The importance of social zeitgeber in paediatric type 1 narcolepsy: What we can learn from the COVID‐19 restrictions adopted in Italy?

Abstract The lockdown due to the new coronavirus pandemic (COVID‐19) has led to unparalleled changes in several aspects of human behaviour. During the lockdown, the general population delayed sleep timing and spent more time in bed; however, little is known on the effects of COVID‐19 restriction on children and adolescents suffering type 1 narcolepsy. In the last months of 2019, we performed follow‐up actigraphy in 18 type 1 narcolepsy children and adolescents under stable pharmacological treatment with sodium oxybate. We contacted these patients for a follow‐up actigraphy during the first Italian lockdown. Actigraphs and the Epworth Sleepiness Scale for children and adolescents (ESS‐CHAD) have been sent to participants’ homes. Differences in motor activity were analysed through functional linear modelling. During lockdown, type 1 narcolepsy children and adolescents went to bed and woke up later, slept more during the daytime and napped more frequently. No difference emerged in time in bed, estimated total sleep time and nocturnal sleep quality. Similarly, no difference emerged in ESS‐CHAD and body mass index. The time‐series analysis of motor activity documented reduced activity during the early morning and in the evening during the lockdown period compared with pre‐lockdown. Our study objectively showed that type 1 narcolepsy children and adolescents delayed the sleep phase and slept more during the daytime during the lockdown. The analysis of type 1 narcolepsy children and adolescents’ behaviour during the lockdown has provided new information that could pave the way to a personalized school programme.

have difficulties in school and in integration with peers (Vignatelli et al., 2019).
In the last months of 2019, a new coronavirus disease , firstly identified in China, quickly spread worldwide forcing several countries to impose national lockdown.
In Italy, the first and strictest lockdown took effect on 11 March and lasted till 18 May 2020: during this period, people were not allowed to leave their home except for limited and documented purposes. Educational activities, from primary school to university, were replaced with remote learning (Decreto del presidente del consiglio dei ministri, 2020).
The restrictions due to the COVID-19 pandemic have led to unparalleled changes in several aspects of human behaviour. The effects of COVID-19 restrictions on sleep have been assessed in several studies, often through retrospective questionnaires. During the lockdown, people delayed sleep phase and spent more time in bed (TIB; Blume, Schmidt, & Cajochen, 2020), but at the same time reported a worsening of sleep quality (Cellini, Canale, Mioni, & Costa, 2020). Suspecting that NT1 children and adolescents might be more susceptible to the COVID-19 lockdown, also in light of the increased ultradian drive to sleep (Filardi, Pizza, Bruni, Natale, & Plazzi, 2016), we aimed to contribute to the knowledge on the impact of COVID-19 restrictions on NT1 children and adolescents by objectively assessing nocturnal sleep, napping behaviour and motor activity pattern through actigraphy, a methodology that has already proved useful in providing information on disease course, and in monitoring effects and adherence to pharmacological treatments (Filardi, Pizza, Antelmi, Ferri, et al., 2018;Filardi, Pizza, Antelmi, Pillastrini, et al., 2018).

| ME THODS
Between November 2019 and January 2020, we performed routine follow-up actigraphy in 30 NT1 children and adolescents (mean age: 13.27 ± 3.47 years) under stable pharmacological treatments with sodium oxybate. Data collection ended on 30 January, 24 days before the closure of schools (23 February) and the subsequent first national lockdown (11 March-18 May 2020).
We therefore decided to follow up these patients during the lockdown period.
Eighteen NT1 children and adolescents (11 males, mean age: 14.44 ± 2.01 years, range 10-17 years) accepted to participate and to wear an actigraph continuously for 14 days. Actigraphs (Micro Motionlogger Watch, Ambulatory Monitoring) have been sent to participants' homes, along with the Epworth Sleepiness Scale for children and adolescents (ESS-CHAD) and the Italian version of the consensus sleep diary, which participants have to fill in daily (Carney et al., 2012;Wang, Benmedjahed, & Lambert, 2017). Moreover, participants were asked to report their height and weight upon receiving the device.
Actigraphic recordings have been processed with the Sadeh algorithm to obtain estimated sleep measures. We computed, separately for weekdays and weekend, bedtime, wake-up time and midpoint of sleep (MS, the middle time-point between bedtime and wake-up time). Social jetlag was computed as the difference between MS of weekend and of weekdays (Wittmann, Dinich, Merrow, & Roenneberg, 2006).
We considered the following estimated nocturnal sleep measures: TIB, total sleep time (eTST), wake after sleep onset (eWASO), sleep efficiency, awakenings and prolonged (lasting more than 5 min) awakenings, longest uninterrupted sleep episode and sleep motor activity (SMA). For the diurnal part of the recording, we considered diurnal motor activity (DMA), estimated diurnal total sleep time (eDTST), estimated nap frequency, and mean nap duration (NapD).
The study was approved by the local health trust's ethics committee (Comitato Etico Interaziendale Bologna-Imola, CE-BI, Prot. Num. 17009), and written informed consent was obtained from the children's parents.

