Cardiovascular risk in adults with Delayed Sleep Phase Syndrome (DSPS) and Attention-Deficit/Hyperactivity Disorder (ADHD)

ABSTRACT Attention-Deficit/Hyperactivity Disorder (ADHD) is associated with a wide variety of sleep problems. The most common sleep disturbance in adults with ADHD is Delayed Sleep Phase Syndrome (DSPS), which leads to sleep insufficiency and social jetlag. ADHD, short sleep, and social jetlag have independently been associated with poorer cardiovascular health. Adults with both DSPS and ADHD may be particularly at risk of developing cardiovascular diseases (CVDs), which are the leading cause of death worldwide. In this study, 24-hour resting-state heart rate variability (HRV) was measured as a biomarker for cardiovascular health in 49 adults (18-55y) with DSPS and ADHD. The prevalence of cardiovascular risk factors obesity, smoking, and hypertension was determined. The majority of participants scored within average ranges for all HRV measures. The prevalence of obesity was normal compared to the general population. Smoking and hypertension were more prevalent, but not related to HRV. In conclusion, we found no evidence for a high risk of CVDs in this group. It however remains important to study cardiovascular risk in adults with DSPS and ADHD using different methodologies. 
Trial Registration FASE, https://www.trialregister.nl/, #NTR3831


