Brain-type Natriuretic Peptide Secretion Following Febrile and Afebrile Seizures—A New Marker in Childhood Epilepsy?


Address correspondence and reprint requests to Dr. Gerhard Luef, MD, Department of Neurology, Medical University Innsbruck, Anichstrasse 35, 6020 Innsbruck, Austria. E-mail:


Summary: Purpose: Markers for epileptic seizures are rare and their use has not been established in the evaluation of seizures and febrile convulsions (FC). Brain-type natriuretic peptide (BNP) is a natriuretic, diuretic, and vasodilator compound first discovered in the hypothalamus but mainly synthesized in the myocardium. The aim of this study was to assess whether epileptic seizures or FC are related to increased secretion of the N-terminal fragment of BNP (NT-proBNP).

Methods: Sixty-five postictal children (43 boys, 22 girls) and 31 children with epilepsy (20 boys, 11 girls) after a seizure-free period for at least 2 months serving as controls were enrolled. Postictal NT-proBNP levels were analyzed and controlled 24–48 h thereafter.

Results: Plasma concentration of NT-proBNP was significantly higher 4 h postictal compared to 24–48 h postictal (p < 0.001). Subgroup analysis revealed increased NT-proBNP levels in children with tonic–clonic seizures and FC compared to children with partial motor seizures (p < 0.001), syncope (SYN; p < 0.01), or control population (p < 0.001).

Conclusions: Our results suggest that elevated plasma NT-proBNP levels are not specific for cardiac dysfunction. Postictal measurement of plasma NT-proBNP seems to be useful in discriminating different types of epilepsy, FC, and SYN in childhood.

Markers for epileptic seizures are rare (e.g., prolactin, neuron-specific enolase (NSE), s-100 protein), their use has not been established in the evaluation of seizures and febrile convulsions (FC); (Malkowicz et al., 1995; Buttner et al., 1999). Atrial natriuretic peptide (ANP), Brain-type natriuretic peptide (BNP), and C-type natriuretic peptide (CNP) have beneficial, compensatory actions including vasodilatation, natriuresis, growth suppression, and inhibition of both the sympathetic nervous system and the renin–angiotensin–aldosterone axis (de Lemos et al., 2003). Measurement of both BNP and NT-proBNP has been shown to be useful in detecting left-ventricular dysfunction (McMurray et al., 1998), particularly after acute myocardial infarction, and is related to poor outcome (Wang et al., 2004). Elevated NT-proBNP levels are associated with major cardiovascular events (nonfatal myocardial infarction, fatal coronary heart disease, unstable angina, heart failure, stroke, and transient ischemic attack) and provide prognostic information of mortality (Kistorp et al., 2005). ANP is synthesized primarily in the cardiac atria, BNP in both the cardiac atria and the ventricles (Minamino et al., 1988). They are also synthesized in the human brain, though the role of natriuretic peptides within the brain and the central nervous system is still unclear (Takahashi et al., 1992). The hypothalamus, hippocampus, and cerebral cortex were found to be the major sites of cerebral NP release (Franzoni et al., 1992; Takahashi et al., 1992). Berendes et al. (1997) suggested that BNP is the major natriuretic peptide responsible for the hyponatremia observed in patients with subarachnoidal hemorrhage and reported an association between plasma levels of BNP and the development of delayed neurological deficits. Moreover, plasma levels of ANP and BNP are elevated in the acute phase of stroke and predict poststroke mortality. High plasma levels of natriuretic peptides predicted mortality after stroke better than any other risk variable, with the risk of death being fourfold among the patients with high NP value (Makikallio et al., 2005).

Little attention has been drawn on NP secretion in epileptic seizures. Status epilepticus, induced in anesthetized and paralyzed rats, showed a conspicuous increase in urine flow, sodium excretion, and plasma ANP (Perrone et al., 1995). Furthermore, altered ANP and BNP secretions caused by an epileptic attack in a 65-year-old woman with partial epilepsy and frequent bouts of polyuria probably were triggered by epileptic stimulation on the diencephalon beyond the focus (Obi et al., 2002). To our knowledge, this study is the first to investigate NP secretion in children with different types of epilepsy, FC, and syncopes (SYNs).

The aim of this study was (i) to assess whether epileptic seizures or FC are related to increased secretion of NT-proBNP and (ii) to evaluate if a subclassification of epileptic seizures by means of NT-proBNP measurement is possible.



