Changes of biochemical biomarkers in the serum of children with convulsion status epilepticus: a prospective study

Status epilepticus (SE) is generally categorized as convulsive status epilepticus (CSE) and non-convulsive status epilepticus (NCSE) [1]. CSE, the main subtype of SE, is a common neurologic emergency with high morbidity and mortality in children each year [2, 3]. It not only causes cerebral ischemia/hypoxia, brain edema, and neuron damage, but also results in cardiac arrests, secondary infection, and internal environment disorders [4]. It is noticed that there were approximately 20–40% of CSE patients with a high risk for secondary brain dysfunction [5], which is manifested as behavioral abnormality, emotional disorder, language retardation, and learning ability decline, or even the serious ones have chronic persistent sequelae, with restricted limb activity, intellectual disorder and cerebral palsy [6].

Lowenstein et al. have shed light on an operational and a conceptual definition of CSE, which is defined as … ≥ 5 min of (i) continuous seizure or (ii) two or more discrete seizures between which there is incomplete recovery of consciousness [7]. In addition, as a basic research (or conceptual) definition of CSE, the International League Against Epilepsy Core Group on Classification group further suggested the following: CSE refers to a condition in which there is failure of the “normal” factors that serve to terminate a typical generalized tonic–clonic seizures [8]. Based on the definition of CSE, the optimal treatment time is within 5 min after seizure and it is likely that a seizure will continue if not treated after this time [9]. If the time is prolonged to 30 min, there may be a risk of long-term consequences [10]. Therefore, accurate early diagnosis and correct treatments are the keys to the successful rescue. Generally, viral encephalitis (VEN), primary generalized epilepsy (PGE), and complex febrile seizures (CFS) are the three main causes of CSE [1, 11, 12] The degrees of brain injury caused by different causes of CSE are different [1, 11, 12]. Recent years, increasing attentions have been paid to the potential efficacies of relevant specific biochemical biomarkers for the diagnosis and prognosis of CSE or related diseases [13‐19]. For instance, Knopman et al. found that the concentration of serum neuron-specific enolase (NSE) in temporal lobe epilepsy (TLE) patients with SE was higher than that of controls [13]. Both Chang et al. and Steinhoff et al. indicated that relatively high level of S-100 calcium binding protein B (S-100B) is determined in TLE patients with higher frequencies of seizures [14, 15]. Madžar et al. have demonstrated that C-reactive protein (CRP) serves as a biomarker in the inflammatory processes of SE and high CRP concentration is strongly associated with in-hospital mortality and poor prognosis [16]. Similarly, hyperneuroinflammation emerging in SE may also increase the level of glial fibrillary acidic protein (GFAP) in mice [17]. In addition, Cho et al. and Nicoletti et al. have revealed that the elevated expression of vascular endothelial growth factor (VEGF) plays a neuroprotective role in SE murine model [18, 19]. Early detection of relevant specific biochemical biomarkers may have guiding significance for diagnosis and prognosis of CSE. Although the researches on the biochemical biomarkers of CSE are not rare, they are mostly the basic research on animal experiments or adults. The common induction factors of SE in children may be different from those in animals and adults, and the significances of these biochemical biomarkers may be also different [20]. Therefore, the measurements for serum biochemical biomarkers in CSE children may be helpful to judge the disease condition and estimate the prognosis.

In this study, the levels of NSE, S-100B, CRP, VEGF, and GFAP in serum of CSE children were explored. In addition, their associations with CSE were also analyzed in logistic regression models.

SE is traditionally defined as any seizure lasting longer than 30 min, whether or not consciousness is impaired or seizures recur without an intervening period of consciousness [21]. It is reported that CSE becomes increasingly resistant to anticonvulsant drugs with the extension of seizure time, it can be associated with extremely high morbidity and mortality if not treated promptly and aggressively [22]. In the current study, we mainly explored the levels of NSE, S-100B, CRP, VEGF, and GFAP in CSE children as well as the associations with CSE to develop a new diagnosis and treatment method for children CSE.