Introduction
Attention-Deficit/Hyperactivity Disorder (ADHD) is a childhood-onset neurodevelopmental disorder characterized by symptoms of inattention, hyperactivity, and/or impulsivity (American Psychiatric Association 2013).In 40-60% of cases, childhood ADHD persists into adulthood (Faraone et al. 2006(Faraone et al. , 2017) ) with a cross-national estimated prevalence of 2.8% in adults (Fayyad et al. 2017).Sleep disturbances and disorders are highly prevalent in children and adults with ADHD (Bijlenga et al. 2019).The most common sleep disturbance in ADHD is due to a delayed circadian rhythm.This delay is on average 1.5 h (Coogan and McGowan 2017) and occurs in up to 78% of adults with ADHD (Van Veen et al. 2010).It is indicative of Delayed Sleep Phase Syndrome (DSPS) (American Psychiatric Association 2013), which commonly leads to chronic sleep loss caused by a mismatch between the internal and external rhythm, termed "social jetlag." Cardiovascular diseases (CVDs) are the leading cause of death worldwide (World Health Organization 2017).High risk of developing CVDs has been associated with ADHD (Bijlenga et al. 2013;Du Rietz et al. 2021;Li et al. 2022), as well as with short sleep (Cappuccio et al. 2011;Knutson 2010Knutson , 2012) ) and social jetlag in particular (Caliandro et al. 2021;Rutters et al. 2014;Sűdy et al. 2019).Major risk factors for CVDs are obesity, hypertension, and smoking (National Health Service 2022; World Health Organization 2017), all of which are strongly related to ADHD (Chen et al. 2018;Cortese 2019;Cortese et al. 2016;Lee et al. 2011;McClernon and Kollins 2008) and to short sleep and social jetlag (Knutson 2010;Meng et al. 2013;Roenneberg et al. 2012;Wittmann et al. 2006).Adults with both DSPS and ADHD may thus be especially at risk of developing CVDs.
The risk of developing CVDs can be evaluated by using heart rate variability (HRV) as a biomarker.In a healthy heart, the interval between consecutive heartbeats varies and HRV represents this fluctuation (Shaffer et al. 2014).An optimal HRV is related to adaptability to changes in the environment and executive functions such as attention and emotion regulation (Shaffer and Ginsberg 2017;Thayer et al. 2009).Lower HRV has been associated with poorer attention (Suess et al. 1994), emotion regulation (Appelhans and Luecken 2006), and executive functioning (Thayer et al. 2009), all of which are often disturbed in ADHD (Sonuga-Barke 2003).HRV can be measured in various ways, using short-term (≤5 minutes) or long-term (24 hours) values, either task-related or in resting state.Regardless of the method, low HRV is associated with higher mortality, although high HRV is not necessarily better as this can be the result of pathological conditions.Because studies use different HRV outcomes, it is difficult to compare findings between studies (Shaffer and Ginsberg 2017).
The relationship between HRV and ADHD has mostly been studied in children, using short-term HRV outcomes.A meta-analysis on short-term restingstate HRV comprising 8 studies in 587 medication-free children reported no differences between those with and without ADHD, but mentions that short-term measures may not be sufficient in detecting the influence of disorders like ADHD on the cardiovascular system (Koenig et al. 2017).Only one study performed longterm (24 h) measurements and found lower restingstate HRV in 23 children with ADHD (of which 11 used stimulant medication), compared to 19 controls, which indicated poorer cardiovascular health.The differences were largest during sleep in unmedicated children with ADHD (Buchhorn et al. 2012).
It is suggested that task-related HRV rather than resting-state HRV may show larger differences between children with and without ADHD in response to cognitively demanding tasks (Koenig et al. 2017;Robe et al. 2019).This has been confirmed by a meta-analysis comprising 13 studies in 1778 medication-free children, showing that task-related short-term HRV during tasks on e.g.attention and emotion regulation was lower in those with ADHD than in controls (Robe et al. 2019).
So despite methodological differences, there is some evidence from studies in children that HRV may be lower in those with ADHD than controls (Buchhorn et al. 2012;Robe et al. 2019), indicating poorer cardiovascular health that may lead to the development of CVDs later in life.Little is known about the relationship between ADHD and HRV in adults.Two small studies in respectively 36 and 42 medication-free adults found no differences in short-term HRV between those with and without ADHD, either in resting-state or taskrelated (Lackschewitz et al. 2008;Oliver et al. 2012).Long-term HRV has not yet been studied in adults with ADHD.It should be noted that ADHD is typically treated with stimulant medication, which increases heart rate and blood pressure (Mick et al. 2013).Nevertheless, there is currently no evidence that longterm stimulant use increases the risk of developing CVDs (Hammerness et al. 2015;Zhang et al. 2022).One pilot study in children with ADHD even suggested a cardio-protective effect of stimulants on HRV (Buchhorn et al. 2012), although this has not yet been confirmed or studied in adults.
Where the literature on HRV in ADHD is relatively scarce, the relationship between HRV and sleep and circadian rhythm has been widely investigated.Sleep and circadian rhythm have a complex, bidirectional relationship with HRV and cardiovascular regulation in general (Tobaldini et al. 2013), an overview of which goes beyond the scope of the current paper.A recent study in adults on the relationship between resting-state HRV during sleep and social jetlag suggested that social jetlag negatively impacts the cardiac regulation of sleep (Sűdy et al. 2019).Various other studies focused on circadian misalignment, either experimentally enforced by sleep restriction or in populations of shift workers.Decreases in HRV during sleep have been reported that suggest that circadian misalignment can increase the risk of developing CVDs (e.g.Cosgrave et al. 2021;Morris et al. 2016).Findings on short sleep duration have been mixed.Insomnia patients with short objective sleep duration had poorer cardiovascular health compared to those with normal sleep (Cosgrave et al. 2021;Jarrin et al. 2018).Furthermore, one study found that adults with shorter objective sleep duration showed a greater reduction in HRV during stress tasks, indicating higher cardiovascular risk (Mezick et al. 2014).Two other studies suggested that poor sleep quality, rather than short sleep duration, is related to poorer HRV, both in adults (Sajjadieh et al. 2020) and in children (Michels et al. 2013).
Taken together, it seems that social jetlag, circadian misalignment, and short sleep duration, all of which are consequences of DSPS, are associated with poorer cardiovascular health.People with both DSPS and ADHD may thus be particularly at risk for developing cardiovascular diseases.The aim of this study is to explore the cardiovascular risk of this group.We measured 24-hour resting-state HRV as a biomarker for cardiovascular health and assessed the prevalence of cardiovascular risk factors obesity, smoking, and hypertension.Despite a higher than average prevalence of smoking and hypertension, we found no evidence for a high risk of CVDs in our study population.Several methodological limitations should however be considered when interpreting the results, and it remains important to study cardiovascular risk further in this group.

Materials and methods
This paper uses baseline data from the PhASE study (Phase Shift in ADHD of Sleep and Appetite), a placebo-controlled randomized clinical trial investigating the effects of chronotherapy in adults with ADHD and DSPS on circadian rhythm, sleep, indications for chronic diseases, and appetite hormones.PhASE was registered in the Netherlands Trial Register, #NTR3831, and approved by the Medical Ethical Committee of Leiden, protocol #NL39579.058.12 (date of approval 22-11-2012).The study protocol complies with the Helsinki Declaration of 1975, as revised in 2008, and international ethical standards for biological rhythms research (Portaluppi et al. 2010).

