Sleep-related drugs utilization and suicide behaviors: a population-based study in China

In recent decades, suicide has been recognized as a significant public health issue globally. The World Health Organization (WHO) estimates that over 700,000 people die by suicide annually, surpassing the death toll from malaria, AIDS, breast cancer, or wars and homicides [1]. Numerous studies have sought to identify factors associated with suicidal behaviors, finding various social, psychological, and behavioral factors to be linked [2]. Among these, sleep disturbances, including insomnia [3‐6], nightmares [7, 8], and others [9, 10], have been identified as risk factors for suicidal behaviors. Consequently, treating sleep disturbances could be an effective strategy for preventing suicidal behaviors. Supporting this, studies among patients with sleep disorders have shown that visits to sleep medicine specialists can play a protective role against suicide attempts [11, 12]. Furthermore, a randomized controlled trial demonstrated that cognitive-behavioral therapy for insomnia can reduce symptoms of insomnia, thereby helping to alleviate and prevent suicidal ideation [13, 14].

Given the positive impact of treating sleep disturbances on suicidal behaviors among patients with sleep disorders, it is conceivable that pharmacological therapy, the most common method for treating sleep disturbances at the population level, could help manage suicidal behaviors [15, 16]. However, a study drawing on two national population-based samples indicated that the use of prescription medications for insomnia was positively associated with suicidal behaviors without being linked to any specific class of medication [17]. These results highlight the complex relationship between the use of SRDU and suicidal behaviors, suggesting that more research is needed to fully understand this connection.

The conflicting associations between treating sleep disturbances and the use of SRDU on suicidal behaviors may be attributable to differences in the populations studied. The protective effect of treating sleep disturbances was supported among sleep disorders patients. However, the positive association between SRDU and suicidal behaviors was supported by population-based studies [17]. For patients with sleep disorders, sleep problems may be a significant risk factor for suicidal behaviors, and addressing these disturbances could effectively control such behaviors. However, in the general community, individuals with suicidal behaviors may be influenced by a variety of factors and might use sleep medications as a way to manage their suicidal tendencies. The positive associations between SRDU and suicidal behaviors observed in community residents could stem from increased use of sleep medications among those considering suicide. On the other hand, the positive correlations between SRDU and suicidal behaviors at the population level might result from the impact of sleep disturbances rather than SRDU itself. Understanding the role of sleep disturbances could provide valuable insights into the relationships between SRDU and suicidal behaviors among community residents.

On the other hand, suicide is a process that ranges from ideation to action, with suicidal behaviors classified into suicidal ideation, suicide planning, suicide attempts, and completed suicide, according to the WHO criteria [2]. Previous population-based studies typically analyzed suicidal behaviors across the entire sample, sometimes including individuals at risk of suicide in the control groups [18‐20]. For instance, in studies examining factors associated with suicide attempts, individuals with only suicidal ideation or plans were often grouped with those without any suicide attempts. While this approach is effective for identifying factors associated with suicide attempts, it may lead to the loss of critical information. Moreover, although suicidal ideation is a significant predictor of subsequent suicidal behaviors, distinctions between ideation and actual suicide attempts have been observed in prior research [21]. To address these nuances, this study categorized suicide risk into general individuals with no suicidal behaviors (GNS), suicidal ideators without plans or attempts (SINPA), suicidal ideators with plans but no attempts (SIP), and those who have attempted suicide (SA), termed conditional suicidal behaviors [22, 23]. This classification acknowledges the stages of suicide risk and, while it may not elucidate causal relationships, it aids in analyzing the differences among individuals at various stages of suicidal behaviors.

In China, sleep medications are categorized into prescription and nonprescription drugs. Prescription medications for sleep issues, such as Zopiclone and Benzodiazepines, are similar to those available in other countries. On the other hand, nonprescription options for sleep disturbances primarily include traditional Chinese medicines like Anshenbunao Syrup and Jieyu Anshen Granules, which have been effective in addressing sleep problems. This diversity in sleep medications may result in different associations between SRDU and suicidal behaviors in China compared to other countries.

Given this context, a population-based study was conducted to investigate the associations between SRDU and suicidal behaviors, adjusting for factors such as insomnia, sleep apnea, RBD, and narcolepsy. The study also examined the relationship between SRDU and various stages of suicidal behavior. These findings not only enhance our understanding of the link between SRDU and suicidal behaviors from a public health standpoint but also offer valuable insights for managing suicidal behaviors among community residents dealing with sleep issues.

