Evaluation of serum and cerebrospinal fluid copeptin as biomarkers in children with febrile seizures, febrile illness without seizures, and meningitis: an exploratory study Authors: Halil Uğur Hatipoğlu, Cevher Kızılırmak, Murat Elevli Published: 2026 ABSTRACT AbstractBackgroundThis study aimed to compare serum and cerebrospinal fluid (CSF) copeptin levels among children with febrile seizures, febrile illness without seizures, and meningitis; to investigate the relationship between serum and CSF copeptin levels; and to assess the potential diagnostic utility of these biomarkers.MethodsThis study included children aged 1 month to 18 years who presented to the pediatric emergency department with fever or seizures and underwent lumbar puncture because of signs of meningeal irritation or the absence of an identifiable source of fever. The patients were classified into three groups: the Meningitis Group, the Febrile Illness Without Seizures Group, and the Febrile Seizure Group. In addition, a control group consisting of healthy children was included. Serum and CSF copeptin levels were measured in the patient groups, whereas only serum copeptin levels were measured in the control group.ResultsSerum copeptin levels were significantly higher in the Febrile Seizure Group than in both the Febrile Illness Without Seizures Group and the Control Group. Univariate linear regression analysis demonstrated a weak but statistically significant positive association between serum and CSF copeptin levels (p = 0.040), indicating that CSF copeptin levels could be estimated from serum copeptin levels using the following formula: CSF copeptin level = 2.02 + 0.05 × serum copeptin level. In multivariable linear regression analysis, C-reactive protein (CRP) and CSF glucose levels were identified as independent predictors of CSF copeptin levels, irrespective of age and sex (p = 0.009 and p = 0.004, respectively). CSF copeptin measurement was not superior to serum copeptin measurement for distinguishing children with febrile seizures, or children with febrile seizures and/or meningitis, from those with febrile illness without seizures.ConclusionSerum copeptin levels were higher in children with febrile seizures than in those with febrile illness without seizures and in healthy controls. However, because of the very small meningitis sample size, the meningitis-related findings should be considered exploratory and interpreted with caution. Larger prospective studies are required to determine whether copeptin has clinically meaningful utility in distinguishing febrile seizures from meningitis.Keywords: Febrile seizure, Meningitis, Copeptin, Cerebrospinal fluid, Biomarker, Child Background Febrile seizures are the most common seizures of childhood and occur in approximately 2%-5% of children. A febrile seizure (FS) is defined as a seizure occurring in a child aged 3 months to 5 years who has no underlying neurological disorder, no intracranial infection, and no previous history of afebrile seizures [1]. Febrile seizures are multifactorial events that may be triggered by infections outside the central nervous system (CNS), as well as by genetic and environmental factors [2]. Fever is one of the most important contributors to their etiopathogenesis. Because of their frequency and tendency to recur, febrile seizures are of considerable importance in pediatric practice [3]. Meningitis is an inflammatory disease of the meninges surrounding the brain and spinal cord. It may develop secondary to infectious or non-infectious causes. Non-infectious causes include certain systemic diseases, medications, and neoplasms, whereas infectious causes include viruses, bacteria, and fungi [4]. Although febrile seizures are typically benign and are generally managed conservatively, bacterial meningitis is associated with a fatality rate of 14.4%. Seizures occur in 23% of meningitis cases. Therefore, distinguishing febrile seizures from meningitis is essential for optimizing patient outcomes; however, this approach may also lead to a considerable number of children with febrile seizures undergoing unnecessary invasive diagnostic procedures [5]. Copeptin is a precursor-derived peptide of arginine vasopressin (AVP), which is secreted by the pituitary gland. AVP plays an important role not only in maintaining fluid balance and vascular tone but also in regulating the endocrine stress response. Because AVP has a short half-life and is difficult to measure directly, its clinical assessment is challenging. Copeptin is synthesized from the same precursor molecule as AVP but is more stable than AVP. Unlike AVP, copeptin can be measured easily using a sandwich immunoassay and therefore indirectly reflects circulating AVP levels [6]. As a biomarker reflecting activation of the vasopressin system and the endogenous stress response, copeptin has been proposed to allow a more direct and sensitive assessment of individual stress burden. It has also been identified as a prognostic biomarker in various acute conditions, including ischemic stroke and myocardial infarction [7]. Published data suggest that copeptin is involved