Participants
Recruitment took place at the specialized PsyQ outpatient adult ADHD clinic in The Hague, The Netherlands.Patients were eligible if they had a clinical diagnosis of both ADHD and DSPS.The most important exclusion criteria were the use of ADHD or sleep medication, presence of major comorbid disorders, and substance use.51 participants were included and signed written informed consent prior to inclusion.The participants did not take antihypertensive medication or any other medication that affected the cardiovascular system.The recruitment process and in-and exclusion criteria have been fully described previously (Van Andel et al. 2021).

Study assessments
Age, sex, and ADHD subtype (inattentive, hyperactive/ impulsive, or combined) were assessed.Participants were also asked whether they smoked and if so, how many cigarettes per day, and their body mass index (BMI) was calculated (weight/height 2 ).
Heart rate variability (HRV) parameters were measured by electrocardiography (ECG) over the course of 24 hours.Four electrodes were placed on the (if necessary, shaven) chest cleaned with alcohol.Measurements were recorded by Multichannel Holter ECG Recorders (Systolé Hardware ©) types H1 and H2 with sample frequencies of 100 Hz and 125 Hz, respectively.
During the same 24 hours, ambulatory blood pressure monitoring (ABPM) additionally measured blood pressure (BP) and heart rate (HR).Participants wore a blood pressure cuff (Welch Allyn ABPM 6100 ©) around their non-dominant upper arm, which measured BP and HR every 20 minutes during the day and every 60 minutes between 23:00 h and 8:00 h.Participants were instructed to keep still during the measurements.In case of an unsuccessful measurement, it was automatically attempted again.The first measurement took place in the presence of the researcher after they had attached the device.Participants were to refrain from showering, bathing, or swimming whilst wearing the devices.They kept a diary during the 24hour assessment, logging their activities and any strong emotions, sleep and wake times, and medication intake or any physical complaints if applicable.If an electrode had come off, participants were instructed to put it back as soon as possible and note in the diary when it had been disconnected.

Processing and analysis of HRV data
HRV variables were calculated from the ECG data using only normal-to-normal (NN) intervals, i.e. interbeat intervals from which artefacts had been automatically removed.Time-domain variables, which quantify the amount of variability in measurements of the interbeat interval, were: standard deviation between NN intervals (SDNN); mean of the standard deviations of all NN intervals for each 5-minute segment (SDNN index); root mean square of successive differences between normal heartbeats (rMSSD) reflecting the beat-to-beat variance in HR; and percentage of adjacent NN intervals that differ from each other by more than 50 ms (pNN50).Frequency-domain variables, which estimate the distribution of absolute power of into frequency bands, were the low-frequency (LF) band (0.04-0.15 Hz) and the high-frequency (HF) band (0.15-0.40 Hz).
It was individually determined per participant whether they scored within or outside normal ranges for HRV variables.Reference values for LF and HF were determined by the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology (1996).The other HRV variables (HR, SDNN, SDNN index, rMSSD, and pNN50) were dependent on both sex and age and the percentages were based on the normal ranges per sex and age reported by Umetani and colleagues (Umetani et al. 1998).

Statistical analysis
Simple descriptives were used to examine general characteristics and study outcomes.The percentages of participants falling outside of the normal range were determined in the whole sample.Additional analyses were performed to make comparisons between those with and without risk factors (smoking, overweight/ obesity, and hypertension) using Fisher's exact tests.Analyses were performed in SPSS version 25 (IBM Statistics).An α-level of .05 was used to indicate statistical significance.

