Furthermore, since parameters of diastolic function have not been systematically assessed, we can not exclude the presence of non-systolic CHF (NS-CHF), which might be associated with high BNP levels 16. Risk factors for NS-CHF primarily include hypertension, older age and female sex, from which only the latter was present in the sepsis group. However, abnormalities in diastolic function have been reported in septic patients with preserved LVEF 17. Due to the retrospective study design no dilutation of BNP concentrations was performed to further assess BNP levels >1300 pg/mL. Four patients in the sepsis group had levels >1300 pg/mL compared to 3 patients of the S-CHF group. Therefore, raw data of BNP levels are presented for visualisation (fig. Fig. 1).