in the thermoregulatory response to fever and seizures and has high diagnostic accuracy in febrile seizures [8]. Several studies have shown that serum copeptin levels are higher in children with febrile seizures than in children with febrile illness without seizures [9]. In the present study, CSF copeptin levels were evaluated in children who underwent lumbar puncture because of fever or febrile seizures, and serum copeptin levels were assessed in all participants. We aimed to determine the relationship between serum and CSF copeptin and to investigate whether copeptin could be used as a biomarker for febrile seizures and meningitis in febrile children. Methods Study design and patient selection This was a single-center prospective cross-sectional study. The study included children aged 1 month to 18 years who presented to the pediatric emergency department of the University of Health Sciences, Sultangazi Haseki Training and Research Hospital between May 10, 2023 and November 10, 2023 with fever or febrile seizures and who underwent lumbar puncture because of signs of meningeal irritation or the absence of an identifiable source of fever. Children presenting with fever and undergoing lumbar puncture because of signs of meningeal irritation or the absence of an identifiable source of fever were evaluated for meningitis. Lumbar puncture was not performed in all febrile children. Rather, it was performed in children with a body temperature of 38.5 °C or higher when no focus of fever could be identified after physical examination and routine evaluation, in children with signs of meningeal irritation, and in children with febrile seizures when meningitis needed to be excluded according to predefined clinical criteria. After lumbar puncture, patients were classified according to CSF findings. Children with CSF findings compatible with meningitis were assigned to the Meningitis Group. Among children without meningitis, those presenting with febrile seizures were assigned to the Febrile Seizure Group, whereas those presenting with fever without seizures were assigned to the Febrile Illness Without Seizures Group. The patient selection and grouping process is summarized in Fig. 1. Flow diagram of patient selection and group allocation Because lumbar puncture is an invasive procedure, no healthy control group could be established for CSF copeptin measurement. However, healthy children who presented to the pediatric outpatient clinic for routine blood testing, who had no complaints, and who met the inclusion and exclusion criteria were enrolled as the Control Group for serum copeptin measurement only. Children younger than 1 month or older than 18 years, as well as eligible children from whom parental consent, or the child’s assent when age-appropriate, could not be obtained, were excluded. In addition, patients with known epilepsy, cerebral palsy, congenital heart disease, diabetes insipidus, diabetes mellitus, fluid-electrolyte imbalance, tubulopathy, or chronic kidney disease were excluded because these chronic conditions could affect copeptin levels. Patients with a urine specific gravity below 1.010 or above 1.030, or a serum osmolality above 295 mOsmol/kg, were also excluded. Meningitis was diagnosed or excluded on the basis of CSF findings, including the presence of cells on direct microscopic examination, elevated CSF protein and decreased CSF glucose on biochemical analysis, detection of a pathogen on the multiplex PCR meningitis panel, or growth on CSF culture. CSF cell count was performed manually by direct microscopic examination. CSF glucose and protein levels were measured using the Roche COBAS c701 analyzer. For the multiplex PCR meningitis panel, the FilmArray ME Panel v1.4 kit was used with the FilmArray system. For bacterial culture, CSF samples were inoculated onto blood agar, chocolate agar (5%–10% CO₂ atmosphere), and MacConkey agar. Sheep blood agar and chocolate agar plates were incubated at 35 ± 2 °C for 72 h, and MacConkey agar plates were incubated at 35 ± 2 °C for 48 h. All serum and CSF copeptin analyses were performed on the same day to avoid possible inter-day and inter-assay variability. Serum and CSF samples were stored at − 80 °C in Eppendorf tubes and analyzed using the BT LAB Human Copeptin ELISA Kit (Bioassay Technology Laboratory, Shanghai Korain Biotech Co., Ltd., Shanghai, China) on an ELx800 platform (BioTek/Elitech). A double-antibody sandwich enzyme-linked immunosorbent assay (ELISA) method was used to measure copeptin levels in the collected samples. All steps of copeptin measurement were performed according to the manufacturer’s instructions. Only baseline laboratory parameters obtained in the emergency department were included in the study. Statistical analysis Statistical analyses were performed by a professional biostatistician and were subsequently reviewed by the authors. Statistical analyses were conducted using SPSS version 25 for Windows. Distribution of variables was assessed using histograms, Q-Q plots, and the Shapiro-Wilk test, whereas homogeneity of variance was assessed using Levene’s test. Descriptive statistics