**3. Results**

A total of 2541 participants were included in this analysis. The majority of participants were male (67.4%) and the mean age was 39.4 years (SD 11.4); 38.3% were Black, 41.3% White, and 14.4% Hispanic. Recent use of alcohol and benzodiazepines was reported by 60.0% and 15.8%, respectively, and 39.0% reported lifetime use of heroin. With regard to co-occurring mental health conditions, 51.6% scored in the depressed range at baseline, and 50.2% indicated that they had at least one preexisting psychiatric diagnosis. The mean baseline CHRT-SR score was 23.9 (SD 8.5; min 11, max 59). A total of 122 (4.8%) either agreed or strongly agreed with the last three items in the CHRT- SR scale and were categorized as suicidal at baseline. Among those who were suicidal, there was a higher proportion of Black/African American (45.1%), followed by White (38.5%) and Hispanic (13.1%) populations. Seventy-five participants (3.0%) had at least one overdose event during their study participation. Of these 49.3% were white, 26.7% were Black/African American and 20.0% were Hispanic. Of those participants who reported suicidal thoughts and intent, only 6 (4.9%) had an overdose event.

Demographic characteristics of the participants can be found in Table 2, and proportions of gender and of race/ethnicity overall, among those who were suicidal, and among those with an overdose event can be found in Figures 1 and 2. Demographic information by study can be found in the primary outcomes' publications [32–39]. Total participants in analyses varied slightly due to occasional missing data.


**Table 2.** Participant characteristics overall, by suicidal yes/no, by overdose yes/no.

\* One participant is missing from the suicidal total due to missing responses to the last 3 items on CHRT.

Preliminary results of model fit revealed that age, gender, and race/ethnicity were not significant in the model, and therefore dropped from the final model. Results of logistic regression show that the continuous CHRT-SR score was associated (*p* = 0.03) with overdose events and the likelihood of overdose increased as the continuous CHRT-SR score increased (OR 1.02; 95% CI = 1.00–1.04). Depression, recent use of alcohol or benzodiazepines, history of psychiatric disorders, and treatment arm were not associated with higher odds of overdose, but lifetime use of heroin was associated (*p* < 0.01) with increased odds of overdose (OR 3.08; 95% CI = 1.93–4.92). Model results can be found in Table 3.

**Figure 1.** Proportion of Female/Male overall, among those who are suicidal, and among those with an overdose event.

**Figure 2.** Proportion of Race/Ethnicity overall, among those who are suicidal, and among those with an overdose event.


**Table 3.** Results of logistic regression/generalized estimating equation assessing CHRT-SR as a continuous predictor of overdose.

## **4. Discussion**

Results of this study demonstrate that suicide propensity, ideation and intent, as assessed by the continuous CHRT-SR score, were associated with overdose events amongs<sup>t</sup> patients seeking treatment in eight clinical trials for SUD. Increases in the total score of CHRT-SR were associated with a higher likelihood of experiencing an overdose event. To our knowledge this is the first study exploring the relationship of suicidality and overdose events in a substance-using, treatment-seeking population.

Surprisingly, only 4.8% of the sample endorsed suicidal ideation or intent at baseline. This is possibly explained by the fact that most studies excluded actively suicidal patients at enrollment. Suicidal ideation and intent have not been previously documented systematically during treatment in patients with SUD or dual disorders. Rather, assessments have been based on suicide mortality in these populations. Esang and Ahmed [57] reported an increased risk of suicide in patients with substance use and psychiatric conditions. Increased opioid overdose risk has been described in patients with OUD [58]. Studies assessing suicidal intent after an overdose report highly variable prevalence of suicidal ideation. For example, while one study documented that 58.0% of patients with OUD who overdosed indicated a desire to die [18], another reported only 6.6% of opioid overdose survivors firmly expressed their intent to die [16]. In addition, these retrospective studies relied on patients' self-reports, where reconstruction of the feelings, thoughts, and desire prior to the overdose event might be misconstrued or be prone to recall bias. Suicidal ideation and intent have been documented in prior studies of patients with acute major depression with non-psychotic features. Based on the type of assessment used, the prevalence of suicidal thoughts and intent ranged from 10.7% to 19.0% at baseline [59]. The prevalence of suicidal ideation or intent documented in our study is therefore lower than that observed in patients in treatment for major depression. Of note, the prevalence of depression in our sample was 51.6%, and 50.2% of the participants reported a previous psychiatric history. While the association between depression and SUD in our sample was high, other studies have demonstrated even higher levels of depression in patients with SUD [60].

Results from this study confirm that lifetime heroin use is associated with an increased likelihood of an overdose event. Brandt and colleagues [58] and Stover and colleagues [19] have demonstrated increased risk of opioid overdose events in patients with a history of heroin use. Surprisingly, and in contrast with the literature, neither depression nor recent alcohol or benzodiazepine use were associated with overdose events in our study. This might be explained by the variability in the approach to the assessment of depression across trials, or that participants were only experiencing mild depressive disorders, and therefore, the association with increased overdose was not observed as expected.

