Serum Ammonia as a Biomarker to 
Differentiate Between True Epileptic and 
Psychogenic Non-epileptic Seizures: 
A Cross-sectional Study

Abstract

Background

A kind of conversion condition known as psychogenic non-epileptic seizures (PNES) is characterised by brief episodes that resemble epileptic fits but lack electroencephalographic (EEG) confirmation. Although the clinical history plays a significant role in the seizure diagnosis, imaging and serum markers are included in the initial evaluation for patients who present to the emergency.

Objectives

To validate the prevalence of serum ammonia as a biomarker in differentiating true epileptic and non-epileptic seizures.

Material   Methods

The study was a hospital-based observational, prospective comparative study spanning over 2 years after surpassing the patient consent, inclusion and exclusion criteria.

Results

The mean of the study participants were middle-aged, males predominant in IES and females in PNES. A slightly higher proportion of the patients in the IES group had episodes of less than 5 minutes in the PNES group, where episodes lasted between 5 to 10 minutes. Awareness of symptoms was observed more frequently in the PNES group. The area under the curve (AUC) for the serum ammonia level at arrival was 0.972 (95% CI: 0.94–0.99) with a P value of < .001.

Conclusion

Mean serum ammonia levels at arrival and after 48 hours of admissions in the idiopathic epileptic seizure group were significantly higher compared to the PNES group with P value < .001, validating its use as a biomarker.

Introduction

Epilepsy is a disorder of the central nervous system (CNS), affecting between 1%–2% of people worldwide. Conventionally, diagnosis of epilepsy can be established when one of the following three criteria is satisfied in an individual, namely (a) individual having two or more unintentional/unprovoked reflex seizures that occur more than 24 hours apart (b) one unintentional/unprovoked reflex seizure with a ten-year recurrence rate of at least 60% and (c) a diagnosis of an epilepsy syndrome.^1–4

Biomarkers are ‘almost anything that measures reflecting an association involving the biological system and a possible risk, which might be either physical, chemical, or biological’, according to the World Health Organisation. Although the clinical history plays a significant role in the seizure diagnosis, imaging and serum markers are included in the initial evaluation for patients who present to the emergency.^5–8

Recently, there has been a lot of focus on the alleged neurological foundations of PNES, with clinicians continually looking for a reliable indicator of PNES presence (or absence). Hormones such as prolactin, catecholamines, creatine kinase, myoglobin, lactate, ammonia and urea were found raised post-convulsions, but they must be tested within two hours of activity, while increased blood ammonia could be tested up to 8 hours after seizure. In a dearth of available literature, we conducted a study to find out the predictive ability of serum ammonia in distinguishing between true epileptic and psychogenic non-epileptic seizures.^9–12 Strong contractions of skeletal muscles raise ammonia concentrations in the blood, outpacing the ability of the liver to help remove ammonia from the circulatory system,^12–17 which could lead to a transient hyperammonaemia, useful as a biomarker.

Aims   Objectives

Primary

To study the prevalence of a transient rise in serum ammonia levels in true epileptic seizure

Secondary

To establish the validity of serum ammonia as a biomarker in differentiating true epileptic and non-epileptic seizures.

Material   Methods

It is a hospital-based observational, prospective comparative study conducted in the outpatient and in-patient ward of the Department of Neurology, Institute of Medical Sciences and SUM Hospital, Bhubaneswar, spanning over a period of two years between February 2021 and January 2023. The study population included all the patients who presented at the Neurology Department of the hospital with seizure or seizure-like activity.

The sample size of the study was time-based because of strict inclusion criteria, footfall of the patients in the tertiary care centre and the restricted period of the study. During the study period, we were able to include 80 subjects (based on patient footfall in the centre), after getting written informed consent.

Inclusion Criteria

Age group between 13–65 years, patients presenting with seizure/seizure-like activity, duration within 8 hours since the last episode, true seizure diagnosed as defined by the International League Against Epilepsy (ILAE) 2014 and PNES diagnosed as defined as a part of DSM V.