are presented as number and percentage for categorical variables and as median [25th percentile–75th percentile] for numerical variables. Because the assumption of normal distribution was not met for comparisons involving more than two groups, the Kruskal-Wallis test was used for numerical variables. Pairwise comparisons among more than two groups were performed using the Dunn-Bonferroni test. Because the assumptions for parametric testing were not met, relationships between numerical variables were evaluated using Spearman correlation analysis. Proportions in independent groups were compared using the chi-square test. Cutoff value analysis was performed using receiver operating characteristic (ROC) curve analysis. ROC coordinates and multiple ROC comparisons were analyzed using the TURCOSA software package (Turcosa Analytics Ltd. Co.). A two-sided p value < 0.05 was considered statistically significant. Because there was no directly comparable reference study evaluating CSF copeptin in children in this specific clinical setting, the a priori sample size calculation was based on serum copeptin data reported by Salam et al. [10]. Based on that study, the standardized effect size was estimated as 1.3, and it was calculated that at least 11 participants per group would be required to achieve 90% power at a 5% significance level. In addition, post hoc power analyses were performed using G*Power version 3.1 (Heinrich Heine University, Düsseldorf, Germany). Based on the serum copeptin results of the 57 participants included in the study, the standardized effect size was 0.73 and the post hoc power was 99% at a 5% significance level. Based on the CSF copeptin results of the 33 participants who underwent lumbar puncture, the standardized effect size was 0.30 and the post hoc power was 30% at a 5% significance level. Results After application of the inclusion and exclusion criteria, 5 patients were assigned to the Meningitis Group, 14 to the Febrile Illness Without Seizures Group, and 14 to the Febrile Seizure Group. In addition, 24 healthy children were enrolled as the Control Group for serum copeptin measurement only. A total of 57 participants were included in the study. Fifteen patients were excluded according to the predefined exclusion criteria (Fig. 1). Meningitis was diagnosed on the basis of CSF findings, including the presence of cells on direct microscopic examination, elevated CSF protein and decreased CSF glucose on biochemical analysis, detection of a pathogen on the multiplex PCR meningitis panel, or growth on CSF culture. Among the 5 patients in the Meningitis Group, enterovirus was identified in 2 patients, Streptococcus pneumoniae in 1 patient, and varicella-zoster virus in 1 patient, while 1 patient had clinical and laboratory findings consistent with bacterial meningitis despite the absence of microbiological confirmation. Four children had age-appropriate vaccination histories, whereas the child in whom Streptococcus pneumoniae was identified was unvaccinated. The 2 patients with enteroviral meningitis received intravenous hydration and antipyretic therapy only. The patient with varicella-zoster virus meningitis received 14 days of intravenous acyclovir. The patient in whom Streptococcus pneumoniae was identified received 14 days of ceftriaxone plus vancomycin. The patient with clinical and cerebrospinal fluid findings consistent with bacterial meningitis despite the absence of microbiological confirmation also received 14 days of ceftriaxone plus vancomycin. All patients included in the Meningitis Group were followed in the inpatient ward. No deaths occurred in the meningitis subgroup, and all patients were discharged without neurological sequelae. Therefore, no analysis of the relationship between mortality and copeptin levels could be performed. Age differed significantly across the study groups (Table 1). Adjusted post hoc pairwise comparisons demonstrated that the Febrile Illness Without Seizures Group was significantly younger than the Control Group, while no other pairwise age differences were statistically significant. There was no statistically significant difference among the groups with respect to sex. Table 1Comparison of age and sex among the study groupsVariableMeningitis Group a(n = 5)Febrile Illness Without Seizures Group b(n = 14)Febrile Seizure Group c(n = 14)Control Group d(n = 24) p Median Age, months [IQR]Mean ± SD27.0 [5.5–156.0](70 ± 79)2.0 [1.3–22.1] d(31 ± 62)25.5 [15.5–38.3](31 ± 21)87.0 [18.5–132.0] b(83 ± 62) 0.001* Female, n (%)3 [60.0]4 [28.6]7 [50.0]10 [41.7]0.553#Male, n (%)2 [40.0]10 [71.4]7 [50.0]14 [58.3]SD Standard deviation. *Kruskal-Wallis test. #Chi-square test. Data are presented as median [25th percentile-75th percentile]. Statistically significant p values are shown in bold. Superscript letters indicate the groups between which a significant difference was observed Comparison of age and sex among the study groups Median Age, months [IQR] Mean ± SD 27.0 [5.5–156.0] (70 ± 79) 2.0 [1.3–22.1] d (31 ± 62) 25.5 [15.5–38.3] (31 ± 21) 87.0 [18.5–132.0] b (83 ± 62) SD Standard deviation. *Kruskal-Wallis