Consistent with the literature, overdose events were more frequent in the White population (49.3%), followed by Black/African American (26.7%) and Hispanic (20.0%) populations, whereas suicidality was more frequent in Black/African American (45.1%), followed by White (38.5%) and Hispanic (13.1%) populations. Gicquelais and colleagues [16] found that active or passive suicidal intent was more prevalent in Whites (86.3%) followed by multiracial individuals (9.3%), whereas only 4.3% of the Black/African American sample intended to die. In our study, the choice of substance used may have accounted for the racial/ethnic differences we found. However, this is speculative, given that most of the information on the overdose was self-reported at the time of documentation and was inconsistently documented in the narrative of overdose events, preventing identification of specific substances involved.

The present study has several strengths. First, it provides an examination across multiple sites and multiple studies expanding on prior single site examinations of suicidality and overdose. Second, it expands on the use of large datasets to rapidly answer practical clinical questions and furthers the understanding of mental health conditions and their association with adverse substance use outcomes.

This study also presents several limitations. First, the design of this study only allowed for examination of associations and not causation. Second, the population enrolled across these eight clinical trials is representative of individuals seeking treatment for SUD who agreed to participate in a research study. Therefore, our results might not be generalizable to the entire population of persons with SUD and other co-occurring mental health conditions. It is noteworthy that while the panel of experts ascertained outcomes using a binary yes/no approach for each of the covariates, the measures used to evaluate these variables differed across trials and may have had slightly different meanings. However, the tradeoff of this heterogeneity might be mitigated by the large sample size accrued across multiple sites and conducted over 10 years. Finally, heroin use was not assessed in one of the trials and is reported as missing data in our results.

This study highlights the relevance of assessing suicidality at baseline for patients entering treatment. As drug overdose deaths continue to rise in the U.S. and worldwide and suicide remains a leading cause of death, particularly in patients with comorbid substance abuse and mental health disorders, multi-pronged approaches covering prevention, early detection, and intervention are needed. Integration of suicide risk reduction should be included as part of an overall prevention strategy in response to the steep climb in overdose fatalities. Entry into treatment for SUD and other co-occurring mental health conditions presents an invaluable opportunity to re-set, evaluate suicide risk, and employ early prevention plans.

**Author Contributions:** V.E.H. formulated the research question and drafted the first version of the manuscript. D.S.-B., K.H., J.F., S.T., V.E.H. and R.D.S. reviewed and decided on measures for all covariates across trials. D.J.F. provided methodological guidance and approach for the model created for data analysis. R.D.S. drafted the methods section and performed the analyses, with support of R.D. R.K. serves as the medical monitor at the Emmes company and provided guidance to the expert panel. D.S.-B., K.H., J.F., S.T., R.K., V.E.H. and R.D.S. are the expert panel members who adjudicated the outcomes. R.M.-N. guided on journal fit and contributed to revisions to the manuscript. S.G. and C.D. conducted literature searches and background review and contributed to the referencing system. All authors have read and agreed to the published version of the manuscript.

**Funding:** Research reported in this publication was supported by the National Institute on Drug Abuse of the National Institutes of Health under Award Numbers UG1DA013720, 2UG1DA049436, U54GM104942, UG1DA013035-09, 2UG1DA049436-01, 75N95019D00013/N01DA-19-2250 (NIDA DSC contract to The Emmes Company, LLC), and 75N95020D00012/N01DA-20-2251 (NIDA CCC

contract to The Emmes Company, LLC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

**Institutional Review Board Statement:** Each of the 8 multisite trials secured approval from their respective Institutional Review Board. However, the current study only used de-identified data and therefore was exempt from ethical review.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** CTN trial data is publicly available on NIDA Data Share: https: //datashare.nida.nih.gov/ (accessed on 15 February 2022). CTN 0064, CTN 0067 and 0068 data has been approved for upload onto NIDA Data Share and will be publicly available soon.

**Acknowledgments:** We would like to acknowledge the Principal Investigators from the 8 CTN trials: Madhukar Trivedi (CTN0037, CTN0068); Lisa Metsch (CTN0049, CTN0064), John Rotrosen (CTN0051); Edward Nunes (CTN0051); Kevin Gray (CTN0053); Walter Ling (CTN0054); Larissa Mooney (CTN0054); and Todd Korthuis (CTN0067) for their support for this study. We also want to acknowledge The Emmes company for provision of coordination, regulatory oversight and data managemen<sup>t</sup> and analysis for all the trials covered by this work. Finally, we would like to acknowledge the Center for the National Drug Abuse Treatment Clinical Trials Network, Betty Tai and her team for the support of the Comorbidity Special Interest Group from which this analysis was developed.

**Conflicts of Interest:** The authors declare no conflict of interest.
