The aim of this study was to identify relations between OSA and suicidality and its prevalence. People with OSA experienced a significantly higher prevalence (p < 0.001) of anxiety (15.9% vs. 5.9%), depression (12.6% vs. 5.3%), suicidal ideation (12.4% vs. 3.7%), suicide attempts (1.5% vs. 0.9%), and death by suicide (0.3% vs. 0.2%) compared to the control population. In comparison, the prevalence of suicidal ideation and suicide attempts in the United States in 2022 was 5.2% and 0.6%, respectively [30]. However, the prevalence of suicidal ideation among patients with OSA is lower than that observed in other chronic conditions, such as Parkinson’s disease and fibromyalgia with rates of 22.2% and 29.6%, respectively [31, 32].

Sleep disturbances, in general, have been associated with an increased risk of mental illness and suicide. A retrospective cohort study of 1,160 U.S. Army service members found that those with pronounced insomnia symptoms had a 3.5-fold higher likelihood of reporting suicidal ideation and were significantly more likely to have depression (12.8% vs. 1.2%) and anxiety (11.6% vs. 1.0%) than controls (p < 0.001) [33]. Theories to explain this association include the hypothesis that dysfunctions in biological sleep-wake systems such as the circadian system may lead to disrupted mood regulation and increased suicidality. Another hypothesis is that cognitive deficits resulting from sleep deprivation could contribute to increased risk-taking and impulsivity, which are known risk factors for suicidal behaviors [34]. The sleep disturbance that arises from OSA could impact mood and cognition in a similar manner, thus increasing the risk of suicidality. Other theories posit that OSA-related sleep architecture disruption and intermittent hypoxemia are linked to increased TNF-α and IL-6 that impair various neuropsychological and affective domains [35,36,37]. Hypoxemia has been linked to alterations in white matter integrity and reductions in gray matter volume. Breakdown of the blood–brain barrier (BBB) coupled with neuroinflammatory processes may compromise neuronal connectivity and accelerate cerebral small vessel pathology [36]. However, the neuropsychological and affective changes have been shown to be reversible with adequate treatment of OSA [38].

A retrospective cohort analysis based on the Taiwan National Health Insurance Research Database identified 3,025 patients diagnosed with OSA from 2000 to 2013 and found that the adjusted hazard ratio for death by suicide was 6.5 (95% CI: 5.5–7.1) compared to the control population (p < 0.001) [39]. In addition, a cohort study of Chinese students found that among those who reported sleep-disordered breathing, the crude odds ratio (OR) for suicidal ideation was 1.8 (1.5–2.1) (p < 0.001), while the adjusted OR was 1.1 (0.9–1.4). A similar trend was observed for suicide attempts, with a crude OR of 1.6 (1.2–2.2) compared to an adjusted OR of 1.0 (0.7–1.4) [40]. Moreover, in a retrospective cohort study using the National Inpatient Sample dataset of adults admitted for major depressive disorder (MDD) from 2006 to 2017, there was a significantly higher prevalence of recurrent depression (77% vs. 69%), moderate to severe depression (72% vs. 68%) and suicidal ideation/attempts (49.5% vs. 41.8%) in patients with MDD who also had OSA compared to the MDD only group (p < 0.001 for all) [41].

Six studies assessed the effect of continuous positive airway pressure (CPAP) treatment on suicidality, two of which were included in our meta-analysis. Chu et al. (2023) reported a lower baseline prevalence of suicide attempts among CPAP users (1.5%) than non-users (2.2%), though the difference was not statistically significant, potentially due to short treatment duration and suboptimal compliance [18]. Udholm et al. (2022) similarly found lower rates of suicide (1.4% vs. 2.1%) and self-harm (1.2% vs. 2.4%) among CPAP users, though adherence and treatment duration were not reported [27]. Wickwire et al. (2024) using device-recorded data, classified adherence based on U.S. Centers for Medicare and Medicaid Services (CMS) criteria and found that patients with sustained CPAP adherence over two years had fewer self-harm events (4.0%) than non-adherent patients (5.0%) [42]. In two cohort studies that administered either the Hamilton Depression Rating Scale (HDRS) or PHQ-9, there was no significant difference in the mean suicidality score before and after receiving CPAP [43, 44]. In a prospective cohort study of 228 OSA patients, the mean score on item 9, which assesses passive suicidal ideation, decreased from 0.28 (SD = 0.58) to 0 (p < 0.001) after 3 months of treatment [45]. These findings suggest adherence and treatment duration may moderate CPAP’s effect on suicidality. On the other hand, sleep surgery significantly reduces depression scores, and this correlates with sleepiness rather than the Apnea-Hypopnea Index (AHI) [46]. We have shown that both CPAP and surgery for OSA can reduce depression scores with surgery showing a higher trend.

