The COMT rs4680 SNP participates in the regulation of the arousal threshold in patients with obstructive sleep apnea

Abstract

Purpose

To investigate whether the functional catechol-O-methyltransferase (COMT) rs4680 polymorphism is related to a predicted low respiratory arousal threshold (LAT) phenotype in adults with polysomnography-confirmed obstructive sleep apnea (OSA), and to assess selected dopaminergic pathway-related polymorphisms.

Methods

In this prospective observational study, 144 adults with OSA underwent polysomnography. LAT was estimated non-invasively using the Edwards criteria based on apnea-hypopnea index < 30 events/h, nadir oxygen saturation > 82.5%, and hypopnea fraction > 58.3%, with LAT defined by fulfillment of at least two criteria. Peripheral blood samples were genotyped for COMT rs4680, COMT rs6269, DRD1 rs686, and a DRD2 variant.

Results

Among the 144 participants, 76 (52.8%) met criteria for LAT and 68 (47.2%) for non-LAT. Compared with non-LAT patients, the LAT group was younger and showed lower OSA severity markers, including lower apnea-hypopnea index, oxygen desaturation index, apnea indices, snoring burden, and time with oxygen saturation < 90%, together with higher mean and minimum oxygen saturation. The arousal index was also lower in the LAT group (4.40 vs. 9.42/h, p = 0.006). A differential distribution of COMT rs4680 genotypes was observed between LAT and non-LAT groups, with G-allele-containing genotypes more frequent in LAT, whereas the AA genotype was more common in non-LAT patients. No association was found for the other analyzed polymorphisms.

Conclusion

These findings suggest that COMT rs4680 may be associated with the predicted LAT phenotype in OSA, supporting the concept that dopaminergic pathways may contribute to OSA endotypic heterogeneity. Given the indirect definition of LAT and the candidate-gene design, the results should be considered hypothesis-generating and require replication.

Introduction

Sleep is a dynamic physiological process regulated by complex interactions between homeostatic sleep pressure, circadian timing, and neurobiological arousal systems. Sleep architecture, including the cyclic organization of non-rapid eye movement and rapid eye movement sleep, reflects coordinated neural and systemic physiological regulation. This broader framework, informed by both contemporary sleep science and historical perspectives on prolonged sleep states, contributes to a more comprehensive understanding of normal sleep and sleep disorders, including obstructive sleep apnea [1]. Disruption of these regulatory mechanisms may impair sleep stability and respiratory control during sleep. Obstructive sleep apnea (OSA) is a highly prevalent sleep disorder characterized by recurrent upper airway collapse during sleep, resulting from complex genetic and anatomical factors and leading to intermittent hypoxemia and sleep fragmentation through cortical microarousals or awakenings [2–5]. Beyond symptom burden, OSA has been linked to serious cardiovascular morbidity, hypertension, stroke, myocardial infarction, heart failure, cardiac arrhythmias, metabolic disorders, and all-cause mortality [6–8]. OSA is a heterogeneous disorder of multifactorial pathogenesis and its four main pathophysiological traits include: anatomical predisposing factors as upper airway easy collapsibility, instability of ventilatory control described as high loop gain; poor muscle responsiveness resulting in ineffective upper airway muscle compensation, and a low respiratory arousal threshold (LAT), which means easy arousal from sleep at a relatively low level of respiratory stimuli [6, 9–12]. LAT is reported as a common trait [6, 11]. In patients with PSG-evaluated OSA, LAT occurred in about half of patients, and its prevalence and distribution varied with OSA severity [5], sleep stage [7], and co-occurrence of sleep bruxism (SB) episodes [13]. A possible pathogenetic impact of low AT in OSA includes the premature termination of apnea events before sufficient time has elapsed for effective upper airway dilator recruitment and stabilization of breathing, which is implicated in the potential conservation of respiratory OSA events and sleep architecture disruption [11, 14]. On the other hand, the dual potential impact of cortical arousal on OSA pathogenesis has also been highlighted, and LAT might become a protective factor in some circumstances when other defensive mechanisms fail [11, 15, 16].

Simultaneously, the notable role of dopaminergic signaling in regulating the sleep-wake cycle has been widely reported in humans and in various experimental models, from rodents to Drosophila [3, 17–23]. Sleep-wake regulation results from the complex interaction of neural, circadian, and neuroendocrine mechanisms. Dopaminergic signaling contributes to arousal regulation and cognitive function, while its effects are closely integrated with circadian rhythmicity and neuroendocrine responses, supporting a multidimensional framework for understanding sleep physiology and its disorders [24]. The dopaminergic effects on the respiratory system and on neural control of muscle tone have been described, supporting the hypothesis that it could interact in breathing stability and upper-airway function [19, 20]. Dopaminergic candidate markers have also been investigated, including exploration of the genetic basis of SB and sleep apnea involving polymorphisms in dopaminergic loci [3, 22].