All children with FC, tonic–clonic seizures (TCS), partial motor seizures (PS), or SYN, admitted to our emergency room during a 6-month period were eligible to participate in the study. Patients with signs of cardiac, hepatic, renal, or endocrine impairment were excluded. The seizure population comprised 65 postictal children (43 boys and 22 girls). The control population (20 males and 11 females) was randomly selected from children with epilepsy after a seizure-free period of at least 2 months without any history of cardiac abnormalities (e.g., ventricular septum defect, atrial septum defect), arrhythmias, renal and pulmonary disease. Seizures were classified according to the guidelines of the International League against Epilepsy (Proposal for revised classification of epilepsies and epileptic syndromes. Commission on Classification and Terminology of the International League Against Epilepsy, 1989). Thirty-three patients were diagnosed as having FC with symmetric generalized tonic–clonic convulsions. Twenty-five children thereof had typical, eight reached the criteria for prolonged FC. Sixteen children with partial epilepsy had PS, 12 children with idiopathic generalized epilepsy had TCS without pyrexia, and four children suffered from SYN. The control population consisted of 11 children with PS, 19 children with TCS, and 1 patient with SYN (Table 1). All patients and controls underwent a comprehensive clinical and cardiological evaluation, standard electrocardiogram (ECG), blood pressure measurement, electroencephalogram (EEG), and biochemical analysis. In patients with high NT-proBNP levels, echocardiography was performed. Moreover, in all patients with generalized and partial seizures, prolonged FC, and in all control patients MRI was done. The study was approved by the Ethical Committee of the Medical University Innsbruck.

Table 1. Baseline clinical characteristics of the study populations
 Patients (n = 65)Seizure-free controls (n = 31)
  1. n, number of subjects; TCS, primary and secondarily generalized tonic–clonic seizure; PS, partial seizure; FC, febrile convulsion; AED, antiepileptic drug; OXC, oxcarbazepine; PB, phenobarbital; VPA, sodium valproate; SUL, sultiame; LTG, lamotrigine; LEV, levetiracetam; SUX, ethosuximide; TPM, topiramate.

Age, years (mean ± SD)4.1 ± 4.06.5 ± 3.7
Sex (female/male)43/2220/11
Types of seizures, n (%)
 PS16 (16.5)11 (35.5)
 TCS12 (12.4)19 (61.3)
 FC typical25 (38.5) 
 FC prolonged 8 (12.3) 
 Syncope4 (4.1)1 (3.2)
AED treatment, n (%)20 (100)27 (100) 
 SUL1 (5) 3 (9.7)
 LTG1 (5) 1 (3.2)
 VPA8 (40)13 (41.9)
 OXC3 (15)3 (9.7)
 PB4 (20) 5 (16.1)
 TPM1 (5) 1 (3.2)
 LEV2 (10) 
 SUX 1 (3.2)
Duration of therapy, months (mean ± SD)


This is a prospective cohort study. Age, weight, blood pressure, heart rate, antiepileptic drugs, and number of seizures experienced from patients were recorded at admission to our hospital and again at discharge. Blood samples for NT-proBNP and biochemical analyses were drawn within 4 h after seizure onset and 24–48 h later, according to NT-proBNP half-life. Each sample of whole blood (1 ml) was assayed within 1 h thereafter. In control children, NT-proBNP and biochemical parameters were analyzed and age, weight, blood pressure, heart rate, antiepileptic drugs, and number of seizures experienced were recorded during a routine examination.

Analytical methods

Plasma concentrations of amino-terminal fragments of BNP were measured using an electrochemiluminescence immunoassay with the Elecsys system 1010/2010, using the proBNP kit (Roche, Mannheim, Germany). Elecsys proBNP contains polyclonal antibodies that recognize epitopes located in the N-terminal part (1–76) of pro BNP (1–108). The assay range is 5–35,000 pg/ml. The assay is unaffected by icterus (bilirubin < 35 mg/dl), hemolysis (Hb < 1.4 g/dl), or lipemia (TG < 4,000 mg/dl). No cross-reactivity (<0.0001%) was observed with ANP, NT-proANP, BNP, CNP, adreomedullin, aldosterone, angiotensin I, II, III, endothelin, renin, urodilatin, or Arg-vasopressin. The plasma half-life of NT-proBNP is about 2 h, reference values for children have been recently published (Mir et al., 2002; Mir et al., 2003; Nir et al., 2004; Albers et al., 2006).

Urea, creatinine, sodium, potassium, calcium, aspartate-aminotransferase (ASAT), alanine-aminotransferase (ALAT), creatine kinase (CK), glucose, and leukocyte counts were measured with standard methods.


