What does data on adverse reactions reveal about benzodiazepines and suicide-related risk?: A disproportionality analysis using FDA FAERS pharmacovigilance data

Abstract

Benzodiazepines (BZDs) are widely used in the treatment of psychiatric disorders, but their association with suicidal/self-injurious behavior is conflicting. This study investigated the relationship between BZDs and suicidal and self-injurious behavior, by analyzing data from the FDA Adverse Event Reporting System. We analyzed FDA Adverse Event Reporting System data from January 2004 and March 2025, to analyze adverse events (AEs) associated with suicidal and self-injurious behavior in BZDs. Thirty-eight BZDs were initially identified via the WHO Anatomical Therapeutic Chemical Code System. After excluding 26 BZDs with insufficient psychiatric disorder-related AEs (<50 reports), 12 (diazepam, chlordiazepoxide, oxazepam, potassium clorazepate, lorazepam, bromazepam, clobazam, alprazolam, flurazepam, triazolam, temazepam, and midazolam) were retained for disproportionality analysis using zolpidem as a control. Reporting odds ratios (RORs) were calculated, and subgroup analysis assessed risk variations by age and gender. Among 52,767 psychiatric disorders AEs, 9474 (17.95%) involved suicidal and self-injurious behaviors. Compared to zolpidem, diazepam (ROR = 1.38), chlordiazepoxide (ROR = 1.62), oxazepam (ROR = 2.14), lorazepam (ROR = 1.11), alprazolam (ROR = 1.64), flurazepam (ROR = 3.26) triazolam (ROR = 1.68), and temazepam (ROR = 1.72) showed significantly elevated risks, while clobazam (ROR = 0.46) and midazolam (ROR = 0.37) demonstrated protective effects. Subgroup analyses revealed higher risks in females using diazepam, oxazepam, lorazepam, alprazolam and temazepam. Adults <65 years who used potassium clorazepate, lorazepam, clobazam, alprazolam, or triazolam faced a significantly higher suicide-related risk than those ≥65 years. Compared to zolpidem, BZDs demonstrate varied suicide-related risks, which necessitated personalized risk-benefit evaluations and increased monitoring for high-risk agents such as flurazepam and alprazolam.

1. Introduction

Benzodiazepines (BZDs) are psychotropic drugs with a wide range of therapeutic applications,^[1] influencing the central nervous system and the brain by enhancing the activity of gamma-aminobutyric acid (GABA) at the GABA-A receptor. They are the most commonly prescribed drugs in psychiatry due to their significant therapeutic effects.^[2–4] The advantages of BZDs lies in their interaction with different receptor subtypes, which exhibit varying half-lives and potencies, thereby allowing for selection based on the specific needs of treatment.^[5] BZDs possess sedative, hypnotic, anxiolytic, anticonvulsant, and muscle relaxant properties, making them suitable for procedural sedation as well as for the treatment of insomnia, anxiety disorders, epilepsy, and alcohol withdrawal syndrome.^[6–8]

The safety and efficacy of BZDs are supported by clinical guidelines, which primarily demonstrate positive efficacy in short-term use.^[4] However, long-term use of BZDs may lead to a range of adverse events (AEs), including tolerance, addiction, central nervous system depression, amnesia, and other AEs,^[5,9,10] resulting in increasing concerns about negative mental health effects. It has been suggested that flunitrazepam, lorazepam, and alprazolam may pose a risk of disinhibition, increasing the likelihood of aggressiveness, impulsivity, intentional self-harm, and suicide risk, other BZDs may exhibit similar adverse effects.^[11–13] A large national observational study leveraging the nationwide French reimbursement healthcare system databases, identified a significant association between BZDs and suicide attempts and suicide, even after controlling for indication bias.^[14] While most studies supposed a link between BZDs and suicide, some research highlighted the potential beneficial role of BZDs in alleviating anxiety symptoms, suggesting that their use could potentially reduce the risk of suicide. Notably, different studies (such as critical commentary, retrospective cohort analysis and a case-control study) had consistently indicated that BZDs use does not significantly elevate suicide risk among high-risk populations.^[15–17] Furthermore, recent systematic reviews have indicated that cautious short-term use of BZDs, when combined with adjunctive therapies, may help reduce suicide risk, which aligned well with the findings of the previously mentioned studies.^[18] Given the mixed findings on the effect of BZDs on suicide risk and further research is warranted.