Catechol-O-methyltransferase (COMT) is one of the key enzymes involved in catecholamine metabolism in mammals and accounts for the majority of dopamine degradation in the prefrontal cortex (PFC) [25]. The functional single-nucleotide polymorphism (SNP) COMT rs4680 variant (Val158Met) is associated with differences in COMT activity and dopaminergic signaling [23, 26–28]. The functional Val158Met SNP occurs in populations worldwide [19, 21], and individuals with the GG allele (Val homozygotic) show higher COMT activity than those with the AA allele (Met homozygotic) [23, 26, 27]. It has been suggested that variation in COMT activity affects the function of the PFC and associated areas [23, 27]. Evidence suggests that COMT may influence executive function, working memory, emotional processing, the risk of depression and psychosis, sleep-wake regulation, sleep behavior, and daytime sleepiness [22, 28]. Available evidence suggests that the regulation of arousal, respiratory drive, and upper airway muscle tone may depend on functional variants in dopaminergic genes encoding dopamine receptor families (DRD1and DRD2), which have been associated with OSA severity [3] and dopamine active transporter (DAT), which affected the EEG after sleep deprivation [29–32].

The aim of this study was to assess the COMT rs4680 (Val158Met) functional polymorphism in patients with PSG-confirmed OSA and LAT to explore the pathogenesis of this phenotype and identify its potential genetic determinants. We additionally assessed selected dopaminergic pathway-related polymorphisms: COMT rs6269, DRD1 rs686, and DRD2 rs1800479, to improve the area of observation.

Patients and methods

The prospective observational study was conducted in the Sleep Laboratory at the Department and Clinic of Diabetology, Hypertension and Internal Medicine, Wroclaw Medical University, Poland. A total of 144 patients with OSA, of whom 81 were men, were recruited for the study from adults diagnosed with OSA who were hospitalized in the Department during 2023 and 2024. The study complied with the Declaration of Helsinki and was approved by the Ethical Committee of Wroclaw Medical University (ID: 199/2023). Written informed consent was obtained from all participants. Information about clinical trial registration is available at www.ClinicalTrials.gov (identifier: NCT0589091). The study was funded by Wroclaw Medical University (grant number SUBK.B160.23.059).

Polysomnography

The polysomnographic examination (PSG) with video recording was conducted in all study participants, and only patients with an OSA diagnosis were recruited. The inclusion criteria were: adult status (at least 18 years of age), a confirmed OSA diagnosis on a PSG, and informed written consent to participate in the study. The exclusion criteria included the presence of certain illnesses: current malignancy or inflammation, decompensated heart and/or respiratory failure, severe asthma, chronic obstructive pulmonary disease (COPD), hypothyroidism, chronic kidney disease, pregnancy, and less than 4 h of recorded sleep during PSG examination. A minimum sleep duration of 4 h was required to ensure adequate signal quality and reliable polysomnographic assessment, although this criterion may have introduced selection bias by excluding patients with highly fragmented or unstable sleep.

Information on research methods has been partially described in our previous studies [3, 15].

In the study, the NOX-A1 polysomnographic system (Nox Medical, Reykjavik, Iceland) was used to perform PSG. The American Academy of Sleep Medicine (AASM) Manual, version 3.0, was used to perform the polysomnographic assessment and to define the sleep disorders [33]. During PSG, the following features were assessed: the bioelectrical activity of the brain by using electroencephalography (EEG), the eye movements by using electrooculography, the airflow by a nasal pressure sensor, the chest and abdomen movements by using inductive plethysmography, and an electromyogram with tibial electrodes demonstrating muscle tension; additionally, the patient’s body position was recorded. In oxygen saturation assessment, a NONIN WristOx2 3150 pulse oximeter (Nonin Medical, Plymouth, MN, USA) was employed. The data obtained from each PSG examination were manually analyzed by a qualified physician at the Sleep Laboratory of Wroclaw Medical University [32, 34].

Because direct measurement of ArTH is labor-intensive, non-invasive clinical estimation approaches based on routine polysomnography have been proposed and applied, including algorithms incorporating apnea–hypopnea burden, oxygen saturation measures, and the proportion of hypopneas [5, 9]. The low AT presence was assessed in the study participants according to the criteria described by Edwards et al. in 2014 for non-invasive estimation of AT [5]. Prior to the provided criteria, the standard invasive examination was used to quantify the AT using an oesophageal or epiglottic pressure catheter. The participants in our study were classified as having low AT based on the following PSG features, according to Edwards et al.: an AHI < 30 events/h, a minimum SpO2 of > 82.5%, and a hypopnea fraction of total respiratory events of > 58.3%. Participants with OSA who met at least two of the above criteria were considered likely to have a low respiratory AT.