test. #Chi-square test. Data are presented as median [25th percentile-75th percentile]. Statistically significant p values are shown in bold. Superscript letters indicate the groups between which a significant difference was observed Leukocyte count was significantly lower in the Control Group than in the Meningitis Group. Hematocrit was significantly lower in the Febrile Illness Without Seizures Group than in the Control Group. Sodium levels were significantly lower in the Meningitis, Febrile Illness Without Seizures, and Febrile Seizure groups than in the Control Group, and potassium levels were significantly lower in the Febrile Seizure Group than in the Febrile Illness Without Seizures Group (Table 2). Table 2Comparison of hematological and biochemical laboratory findingsVariableMeningitis Group aFebrile Illness Without Seizures Group bFebrile Seizure Group cControl Group dp*Leukocyte count (/µL)14,590[11120–17285] d8000[6535–10250]9845[7600–12963]7365[6823–9720] a 0.010 Hemoglobin, g/dL12.20 [11.10–12.90]11.00 [10.10–12.30]11.60 [10.40–12.10]12.40 [11.30–12.90]0.083Hematocrit, %35.80[33.40–37.40]31.20[28.50–34.70] d32.80[30.90–35.50]36.60[34.40–38.90] b 0.001 Platelet count (/µL)374,000[296500–389000]293,500[208500–363250]256,000[177000–342000]327,500[261250–402000]0.133Neutrophil count (/µL)8800[7015–12770] d4215[1975–7405]7365[5015–8970] d3790[2595–5065] a, c 0.002 Lymphocyte count (/µL)3720 [910–6605]3995 [1705–5320]1735 [1390–2480]2845 [2153–4395]0.081Glucose, mg/dL112.00[102.00-118.50] d103.00[93.80-116.80]100.50[87.00-135.50]90.00[83.00-100.00] a 0.026 Urea, mg/dL28.00[12.50–37.30]11.5[7.60–21.50]17.00[11.50–23.50]19.00[14.30–23.90]0.124Creatinine, mg/dL0.30 [0.27–0.63]0.30 [0.20–0.40]0.30 [0.20–0.30]0.45 [0.29–0.54]0.106Sodium, mmol/L133.00[131.00-135.00] d135.50[134.00-139.00] d137.00[134.80-138.30] d140.00[138.00-142.00] a, b, c < 0.001 Potassium, mmol/L4.80 [4.10–5.40]5.00 [4.60–5.20] d4.30 [4.00-4.60]4.60 [4.40–5.10] b 0.019 Chloride, mmol/L97.50[94.80-101.80]101.50[100.00-103.30]102.00[100.00-103.60]102.00[101.00-103.00]0.251Calcium, mg/dL9.8 [8.9–10.3]9.9 [9.3–10.1]9.4 [9.2–9.5]9.6 [9.3–10.0]0.113C-reactive protein, mg/L26.00[15.40–35.70] d20.90[4.20–68.30] d15.00[2.40–21.30] d1.40[0.50–8.70] a, b, c < 0.001 Serum copeptin, pmol/L2.15[1.00-4.80]1.50[1.10–1.60] c2.38[1.60–5.36] b, d1.42[1.22–1.60] c 0.005 CSF copeptin, pmol/L2.18 [1.90–2.40]2.10 [1.80–2.20]2.25 [1.99–2.55]-0.285* Kruskal-Wallis test. Data are presented as median [25th percentile-75th percentile]. Statistically significant p values are shown in bold. Superscript letters indicate the groups between which a significant difference was observed Comparison of hematological and biochemical laboratory findings 14,590 [11120–17285] d 8000 [6535–10250] 9845 [7600–12963] 7365 [6823–9720] a 35.80 [33.40–37.40] 31.20 [28.50–34.70] d 32.80 [30.90–35.50] 36.60 [34.40–38.90] b 374,000 [296500–389000] 293,500 [208500–363250] 256,000 [177000–342000] 327,500 [261250–402000] 8800 [7015–12770] d 4215 [1975–7405] 7365 [5015–8970] d 3790 [2595–5065] a, c 112.00 [102.00-118.50] d 103.00 [93.80-116.80] 100.50 [87.00-135.50] 90.00 [83.00-100.00] a 28.00 [12.50–37.30] 11.5 [7.60–21.50] 17.00 [11.50–23.50] 19.00 [14.30–23.90] 133.00 [131.00-135.00] d 135.50 [134.00-139.00] d 137.00 [134.80-138.30] d 140.00 [138.00-142.00] a, b, c 97.50 [94.80-101.80] 101.50 [100.00-103.30] 102.00 [100.00-103.60] 102.00 [101.00-103.00] 26.00 [15.40–35.70] d 20.90 [4.20–68.30] d 15.00 [2.40–21.30] d 1.40 [0.50–8.70] a, b, c 2.15 [1.00-4.80] 1.50 [1.10–1.60] c 2.38 [1.60–5.36] b, d 1.42 [1.22–1.60] c * Kruskal-Wallis test. Data are presented as median [25th percentile-75th percentile]. Statistically significant p values are shown in bold. Superscript letters indicate the groups between which a significant difference was observed CSF copeptin levels did not differ significantly among the three lumbar puncture groups (Meningitis Group, Febrile Illness Without Seizures Group, and Febrile Seizure Group; p = 0.285), whereas serum copeptin levels differed significantly across the four study groups, including the Control Group (p = 0.005) (Table 2). Adjusted pairwise comparisons showed significant differences between the Febrile Seizure Group and the Febrile Illness Without Seizures Group (p = 0.038), and between the Febrile Seizure Group and the Control Group (p = 0.003), while all other pairwise comparisons were not statistically significant. When all cases were analyzed together, CSF copeptin showed a weak negative correlation with C-reactive protein (CRP), a moderate positive correlation with CSF glucose, and a weak positive correlation with serum copeptin (rho=-0.343, p = 0.050; rho = 0.455, p = 0.009; and rho = 0.359, p = 0.040, respectively). Serum copeptin showed a weak negative correlation with creatinine (rho=-0.272, p = 0.041). In the Meningitis Group, serum copeptin was very strongly and negatively correlated with neutrophil count (rho=-0.900, p = 0.037). In the Febrile Illness Without Seizures Group, CSF copeptin was moderately and positively correlated with glucose, strongly and positively correlated with creatinine, and