To address these seemingly contradictory perspectives, this study aimed to explore the potential association between benzodiazepines and suicidal and self-injurious behavior by utilizing real-world data. The reported odds ratio (ROR) for risk was calculated using spontaneous AE reports from the FDA Adverse Event Reporting System (FAERS). FAERS is a designated FDA database for post-marketing drug safety surveillance that adheres to international standards (ICH E2B) and aggregates global reports. It is anticipated that the findings will contribute to guiding rational clinical practices concerning benzodiazepine use, ultimately aiding in risk mitigation.

2. Methods

2.1. Data source

Researchers can utilize the spontaneous reporting data in FAERS for signal detection and to assess the relationship between drugs and AEs. It includes AEs data reported globally by manufacturers, consumers, and healthcare professionals following medication use, covering populations exposed to FDA-regulated products worldwide, though with variable reporting completeness. The database is continuously updated and can be downloaded from the FDA website (https://www.fda.gov/drugs/drug-approvals-and-databases/fda-adverse-event-reporting-system-faers-database).^[19–21] This study used FAERS data from January 2004 to March 2025. Data were collected by identifying the generic names of BZDs, followed by an analysis that involved the removal of duplicate reports. We evaluated the associations between these drugs and suicidal and self-injurious behavior, which is high-level term within the FAERS. An ethical review was not required, as personal information in the FAERS database is de-identified.

2.2. Benzodiazepines and control drug definition

According to the Anatomical Therapeutic Chemical Coding System (https://www.atccode.com), a total of 38 BZDs were identified and classified as benzodiazepine derivatives anxiolytics (e.g., alprazolam, diazepam) and benzodiazepine derivatives (e.g., midazolam, clobazam). Anatomical Therapeutic Chemical is an official drug classification system established by the World Health Organization (WHO). Twenty-six of the BZDs were excluded due to an insufficient number of psychiatric disorders AEs (<50). Ultimately, 12 BZDs were remained: diazepam, chlordiazepoxide, oxazepam, potassium clorazepate, lorazepam, bromazepam, clobazam, alprazolam, flurazepam, triazolam, temazepam, and midazolam. This refined selection allowed enabled further examination of the potential association between these specific benzodiazepines and the risk of suicidal and self-injurious behavior AEs.

Zolpidem, an activator of GABA-A receptors, was employed as a control in this study. It is one of the most commonly prescribed non-BZD hypnotics for the treatment of insomnia, and it is a psychoactive substance similar to BZDs. Existing epidemiological studies and meta-analysis have indicated that the use of zolpidem is associated with an increased risk of suicide-related outcomes.^[22–24] However, the REST-IT randomized trial demonstrated that zolpidem significantly reduced suicidal ideation in severely depressed insomniacs when measured by the Columbia-Suicide Severity Rating Scale, with effects most pronounced in patients with baseline severe insomnia.^[25,26] The evidence shows zolpidem’s intrinsic suicide risk while highlighting its therapeutic potential in targeted populations, making it an ideal comparator for evaluating BZD-related suicide risks. The data processing procedure is illustrated in Figure 1.

Figure 1.

Flowchart of data of BZDs processing procedure on FAERS. AEs = adverse events, ATC = Anatomical Therapeutic Chemical Coding System, BZDs = benzodiazepines, FAERS = the Food and Drug Administration Adverse Event Reporting System.

To further ensure the robustness of disproportionality signals, lithium was incorporated as an alternative control in sensitivity analysis. Details are provided in Supplementary Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q186. This selection is grounded in its established pharmacological independence from GABAergic pathways (distinct from BZDs/zolpidem) and its well-documented anti-suicidal properties, as evidenced by meta-analyses of RCTs showing lithium significantly reduces suicide incidence versus placebo in mood disorders.^[27–29] Details are provided in Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q186.