Exclusion Criteria

Patients who presented after 8 hours of the last seizure activity, on prior continuing anti-seizure medications (valproic acid, carbamazepine, topiramate), on drugs - salicylates, sulfadiazine, with prior comorbidities including liver, renal diseases, with infections (particularly Proteus, E.coli, Klebsiella), haematological disorders - multiple myeloma, acute leukaemia, patients with inborn errors of metabolism and non-consenting individuals.

Study Methodology

Each patient presented with seizure or seizure-like activity was approached to be included in the study. Once the patients were enrolled, they were subjected to data collection through a semi-structured questionnaire which included the basic socio-demographic characteristics, past clinical history and current history of the seizure activity.

The enrolled subjects were then subjected to a battery of tests including:

• Serum ammonia level, using Siemens Dimension 1500 Vista R system (within 8 hrs of seizure, repeat sample after 48 hrs)

• Liver function test

• Kidney function test

• Electrocardiogram

• Routine electroencephalogram

• Computed Tomography (CT) Scan of Brain) - (if MRI not feasible)

• Magnetic resonance imaging (MRI) of the brain to rule out pathological conditions (both plain and contrast-enhanced)

Selection biases were minimised during patient recruitment and data collection by a designed questionnaire. All the data were recorded and cleaned for further analysis.

Statistical Analysis

The data were gathered, put into Microsoft Excel 2007 for additional analysis and then transferred to IBM Corporation’s SPSS programme, version 27.0. Numbers and percentages were used to express categorical variables. Fischer exact test or the Chi-squared test was used to examine the association between two category variables. The mean and standard deviation were used to express all of the quantitative variables. Depending on the distribution, either a Mann-Whitney U test or an independent sample t-test was used to compare the means of the two groups. Prior to running the statistical test, the Shapiro-Wilks test was used to determine the normality of the weather and the continuous variables. The Receiver Operating Characteristic Curve (ROC) was plotted to determine the discriminative ability of serum ammonia at the admission and 48 hours for the true seizure and PNES. P value less than .05 was considered statistically significant.

Results

The present study was an attempt to find out the predictive ability of serum ammonia in distinguishing between true epileptic and psychogenic non-epileptic seizures. For the final analysis, we included 40 patients in IES and PNES groups each.

The mean age of the study participants was 28.96 ± 15.52 years with a minimum of 13 years and a maximum of 65 years. The mean age of the patients in the IES group was 30.35 ± 17.26 years, while the mean age in the PNES group was 27.58 ± 13.64 years (Figure 1). This mean difference was not statistically significant (P value = .427). The majority of the patients in the IES group were male (62.5%), while the majority of the patients in the PNES group were females (65%). This difference in proportion was statistically significant with a P value of .014.

Figure 1. Histogram Showing Age Distribution of the Study Participants.

The majority of the study participants were students (43.8%) and were almost equally distributed in the IES group (47.5%) and PNES group (40%) (Table 1). The majority of the homemakers were in the PNES group (42.5%) compared to the IES group (20%). The distribution of the occupation did not show any statistically significant difference between the groups with a p-value of 0.171. Generalised tonic-clonic seizure was seen in 67.5% of the subjects in the IES group compared to 52.5% of the PNES group. Jaw clenching as a presenting symptom was found in higher proportions in the PNES group (30%) compared to only 5% in the IES group. This difference was statistically significant (P value = .003). Uprolling of the eyeball (72.5%), tongue bite (27.5%), urinary/bowel incontinence (65%), loss of consciousness (80%) and postictal confusion were observed as being significantly higher in the IES group compared to the PNES group (P value < .05). Awareness symptoms were observed more frequently in the PNES group (62.5%) compared to the IES group (17.5%), which was statistically significant with P value of < .001 (Table 2). We did not find any statistically significant difference concerning loss of consciousness and generalised stiffening as presenting symptoms. A slightly higher proportion of the patients in the IES group had episodes of less than 5 minutes (50%) compared to 40% in the PNES group. A higher proportion of patients in the PNES group had episodes lasting between 5 to 10 mins (47.5%) compared to only 27.5% in the IES group (Figure 2).