strongly and negatively correlated with CRP (rho = 0.546, p = 0.043; rho = 0.610, p = 0.021; and rho=-0.675, p = 0.008, respectively). In addition, no statistically significant correlation was found between CSF copeptin and CSF protein levels, either in the overall lumbar puncture cohort (rho=-0.317, p = 0.077) or within the individual study groups. Univariate linear regression analysis demonstrated a weak but statistically significant positive relationship between serum copeptin and CSF copeptin (p = 0.040). This relationship indicated that CSF copeptin could be estimated from serum copeptin using the following equation: CSF copeptin level = 2.02 + 0.05 × serum copeptin level (Fig. 2). Univariate linear regression plot showing the relationship between serum copeptin and CSF copeptin levels To distinguish children with febrile seizures from those with febrile illness without seizures, a serum copeptin level ≥ 1.58 pmol/L yielded 92.9% sensitivity and 57.1% specificity (AUC, 0.793; 95% CI, 0.599–0.922; p < 0.001), whereas a CSF copeptin level ≥ 1.93 pmol/L yielded 92.9% sensitivity and 42.9% specificity (AUC, 0.679; 95% CI, 0.477–0.841; p = 0.091). DeLong’s test showed no superiority of either biomarker over the other for discriminating febrile seizures (p = 0.393) (Table 3). Table 3DeLong test comparing the ROC curves of CSF and serum copeptinComparisonMarker 1 [A]Marker 2 [B]AUC [A]AUC [B]Difference [A, B]Standard Error of Difference [A, B]z p 1CSF copeptinSerum copeptin0.6790.7934-0.1150.135-0.8540.3932CSF copeptinSerum copeptin0.6500.733-0.0830.119-0.6940.4883CSF copeptinSerum copeptin0.5860.614-0.0290.130-0.2200.826Comparison 1: discrimination of febrile seizures from febrile illness without seizures. Comparison 2: discrimination of febrile seizures and/or meningitis from febrile illness without seizures. Comparison 3: discrimination of febrile seizures from meningitis DeLong test comparing the ROC curves of CSF and serum copeptin Comparison 1: discrimination of febrile seizures from febrile illness without seizures. Comparison 2: discrimination of febrile seizures and/or meningitis from febrile illness without seizures. Comparison 3: discrimination of febrile seizures from meningitis To distinguish children with febrile seizures and/or meningitis from febrile children, the optimal cutoff values were a serum copeptin level ≥ 1.576 pmol/L, with 63.2% sensitivity and 85.7% specificity (AUC, 0.733; 95% CI, 0.551–0.871; p = 0.011), and a CSF copeptin level ≥ 2.31 pmol/L, with 47.0% sensitivity and 86.0% specificity (AUC, 0.650; 95% CI, 0.465–0.807; p = 0.131). DeLong’s test again showed no superiority of either biomarker over the other for this distinction (p = 0.488) (Table 3). In an additional exploratory analysis directly comparing the Febrile Seizure Group and the Meningitis Group, neither serum copeptin nor CSF copeptin differed significantly between the two groups (median serum copeptin: 2.38 vs. 2.15 pmol/L, p = 0.487; median CSF copeptin: 2.25 vs. 2.18 pmol/L, p = 0.622). ROC analysis showed limited discriminatory performance for both serum copeptin (AUC, 0.614; 95% CI, 0.243–1.000; cutoff ≥ 1.30 pmol/L; sensitivity 100.0%, specificity 40.0%) and CSF copeptin (AUC, 0.586; 95% CI, 0.278–0.875; cutoff ≥ 2.49 pmol/L; sensitivity 28.6%, specificity 100.0%). DeLong testing showed no significant difference between the AUCs of serum and CSF copeptin for distinguishing febrile seizures from meningitis (z=-0.220, p = 0.826) (Table 3). To identify independent factors affecting CSF copeptin levels, a multivariable linear regression model was constructed including age, sex, serum copeptin, creatinine, sodium, potassium, CRP, leukocyte count (x103/mm3), and CSF glucose. Independent of age and sex, CRP and CSF glucose were identified as independent predictors of CSF copeptin levels (p = 0.009 and p = 0.004, respectively) (Table 4). Table 4Multivariable linear regression analysis of factors independently associated with CSF copeptin levelsVariableBBeta p Constant1.500Leukocyte count (x103/mm3)0.020.3010.057CRP-0.005-0.425 0.009 CSF glucose0.0090.460 0.004 CSF Cerebrospinal fluid; CRP: C-reactive protein; R2 = 0.409. Statistically significant p values are shown in bold Multivariable linear regression analysis of factors independently associated with CSF copeptin levels CSF Cerebrospinal fluid; CRP: C-reactive protein; R2 = 0.409. Statistically significant p values are shown in bold Discussion Distinguishing febrile seizures from non-ictal fever-related events, such as shivering, remains a major clinical challenge. Consequently, the diagnosis of febrile seizures after paroxysmal events occurring in the context of fever is often complicated by the absence of an objective biomarker. One of the biomarkers investigated for this purpose is copeptin [8–10]. However, no previous study evaluating CSF copeptin levels in children with febrile seizures was identified. Moreover, to the best of our knowledge, no study has investigated serum and CSF copeptin levels in children with meningitis. Published data indicate that copeptin is involved in the thermoregulatory response