2.3. Statistical analysis

In this research, the ROR was employed for adverse reaction signal mining, recognized as the most prevalent technique in contemporary pharmacovigilance research.^[30,31] It served as a measure of the odds ratio for the occurrence of suicidal and self-injurious behavior associated with BZDs and zolpidem. The 95% confidence intervals (CIs) were calculated using a 5% alpha level to assess statistical significance. Disproportionate reporting was identified when the lower limit of the 95% CI exceeded 1.0, indicating positive signal, with P < .05 considered statistically significant. To further explore the distribution of complete suicide risk across age and gender, subgroup analysis were conducted.

Data analysis was employed Microsoft Excel 2019 (Microsoft Corporation, Redmond), SAS version 9.4 (SAS Institute Inc., Cary), and R version 4.3.2 (R Core Team, hosted by Friedrich‐Alexander‐Universität Erlangen‐Nürnberg, Erlangen, Bavaria, Germany).

3. Results

Between January 2004 and March 2025, there were 52,767 psychiatric disorders AEs, with 9474 (17.95%) being classified suicidal and self-injurious behavior. A total of 8 benzodiazepine derivatives anxiolytics and 4 benzodiazepine derivatives met the criteria, as detailed in Table S2, Supplemental Digital Content, https://links.lww.com/MD/Q186. Among these, benzodiazepine derivatives anxiolytics accounted for 12 cases (bromazepam) to 4890 cases (alprazolam), while benzodiazepine derivatives ranged from 39 cases (midazolam) to 235 cases (triazolam). Using zolpidem as control drug, the analysis of suicidal and self-injurious behavior with BZDs were illustrated in Table 1.

Table 1.

Statistical analysis of suicidal and self-injurious behavior with BZDs.

Drugs	Number of cases (n)	Total cases of psychiatric disorders (N)	ROR	95% CI lower	95% CI upper	χ 2	P-value
Benzodiazepine derivatives anxiolytics
Diazepam	1859	10,525	1.38	1.29	1.47	93.55	<.001
Chlordiazepoxide	41	204	1.62	1.15	2.29	7.61	.006
Oxazepam	145	581	2.14	1.77	2.59	62.66	<.001
Potassium clorazepate	17	169	0.72	0.44	1.19	1.66	.197
Lorazepam	1821	12,383	1.11	1.04	1.18	9.70	.002
Bromazepam	12	97	0.91	0.50	1.66	0.10	.755
Clobazam	94	1411	0.46	0.37	0.57	53.59	<.001
Alprazolam	4890	24,114	1.64	1.55	1.72	342.06	<.001
Benzodiazepine derivatives
Flurazepam	84	249	3.28	2.51	4.27	85.36	<.001
Triazolam	235	1134	1.68	1.45	1.95	47.27	<.001
Temazepam	224	1061	1.72	1.48	2.01	49.28	<.001
Midazolam	39	712	0.37	0.27	0.52	38.18	<.001
Zolpidem*	2507	18,633	1	–	–	–	–

BZDs = benzodiazepines, CI = confidence interval, ROR = reporting odds ratio.

^*Control drug.

3.1. Comparison of benzodiazepines to zolpidem

Compared to zolpidem, the ROR for suicidal and self-injurious behavior for each selected BZD were: diazepam (ROR = 1.38, 95% CI 1.29–1.47, P < .001), chlordiazepoxide (ROR = 1.62, 95% CI 1.15–2.29, P = .006), oxazepam (ROR = 2.14, 95% CI 1.77–2.59, P < .001), lorazepam (ROR = 1.11, 95% CI 1.04–1.18, P = .002), alprazolam (ROR = 1.64, 95% CI 1.55–1.72, P < .001), flurazepam (ROR = 3.28, 95% CI 2.51–4.27, P < .001), triazolam (ROR = 1.68, 95% CI 1.45–1.95, P < .001) and temazepam (ROR = 1.72, 95% CI 1.48–2.01, P < .001). There were 2 BZDs with RORs <1, which were clobazam (ROR = 0.46, 95% CI 0.37–0.57, P < .001) and midazolam (ROR = 0.37, 95% CI 0.27–0.52, P < .001).