In this study, a total of 21,376 participants were analyzed. The prevalence rates of suicidal ideation, planning, and attempts were found to be 1.4%, 0.3%, and 0.2%, respectively. Additionally, 575 participants (2.7%) reported using sleep-related drugs in the past year. Factors associated with suicidal ideation included gender (\(\:{\chi\:}^{2}\)=33.93, p < 0.001), age (t = 5.51, p < 0.001), education (\(\:{\chi\:}^{2}\)=33.57, p < 0.001), living alone (\(\:{\chi\:}^{2}\)=31.84, p < 0.001), insomnia (t = 21.26, p < 0.001), sleep apnea (\(\:{\chi\:}^{2}\)=5.69, p < 0.05), RBD (t = 15.81, p < 0.001), and SRDU (\(\:{\chi\:}^{2}\)=351.61, p < 0.001). Factors associated with making a suicide plan were age (t = 2.20, p < 0.05), insomnia (t = 6.02, p < 0.001), RBD (t = 8.58, p < 0.001), narcolepsy (t = 4.26, p < 0.001), and SRDU (\(\:{\chi\:}^{2}\)=71.72, p < 0.001). Factors associated with suicide attempts included RBD (t = 5.38, p < 0.001), and SRDU (\(\:{\chi\:}^{2}\)=23.20, p < 0.001 Detailed information regarding sociodemographic characteristics and individual analyses can be found in Table 1.

In Table 2, logistic regression analyses were performed to examine the associations between SRDU and suicidal behaviors, with and without controlling for sleep disorders. After adjusting for age, gender, ethnicity, educational level, marital status, region, and living alone, SRDU was significantly associated with suicidal ideation, as shown in Model 1 A (OR = 7.82, p < 0.001). Model 1B, which further controlled for insomnia, sleep apnea, RBD, and narcolepsy, found that suicidal ideation was significantly associated with SRDU (OR = 3.02, p < 0.001), insomnia (OR = 1.15, p < 0.001), sleep apnea (OR = 1.53, p < 0.05), RBD (OR = 1.04, p < 0.001), and narcolepsy (OR = 1.06, p < 0.05). In Model 2 A, the results supported that suicide planning was associated with SRDU (OR = 7.88, p < 0.001), after controlling for age, gender, ethnicity, educational level, married status, region, and living alone. When sleep disorders were controlled for, suicide planning was significantly associated with SRDU (OR = 4.78, p < 0.001), RBD (OR = 1.04, p < 0.001), and narcolepsy (OR = 1.16, p < 0.001). In Model 3 A, SRDU was significantly associated with suicide attempts (OR = 7.89, p < 0.001). After controlling for sleep disorders, suicide attempts were significantly associated with SRDU (OR = 6.40, p < 0.001) and RBD (OR = 1.05, p < 0.001).

In Table 3, conditional suicidal behaviors were further analyzed and compared with GNS. Consistent with the models presented in Table 2, variables such as age, gender, ethnicity, educational level, marital status, region, and living status (living alone or not) were controlled across all models. In Model 4 A, SRDU was significantly associated with suicidal ideation (OR = 7.75, p < 0.001). After adjusting for related sleep disorders, SINPA was found to be associated with SRDU (OR = 2.46, p < 0.001), insomnia (OR = 1.18, p < 0.001), sleep apnea (OR = 1.55, p < 0.05), and RBD (OR = 1.04, p < 0.001). In Model 5 A, SRDU showed a strong association with SIP (OR = 8.20, p < 0.001). When controlling for sleep disorders in Model 5B, SIP was associated with SRDU (OR = 5.11, p < 0.001), insomnia (OR = 1.08, p < 0.05), and narcolepsy (OR = 1.19, p < 0.001). Model 6 A demonstrated that SA was significantly associated with SRDU (OR = 8.72, p < 0.001). In Model 6B, after further adjustments, SA was associated with SRDU (OR = 6.93, p < 0.001) and RBD (OR = 1.05, p < 0.001).

Further logistic regression analyses were conducted to compare the differences in SRDU among various conditional suicidal behaviors. In Models 7 A, 8 A, and 9 A, no significant differences in SRDU were observed among the conditional suicidal behaviors (all p > 0.05). This finding remained consistent even after controlling for related sleep disorders, as indicated in Models 7B, 8B, and 9B, where no significant differences in SRDU among the conditional suicidal behaviors were found (all p > 0.05). Detailed information can be found in Table 4.