Data obtained were analyzed by SPSS for Windows (SPSS, Inc., Chicago, IL, U.S.A.). NT-proBNP concentration was tested for normal distribution. Since NT-proBNP did not fulfill the criteria for normal distribution, univariate comparisons of NT-proBNP between groups were done using Kruskal–Wallis test. In order to control for a potential age confounding, logistic regression analyses were performed to adjust the effect of NT-proBNP for age. Regression analyses were calculated on logarithmic-transformed data. Wilcoxon test was used to compare repeated measures within seizure group. Multiple comparisons were performed using Bonferroni type I error correction. Spearman's correlation was used to analyze the relation between NT-proBNP levels and duration of seizure, body temperature, and number of seizures experienced. In addition, partial correlation coefficients adjusted for age were calculated. A value of p < 0.05 was considered statistically significant.


A total of 65 children (43 boys and 22 girls) fulfilled the inclusion criteria of the study. Baseline characteristics of postictal children and controls are presented in Table 1. The mean plasma concentration of NT-proBNP was 235.62 ± 239.8 pg/ml (median: 130.0 pg/ml; range: 13–1,056) 4 h postictal and 138.2 ± 125.3 pg/ml (median: 83.0 pg/ml; range: 13–509) 24–48 h postictal. In the control population NT-proBNP plasma levels reached only 74.26 ± 48.8 pg/ml (median: 62.0 pg/ml; range: 17–187). Differences in NT-proBNP levels 4 h postseizure versus controls were significant either in univariate comparisons (p < 0.001) as well as in age-adjusted analyses (p = 0.008). Subanalyses revealed significant differences between TCS, PS, FC, SYN (4 h postictal), and controls. NT-proBNP plasma concentrations were significantly higher among patients with TCS (mean: 262.9 ± 217.6 pg/ml; median: 219.5 pg/ml, range: 84–887) and FC (mean: 338.1 ± 255.6 pg/ml; median: 269.0 pg/ml, range: 72–1,056) compared to patients with PS (mean: 51.4 ± 35.8 pg/ml; median: 45.5 pg/ml, range: 13–154), SYN (mean: 45.0 ± 44.7 pg/ml; median: 24.0 pg/ml, range: 20–112) and controls (Fig. 1). Differences between subgroups and controls remained significant (p < 0.01) after age adjustments. Comparison of NT-proBNP levels between children with PS, SYN and controls as well as NT-proBNP levels in children with FC compared to children with TCS did not reach statistical significance (Fig. 1). A statistically significant decrease of NT-proBNP levels in children with TCS and FC was observed after the 24–48 h follow-up (Figs. 2 and 3). NT-proBNP plasma levels of children with prolonged FC were significantly higher compared to children with typical FC (p < 0.001). Univariate correlation analyses of NT-proBNP with age, sex, duration of seizures, heart rate, and body temperature are given in Table 2. The corresponding age-adjusted partial correlation coefficients revealed a significant correlation only for NT-proBNP levels with body temperature (R = 0.336, p = 0.034). No significant relationship between NT-proBNP levels and the number of FC experienced was observed (R = 0.245, p = 0.162). No sex differences in NT-proBNP plasma concentrations were found in both groups. Correlation analysis did not reveal any significance between NT-proBNP levels and antiepileptic medication, dose, and serum concentration in both groups.

Figure 1.

NT-proBNP concentrations in postictal children with generalized tonic–clonic seizures (TCS), febrile convulsions (FC), partial motor seizures (PS), and syncope (SYN) compared to controls. The box plots show median, interquartile values, and 10th and 90th percentiles. *Significant differences compared with TCS and FC at the level of p < 0.01. **Significant differences compared with TCS and FC at the level of p < 0.001.

Figure 2.

NT-proBNP concentrations in postictal children with generalized tonic–clonic seizures 4 h postictal compared to 24–48 h postictal. Box plots show median, interquartile values, and 10th and 90th percentiles.

Figure 3.

NT-proBNP concentrations in postictal children with febrile convulsions 4 h postictal compared to 24–48 h postictal. Box plots show median, interquartile values, and 10th and 90th percentiles.

Table 2. Correlation analyses of NT-proBNP with age, sex, duration of seizure, heart rate, and body temperature
 NT-proBNP PatientsControls
 p-value<0.001 0.04
 R−0.092  0.128
 p-value  0.467  0.492
Duration of seizure
 R  0.305 
 p-value  0.033 
Heart rate (postictal)
 R  0.531 
Body temperature
 R  0.626 

Heart rate, serum sodium, and leukocyte counts were found to differ significantly between patients and controls (Table 3). Moreover, analysis of blood pressure, ALAT, ASAT, urea, creatinine, sodium, potassium, and calcium between patients and controls did not reveal any statistical significance (Table 3).