3.2. Subgroup analysis of completed suicide

Subgroup analysis of completed suicide was conducted according to gender and age (Fig. 2). Differences in risk of AEs across BZDs were observed. Regarding gender subgroups, females exhibited a higher risk compared to males in some BZDs, with ROR values exceeding 1. These BZDs were diazepam (ROR = 1.83, 95% CI 1.64–2.04, P < .001), oxazepam (ROR = 1.98, 95% CI 1.31–3, P = .001), lorazepam (ROR = 1.2, 95% CI 1.08–1.34, P = .001), alprazolam (ROR = 1.24, 95% CI 1.16–1.32, P = .005) and temazepam (ROR = 1.7, 95% CI 1.23–2.34, P = .001). In terms of age subgroups, the drug associated with a higher risk of completed suicide in younger individuals were potassium clorazepate (ROR = 3.77, 95% CI 1–14.17, P = .038), lorazepam (ROR = 1.53, 95% CI 1.34–1.75, P < .001), clobazam (ROR = 4.16, 95% CI 1–17.33, P = .034), alprazolam (ROR = 1.17, 95% CI 1.06–1.28, P = .002), and triazolam (ROR = 2.46, 95% CI 1.72–3.5, P < .001).

Figure 2.

RORs of BZDS for suicidal and self-injurious behavior by gender and age. BZDs = benzodiazepines, ROR = reporting odds ratio.

4. Discussion

In this study, we conducted a disproportionality analysis utilizing the FAERS to explore the potential association between BZDs and the risk of suicidal and self-injurious behavior. Anxiety and insomnia are often regarded as modifiable risk factors for suicide, with BZDs playing a significant role in the treatment of these symptoms.^[32] However, the relationship between BZDs and suicidal and self-injurious risk remains inadequately substantiated, necessitating further investigation. This study aimed to leverage real-world data to identify potential safety signals and provide evidence to support clinical decision-making.

The study found suicidal Ideation AEs occurred 1146 times, while suicide attempt and completed suicide AEs occurred 2714 and 5069 times, respectively, which could be seen in Table S1, Supplemental Digital Content, https://links.lww.com/MD/Q186. The AEs of suicidal ideation were much smaller than the AEs of actually performing the suicidal act. BZDs had been shown to be associated with suicidal ideation.^[33] However, the actual act of suicide usually preceded suicidal ideation, there was no linear relationship between suicide-related events, and the majority of suicidal individuals did not commit suicide beforehand.^[34,35] The association between BZDs and suicide may be influenced by various factors, including the mental health conditions of patients, the drugs’ mechanisms, dosages, duration of use and so on.

The study demonstrated that diazepam, chlordiazepoxide, oxazepam, lorazepam, alprazolam, flurazepam, triazolam, and temazepam were found to be at higher risk than Zolpidem. They are mainly used for the treatment of insomnia and anxiety disorders. Insomnia is a potential risk factor for suicide,^[34] while anxiety is also strongly linked to suicidal behavior. BZDs may influence neurotransmission in the brain, particularly in neural pathways related to emotional and behavioral regulation, which can lead to increased impulsivity and aggression in certain individuals.^[36,37] GABA is the major inhibitory neurotransmitter in the brain, and changes in its levels have been associated with the development of a variety of psychiatric disorders such as anxiety, depression, epilepsy, and so on. By modulating GABA levels, BZDs may impact emotions and behaviors associated with suicidal tendencies.^[5] For instance, heightened impulsivity or aggression may elevate the risk of suicide.^[32] The precise mechanisms underlying these effects require further investigation.

