A prospective observational study including all living CLN3 patients in Denmark > 15 years of age.

Patients

In Denmark, all individuals with NCL are allocated to the Centre for Rare Diseases, Aarhus University Hospital, and are observed at the NCL clinic regularly at least every 6 months. In May 2022, the number of alive CLN3 patients were 20 (7 male, 13 females, 4–39 years of age). Since the periods of anxious, fearful behavioral symptoms almost exclusively occur beyond 14 years of age [9, 10], only the nine patients ≥ 15 years of age (3 male and 6 female) were offered to participate.

Standard protocol approval, registrations, and patient consent

The study was approved by the Danish Ethics Committee (report number: 87469) and individual written informed consents were obtained.

Clinical assessment of the patients

Clinical assessment at inclusion of the study was evaluated using the slightly moderated Hamburg Kohlschütter scale [14, 15]. Originally, the Hamburg-Kohlschütter scale [15] adds up to a total of 15 points in 5 categories: motor, vision, language, intellect and epilepsy. Each category ranks from 0 to 3 with 3 points representing a healthy, age-appropriate score and 0 points referring to no residual function. Since epilepsy score is strongly influenced by each patient’s anticonvulsive medication, we excluded it from the total clinical score which therefore resulted in a maximum score of 12 [14]. The scoring was assessed in our NCL clinic and from video-recordings by CB + MMH + JRO who have significant experience in clinical assessment of NCL patients.

Assessment methods of the autonomic activity

Heart Rate Variability (HRV) is a biomarker of the heart-brain interaction and the ongoing activity of the autonomic nervous system (ANS) [16, 17]. Traditionally, HRV is measured by a multi-lead electrocardiography (ECG) method using a Holter monitoring system [9, 16, 17]. Recent studies recommend use of the single-lead chest belt, Polar H10, when assessing HRV during intense activity with strong body movements [18,19,20]. It has a high correlation (r = 0.99) to the Holter monitor, and outperforms the Holter monitoring during intense motor activity [17].

The participants continuously carried a Lifecard CF Holter and a Polar H10 chest belt paired with a Polar Vantage M watch for 24 h in seven days. Time and frequency domains and non-linear HRV parameters from both ECG-devices were analysed using Kubios OY HRV Premium 3.5.0 software according to standardized techniques [16].

Time domain parameters included the standard deviation of normal-to-normal intervals (SDNN), primarily reflecting sympathetic activity, the mean RR interval (\(\overline }\)), and the root mean square of successive differences of the RR interval (RMSSD), a measure indicating the parasympathetic activity [16]. Frequency domain parameters included Low Frequency (LF) mainly implicating the tonic sympathetic influence, and High Frequency (HF) which communicates the parasympathetic impact. LF/HF ratio quantifies the sympathetic/parasympathetic balance [16]. Questions have been raised as to whether LF represents pure sympathetic activity or should be perceived as composed of a more complex and not so easily discernible mixture of sympathetic, parasympathetic, and other unidentified factors, and instead, the use of the Non-linear measurements SD1,SD2, SD2/SD1 ratio and a Poincaré plot has increased and has been recommended as a simple and more accurate way of expressing short and long term HRV [21,22,23].The Poincaré plot is presented as a graph where each RR interval is plotted against the preceding RR interval resulting a formation of distributed point cloud in the shape of an ellipse extended along the diagonal identity line y = x. Points above the line indicate R-R intervals that are longer than the preceding R-R interval, and points below the line indicate a shorter R-R interval than the previous one. Accordingly, the dispersion of points perpendicular to the line of identity (the “width”, SD1) reflects the level of short-term variability, thus correlating to RMSSD, and the dispersion of points along the line of identity (the “length”, SD2) reflects the level of long-term variability, a measurement correlating to LF. The SD2/SD1 ratio is an assessment of the autonomic balance [16]. A further important benefit of Poincaré plots is of visual nature [23]. As SD1 is mainly determined by parasympathetic activity, a smaller value of SD1 indicates weakening of the parasympathetic regulation. SD2 is related more strongly to sympathetic activity, and when the sympathetic activity is increased, the SD2 value increases. The position of the cluster gives valuable information as well. If the cluster is at the bottom of the graph, the sympathetic pathways prevail. A summary of the different used HRV parameters, their definitions and interpretations are shown in Table 1.

Table 1 Summary of the different used HRV parameters, their definitions and interpretations

A particular focus of the described HRV parameters were obtained in close temporal context to the clinically recurrent episodes of the anxious, fearful behavior in order to determine their temporally association to the activity of the autonomic nervous system. Samples were chosen with a duration of 30 min 2.5 h, 2 h, 1.5 h and 1 h before the episode. In order to increase transparency of eventual fluctuations during the episodes, the interval of measurements was decreased to every 5 min 15 min before start of every episode, during the entire length of the episodes, and every 15 min for 1 h following the end of the attacks.

Clinical grading assessment of the recurrent episodes

Since the recurrent anxiety-like episodes both phenotypically and patho-anatomically to a great extend resemble PSH following TBI [7, 9] the clinical assessment score of PSH described by Baguley and co-workers [24], including a Clinical Feature Scale (CFS) and a Diagnosis Likelihood Tool (DLT), was used to assess the likelihood of the PSH diagnosis for each episode. The CFS evaluates 6 parameters: heart rate, respiratory rate, systolic blood pressure, temperature, sweating and posture during episodes. For each parameter, a score from 0 referring to no increase/normal to a maximum of 3 representing the highest increase/severity was assessed. Eighteen is thus the maximum score. Since it was not possible to measure the systolic blood pressure and the body temperature regularly during all the episodes, we modified the CFS to a maximum of 15 and 12, respectively. The modification was done percentage wise leading to cut-off values of 1–5 (mild), 6–10 (moderate) and ≥ 11 (severe) in cases without either blood pressure or body temperature, and 1–4; 5–8 and ≥ 11 during episodes without both body temperature and blood pressure measurements [24]

The DLT evaluates the probability of the diagnosis using 11 significant clinical items; each feature counts for 1 point (for review, see ref. 24). We exchanged the term “brain injury” with the term “clinically notable brain degeneration” since there exists no exact moment of brain injury in CLN3 disease, as the brain degenerates progressively during years. Without blood pressure measurements the combined modified total score (CSF + DLT) summarizes to 26 and with the following cut-off values: < 7 (PSH diagnosis is unlikely), 7–14 (the diagnosis of PSH is possible), and when the total score summarized beyond 15, a diagnosis of PSH is signed to be probable. In patients without blood pressure and/or body temperature measurements the total score, and the scores for diagnosis unlikely, possible, and probable were < 6, 7–12 and ≥ 13, respectively [24]. The caregivers were instructed to use the Clinical Feature Scale (CFS) and the signs and severity of the individual seizures were described in a diary and by answering a questionnaire for each attack. The grading of the clinical assessment was further examined and assessed by the primary investigator (CB) and a senior physician (MHH) by using the video documentation supplemented by the caregivers reported written information of each episode. The grading, assessment, and use of the Diagnosis Likelihood Tool (DLT) were performed before CB and MHH were aware of the results of the HRV measurements. Heart rate frequencies were obtained from the ongoing ECGs.

Statistical analysis

Statistical analysis was performed using Stata Basic Edition 17 (64-bit) statistical software package. The relationship between age and HR or HRV parameters was estimated with linear regression without violation of model conditions. All data was inspected with normal quantile plots and were all found to be normal distributed. Differences in HRV parameters 30 min before and during PSH-like episode were tested using a paired Student’s t-test. A significance level of 0.05 was used.