Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 18.
Published in final edited form as: Suicide Life Threat Behav. 2018 Apr 25;49(3):635–642. doi: 10.1111/sltb.12468

Poisoning Severity Associated with a Range of Medications in Suicide Attempts by Ingestion

Kenneth R Conner 1, Timothy J Wiegand 1, Rachel Gorodetsky 1, Rachel F Schult 1, Kimberly Kaukeinen 1
PMCID: PMC8372239  NIHMSID: NIHMS1733139  PMID: 29693268

Abstract

There are limited data on the medical severity of suicide attempts by intentional self-poisoning (ISP) associated with ingestion of differing classes of medications and meager data on specific agents. The purpose of the study was to address these gaps. This was an analysis of a consecutive series of ISP cases ages 13 to 65 treated at a U.S. university medical center (N = 671). The outcome, poisoning severity, was dichotomized as “moderate-severe” and “low” (reference) based on a standard measure. Class of medication (e.g., opiate) and specific agents ingested were the predictors of interest. Covariates were age, sex, and the ingestion of multiple classes of medications. Data were analyzed using multivariate logistic regression models. At the class level, ingestion of opiate was uniquely associated with increased risk for moderate-severe ISP at a statistically significant level, adjusted odds ratio (95% CI) = 2.97 (1.69, 5.21), p = .0002. Several specific agents were also associated with moderate-severe ISP. Along with the key role of opiate medications in unintentional overdose morbidity and mortality, opiate medications may also play an important and largely unrecognized role in medically serious suicidal behavior. Results also underscore the variability in toxicity of specific agents within drug classes.

Worldwide, intentional self-poisoning (ISP) is the most common form of suicide attempt and the most common method that results in medical hospitalization (World Health Organization, 2014). ISP is a leading method of suicide including the third leading method in the United States and the second among women (Centers for Disease Control and Prevention, 2016). Individuals who come to medical attention for ISP are at marked risk for eventual suicide (Finkelstein et al., 2015).

Intentional self-poisoning in the United States and other Western, industrialized countries typically involves the ingestion of one or more pharmaceutical agents (Bhaksaran et al., 2015; World Health Organization, 2014). Pharmaceutical agents differ widely in their toxicity (Rhyee, Farrugia, Campleman, Wax, & Brent, 2015). This observation has been instrumental in the lowered use of some medications such as tricyclic antidepressants (TCAs) that are difficult to manage in overdose (Woolf et al., 2007) and have high potential for lethality (Shah, Uren, Baker, & Majeed, 2001). Lowered prescription rates for TCAs are reflected in decreased presence of TCAs in ISP events over time (Bhaksaran et al., 2015). There have also been selected suicide prevention efforts targeting specific agents, for example through more restrictive packaging of paracetamol (acetaminophen) in the United Kingdom (Hawton et al., 2013).

There are limited data on the medical severity of ISP events associated with a broad range of classes of medications that may inform additional suicide prevention and treatment efforts, and data are particularly limited when considering specific agents within classes. An exception comes from a study that estimated the potential for lethality of a range of medications at differing doses (Berman, Shepherd, & Silverman, 2003). However, the study was a review based on available reports from human and animal studies, not original data. U.S. National Violent Death Reporting System data have been used to estimate the percentage of suicide decedents with various drugs in their system based on postmortem toxicology tests, for example a report of suicides in 16 states reporting 24% of decedents tested positive for one or more antidepressants and 20% for opiates (Parks, Johnson, McDaniel, & Gladden, 2014). Such research is limited, however, because it provides no comparison of the relative toxicity of different drugs. Moreover, although a focus on lethality is important, the large majority of ISP events are survived and nonlethal poisoning events differ widely in their severity (Persson, Sjöberg, Haines, & Pronczuk de Garbino, 1998), yet there are few data comparing the severity of nonlethal ISP events in differing medications.

To address these limitations, the primary purpose of this study was to compare the acute medical severity of hospital treated ISP cases across a broad range of medication classes and, as data allowed, across specific agents. In considering confounding factors, ISP commonly involves co-ingestion of different classes of medications and/or co-ingestion with alcohol (Bhaksaran et al., 2015), potentially increasing lethality. The medical severity of ISP events is also increased with older age and in males (Oh, Kim, Kim, Kim, & Park, 2015). Therefore, analyses were adjusted for the ingestion of multiple classes of medications or alcohol, age, and sex.