DNA isolation and genotyping

Peripheral blood samples (10 mL) were collected from each participant in the morning hours following the PSG. The Vacutainer^® blood sampling system (Becton Dickinson, Franklin Lakes, NJ) was used. Samples were collected into EDTA-containing tubes (ethylenediaminetetraacetic acid) and frozen at − 20 °C until further use. DNA was extracted using the NucleoSpin Blood L Kit (Macherey-Nagel, Düren, Germany), following the manufacturer’s instructions. DNA concentration and purity were measured using a DeNovix DS-11 spectrophotometer (DeNovix Inc., Wilmington, DE). The extracted DNA was stored at − 20 °C until SNP genotyping. The genotyping of COMT rs4680 (and also COMT rs6269 A/G; DRD1 rs686 A/G; DRD2 rs1800479 C/T) was performed using the LightSNiP typing assay (TIB-MolBiol, Berlin, Germany), which involved amplification by quantitative polymerase chain reaction (qPCR) and subsequent melting curve analysis. Reactions were conducted on a LightCycler 480 II system (Roche Diagnostics, Rotkreuz, Switzerland) in accordance with the manufacturer’s protocol. The reaction program included an initial denaturation step at 95 °C for 10 min, followed by 45 cycles of denaturation at 95 °C for 10 s, annealing at 60 °C for 10 s, and extension at 72 °C for 15 s. Melting curve analysis was performed, starting with a 30-s incubation at 95 °C, followed by 2 min at 40 °C, with continuous data acquisition during the melting process between 40 and 75 °C at a ramp rate of 1.5 °C/s. This was in accordance with the methods used in our previous study [3].

Statistical analysis

The dataset, comprising data from 144 individuals, was analyzed using Statistica 13.3 (StatSoft, Poland). Descriptive statistics are presented as mean and standard deviation. The choice of statistical tests was guided by the distribution of variables and the fulfillment of relevant statistical assumptions. Normality was assessed using the Shapiro-Wilk test, and homogeneity of variance was evaluated using Levene’s test. For variables meeting parametric assumptions, between-group comparisons were performed using the independent-samples t-test. For variables that did not meet parametric assumptions, the Mann-Whitney U test was used.

Genotype distributions were assessed for conformity with Hardy-Weinberg equilibrium using the chi-square test with 1 degree of freedom. Categorical variables, including genotype and allele frequency comparisons between groups, were analyzed using Pearson’s chi-square test. When the assumptions of this test were not met, defined as expected cell counts below 5, Fisher’s exact test was applied. Effect sizes for genetic association analyses were reported as odds ratios (ORs) with 95% confidence intervals (95% CIs). To assess the association between covariates and the LAT phenotype, logistic regression analysis was performed. Statistical significance was set at p < 0.05. Very small p-values were reported as p < 0.001 rather than p = 0.000.

Results

The characteristics of the study group have been presented in Table 1.

Table 1.

Characteristics of the study group, all OSA participants

Variable	Study group (n = 144)
	Average	Minimum	Maximum	SD
Age (years)	50.94	20.00	79.00	14.33
AHI (/h)	23.22	0.00	87.40	19.57
ODI (/h)	21.91	0.00	88.40	19.03
Snore (%)	28.80	0.00	87.90	21.93
AI (/h)	9.41	0.00	77.50	15.21
OA (/h)	7.93	0.00	71.20	13.50
MA (/h)	0.53	0.00	25.20	2.47
CA (/h)	0.72	0.00	10.90	1.55
HI (/h)	13.79	0.00	54.20	10.26
Average O2 saturation (%)	92.32	74.60	96.50	2.78
Minimal O2 saturation (%)	81.79	51.00	95.00	8.25
O2 Saturation < 90% (%)	12.66	0.00	86.60	19.63
Arousals (/h)	7.30	0.00	88.30	11.74

Abbreviations: AHI apnea/hypopnea index (AHI = AI (apnea index) + HI (hypopnea index)), ODI oxygen desaturation index, AI apnoea index, OA obstructive apnoea, MA mixed apnoea, CA central apnoea, HI hypopnea index, (/h) per hour, SD standard deviation, n number of participants

