Heart rate variability measures as biomarkers in patients with psychogenic nonepileptic seizures: Potential and limitations

Abstract

Heart rate variability (HRV) metrics provide reliable information about the functioning of the autonomic nervous system (ANS) and have been discussed as biomarkers in anxiety and personality disorders. We wanted to explore the potential of various HRV metrics (VLF, LF, HF, SDNN, RMSSD, cardiovagal index, cardiosympathetic index, approximate entropy) as biomarkers in patients with psychogenic nonepileptic seizures (PNES). HRV parameters were extracted from 3-minute resting single-lead ECGs of 129 subjects (52 with PNES, 42 with refractory epilepsy and 35 age-matched healthy controls). Compared with healthy controls, both patient groups had reduced HRV (all measures P < 0.03). Binary logistic regression analyses yielded significant models differentiating between healthy controls and patients with PNES or patients with epilepsy (correctly classifying 86.2 and 93.5% of cases, respectively), but not between patients with PNES and those with epilepsy. Interictal resting parasympathetic activity and sympathetic activity differ between healthy controls and patients with PNES or those with epilepsy. However, resting HRV measures do not differentiate between patients with PNES and those with epilepsy.

Introduction

Psychogenic nonepileptic seizures (PNES) resemble epileptic seizures, but are not associated with ictal discharges in the brain. PNES have psychosocial causes and are categorized as a manifestation of dissociative or somatoform (conversion) disorder in the current diagnostic manuals [1], [2]. This means that PNES are interpreted as an involuntary response to distress [3]. Although PNES are not thought to be explained by pathophysiological changes in the brain, cognitive, electrical, and structural brain abnormalities are found more commonly in patients with PNES than in healthy individuals [4], [5], [6]. The relationship between PNES and the functioning of the autonomic nervous system has not previously been studied in detail.

A large body of research shows that there is a complex interplay between the resting tone of the two components of the autonomic nervous system (ANS) and the body's ability to engage with and respond to external stimuli [7]. Most attention has been directed to the sympathetic component of the stress response. However, some studies have also focused on the parasympathetic nervous system. Most specifically, in adults, a lower resting vagal tone (VT) has been associated with higher rates of physical illness and poorer recovery from invasive neurosurgical procedures [8]. Individuals with a lower resting VT also show greater attachment anxiety [9], increased defensiveness, and lower behavioral activation sensitivity [10].

The parasympathetic nervous system appears to act as a buffer, allowing people to react more appropriately to external stressors. A low resting VT has been found in patients with panic disorder [11], generalized anxiety disorder [12], posttraumatic stress disorder (PTSD) [13], and borderline personality disorder [14] and in a small group of women with PNES (N = 20), most of whom had experienced sexual abuse [15].

In the majority of the studies cited, VT measures were derived from the analysis of RR interval changes in the electrocardiogram. These measures are based on the observation that the vagal nerve is responsible for the heart rate changes associated with respiration (respiratory sinus arrhythmia [RSA]). Whereas pharmacological blockade of vagal synapses at the sinoatrial node of the heart nearly abolishes the coupling of heart rate and respiration, the interruption of sympathetic inputs via β-adrenergic blockade does not [16]. In mathematical terms, RSA can be defined as the natural log of band-limited variance (0.15–0.40 Hz) of beat-to-beat interval time series. Although there are unresolved controversies about the use of RSA-derived measures of VT and their interpretation (especially in relation to making adjustments for the respiratory rate) [17], [18], there are high levels of correlation between RR interval-derived parameters and other indicators of VT (such as the fast phase of the pupillary light response) [19], [20], and a combined task force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology concluded that ECG-derived measures of VT are suitable for future studies [21].

Given the findings of a low VT in similar mental disorders, this study was undertaken to examine the potential of VT and other heart rate variability (HRV) parameters as biomarkers in a consecutive group of patients with PNES. To evaluate whether it helps in the differentiation of people who have epilepsy, the findings in patients with PNES were compared with those of healthy controls and patients with epilepsy.