METHODS

Procedure

The sample was a consecutive series of ISP patients ages 13 to 65 presenting to a large university medical center in the northeastern United States and treated by a multi-disciplinary consultation team between January 7, 2014 and July 28, 2016 (N = 673). The team is led by one of the authors (TJW) and provides toxicology consultation on ISP and a variety of other poisoning cases treated throughout the medical center including the emergency department, intensive care, and other settings. The team forms a site for the Toxicology Investigators Consortium (ToxIC) that maintains a large poisoning case registry (ToxIC Registry) which collects data on poisoning cases seen by medical toxicology specialists at its various sites (Rhyee et al., 2015). As they were treated, the cases herein were entered by the study team into a standard, de-identified, central database managed by the ToxIC Registry, and were subsequently downloaded by the study team for analysis. Therefore, the current analyses focus on local site data. The study was conducted with the approval of the local university’s human subjects’ review committee.

Variables

Poisoning Severity Score (PSS) served as the assessment of acute medical severity (Persson et al., 1998). PSS was determined by American Board of Applied Toxicology boarded PharmD clinical toxicologists who are permanent members of the consultation team, or by pharmacy students on a research internship who were trained by and under direct supervision of the clinical toxicologists. Evidence of validity of PSS include predicting mortality in poisoning patients presenting to hospital (Peter, Thomas, & Graham, 2013; Thanacoody et al., 2016), and it is correlated with other validated measures of acute medical morbidity (Akdur et al., 2010; Churi, Ramesh, Bhakta, & Chris, 2012; Peter et al., 2013). PSS ranges from 0 to 4: none (0), minor (1), moderate (2), severe (3), and fatal (4). For analyses, PSS was categorized as moderate–severe (scores 2–4) or low (scores 0–1). The label moderate–severe (without mention of “fatal”) for this study is sufficient because a small percentage of ISP events receiving hospital-based medical intervention are fatalities (Rhyee et al., 2015).

The medications ingested in ISP were determined based on clinical judgment using available information including patient interview, observations of family members and emergency medical technicians (e.g., empty pill bottles), and toxicology test results. The substances were categorized into classes of pharmaceutical agents (analgesic, etc.) as established for ToxIC. Substances within each class were further subcategorized into specific pharmaceutical agents (e.g., acetaminophen) per ToxIC. There were three covariates including age, sex, and co-ingestion, with age dichotomized as 19–65, 13–18 (reference), sex as male, female (reference), and co-ingestion as use of multiple drug classes, single class (reference). The determination of suicide attempt using ISP (vs. unintentional overdose or nonsuicidal self-injury) was a clinical judgment whether the patient had any intent to die, consistent with a consensus definition of suicidal behavior (Chappell et al., 2017).

Analyses

Unadjusted and adjusted logistic regression analyses (Agresti, 2002; Hosmer & Lemeshow, 2000) were used to regress poisoning severity, the outcome, on the presence or absence (reference) of ingestion of a series of categories of medications and specific agents, the variables of interest. Each multivariate model included the medication variable of interest and the three covariates. Unadjusted results were reported as odds ratio, OR (95% confidence interval, 95% CI), and adjusted results as adjusted odds ratio, AOR (95% CI). Confidence intervals that did not cross unity were considered statistically significant. Analyses were performed for all drug classes and for specific agents that were ingested by 15 or more ISP patients. Analyses used SAS (SAS Institute, Inc., Cary, NC).

In sensitivity analyses, we used continuous PSS scores and re-ran the analyses using linear regression models (Cohen, Cohen, West, & Aiken, 2003). We also compared the acute hospital treatments necessitated in moderate–severe versus low severity ISP cases to serve as a surrogate for the medical severity of ISP events based on the idea that more severe cases should require more intensive treatment. In these models, we regressed poisoning severity (low, reference) on the receipt of four common hospital-based treatments for acute poisoning (use of antidote, etc.), after adjusting for age, sex, and co-ingestion.

RESULTS

Descriptive Data and Covariates

There were 436 (65%) female patients and 237 (35%) males, and 218 (32%) aged 13–18 and 455 (68%) aged 19–65. Multiple classes of drugs were ingested by 330 (49%) patients and 343 (51%) ingested a single class. Two subjects were eliminated due to missing PSS data, yielding 671 patients analyzed including low severity cases (n = 390, 58%) and moderate–severe cases (n = 281, 42%). The latter category included four deaths.

In univariate models, each covariate was associated with greater poisoning severity at a statistically significant level: age 19–65 (vs. 13–18), OR (95% CI) = 2.36 (1.66, 3.33), p < .001; male (vs. female), OR (95% CI) = 1.48 (1.08, 2.04), p = .016; co-ingestion (vs. single ingestion), OR (95% CI) = 1.94 (1.42, 2.65), p < .001 (data not shown). The most common co-ingested classes of medications were antidepressant and sedative-hypnotic/muscle relaxant (n = 55, 8%), sedative-hypnotic/muscle relaxant and alcohol (n = 55, 8%), analgesic and anticholinergic/antihistamine (n = 44, 7%), and antidepressant and alcohol (n = 42, 6%) (data not shown).