All 144 patients, Caucasians with PSG-diagnosed OSA (81 were men), had a mean age of 50,94 (± 14,33 y) years; the youngest participant was 20 years old, and the oldest was 79 years old. In the whole study group (n = 144), 76 persons (52.8%) were identified as having a low arousal threshold (LAT), while 68 persons (47.2%) were identified as having a high arousal threshold (non-LAT). The LAT group was visibly younger than the non-LAT group (43.95 vs. 53.11 years; p < 0.001) and had relatively lower levels of the following OSA polysomnographic markers than non-LAT group: AHI apnoea/hypopnea index, ODI oxygen desaturation index; HI hypopnea index AI: apnoea index; OA obstructive apnoea and MA: mixed apnoea, as well as the percentage of SpO₂ <90% which highlights the less severe OSA stage than in non-LAT participants. As expected, the average and minimum oxygen saturation values were higher in LAT patients than in non-LAT patients. The arousal index, reflecting arousal frequency rather than respiratory arousal threshold itself, was lower in the LAT group than in the non-LAT group (4.40 vs. 9.42 events/h, p = 0.006). The detailed comparison of OSA variables between the LAT and non-LAT groups is presented in Table 2.

Table 2.

Comparison of LAT and non-LAT OSA patient groups in the study

Variable	non-LAT (n = 68)		LAT (n = 76)		compare
	Average	SD	Average	SD	p-value
Age (years)	53.11	14.65	43.95	13.88	p < 0.001
AHI (/h)	29.58	24.27	7.43	6.74	p < 0.001
ODI (/h)	27.89	23.44	7.71	6.92	p < 0.001
Snore (%)	29.91	21.10	17.26	20.60	p < 0.001
AI (/h)	13.97	18.68	1.71	2.63	p < 0.001
OA (/h)	12.59	17.19	1.15	2.25	p < 0.001
MA (/h)	0.57	1.88	0.06	0.24	0.009
CA (/h)	0.81	2.05	0.43	0.66	0.097
HI (/h)	15.51	12.78	5.77	5.64	p < 0.001
Average O2 saturation (%)	91.49	3.76	93.83	2.16	p < 0.001
Minimal O2 saturation (%)	76.09	8.45	88.25	3.32	p < 0.001
O2 Saturation < 90% (%)	21.16	23.37	2.90	7.58	p < 0.001
Arousals (/h)	9.42	16.77	4.40	4.27	0.006

Abbreviations: LAT low respiratory arousal threshold, non-LAT non low repiratory arousal threshold, AHI apnoea/hypopnea index (AHI = AI (apnea index) + HI (hypopnea index)), ODI oxygen desaturation index, AI apnoea index, OA obstructive apnoea, MA mixed apnoea, CA central apnoea, HI hypopnea index, (/h) per hour, SD standard deviation, n number of participants

The distribution of COMT rs4680 genotypes differed between LAT and non-LAT patients. Among the 76 patients classified as LAT, the GG, AG, and AA genotypes were observed in 20 (26.3%), 40 (52.7%), and 16 (21.0%) participants, respectively. Among the 68 non-LAT patients, the corresponding genotype counts were 19 (28.0%), 24 (35.3%), and 25 (36.8%). The distribution of COMT rs4680 genotypes according to the predicted LAT phenotype is presented in Table 3.When allele-carrier status was analyzed in a 2 × 2 format, the presence of the G allele of COMT rs4680 was significantly associated with the predicted LAT phenotype (p = 0.037). G allele carriers had approximately 2.18-fold higher odds of being classified as having LAT compared with non-carriers (OR = 2.18; 95% CI: 1.04–4.57). In contrast, no significant association was observed for the presence of the A allele of COMT rs4680 (OR = 1.086; 95% CI: 0.52–2.26; p = 0.826). The genotype distribution and allele-carrier effect size analyses are presented in Tables 3 and 4.

Table 3.

Distribution of COMT rs4680 genotypes according to predicted low respiratory arousal threshold status

Variable	COMT rs 4680 allelic variant A/G
	GG	AG	AA
LAT (n = 76)	20 (51%)	40 (62%)	16 (39%)
non-LAT (n = 68)	19 (49%)	24 (38%)	25 (61%)
Total (n = 144)	39 (100%)	64 (100%)	41 (100%)

Abbreviations: COMT catechol-O-methyltransferase, A adenine, G guanine, LAT low respiratory arousal threshold, non-LAT non-low respiratory arousal threshold

Table 4.