On assessing suicidality in OSA by gender, Chu et al. (2023) reported that the prevalence of suicide attempts among the OSA population was 1.7% overall, 1.6% among males, and 2.0% among females [18]. Hoier et al. (2022) reported that prevalence of death by suicide in the OSA population was 0.21% in males and 0.03% in females [21]. In Rod et al. (2017), the prevalence of death by suicide was also higher in males with OSA compared to females (0.15% vs. 0.09%) [25]. In two studies that assessed suicidal ideation, the prevalence was higher in females than males: 25.0% vs. 17.8% and 5.1% vs. 4.0%, respectively [26, 29]. Although the rate of suicide attempts and suicidal ideation was higher among females with OSA, the rate of death by suicide was higher among males. These trends are similar to the overall data comparing women to men in terms of suicidal ideation (4.5% vs. 4.1%) and suicide attempts (0.6% vs. 0.5%) in the United States [47]. In addition, the suicide rate among males in 2022 was nearly four times higher than that of women, with an age-adjusted rate of 23.0 per 100,000 people compared to 5.9 per 100,000, respectively [48].

It is imperative for OSA-managing providers (otolaryngologists, sleep physicians, dentists, and surgeons) to be aware of the clinical implications of these findings. A standardized psychological questionnaire is an important screening tool to detect suicidality in the clinical setting, since many patients may not otherwise disclose their distress. However, to triage the urgency of the situation and determine the resources that will be most helpful, a psychiatric interview is necessary. Upon identifying patients experiencing suicidality, it is essential to connect them with mental health professionals who can address not only their suicidality but any comorbid mental health conditions.

Several limitations should be acknowledged. There was substantial heterogeneity across studies in terms of OSA populations, comorbidities, suicidality assessment, and control groups. Most data originated from the United States, which has a higher-than-average suicide mortality rate, potentially limiting generalizability [49]. Several studies had predominantly male samples, introducing sampling bias.

Psychiatric comorbidities such as depression and anxiety were inconsistently reported. Only a subset of studies systematically assessed these conditions, and those that did often used different tools and diagnostic criteria. Most did not adjust for mental health status, introducing potential confounding. Moreover, no studies stratified suicidality outcomes by presence of mood disorders, precluding subgroup or meta-regression analyses. We recommend that future studies consistently assess and report mood disorder comorbidities to allow for more nuanced analysis.

Suicidality was measured using a range of approaches, including validated psychiatric scales (e.g., BDI-II, PHQ-9), administrative codes (ICD-9/10), medical records, and self-report questionnaires. This variability in timeframes and measurement methods likely contributed to heterogeneity in prevalence estimates. Subgroup analysis by assessment method was not feasible due to a limited number of studies using the same tools, overlapping measurement strategies, and inconsistent reporting practices across studies. Furthermore, self-reported data may underestimate prevalence due to underreporting [50].

OSA diagnosis methods also varied. Twelve studies used clinical tools such as polysomnography or ICD codes, while three relied on self-report. Our subgroup analysis showed a higher prevalence of suicidal ideation in clinically diagnosed patients (17.4%) compared to those with self-reported OSA (7.3%). However, clinically diagnosed cases represented only 601 individuals with suicidal ideation data, while self-reported cases represented 8,459. Notably, only one self-reported study (1,155 participants) examined suicide attempts; all other self-reported studies focused on suicidal ideation. Consequently, data on suicide attempts and death by suicide were almost entirely drawn from studies using clinical diagnoses, which accounted for 215,500 of the 223,959 total OSA or SDB cases. This distinction likely influences the overall pattern of findings and should be considered in interpretation.

Sex-stratified data were also limited. While five studies provided sex-specific suicidality data (two for suicidal ideation, two for suicide death, and one for suicide attempts), this was insufficient for meta-analysis. We have narratively synthesized these findings, which underscore the need for consistent sex-stratified reporting in future studies, given established sex differences in both OSA presentation and suicidality in the general population.

Other demographic variables like age and socioeconomic status could not be analyzed due to lack of subgroup data. Finally, the cross-sectional nature of most studies prevents causal inference. While subgroup analysis helped clarify some heterogeneity, residual confounding is likely. Longitudinal research with standardized tools is needed.

There is a need for further research on suicidality and depression in patients with OSA using standardized, validated screening tools. In addition, the fluctuating nature and spectrum of suicidality cannot be captured with a single question; rather, it is important to assess suicidality at various time points and with follow-up questions. Moreover, gathering more data on potential risk factors including but not limited to socioeconomic status, mental health comorbidities, OSA severity and phenotypes that impact suicidality can improve clinical management in these patients.