to fever and seizures, has high diagnostic accuracy for febrile seizures, and is elevated in children with febrile seizures compared with children with febrile illness without seizures [8, 10]. Consistent with these reports, we found that serum copeptin levels were significantly higher in children with febrile seizures than in febrile children without seizures and healthy controls (Table 2). This finding supports the potential role of serum copeptin as an adjunctive biomarker in febrile children with suspected febrile seizures, particularly when the history and clinical presentation are inconclusive. In addition, the observation that copeptin levels in febrile children without seizures were similar to those in healthy controls suggests that the increase in copeptin is more closely related to febrile seizures than to fever itself. One possible contributing factor to the copeptin differences observed between groups is the significantly younger age of the Febrile Illness Without Seizures Group. Developmental differences in hypothalamic–pituitary maturation may have influenced copeptin regulation and should therefore be considered when interpreting the findings. Numerous studies have investigated serum copeptin levels in neurological disorders, fluid-electrolyte disturbances, sepsis, and cardiovascular disease [11–13]. In contrast, only a limited number of studies have evaluated copeptin in CSF. All of these studies were conducted in adults and focused on conditions such as intracranial hemorrhage, subarachnoid hemorrhage, aneurysm, and delayed ischemic neurological deficit [14–16]. To our knowledge, the first study addressing CSF copeptin was conducted by Mindt et al. in patients with aneurysmal subarachnoid hemorrhage [14]. We were unable to identify any study evaluating CSF copeptin levels in children. To the best of our knowledge, this is the first study to evaluate CSF copeptin in children in this clinical context. In our cohort, comparison of CSF copeptin levels among the groups did not reveal a significant difference (Table 2). However, the post hoc power analysis based on the CSF copeptin levels obtained in this study yielded a power of only 30%, indicating that the sample size was insufficient. In particular, the small number of patients with meningitis and the lack of CSF copeptin data from healthy controls limit the generalizability of this finding. Larger studies including more patients and, if ethically feasible, an appropriate control framework are needed to validate these results. When the relationships between serum and CSF copeptin levels and blood and CSF parameters were examined in all patients included in the study, CSF copeptin was found to be negatively correlated with CRP. The literature on CSF copeptin measurement is very limited, and, to the best of our knowledge, no previous study has compared CSF copeptin levels with blood parameters. In contrast, one sepsis study reported significantly increased CRP and serum copeptin levels [17], and Jiang et al. similarly reported a positive correlation between serum copeptin and CRP [18]. Unlike those studies, we did not find a correlation between serum copeptin and CRP; instead, we observed a negative relationship between CSF copeptin and CRP. Additional studies are required to explain these apparently discordant associations between CRP and copeptin in serum and CSF. One possible explanation is that we did not distinguish viral from bacterial infections among febrile children who underwent lumbar puncture, and not all children met criteria for sepsis. When all groups were considered together, serum copeptin showed a weak but statistically significant negative correlation with creatinine. This finding differs from previous studies reporting a positive correlation between serum copeptin and creatinine [19, 20]. However, those studies were performed in patients with renal impairment. Because all participants in our study had normal renal function, this difference may account for the discrepant findings. We did not observe a statistically significant correlation between CSF copeptin and CSF protein, either overall or within the individual study groups. In children diagnosed with meningitis, serum copeptin showed a strong inverse correlation with neutrophil count. Similarly, in a study by Kuluozturk et al. investigating the role of copeptin in distinguishing COVID-19 pneumonia from community-acquired pneumonia, copeptin levels were negatively correlated with neutrophil counts [21]. Further studies are needed to clarify the mechanism underlying this association. Nevertheless, because only five patients were included in the meningitis group, this correlation should be interpreted cautiously. The result is better viewed as hypothesis-generating and warrants validation in larger, adequately powered studies. Because copeptin reflects activation of the endogenous stress response, the observed copeptin levels may be influenced by overall illness severity and systemic stress rather than by meningitis itself alone. This should be taken into account when interpreting the findings related to the Meningitis