| Statistical analysis
Continuous and categorical data were explored with descriptive statistics (mean ± SD). Differences between pre-lockdown and lockdown period in ESS-CHAD, body mass index (BMI) and actigraphic parameters were assessed by means of paired samples t-test followed by effect size computation (Hedges' g) in case of statistically significant results. To assess changes in motor activity profile, we processed the time-series of raw motor activity data through functional linear modelling (FLM; Wang et al., 2011). Firstly, we extracted raw motor activity data through the Action-4 software (Ambulatory Monitoring); to avoid variations related to the weekend, we considered solely data from Sunday midnight to Friday evening (corresponding to the regular school week in Italy). The five vectors of motor activity data were averaged into a single activity profile and fitted using a Fourier expansion model with n = 19 basis permutations. Differences in motor activity were assessed through non-parametric permutation F-test. Finally, we fitted a FLM on the motor activity profile of NT1 children and adolescents during the lockdown period according to differences in the ESS-CHAD scores between the two evaluations (ᐃESS-CHAD: ESS-CHAD of the lockdown minus ESS-CHAD of the pre-lockdown). Statistical analyses were performed with R and SPSS 19.0 (SPSS); p-values < .05 were considered statistically significant.

| RE SULTS
Actigraphic parameters, ESS-CHAD score and BMI values are reported in Table 1.
None of the patients discontinued treatment with sodium oxybate during the lockdown, but we observed a slight increase in daily intake (7.25 ± 0.75 versus 7.42 ± 0.72, t (17) = −2.06, p = ns). No statistically significant differences were observed in BMI (t (17) = 0.30, p = ns) and ESS-CHAD (t (17) = 0.27, p = ns). Similarly, no differences were observed in actigraphic measures reflecting sleep quality (eTST, eWASO, estimated sleep efficiency, longest uninterrupted sleep episode, awakenings and prolonged awakenings). Significant differences emerged in sleep timing with NT1 children and adolescents who went to bed and woke up significantly later during the lockdown period on both weekdays and weekend (all p < .005). Social jetlag decreased during the lockdown period (t (17) = 2.36, p < .05).
The motor activity profile of NT1 children and adolescents prior and during the lockdown period is reported in Figure 1

| DISCUSS ION
Our study is the first to have objectively assessed the effects of Unique to our study design is the possibility to objectively study changes in diurnal sleep pattern and motor activity profile of young NT1 patients while confined at home.
Overall, NT1 children and adolescents slept more during the daytime and napped more frequently during the lockdown, but the duration of naps remained similar to the pre-lockdown period.
On the other hand, mean motor activity levels decreased during the lockdown. Second, promote a regular schedule of at-home physical exercise; albeit in our study the decrease in motor activity was not associated with an increase in BMI, probably also due to the proximity between observations; the promotion of regular physical exercise is of the utmost importance in a population that is prone to weight gain (Filardi, Pizza, Antelmi, Ferri, et al., 2018;Filardi, Pizza, Antelmi, Pillastrini, et al., 2018).
Our study also suffers several limitations. First, the sample size is relatively small, which makes problematic the comparison with previous studies (Postiglione et al., 2020;Rodrigues Aguilar et al., 2020). Second, we could not compare the severity of NT1 symptoms

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available from the corresponding author upon reasonable request.