Table 3. Laboratory findings, blood pressure, and heart rate of the study populations
 4 h Postictal (mean ± SD)24–48 h Postictal (mean ± SD)Seizure-free controls (mean ± SD)p-value†p-value‡
  1. ALAT, alanine-aminotransferase; ASAT, aspartate-aminotransferase; Syst, systolic blood pressure; Dia, diastolic blood pressure.

  2. †p values are given for the comparison of patients 4 h postictal to 24–48 h postictal.

  3. ‡p values are given for the comparison of the patients 4 h postictal to seizure-free controls.

Blood pressure
 Syst111 ± 17104 ± 11107 ± 210.4690.438
 Dia 61 ± 13 55 ± 10 56 ± 120.0110.376
Heart rate/min124 ± 29103 ± 18 87 ± 13<0.0010.001
ALAT (U/l)17.2 ± 4.6 18.4 ± 8.0 0.837
ASAT (U/l)37.1 ± 8.6 34.0 ± 8.4 0.232
Urea (md/dl)24.0 ± 7.5 27.1 ± 7.7 0.071
Creatinine (md/dl) 0.40 ± 0.12  0.46 ± 0.17 0.075
Sodium (mmol/l)136.9 ± 2,8  139.2 ± 1.8  <0.001  
Potassium (mmol/l) 4.44 ± 0.44  3.96 ± 0.22 0.542
Calcium (mmol/l) 2.45 ± 0.10  2.44 ± 0.10 0.773
Leukocytes (G/l)11.5 ± 5.2  6.7 ± 2.0 <0.001  


The diagnosis of epilepsy on the basis of clinical signs, symptoms, or routine laboratory tests is not always an easy task. The variety of ways in which different types of seizures and syncopal attacks are expressed complicate their diagnosis. Clinical examination, physiological and biochemical tests can be used to discriminate TCS, PS, FC, and SYN. Most of the features mentioned, raise important medical issues, because the urgency and course of treatment are quite different depending on the diagnosis. Investigations such as the EEG may aid the diagnostics, but are never diagnostic solely. Previous studies have evaluated the utility of prolactin, NSE, and s-100 protein as markers of epileptic seizures in children and adults (Buttner et al., 1999; Chen et al., 2005). Serum levels of prolactin may increase as a consequence of epileptic seizures (grand mal and complex partial seizures), but the increase is transient. Usually, prolactin blood levels increase within 30 min after epileptic seizures and return to normal values within 1 h (Malkowicz et al., 1995). Therefore, measurement of serum prolactin is a reliable confirmatory test, but only modestly effective as a screening test for seizures or SYNs (Chen et al., 2005). S-100 protein (s-100) and NSE are proteins that are distributed mainly in the central nervous system. Both proteins are able to provide information about the extent of brain damage after stroke, cerebral hypoxia, and head trauma (Persson et al., 1987). S-100 protein seems not to be a suitable marker for epileptic seizures, whereas NSE as a marker for epileptic seizures and FC is discussed controversly (Buttner et al., 1999; Tanabe et al., 2001).

In the present study, age-adjusted analyses show elevated values of plasma NT-proBNP levels 4 h postictal in patients with TCS and FC but not with PS and SYN compared to controls. Since NT-proBNP levels vary with age (Mir et al., 2002; Mir et al., 2003; Nir et al., 2004; Albers et al., 2006), one can assume that elevated NT-proBNP levels 4 h postictal compared to controls might be, to some extent, due to age difference. In our opinion, this seems unlikely simply because: (i) high postictal levels return to lower levels within 24–48 h (Figs. 2 and 3), (ii) a significant correlation of NT-proBNP with age was seen also in controls, and (iii) the pronounced difference was found only for TCS and FC but not for PS and SYN.

Two mechanisms could be responsible for the upregulation of NT-proBNP levels in postictal patients with TCS and FC.