Effect size for the association between COMT rs4680 G allele carriage and the predicted low respiratory arousal threshold phenotype

Genetic comparison	non-LAT, n	LAT, n	OR	95% CI	p-value
G allele carriers vs. non-carriers	43 vs. 25	60 vs. 16	2.18	1.04–4.57	0.037
A allele carriers vs. non-carriers	49 vs. 19	56 vs. 20	1.086	0.52–2.26	0.826

Abbreviations: LAT low respiratory arousal threshold, non-LAT non-low respiratory arousal threshold, OR odds ratio, CI confidence interval. For the G allele analysis, carriers included participants with GG or AG genotypes, and non-carriers included participants with the AA genotype. For the A allele analysis, carriers included participants with AG or AA genotypes, and non-carriers included participants with the GG genotype

None of the single-nucleotide polymorphisms simultaneously analyzed (COMT rs 6269 A/G; DRD1 rs 686 A/G; DRD2 rs 1800479 C/T) showed any dominant prevalence or association in the LAT or non-LAT patient groups. Genotype distributions were assessed for conformity with Hardy-Weinberg equilibrium. No significant deviations from Hardy-Weinberg equilibrium were observed for COMT rs4680 (χ² = 1.772; p = 0.183), COMT rs6269 (χ² = 0.956; p = 0.328), DRD1 rs686 (χ² = 1.127; p = 0.288), or DRD2 rs1800479 (χ² = 0.001; p = 0.973) (Table 5).

Table 5.

Hardy-Weinberg equilibrium analysis of the investigated polymorphisms

Polymorphism	Genotypes, n	χ²	p-value
COMT rs4680	GG = 39; AG = 64; AA = 41	1.772	0.183
COMT rs6269	GG = 23; AG = 62; AA = 59	0.956	0.328
DRD1 rs686	AA = 52; AG = 74; GG = 18	1.127	0.288
DRD2 rs1800479	CC = 95; CT = 44; TT = 5	0.001	0.973

HWE Hardy-Weinberg equilibrium, χ² chi-square statistic. Hardy-Weinberg equilibrium was assessed using the chi-square test with 1 degree of freedom

To account for potential confounding by age and sex, a multivariable logistic regression analysis was performed with the presence of the LAT phenotype as the dependent variable. Age, sex, and the presence of the COMT rs4680 G allele were included as independent variables. Age was significantly associated with LAT (B = − 0.058; OR = 0.944; p < 0.001), indicating that each additional year of age was associated with lower odds of LAT. Sex was also significantly associated with LAT (B = 0.837; OR = 2.309; p = 0.033). After adjustment for age and sex, the presence of the COMT rs4680 G allele remained independently associated with the LAT phenotype (B = 0.899; OR = 2.458; p = 0.034), corresponding to approximately 2.5-fold higher odds of LAT among G allele carriers. The multivariable logistic regression analysis is presented in Table 6.

Table 6.

Multivariable logistic regression analysis for predictors of the low respiratory arousal threshold phenotype

Variable	B	OR	p-value	Interpretation
Intercept	2.032	7.630	0.012	—
Age	−0.058	0.944	< 0.001	Each additional year of age was associated with lower odds of LAT
Sex	0.837	2.309	0.033	Sex was significantly associated with LAT
COMT rs4680 G allele	0.899	2.458	0.034	G allele carriers had approximately 2.5-fold higher odds of LAT

LAT low respiratory arousal threshold, OR odds ratio, COMT catechol-O-methyltransferase. The model included age, sex, and the presence of the COMT rs4680 G allele as independent variables, with the LAT phenotype as the dependent variable. G allele carriers included participants with GG or AG genotypes; non-carriers included participants with the AA genotype

Discussion

The most important result in this prospective study of patients with PSG-confirmed OSA is the link between COMT rs4680 (Val158Met) genotype distribution and the low respiratory arousal threshold (LAT). We observed the higher prevalence of the G allele of the COMT rs4680 variant in patients with LAT and OSA than in those with non-LAT and OSA. These findings are consistent with the concept that inter-individual variability in physiological traits and OSA expression might have a measurable genetic component. These findings are consistent with the concept that inter-individual variability in OSA expression may have a genetic component, although our data do not establish causality.