Group. Among all children who underwent lumbar puncture, CSF copeptin was positively correlated with glucose. To our knowledge, no comparable study has been reported in the literature. Nevertheless, this relationship may be explained by the mechanism proposed by Enhorning et al., who suggested that low water intake increases vasopressin levels and that stimulation of vasopressin 1a and 1b receptors may subsequently increase plasma glucose levels; this effect may be reversible with water intake through a reduction in vasopressin [22]. Mindt et al. reported no statistically significant correlation between serum and CSF copeptin levels [14]. In contrast, we found a statistically significant relationship between serum and CSF copeptin. In our univariate linear regression analysis, this relationship was weak, positive, and statistically significant, and we showed that CSF copeptin levels could be estimated from serum copeptin levels using the formula CSF copeptin level = 2.02 + 0.05 × serum copeptin level (Fig. 2). This finding represents one of the key observations of the present study. In the study by Stocklin et al., which included 83 children with febrile seizures and 69 febrile children, a serum copeptin level ≥ 13.0 pmol/L discriminated children with febrile seizures from febrile children with 67.5% sensitivity and 88.4% specificity (AUC, 0.824; 95% CI, 0.753–0.881; p < 0.001) [8]. In another study by Abd El-Moneim et al., a serum copeptin level ≥ 1.3 pmol/L discriminated the same groups with 90% sensitivity and 60% specificity (AUC, 0.702; 95% CI, 0.580–0.824; p = 0.007) [23]. In our literature review, we found no study reporting a cutoff value for CSF copeptin. In our study, to distinguish children with febrile seizures from those with febrile illness without seizures, a serum copeptin level ≥ 1.58 pmol/L yielded 92.9% sensitivity and 57.1% specificity (AUC, 0.793; 95% CI, 0.599–0.922; p < 0.001), whereas a CSF copeptin level ≥ 1.93 pmol/L yielded 92.9% sensitivity and 42.9% specificity (AUC, 0.679; 95% CI, 0.477–0.841; p = 0.091). DeLong’s test showed no statistically significant difference between the areas under the ROC curves for serum and CSF copeptin in discriminating febrile seizures (Table 3). To distinguish children with febrile seizures and/or meningitis from febrile children, the cutoff values were a serum copeptin level ≥ 1.576 pmol/L, with 63.2% sensitivity and 85.7% specificity (AUC, 0.733; 95% CI, 0.551–0.871; p = 0.011), and a CSF copeptin level ≥ 2.31 pmol/L, with 47.0% sensitivity and 86.0% specificity (AUC, 0.650; 95% CI, 0.465–0.807; p = 0.131). Again, DeLong’s test showed no superiority of one test over the other (Table 3). Taken together, these findings suggest that CSF copeptin is not superior to serum copeptin for distinguishing either children with febrile seizures alone or children with febrile seizures and/or meningitis from febrile children without seizures. In this context, serum copeptin alone may represent the more practical adjunctive biomarker for these distinctions. To improve the clinical relevance of the study, we performed an additional exploratory direct comparison between children with febrile seizures and those with meningitis. In this analysis, neither serum copeptin nor CSF copeptin showed significant intergroup differences, and ROC analyses demonstrated only limited discriminatory performance. Moreover, DeLong testing showed no significant difference between the AUCs of serum and CSF copeptin for this distinction. Given the very small meningitis sample size, these findings should be interpreted with caution and regarded as hypothesis-generating rather than definitive. Based on the present findings, copeptin cannot yet be considered a definitive biomarker for improving the diagnosis of meningitis or for reliably discriminating life-threatening disease from febrile seizures. Although serum copeptin may have adjunctive value in the evaluation of febrile seizures, larger prospective studies are required before any diagnostic role in meningitis can be established. We were unable to identify any study investigating independent determinants of CSF copeptin levels. In our multivariable linear regression analysis, which included age, sex, serum copeptin, creatinine, sodium, potassium, CRP, leukocyte count (x103/mm3), and CSF glucose, CRP and CSF glucose were identified as independent predictors of CSF copeptin, regardless of age and sex (Table 4). Although bacterial and viral meningitis can occur at any age, viral meningitis is particularly common in younger children, and bacterial meningitis is a CNS infection that may follow a severe and often fatal course [24]. Several studies have reported that meningitis may cause substantial mortality and morbidity, especially in early infancy. Although viral infections are the most common cause of meningitis in infants younger than 3 months, bacterial meningitis may also occur and can be fatal [25, 26]. In our study, the younger age of the Febrile Illness Without Seizures Group compared with the Control Group most likely reflects the fact that lumbar puncture is performed more