Results
Baseline data were available for N = 49 participants, since 2 of the 51 included participants dropped out before the baseline measurement.ABPM and ECG measurements were available and valid for N = 38 for ABPM and N = 34 for ECG/HRV.
Table 1 shows general characteristics and study outcomes.41.7% of the study population smoked, with an average of 8.58 (SD 9.68) cigarettes per day.More than half (54.2%) had a BMI above 25, of which nearly one third had obesity (BMI >30).The average 24-hour blood pressure was 126/72 mmHg and 25.6% had hypertension, defined as having a systolic blood pressure of ≥140 mmHg and/or a diastolic blood pressure of ≥90 mmHg.People with hypertension had an average systolic blood pressure of 145.70 (SD 4.97) mmHg and diastolic blood pressure of 81.30 (SD 6.27) mmHg (not shown in Table 1).71.4% of people with obesity had hypertension, compared to 20.0% of people with overweight (χ 2 (1,22) = 5.46, p = 0.020) and 12.5% of people with normal weight (χ 2 (1,23) = 7.99, p = 0.005).There was no relationship between smoking and hypertension (χ 2 (1,38) = 1.28, p = 0.258), and no differences between sexes on any of the outcomes.These numbers are not shown in Table 1.
Figure 1 displays how many participants scored within and outside normal ranges for HRV variables.Low HRV is associated with higher mortality, although high HRV can be the result of pathological conditions (Kemp and Quintana 2013;Shaffer and Ginsberg 2017;Thayer et al. 2010); optimal HRV is reflected by scores within normal ranges.The majority of participants (67.6-82.4%)scored within age-and sex-specific normal ranges for SDNN, SDNN index, rMSSD, pNN50, and heart rate.LF was average in 35.3% and HF in 17.6%, for which most people had higher than average values (67.6%).There were no differences in percentages between smokers and non-smokers, people with normal and high BMI, or people with and without hypertension, apart from a larger percentage of high SDNN in smokers vs. non-smokers (Supplementary Table S1).