The COMT rs4680 guanine-to-adenine SNP is responsible for valine (G allele) to methionine (A allele) substitution. It has been confirmed that the G allele is a major factor determining 3- to 4-fold higher COMT activity in human PFC tissue, which leads to lower synaptic dopamine levels and relatively impaired prefrontal function [23, 27]. There is a growing body of literature assessing the effect of the functional COMT rs4680 variant on behavioral phenotypes and treatment response [35]. The loss of normal dopaminergic function in humans contributes to sleep disorders, especially restless legs syndrome, reported in patients with Parkinson’s disease [30], and the COMT rs4680 SNP has been specifically suggested to affect sleep-wakefulness cycle regulation as well as its associated pathologies [23]. We previously noted that patients with OSA had elevated serum dopamine levels compared with those without OSA, which might serve as a bridge to our present finding of COMT variants and low AT [3]. The large-sample data highlighted that genetically determined variation in COMT activity has neurobiological effects specific to the prefrontal cortex (PFC) [36]. In healthy adults, the GG allele of the COMT rs4680 variant was connected with poorer performance on working memory tests [31], and a dopamine-boosting drug maintained executive functioning performance and alert attention throughout sleep deprivation, especially in subjects with the GG allele [23, 37]. In the other study, persons with COMT genotypes did not differ in performance across various executive function and cognitive tasks [38]. COMT rs4680 SNP has been also linked to inter-individual differences in sleep physiological responses to sleep loss and in EEG-related sleep markers, including reports on sleep rebound, alpha oscillatory characteristics, and sleep spindle density [38–40]: in COMT rs4680 snp homozygous GG allele carriers alpha peak frequency in wakefulness was slower than in AA carriers [39], in healthy humans presence of G allele was linked to differentially smaller decline in EEG slow-wave sleep, smaller reductions in stage 2 sleep during partial sleep deprivation, and more stage 1 sleep across nights and shorter baseline REM sleep latency [38]. The data about daytime sleepiness in response to sleep loss and the COMT rs4680 SNP are divergent. In a large sample of healthy Colombian students, the association of COMT Val158Met with daytime sleepiness has been reported and further supported by differences in genotype frequencies between categories of diurnal sleepiness for the COMT gene [41]. In another study, however, the COMT Val158Met genotypes showed comparable increases in subjective and physiological sleepiness in response to chronic sleep loss [38]. A recent long-term actigraphy study in humans revealed that (GG) and (AA) homozygotes habitually prolonged sleep on resting days compared to workdays, whereas (GA) heterozygotes did not significantly extend their sleep length on weekends [42]. These data are consistent with an inverted U-shaped relationship between COMT genotype-dependent differences in PFC dopamine and the accumulation of sleep debt during workdays [23]. We hypothesize that respiratory arousal responses may follow an inverted U-shaped relationship with dopamine availability, whereby both insufficient and excessive dopaminergic signaling impair the appropriate cortical response to increasing respiratory stress. Within this model, COMT rs4680-dependent differences in dopamine metabolism could shift the respiratory arousal threshold by altering the balance between arousal responsiveness, ventilatory stability, and sleep continuity during obstructive events. Recent behavioral intervention data suggest that sleep duration is a modifiable factor influencing subsequent cognitive performance, with even a single night of sleep extension improving cognitive function in healthy individuals [43]. These observations support the concept that genetic factors, including COMT-related variability in dopaminergic signaling, may interact with sleep quantity and quality to influence cognitive and behavioral outcomes, although this hypothesis requires further investigation in patients with OSA. In the recent literature, there are data on phenotypic differences in EEG fast sleep spindles (a hallmark of NREM stage 2 sleep) in healthy humans dependent on the COMT Val158Met polymorphism, suggesting that genetically determined dopaminergic neurotransmission might modulate spindle oscillations during NREM sleep. The COMT Val allele (G) correlated with reduced fast spindle density [40]. However, current evidence linking dopaminergic candidate markers to OSA phenotypes is primarily associative and does not establish a direct mechanistic role in the regulation of respiratory arousal.

We reported in a previous study that the presence of the G allele of COMT rs4680 was a candidate protective factor against higher OSA levels, which is in line with our present observation [3]. Since it has been shown that the COMT Val158Met polymorphism may predict differences in brain EEG oscillations and in sleep rebound after partial sleep loss in healthy humans [38, 42], we hypothesize that it may also influence the arousal threshold phenotype in patients with OSA adding another puzzle in the pathophysiological trait in OSA [15]. In terms of its clinical expression, the COMT SNP Val158Met might influence the phenotype of OSA, with the G allele primarily reducing the arousal threshold, potentially providing a protective effect against severe OSA. This would create the possibility of new pharmacological therapeutic strategies.