frequently in infants, particularly those younger than 3 months, who present with fever without an identifiable source, both to exclude meningitis and to allow early diagnosis and treatment. In addition, because this was a cross-sectional study, all patients who underwent lumbar puncture during the study period and met the inclusion and exclusion criteria were enrolled, and complete age matching among groups could not be achieved. Conclusion Serum copeptin levels were higher in children with febrile seizures than in febrile children without seizures and healthy controls. However, because of the very small meningitis sample size, the meningitis-related findings should be considered exploratory and interpreted with caution. Larger prospective studies are required to determine whether copeptin has clinically meaningful utility in distinguishing febrile seizures from meningitis. Strengths and limitations This study has several limitations and strengths. The major limitation was the inability to establish a healthy control group for CSF copeptin because lumbar puncture is an invasive procedure and could not ethically be performed in healthy children. Additional limitations include the relatively short study period, the small number of patients with meningitis, the single-center design, and the limited overall sample size. The very small meningitis subgroup substantially limits inference regarding this population. It should also be noted that several meningitis patients requiring intensive care were not included because they met predefined exclusion criteria involving comorbid conditions that could potentially affect copeptin levels, such as hydrocephalus, chronic kidney disease, heart disease requiring diuretic use, and other conditions affecting fluid-electrolyte balance. Therefore, the present findings may not be generalizable to critically ill meningitis populations. The duration of fever before lumbar puncture was not systematically recorded and therefore could not be analyzed. Vaccination status was not systematically recorded for the Febrile Illness Without Seizures Group, the Febrile Seizure Group, or the Control Group. Although follow-up hemogram and CRP measurements were generally performed at approximately 48–72 h according to clinical status, serial follow-up laboratory values were not part of the study design and were therefore not systematically recorded for analysis. ESR was not available at baseline in the emergency department in our setting. No follow-up lumbar punctures or post-treatment copeptin measurements were performed. Because the post hoc power analysis for the CSF copeptin findings yielded a power of only 30%, the CSF-related results should be interpreted cautiously and may not be generalizable. On the other hand, given that the available literature on CSF copeptin is limited to a small number of adult studies in stroke and subarachnoid hemorrhage, this is, to the best of our knowledge, the first study to evaluate CSF copeptin in children in this clinical context, which represents an important strength. Other strengths include the demonstration of a correlation between serum and CSF copeptin levels, the identification of a formula enabling point estimation of CSF copeptin from serum copeptin, and the determination of age- and sex-independent predictors of CSF copeptin levels. Acknowledgements The authors thank all patients and their families for their participation in the study. Use of AI and AI-assisted technologies The authors used AI-assisted tools for language editing only. All content was critically reviewed by the authors, who take full responsibility for the final version of the manuscript. Abbreviations Arginine vasopressin Confidence interval Central nervous system C-reactive protein Cerebrospinal fluid Enzyme-linked immunosorbent assay Febrile seizure Lumbar puncture Polymerase chain reaction Receiver operating characteristic Area under the curve Interquartile range Standard deviation Authors’ contributions HUH conceived and designed the study, collected data, performed the literature review, interpreted the data, and drafted the manuscript. CK contributed to data collection, data interpretation, and critical revision of the manuscript. ME supervised the study, contributed to study design and interpretation of the findings, and critically revised the manuscript. All authors read and approved the final manuscript. Funding No external funding was received for this study. The copeptin assay kits were purchased by the authors. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study was approved by the institutional ethics committee in accordance with the principles of the Declaration of Helsinki. Ethical approval was obtained from the Clinical Research Ethics Committee of Istanbul Haseki Training and Research Hospital, University of Health Sciences, prior to the start of the study (Decision No. 88-2023, dated 10 May 2023). Written informed consent was obtained from the parents or legal guardians of all participants, and assent was obtained from the children when age-appropriate. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.