First, increased BNP secretion might be of cardiac origin, secondary to an increase in stress and noradrenaline release during epileptic seizures. NP plays a role in modulating the autonomic control of the heart (Richards et al., 2003). BNP is known to augment parasympathetic vagal neurotransmission but high concentrations of NP can directly stimulate cardiac pace-making (Herring et al., 2001). During an epileptic attack blood pressure is almost doubling, heart rate changes occur frequently and around the time or even before the earliest ECG or clinical change (Zijlmans et al., 2002). Rising evidence suggests cardiac ischemia during seizures, possibly related to the development of sudden unexplained death in epilepsy (SUDEP) in adults (Tigaran et al., 2003). Different authors support the theory of heart damaging by recurrence of seizures (P-Codrea Tigaran et al., 2005). In the current study no significant correlation between NT-proBNP levels in patients with FC and the number of seizures experienced could be demonstrated. In contrast, plasma NT-proBNP levels in patients suffering from prolonged FC were significantly higher compared to typical FC. In line with these observations, higher levels of postictal leukocytes and heart rate in patients compared to controls and correlation of NT-proBNP levels with postictal heart rate might indicate a crucial role of stress as the underlying factor. ECG and a careful clinical examination by a cardiologist did not reveal any cardiac abnormalities in our patients. Furthermore, in patients with highest NT-proBNP levels echocardiographic evaluation was performed, revealing no cardiac pathology (e.g., ventricular septum defect, atrial septum defect). Previous studies suggested cardiac ischemia and ventricular dilatation as the trigger for cardiac BNP release (McMurray et al., 1998; Wang et al., 2004). To the best of our knowledge, NT-proBNP levels are not elevated by tachycardia exclusively. Although prolonged tachycardia possibly leads to relative ischemia of the myocardium, we do not believe in tachycardia as the only underlying factor for the pronounced elevation of NT-proBNP levels in our children.

Alternatively, these findings may indicate increased BNP secretion from the brain. BNP may be part of a central mechanism for control of blood volume, blood pressure, and electrolyte composition. This notion is supported by the observation that BNP increases in proportion to increases in intracranial pressure (Berendes et al., 1997). NPs are produced in the hypothalamus and their production is mediated and induced by catecholamine triggers, endothelin, and arginine vasopressin (Levin et al., 1998). The systemic effects of BNP are a reduction of blood pressure, peripheral vascular resistance, and sympathetic tone in the peripheral vascular tissue by dumping baroreceptors, suppressing catecholamine release from autonomic nerve endings, and suppressing sympathetic outflow from the central nervous system (Yang et al., 1992). In addition, NPs have direct vasodilator effects on human vascular tissue (Protter et al., 1996). High concentrations of plasma BNP have been associated with the development of cerebral ischemia and neurological deficits (Sviri et al., 2003), suggesting that BNP release is associated with the intensity of brain tissue ischemia, reflecting increased biosynthesis and secretion from ischemic brain tissue, especially from the hypothalamus (Franzoni et al., 1992). Furthermore, experimental findings in anesthetized and paralyzed rats showed that plasma NP levels increased approximately threefold over a 30-min period of generalized seizures, indicating that NP release from the brain might be triggered by epileptic seizures (Perrone et al., 1995). In line with this experimental findings, altered ANP and BNP secretions caused by an epileptic attack in a 65-year-old woman with partial epilepsy and frequent bouts of polyuria probably were triggered by epileptic stimulation on the diencephalon beyond the focus (Obi et al., 2002). This suggests a possible contribution of the brain to elevated BNP levels after epileptic seizures. Intracerebral epileptic activity might trigger BNP secretion by heart or by brain. Interestingly, in children with PS and SYN, NT-proBNP levels remained stable. This might be traced back to lower cardial burden or to a minor cerebral stimulus to trigger BNP secretion by heart or brain, respectively. Additionally, serum sodium concentrations 4 h postictal were significantly lower than that in seizure-free controls. A likely explanation for this disturbance is an increased urinary excretion of sodium triggered by BNP and seizure.

We conclude from this study that elevated plasma levels of NT-proBNP are not solely a specific marker of cardiac dysfunction. In contrast to other markers of epileptic seizures in children and adults (prolactin, NSE, and s-100 protein), NT-proBNP levels increase rapidly after epileptic seizures and FC and, corresponding to the half-life of NT-proBNP, return to normal values within hours. Measurement of plasma NT-proBNP seems to be useful in discriminating TCS and FC from PS and SYN in childhood.

Nevertheless, several aspects of NT-proBNP release following epileptic seizures and FC remain unanswered: Do NT-proBNP levels increase during seizures or thereafter? When is the time-point of maximum release? Is the NT-proBNP decrease gradually and when are normal values reached again?

Further studies with serial NT-proBNP measurements in larger cohorts of children and adults are needed to discover the detailed mechanisms of NP secretion in epileptic attacks and FC, to evaluate the impact of ictal tachycardia and blood pressure on NT-proBNP levels and to evaluate the implications of high plasma BNP-levels following epileptic seizures on cardiac events and SUDEP in adulthood.