There is still debate in the literature about whether respiratory AT is a cause or a consequence of OSA [44]. Consistent with prior data, low AT has been shown to be particularly prevalent in the non-severe OSA population [15, 16]. Moreover, a possible protective role of LAT has been postulated in the population of adult patients with OSA and coexisting sleep bruxism (SB) [15, 45] and asthma [16]. Also, the possible protective role of SB in the population of adult patients with OSA has been highlighted by our team previously and discussed in the literature [46]. In the present study, similarly to the data so far, the LAT group had visibly lower levels of the traditional OSA severity clinical markers than the non-LAT group, including: AHI apnoea/hypopnea index, ODI oxygen desaturation index; HI hypopnea index, AI apnoea index; OA obstructive apnoea, and MA: mixed apnoea, as well as the percentage of reduced < 90% O₂ Saturation. And, as expected, the average and minimal oxygen saturation values were higher in the LAT than in non-LAT patients, highlighting a less severe OSA stage in LAT participants. This is very much in line with the recent pilot study, which also demonstrated a high prevalence of low AT in patients with mild OSA and a low, almost absent, prevalence of low AT in patients with severe OSA and asthma [16]. Consistent with our results, these data also revealed higher O2 saturation and lower AHI in participants with a LAT, compared with those with a high AT [16]. Our previous observations also suggested that patients with low AT may present with less severe OSA, although this should not be interpreted as evidence of a protective effect independent of disease severity and other clinical characteristics. LAT patients were also characterized by visibly higher saturation and more hypopneas than apneas in PSG screening [15]. This is also in line with Edwards et al.‘s study, which found that the proportion of patients with a LAT decreased as the severity of OSA increased [11]. It is worth noting that in the 2014 canonical AT study, in which LAT was assessed both invasively and noninvasively, most non-OSA control subjects had a low respiratory arousal threshold (86.4%) compared with patients with OSA, in whom LAT occurred in 50.5% [11]. Our present observation closely aligns with this: across the entire OSA study group, 52.8% (76 patients) were identified as having a low arousal threshold, while 47.2% (68 patients) were identified as having non-LAT. This confirms both the adequacy of the group selection and the stable prevalence of the LAT phenotype in patients with OSA. There has also been a concept suggesting that an elevated arousal threshold (non-LAT) might be one of the consequences of OSA severity in many individuals, which could partially reflect the higher occurrence of LAT in the healthy (non-OSA) population compared with people with OSA [11]. There are also suggestions in the literature that low AT may be a joint pathophysiological root to both OSA and asthma [16].

The observed differences in age and OSA severity between the study groups should be considered when interpreting our findings. Patients with a predicted low respiratory arousal threshold were younger and had milder OSA, which may have influenced the observed association between COMT rs4680 and the LAT phenotype. However, these characteristics are consistent with previous studies showing that a low respiratory arousal threshold is more frequently observed in patients with less severe OSA, whereas increasing OSA severity may be associated with a higher respiratory arousal threshold, possibly due to repeated exposure to respiratory events and greater accumulation of respiratory stimuli before arousal [11, 16]. Nevertheless, because age and disease severity may act as confounding factors, the observed association should be interpreted with caution. Future studies involving larger, age- and OSA-severity-matched cohorts, as well as adjusted multivariable analyses, are warranted to confirm whether COMT rs4680 contributes independently to the predicted LAT phenotype.

It is important to distinguish respiratory arousal threshold from arousal frequency, as these terms describe different aspects of sleep physiology. Respiratory arousal threshold is a physiological endotype of OSA and reflects the magnitude of respiratory stimulus required to induce cortical arousal during an obstructive event [47]. In contrast, arousal frequency is a polysomnographic measure that quantifies the number of arousals occurring during sleep. In the present study, patients classified as having a predicted low respiratory arousal threshold had a lower arousal index than non-LAT patients. This finding should not be interpreted as contradictory, because the LAT phenotype was estimated using the Edwards criteria, based on apnea-hypopnea index, nadir oxygen saturation, and hypopnea fraction, rather than on the arousal index itself [11]. Although a low respiratory arousal threshold may be associated with increased sleep fragmentation in some clinical contexts, respiratory arousal threshold and arousal frequency are not synonymous. Arousal frequency is influenced by multiple factors, including OSA severity, upper airway collapsibility, ventilatory control instability, sleep architecture, comorbid sleep disorders, and neurobiological mechanisms involved in sleep-wake regulation [23]. Therefore, respiratory arousal threshold provides mechanistic insight into the propensity to awaken in response to respiratory stimuli, whereas arousal frequency reflects the overall burden of sleep disruption. The lower arousal index observed in the LAT group in our cohort may therefore reflect the lower severity of OSA in this group rather than a direct measure of the intrinsic arousal threshold. Although non-invasive algorithms provide a practical approach to estimating respiratory arousal threshold, their performance may be influenced by differences in sensor technology, signal quality, and proprietary scoring algorithms. Therefore, standardization and transparency of sensor-derived physiological assessments remain important for improving the reproducibility and comparability of future studies [48]. In this context, previous interventional data showing that effective CPAP treatment may reduce respiratory arousal threshold further support the concept that arousal threshold is dynamic and may be influenced by OSA severity and treatment [49].

On the other hand, hypnotics have been shown to modestly increase the respiratory arousal threshold while not changing OSA severity [44]. This highlights the rationale for phenotyping patients with OSA to qualify them for potential therapeutic intervention according to AT manipulation. Future mechanistic physiological studies needed.

A low arousal threshold has been associated with greater sleep instability and more frequent arousals, which may contribute to autonomic activation [16, 50, 51]. Taken together with evidence linking autonomic activity to arousals, sleep bruxism, and upper airway obstruction, as well as emerging observations of coexisting sleep bruxism and asthma in patients with low arousal threshold and OSA, these data support the hypothesis of a complex interaction among low arousal threshold, OSA, and autonomic dysregulation [15, 16].