The FDA requested an update to the Boxed Warning for BZDs in September 2020, which is the FDA’s most serious warning, updating the risk of misuse, abuse, physical dependence, addiction, and withdrawal reactions to BZDs. Such as an overdose of temazepam was likely to cause death.^[38] High-risk individuals may be affected by prolonged or high-dose use of BZDs, potentially leading to suicidal action.^[39] Furthermore, long-term use of BZDs may result in physical dependence and heighten the risk of substance abuse.^[2] The abuse of BZDs was often combined with other substances such as opioids or alcohol, which may further increase the risk of suicide.^[40] This combination can induce a state of severe intoxication, impairing judgment and heightening the risk of impulsive behavior. However, abruptly discontinuing BZDs may lead to withdrawal symptoms, including anxiety, insomnia, and seizures. These withdrawal symptoms can exacerbate existing mental health conditions and may increase suicidal thoughts or behaviors.^[1]

The association between BZD use and suicide risk appears multifactorial, influenced by pharmacological properties, clinical indications, and patient populations. Notably, clobazam and midazolam exhibited lower risks compared to zolpidem, likely due to their distinct mechanisms and usage contexts. Clobazam preferential binding to GABA-A α2 subunits may mitigate disinhibition-related impulsivity,^[5] while its primary use in epilepsy under strict monitoring reduces long-term exposure. Midazolam, typically administered for acute procedural sedation, minimizes cumulative risks associated with chronic use. In contrast, zolpidem’s widespread prescription for insomnia may exacerbate vulnerability in high-risk populations.

In subgroup analysis of completed suicide, females exhibited a higher risk than males for the use of diazepam, oxazepam, lorazepam, alprazolam, and temazepam. The study suggested a relationship between gender and suicidal behavior.^[41] Some studies had shown that women had a higher risk of suicide.^[42,43] May be related to gender role expectations, coping strategies to deal with stress. In general, younger people demonstrate relatively higher risk of suicide, particularly with regard to the use of potassium clorazepate, lorazepam, alprazolam, and triazolam. Increasing suicidal ideation and suicide attempt among adolescents,^[44,45] which may be related to factors such as depression, substance use and adverse experiences. Suicidal behavior may behave differently in various age groups and more in-depth studies are needed, and targeted prevention strategies should be developed.

While our study provided insights into the association between BZDs and suicide risk, several limitations and considerations remained. Firstly, the FAERS database relies on voluntary reporting, which may introduce reporting bias, and not all AEs are recorded, limiting the completeness and reliability of the data. Secondly, the study failed to control for other variables that could influence suicide risk, such as dosage, duration of medication, and patients’ psychiatric history. Lack of dosage and treatment duration data in FAERS further limits the assessment of dose-dependent suicide risk and the impact of long-term BZD use. Furthermore, the study design was unable to establish causality because the observed association may be confounded by other unmeasured variables or factors. Additionally, sample selection may be biased, as the reports may not represent all patient groups using BZDs. Lastly, the findings may not be generalizable to all populations, especially given that BZD use may vary across regions and healthcare systems. Therefore, more rigorous study designs and comprehensive data are needed to further validate these findings.

5. Conclusion

In conclusion, the relationship between BZDs and suicide is complex and may be influenced by a variety of factors. Although BZDs are associated with suicide risk, they are safe and effective when used appropriately. Healthcare professionals must conduct a thorough assessment of the patient’s overall health, carefully weighing the potential risks against the benefits. Timely monitoring of the patient’s mental status and behavioral changes is essential to ensure the safe use of BZDs. Patients should follow their doctor’s instructions and contact a healthcare professional if they experience any concerning symptoms. Furthermore, special caution is warranted with BZDs due to the potential risks of physical dependence and abuse.

Author contributions

Data curation: Weiling Guo.

Formal analysis: Weiling Guo.

Project administration: Weiling Guo.

Resources: Liang Yan.

Software: Weiling Guo.

Supervision: Chenyu Zhang, Ren Yang, Liang Yan.

Validation: Chenyu Zhang, Ren Yang.

Writing – original draft: Weiling Guo.

Writing – review & editing: Chenyu Zhang, Ren Yang, Liang Yan.

Supplementary Material

medi-104-e44797-s001.pdf ^{(488KB, pdf)}