Table 1. Distribution of the Study Participants with Respect to Occupation.

	IES Group	PNES Group	Total	P Value
	N (%)	N (%)	N (%)
Unemployed	2 (5.0)	0 (0)	2 (2.5)	.171
Student	19 (47.5)	16 (40.0)	35 (43.8)
Business	5 (12.5)	3 (7.5)	8 (10.0)
Homemaker	8 (20.0)	17 (42.5)	25 (31.3)
Farmer	6 (15.0)	4 (10.0)	10 (12.5)

Table 2. Distribution of Presenting Signs Among the Study Population.

	IES Group N (%)	PNES Group N (%)	Total N (%)	P Value
Uprolling of eyeballs	29 (72.5)	16 (40.0)	45 (56.3)	.003
Tongue bite	11 (27.5)	2 (5.0)	13 (16.3)	.006
Urinary and bowel incontinence	26 (65.0)	0 (0)	26 (32.5)	<.001
Awareness symptoms	7 (17.5)	25 (62.5)	32 (40.0)	<.001
Loss of consciousness	32 (80.0)	24 (60.0)	56 (70.0)	.051
Post-ictal confusion	19 (47.5)	0 (0)	19 (23.8)	<.001

Figure 2. Pie Chart Showing Distribution of the Study Participants with Respect to Duration of Episodes.

The mean serum ammonia level at arrival in the IES group was 90.0 ± 20.63, while it was 47.83 ± 12.25 in the PNES group. Similarly, at 48 hours the mean serum ammonia level in the IES group was 42.68 ± 13.0, while it was 33.9 ± 7.39 in the PNES group. At both time points serum ammonia levels in the IES group were significantly higher compared to the PNES group with P value < .001 as shown in the boxplot (Figure 3) and the receiver operating characteristic curve (Figure 4) of the serum ammonia level at arrival and 48 hours for predicting idiopathic true epileptic seizure. The area under the curve (AUC) for the serum ammonia level at arrival was 0.972 (95% CI: 0.94–0.99) with a P value of < .001. The area under the curve (AUC) for the serum ammonia level at 48 hours was 0.722 (95% CI: 0.61–0.83) with a P value of .001. This shows a better predicting ability of serum ammonia for idiopathic true epileptic seizure. Higher serum ammonia levels correlated directly proportional to the area and the severity of the motor convulsions.

Figure 3. Boxplot Showing Comparison of Serum Ammonia Level Between the Groups at Different Time Points.

Figure 4. ROC Curve Showing Predictive Ability of the Serum Ammonia at 48 Hours of Arrival for IES and PNES.

Discussion

Hyperammonaemia, which we here define as high plasma ammonia levels (> 80 µmol/L), has been noted in more than 50%–60% of patients who have had a convulsive seizure and is occasionally tested for the detection of seizures in the emergency room.^16–18 Clearly visible changes to metabolism result after convulsions. The neuroendocrine system is stimulated by maximum neural stimulation to produce hormones such as prolactin and catecholamines. As agitated muscle cells release the enzyme creatine kinase and myoglobin, overworked cells emit compounds such as lactate, ammonia and urea. Following this, there is an inflammatory reaction involving the release of cytokines and leukocytosis (Table 3). Increased blood ammonia was found between 3–8 hours after seizure, in contrast to other biomarkers like Prolactin, which has to be tested within two hours. This suggests that blood ammonia might be easier to access across a wider time window following a seizure event.^{19, 20}

Table 3. Laboratory Test Differentiating Epileptic and Non-epileptic Seizure.