Unlike its well-established effects on prefrontal cognitive and executive processes, COMT-related dopaminergic signaling may influence respiratory arousal through modulation of sleep-wake neural circuits, ventilatory control, and upper-airway motor responsiveness, although the specific mechanisms remain to be determined.

From a clinical perspective, identification of COMT rs4680-related differences in respiratory arousal threshold may support future integration of genetic markers with physiological endotyping to improve OSA phenotyping and guide individualized therapeutic strategies. Further studies should replicate these findings in larger, ethnically diverse cohorts, include direct measurements of respiratory arousal threshold, and assess whether COMT genotype predicts treatment response or long-term clinical outcomes.

Limitations of the study

Some limitations of our study should be considered when interpreting the findings. The study focused on a Caucasian population, which limits the generalizability of the results to other ethnic groups, although the prevalence of COMT allelic variants has previously been assessed in the general population [23, 26]. The lack of data on body mass index (BMI), inflammatory marker levels, sleep quality, psychoemotional disturbances, comorbidities, and OSA treatment should also be considered a limitation of this study. Specifically, the absence of BMI data prevented us from assessing whether adiposity differed between the LAT and non-LAT groups or contributed to the observed associations. Because obesity is a major clinical determinant of OSA phenotype, upper airway mechanics, and respiratory load, BMI may have acted as an unmeasured confounding factor and may have influenced the interpretation of the predicted respiratory arousal threshold phenotype. In particular, BMI may indirectly influence the polysomnographic variables used in the Edwards criteria, including apnea-hypopnea index, oxygen desaturation severity, and the relative predominance of hypopneas versus apneas. The absence of BMI data may not only confound the observed association but may also affect the accuracy and generalizability of the Edwards criteria when applied to a cohort with a potentially different anthropometric profile. Therefore, although BMI is not a formal component of the Edwards prediction model, the absence of BMI data may have affected the non-invasive classification of the predicted LAT phenotype and prevented BMI-stratified or BMI-adjusted sensitivity analyses. Available data on the relationship between BMI and respiratory arousal threshold are inconsistent [11, 16, 52]. Some studies suggested that arousal threshold does not differ according to BMI, whereas others reported that LAT was more frequent in non-obese than obese patients with OSA, or conversely that LAT was more prevalent in obese than non-obese patients with OSA [16]. Importantly, BMI is not included in the non-invasive Edwards criteria used to estimate respiratory arousal threshold, which are based on apnea-hypopnea index, nadir oxygen saturation, and hypopnea fraction [11]. Therefore, the absence of BMI data does not alter the formal application of the Edwards prediction model, but it limits the clinical characterization of the cohort, prevents assessment of adiposity-related confounding, and should be considered when interpreting both the estimated LAT phenotype and the observed genetic association. The use of this cohort was nevertheless considered appropriate because all participants had PSG-confirmed OSA, LAT was estimated using the predefined non-invasive Edwards criteria based on routinely available polysomnographic parameters, and genotyping data were available for the investigated dopaminergic pathway-related polymorphisms. Similarly, previous data on sex-related differences in arousal threshold are not fully consistent, with some studies reporting no sex-related difference in AT, whereas others have highlighted sex-dependent differences [37]. Future studies should include complete anthropometric data and assess whether BMI modifies or confounds the association between COMT rs4680 and the predicted LAT phenotype. No formal correction for multiple testing was applied because of the exploratory and hypothesis-generating character of the study. Therefore, the reported genetic associations should be interpreted with caution and require confirmation in larger independent cohorts.

Conclusions

1. Our results support the concept that OSA is a pathogenetically heterogeneous disorder with a genetic component.

2. Our data suggest an association between COMT rs4680 genotype distribution and the low respiratory arousal threshold phenotype in PSG-diagnosed OSA.

3. This observation supports the integration of genetic analyses with physiological endotyping to identify potential targets for personalized OSA treatment.

Funding

The study was funded by Wroclaw Medical University (grant number SUBK.B160.23.059).

Data availability

The datasets generated and analyzed during the current study are not publicly available because they contain sensitive participant-level clinical, polysomnographic, and genetic data, and public deposition could compromise participant privacy. No publicly available datasets were analyzed in this study. A minimal de-identified dataset supporting the findings of this article may be made available from the corresponding author upon reasonable request, provided that the proposed use is scientifically justified and subject to approval by the relevant institutional and/or ethics committee, compliance with applicable data protection regulations, and completion of an appropriate data-sharing agreement.

Declarations

Ethical approval

The study was approved by the Ethical Committee of Wroclaw Medical University (ID: 199/2023).

Conflict of interest

The authors declare that they have no conflict of interest regarding the subject of this manuscript.