Discussion
The current study investigated HRV as a biomarker for cardiovascular health and the prevalence of known risk factors for developing CVDs in adults with both ADHD and DSPS.Our results could not be compared with earlier findings on HRV in adult ADHD, as these numbers are unknown due to large methodological differences regarding HRV measurement and study population between previous studies.Our aim was therefore to gain insight into the cardiovascular risk of this population in an exploratory manner, by evaluating how many people scored outside reference ranges for various HRV measurements and by assessing the prevalence of known cardiovascular risk factors.The majority of participants (67.6-82.4%)scored within ageand sex-specific normal ranges (Umetani et al. 1998) for most HRV parameters.It is unfortunately unknown how this compares to the distribution in the general population (or to DSPS or ADHD without DSPS populations), as these numbers are unknown, although these norm values were determined on the basis of data from a large sample of healthy adults of all ages (Umetani et al. 1998).Since the majority of our study population scored within these normal ranges, we believe the current results do not give any strong indications for poor cardiovascular health or a high risk of developing CVDs in this group of adults with ADHD and DSPS.We should however be careful in assuming that adults with ADHD and DSPS are not at particular risk for CVDs, as we could not compare our findings to ADHD: Attention-Deficit/Hyperactivity Disorder.BMI = Body Mass Index.ABPM = ambulatory blood pressure monitoring.Hypertension: systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg.SDNN = standard deviation between normal-to-normal (NN) intervals.SDNN index = standard deviation of all NN intervals for each 5-minute segment.rMSSD = root mean square of successive differences between normal heartbeats.pNN50 = percentage of adjacent NN intervals that differ from each other by more than 50 ms.LF = low-frequency band (0.04-0.15 Hz).HF = high-frequency band (0.15-0.40 Hz).
a control group in which the exact same methodology was used.Much previous work, as discussed in the Introduction, suggests that cardiovascular risk is associated with both ADHD and sleep problems, and there are several reasons imaginable why we would not have been able to pick up a potential high risk in the current study.
One explanation is that our study population was relatively young (18-53 y, 29.5 y on average with a median of 26), whereas CVDs generally do not develop until later in life.However, most previous work on HRV in relation to ADHD or sleep included children and young adults.A more plausible explanation is the manner in which HRV was measured.Prior work on HRV and sleep measured HRV during sleep only, where we used 24-hour measurements and did not have sufficient data to look at HRV during sleep specifically.Furthermore, it has been suggested that, at least in children, differences in HRV between those with and without ADHD may occur mainly on short-term measures in response to cognitively demanding tasks (Koenig et al. 2017;Robe et al. 2019).We currently looked at long-term (24-hour) resting-state HRV in adults and interpreted findings relative to normal ranges rather than made comparisons with a control group and found no evidence for a high risk of CVDs.Measuring short-term HRV in response to cognitively demanding tasks and during sleep only may reveal more deviations from the normal ranges in this group that would indicate a high risk of developing CVDs that we were currently unable to identify.
One HRV parameter did however show deviations in a large part of participants; HF was decreased in 67.6%, which could indicate poorer ability to respond flexibly to changes in the environment (Forte et al. 2019).HF is strongly related to other measures of HRV (Kleiger et al. 2005) that did not show such large deviations, making it difficult to interpret this finding.Larger studies looking at HRV during cognitive tasks or during sleep specifically are necessary to determine if the decrease in HF was a chance finding or whether it is clinically relevant.
We also assessed the prevalence of obesity, hypertension, and smoking, as these are major risk factors for CVDs (National Health Service 2022; World Health Organization 2017) and compared these to numbers from the general population and known ADHD and DSPS populations.Statistical comparisons were unfortunately not possible.The BMI distribution seemed similar to the general Dutch population (Rijksinstituut voor Volksgezondheid en Milieu 2022) and control  (1996).Low HRV is associated with higher mortality, although high HRV can be the result of pathological conditions; optimal HRV is reflected by scores within normal ranges.
groups in previous studies (e.g.Van Andel et al. 2021).Hypertension appeared to occur more often in our study population than in the general Dutch population (25.6% vs. 16%) (Centraal Bureau voor de Statistiek 2022).The percentage of smokers also seemed higher in our study population than in the general Dutch population (42% vs. 20%) (Trimbos-instituut 2022).This is similar to previous reports on adult ADHD populations (Lambert and Hartsough 1998;Pomerleau et al. 1995), yet not as high as the 61% found in a large sample of adolescents with DSPS (Saxvig et al. 2012).Interestingly, HRV outcomes did not differ between people with normal or high BMI, smokers and nonsmokers, or people with or without hypertension.So despite high percentages of smoking and hypertension, these risk factors do not appear to be related to a high risk of developing CVDs in our study population.
Several limitations should be mentioned.First, the current study did not include a control group to which findings obtained using the same methodology could be compared.Instead, the results were interpreted relative to findings and norm values from known study populations, which was complicated due to methodological differences, and therefore were difficult to compare.Second, we used 24-hour resting-state HRV measurements, which was a novel assessment in our study population, yet difficult to compare to earlier work that looked at HRV during sleep specifically or task-related HRV.Future studies should measure HRV in different ways (i.e.shortand long-term, resting-state and task-related) and include control groups to allow for comparisons of HRV variables obtained by the same method.Third, the sample size was relatively low.Only exploratory comparisons were made as any statistical comparisons were not sufficiently powerful due to the small sample size.Larger samples would increase the statistical power of the comparisons between participants with and without risk factors.Lastly, cardiovascular measures were secondary outcomes of our PhASE study, which were included to get an indication of the cardiovascular health of adults with ADHD and DSPS.Unfortunately, cardiovascular measures were not available or successful for all participants due to technical or logistic problems, resulting in a smaller HRV data set.However, it was a strength of the study that measurements took place in the participants' naturalistic home environment during their normal day-to-day life.
We found no evidence of poor HRV or high risk of developing CVDs in our population of adults with ADHD and DSPS, despite a high prevalence of smoking and hypertension.Future research should ultimately determine any cardiovascular risk in adults with both ADHD and DSPS in more detail using larger samples and different methods of measuring HRV.

Figure 1 .
Figure1.Percentages of people scoring below, within, or above reference ranges for HRV (heart rate variability) variables (N = 34).SD = standard deviation.M = mean.SDNN (standard deviation between NN (normal-to-normal) intervals), SDNN index (mean of the standard deviations of all NN intervals for each 5-minute segment), rMSSD (root mean square of successive differences between normal heartbeats), pNN50 (% of adjacent NN intervals that differ from each other by more than 50 ms), and HR (heart rate) are based on reference values per sex and age group(Umetani et al. 1998, Table 4); LF (low-frequency band) and HF (high-frequency band) are based on reference values by the Task Force of the European Society of Cardiology the North American Society ofPacing and  Electrophysiology (1996).Low HRV is associated with higher mortality, although high HRV can be the result of pathological conditions; optimal HRV is reflected by scores within normal ranges.

Table 1 .
General characteristics and study outcomes for total study population of adults with DSPS and ADHD (N = 49).