Numerous studies are focused on the possible role of relevant specific biochemical biomarkers during the period of seizure [13‐19], which can be the potential biomarkers for further researches and future diagnosis. NSE is located in the mature neurons and neuroendocrine cells of the nervous system with the highest content in the brain and it will enter cerebrospinal fluid and blood when blood–brain barrier is destroyed and permeability increases [23]. Approximately 96% of S-100B protein is existed in the brain, which level reflects the injury and severity of astrocytes [24]. CSE is generally accompanied with the occurrence of inflammation, and CRP is one of the most sensitive biomarkers for inflammation [16]. VEGF is a specific mitogen of endothelial cells, which is involved in the neural processes such as growth, differentiation, survival, regeneration, and neuroprotection [25]. GFAP is an important cytoskeletal component and a specific marker of astrocytes [26]. Summarily, the concentrations of these biochemical biomarkers in blood can reflect the severity of brain injury. In this study, we initially explored the associations between biochemical biomarkers levels and the occurrence of CSE. Interestingly, we demonstrated that the levels of NSE and VEGF were significantly elevated in serum of children at 1^st and 4^th day of CSE compared to that of controls, which was consistent with the previous results [13, 18, 19]. In addition, we further found sharply decreased NSE and VEGF after seizure for 10 day. Logistic regression analysis also confirmed the levels of NSE and VEGF at 1^st and 4^th day of CSE may be risk factors for increased seizure frequency. For the data of NSE, DeGiorgio et al. uncovered the similar results, which indicated that serum NSE level for the 19 subjects within 24 h after SE was significantly elevated compared with the levels for normal and epileptic controls [27]. In addition, they also demonstrated that serum NSE level was highest in those with an acute neurologic insult, and the level of serum NSE was also significantly elevated compared with control levels in 11 patients without acute neurologic insults other than SE [27]. Therefore, they concluded that in the absence of an acute insult or lesion, NSE was elevated as a direct result of SE [27]. Based on this previous experiment data, we believed that NSE may be also an effective specific biomarker of CSE. Additionally, the results further implied that VEGF may have no diagnostic values in advanced CSE (10 day later). S-100B level was also increased at 1^st day of CSE, which was strongly associated with CSE. Interestingly, a recent study on epilepsy seizure of TLE patients conducted by Chang et al. indicated a relatively high level of S-100B in TLE patients with higher frequencies of seizures [14]. Meanwhile, Gunawan et al. also revealed that there is an increase in serum S-100B level within 24 h after seizure in patients with SE, and this has a strong positive correlation with brain damage seen in head MRI and diffusion tensor imaging (DTI) [28]. We speculated that S-100B may be used as a more sensitive biochemical marker of brain injury in early stage of CSE, especially within one day. Similar to the results of S-100B, CRP is also a biomarker elevated in the early stage of CSE (at 1^st day of CSE). Given that CRP is a sensitive marker in the inflammatory processes of SE [16], we speculated the early stage of anti-inflammatory therapies may have significance in the treatment of CSE to some extent. However, Sutter et al. have voiced concern that single level of CRP is not reliable for diagnosis of infections during SE [29] and not independently associated with refractory epileptic activity and death [30]. The data implied that further studies are needed to assess the potential of CRP for inclusion in prediction models allowing to identify patients with poor outcome of SE. Combined our results obtained above, we further considered using linear changes of CRP level to assess its role in CSE in future studies. In addition, Wang et al. reported that hyperneuroinflammation leads to an increase of GFAP level in a SE mouse model [17]. Similarly, a higher level of GFAP in CSE children than control was found in this study, and the time point was at 4^th day of CSE. At the same time, the level of GFAP at 4^th day of CSE was also determined a risk factor in the occurrence of CSE using a logistic regression analysis. Interestingly, Gurnett et al. found that the level of cerebrospinal fluid (CSF) GFAP is increased in SE children, and the increased CSF GFAP is strongly correlated with seizure duration [31]. Evidence has demonstrated that biomarkers are firstly released into the CSF compartment prior to release into serum when SE occurs [32]. Furthermore, Saija et al. believed that the elevation in the serum compartment that occurs after SE may reflect the increase in permeability of the blood–brain barrier that is known to occur after SE. We speculated that the high level of serum GFAP may be used as an effective biomarker to reflect the brain injury after CSE.

Additionally, we further investigated the associations between these biochemical biomarkers and VEN/PGE/CFS using logistic regression models. We found that NSE level in VEN children was relatively high than that of PGE children or CFS children. Also strong association between NSE level at 4^th day of CSE and VEN was determined by logistic regression analysis. Our results were consistent with one of mechanisms for the increased NSE level in CSE children: direct invasion of neurons by pathogens leads to edema, necrosis and myelin breakdown of neurons, destruction of blood brain barrier, and release of NSE in cytoplasm into cerebrospinal fluid and blood [33]. At the same time, increased S-100B at 1^st and 4^th day of CSE was observed in VEN children compared to the PGE children or CFS children, which was in line with the previous results [33]. However, further logistic regression analysis demonstrated the level of S-100B at 4^th day of CSE was a risk factor for VEN. Similar to the data of the above two biomarkers, VEGF level was upregulated at 4^th day of CSE. Interestingly, a recent study on VEN children also reported that the serum VEGF level in severe VEN was higher than that in normal control group and mild VEN group [34]. Combined the results of the above three biomarkers, we further speculated that the brain injury of patients with central nervous system infection such as VEN within one week after seizure is more serious than that of PGE and CFS.

Additionally, there are also some limitations of this study such as a less varied population in which research cases limited to CSE caused by VEN, PGE and CFS, so the results cannot be generalized to other disorders. Meanwhile, the relatively small sample size may also affect the accuracy of the acquired data. In addition, the most important limitation of this study is the fact that VEN was not fully excluded and is a major confounder, as encephalitis can increase the biomarkers and lead to more severe CSE and poor outcomes.

In a word, in the early stage of CSE (within one week), the increased levels of NSE, S-100B, and VEGF may serve as effective diagnostic markers for CSE, and relatively high level of CRP represents hyperneuroinflammation. Meanwhile, GFAP can be used as an effective biomarker to reflect the brain injury after CSE. In addition, we indicated that the increased levels of NSE, S-100B, and VEGF may be risk factors for VEN, which is considered as one of the main induction factors of CSE.