Sl No.	Laboratory Tests
1	Hormones:  Prolactin  Arginine vasopressin and Copeptin  Other Pituitary hormones like ACTH, cortisol
2	Muscle enzymes:  Creatinine kinase  Myoglobin
3	Metabolites:  Ammonia  Lactate  Uric Acid
4	Cerebral stress markers:  Neuron specific Enolase  Protein S100-B

Source: Parvareshi Hamrah et al.31

Epileptic and nonepileptic seizures are nearly usually diagnosed purely based on the medical record, which must be skilfully gathered and frequently necessitates protracted questioning of the patient and witnesses.^{21, 22} Common diagnosing presentations of PNES include: Frequently brought on by demanding situations and in reaction to advice, taking place when you are awake and in front of witnesses, are not stereotypical, accelerating and decelerating, asymmetrical, waxing and waning, asynchronous convulsive-like motions, sometimes accompanied by pelvic thrusts, flailing and shaking hands are what are referred to as ‘convulsions’.^22–24 These may be halted or unyielding under control, the eyelids tend to thicken when people try to open their eyes passively, the level of awareness may remain constant or exhibit noticeable changes, absence of postictal confusion, the patient may get upset and cry and be resistant to antiepileptic drugs. Numerous theories have been put forth in recent years to clarify the fundamental aetiology of seizures, including degeneration of neurons, amygdala abnormalities, oxygen deprivation, disruption of the brain-blood barrier (BBB), changes to the glutamatergic framework and epigenetic DNA alteration, and a vast majority of research on inflammation and seizures shows that markers of inflammation play a significant role in epilepsy development via the complement pathway or the irregularities of cytokine equilibrium.^25–27 Ammonia, which is a precursor to glutamine, is often converted to the readily excreted compound urea in the liver. As adenosine triphosphate (ATP) is used up during extended epileptic seizures, strong contractions of skeletal muscles raise ammonia concentrations in the blood, outpacing the ability of the liver to help remove ammonia from the circulatory system. High quantities of ammonia cause it to pass the blood-brain barrier, where it is transformed into glutamine, which is an osmotic agent that encourages cerebral oedema, and it continues to be neurotoxic.^{22, 28–30}

Study Limitations

The sample size is small and the study is a single-centre design, which might have diagnostic analyser bias and limitations. Also, potential confounders like undiagnosed metabolic conditions are yet to be ascertained in such cases.

Conclusion

A significant number of the patients in the PNES group were females, which shows that PNES majorly affects the female gender. Uprolling of the eyeball, tongue bite, urinary/bowel incontinence, loss of consciousness and post ictal confusion were observed as being significantly higher in the idiopathic epileptic seizure group compared to the PNES group (P value < .05), while awareness symptoms were observed more frequently in the PNES group. Mean serum ammonia level at arrival and after 48 hours of admissions in the idiopathic epileptic seizure group were significantly higher compared to the PNES group with P value < .001.

Serum ammonia could be immediately implemented as a clinical biomarker but also further studies are needed to validate findings in larger, more diverse populations, or explore serum ammonia in other seizure subtypes. There is minimal literature regarding the same. We aim to add to valuable data to existing literature to promote serum ammonia as a screening test in emergency admissions.

Abbreviations

CNS, Central nervous system; BBB: Brain-blood barrier; DNA, Deoxyribo Nucleic Acid; ATP, Adenosine triphosphate; EEG, Electroencephalograph; PNES, Psychogenic Non-epileptic Seizures; IES, Idiopathic Epileptic Seizures; ILAE, International League Against Epilepsy; GTC, Generalised tonic-clonic; WHO, World Health Organisation.

Authors’ Contributions

Sumirini Puppala: Primary draft writing. Srikanta Sahoo: Final draft revision and editing. Srimant Pattnaik: Methodology. Surjyaprakash Choudhury: Review of literature. Abhijit Acharya: Tables and images compilation. T. R. Sirohi: Descriptive analysis.

Authorship Statement

All authors have agreed to conditions of authorship.

Informed Patient Consent

All the patients were informed about the purpose of the study using patient information sheets in both Odia and English language. The participation was purely voluntary, and the subjects had the option to opt out of the study at any point in time without affecting their routine care. Only subjects who gave written informed consent were included in the study.

Responsibility

Sumirini Puppala takes full responsibility for the data.

Statement of Ethics

The study was approved by the Institutional Ethics Committee with Ref.no. IEC/IMS.SH/SOA/2022/291.