This study revealed several key findings. Firstly, the self-reported lifetime prevalence rates of suicidal ideation, planning, and attempts were 1.4%, 0.3%, and 0.2%, respectively. Secondly, SRDU was positively associated with suicidal ideation, planning, and attempts, even after controlling for insomnia, sleep apnea, RBD, and narcolepsy. Additionally, compared to GNS, SRDU was also positively associated with conditional suicidal behaviors. However, among individuals who had considered suicide, SRDU was not found to be associated with any further suicidal behaviors.

Regarding the first point, the prevalence rates of suicidal ideation, planning, and attempts in this study were significantly lower than those reported in previous research. A meta-analysis indicated that the prevalence rates of suicidal ideation and attempts among the Chinese community population were 3.9% and 0.8%, respectively [47]. The lower prevalence of suicidal behaviors observed in this study could be attributed to the decreasing trend of suicide rates in China. Recent years have seen a significant reduction in suicide rates in the country [48], which may also reflect a concurrent decline in the prevalence of suicidal ideation and attempts.

This study further confirmed that SRDU was positively associated with suicidal behaviors, even after adjusting for insomnia, sleep apnea, RBD, and narcolepsy. As mentioned in the Introduction, similar findings were supported by two national population-based samples without adjusting for sleep disturbances [17]. Several explanations might help us understand the link between SRDU and suicidal behaviors. The first explanation could be mental or sleep disorders. A study utilizing the Norwegian Prescription Database found that psychotropic medications were commonly prescribed before suicide, particularly antidepressants, hypnotics, and sedatives [49]. SRDU indicates more severe sleep problems, which are significant risk factors for suicidal behaviors [50]. The second explanation might be the adverse effects of sleep-related drugs. Previous studies have identified several adverse reactions to these drugs, such as cognitive impairment, respiratory depression, and suicidal ideation [51, 52], which are also risk factors for suicidal behaviors [52, 53]. The final explanation could be the misuse of sleep-related drugs. Overdosing on sleep medications is also a method of suicide [54], suggesting a direct link between SRDU and suicidal behaviors.

This study also examined the use of SRDU across different stages of suicide risk, finding that participants at any level of suicide risk were more likely to use sleep-related drugs compared to GNS. These results suggest that SRDU may be positively associated with all types of conditional suicidal behaviors. The reasons for these positive associations are likely similar to those identified in the relationship between SRDU and overall suicidal behaviors.

This study found no statistical differences in SRDU among individuals with SINPA, those with SIP, and SA, suggesting that SRDU does not influence the progression from suicidal ideation to attempt. While suicide is a continuum from ideation to action, and ideation is a known predictor for subsequent suicidal behaviors [55, 56], distinctions between ideation and action have been supported by other research [57, 58]. Investigations have also been conducted to explore factors that contribute to the transition from ideation to attempt [59]. To our knowledge, no study has documented differences in SRDU among these conditional suicidal behaviors. These findings suggest that SRDU may not act as a catalyst for moving from ideation to attempt; once individuals have considered suicide, the impact of SRDU appears limited, at least at the population level. Indeed, the marginal benefit of improving sleep quality in reducing suicidal behaviors was also observed in a previous population-based study [60].

Several limitations should be considered for interpreting the findings in this study. Firstly, because of the cross-sectional design, any causal relationships between SRDU and suicidal behaviors cannot be inferred. Secondly, the specific sleep drugs were not listed in the questionnaire, and the associations between SRDU and suicidal behaviors cannot be inferred into the particular sleep-related drugs. Thirdly, although all the scales about sleep distributions could evaluate the symptoms of sleep distributions with nice reliability and validity, any diagnoses cannot be made in this study. Finally, the data was collected by self-report, and the participants’ report may be not completely accurate. It may also cause some bias for the findings in this study.

This population-based study found that the prevalence of suicidal behaviors was low, yet SRDU was positively associated with suicidal ideation, planning, and attempts, even after controlling for insomnia, sleep apnea, RBD, and narcolepsy. However, SRDU did not impact the progression from suicidal ideation to attempts. Although a causal link between SRDU and suicidal behaviors cannot be established, the findings highlight that individuals using sleep-related drugs are at a higher risk of suicidal behaviors. Nonetheless, managing SRDU may not effectively prevent subsequent suicidal behaviors once individuals have considered suicide.

While the reasons for the positive association between SRDU and suicidal behaviors are difficult to pinpoint, this study suggests that SRDU could serve as an indicator of suicidal risk. This underscores the importance of monitoring individuals with SRDU for suicidal behaviors. Since this study did not specify particular sleep medications, future research could investigate the relationships between specific sleep medications and suicidal behaviors. This would further our understanding of the underlying reasons and mechanisms behind the association between SRDU and suicidal behaviors.