Drug Classes

Results for drug classes are shown in Table 1. The most commonly ingested classes of medications were antidepressants and sedative-hypnotic/muscle relaxants. After adjustment for covariates, ingestion of opioid was uniquely associated with increased probability of moderate–severe poisoning. Ingestion of antipsychotic, sedative-hypnotic/muscle relaxant, and sympathomimetic was associated with higher odds of moderate–severe poisoning in unadjusted models, but not after adjustment. Ingestion of analgesic was associated with lower odds of moderate–severe poisoning in the unadjusted model, but not after statistical adjustment.

TABLE 1.

Drug Class and Poisoning Severity in ISP Cases (N = 671)

Drug class Low severity n (%) Mod–high severity n (%) Odds ratio (95% CI) Adjusted odds ratio (95% CI)
Alcohol 65 (17%) 56 (20%) 1.24 (0.84, 1.85) 0.66 (0.42, 1.04)
Analgesic 115 (29%) 61 (22%) 0.66 (0.46, 0.95) 0.70 (0.48, 1.02)
Anticholinergic/Antihistamine 77 (20%) 65 (23%) 1.22 (0.84, 1.78) 1.13 (0.76, 1.68)
Anticonvulsant 23 (6%) 23 (8%) 1.42 (0.78, 2.59) 1.11 (0.60, 2.06)
Antidepressant 115 (29%) 94 (33%) 1.20 (0.86, 1.67) 1.04 (0.73, 1.48)
Antipsychotic 42 (11%) 48 (17%) 1.71 (1.09, 2.67) 1.40 (0.89, 2.23)
Cardiovascular 38 (10%) 27 (10%) 0.99 (0.59, 1.66) 0.80 (0.46, 1.37)
Lithium 11 (3%) 9 (3%) 1.14 (0.47, 2.79) 0.80 (0.32, 2.00)
Opioid 20 (5%) 48 (17%) 3.81 (2.21, 6.58) 2.97 (1.69, 5.21)
Sedative-hypnotic/Muscle relaxant 90 (23%) 92 (33%) 1.62 (1.15, 2.29) 1.17 (0.81, 1.69)
Sympathomimetic 30 (8%) 36 (13%) 1.76 (1.06, 2.94) 1.35 (0.79, 2.32)
Open in a new tab

ISP, intentional self-poisoning. Column percentages shown. PSS, poisoning severity score. Low severity cases (n = 390) = PSS Scores 0, 1. Moderate–high severity cases (n = 281) = PSS Scores 2, 3, 4. Statistically significant p-values bolded. CI, confidence interval. Adjusted odds ratio, with adjustment for age, sex, and co-ingestion of two or more classes of drugs. Reference group = low severity cases in all analyses.

Specific Agents

Results for specific agents are also shown in Table 2. The two most commonly ingested medications were acetaminophen, an analgesic, and diphenhydramine, an anticholinergic/antihistamine. There were multiple statistically significant results in the adjusted analyses including greater odds of moderate–severe poisoning severity with ingestion of diphenhydramine, the antidepressant amitriptyline, the antipsychotic quetiapine, the cardiovascular medication clonidine, the opioid oxycodone, and the sedative-hypnotic/muscle relaxants cyclobenzaprine and zolpidem. Two agents were associated with lower probability of moderate–severe poisoning in adjusted models including the antidepressant fluoxetine and the sedative-hypnotic/muscle relaxant clonazepam.

TABLE 2.

Specific Agents and Poisoning Severity in ISP Cases (N = 671)

Agent (class) Low severity n (%) Mod-high severity n (%) Odds ratio (95% CI) Adjusted odds ratio (95% CI)
Acetaminophen (analgesic) 73 (19%) 41 (15%) 0.74 (0.49, 1.13) 0.79 (0.51, 1.22)
Ibuprofen (analgesic) 37 (9%) 14 (5%) 0.50 (0.27, 0.95) 0.59 (0.30, 1.15)
Diphenhydramine (antichol/antihist) 28 (7%) 38 (14%) 2.02 (1.21, 3.38) 2.09 (1.23, 3.55)
Hydroxyzine (antichol/antihist) 37 (9%) 19 (7%) 0.69 (0.39, 1.23) 0.57 (0.31, 1.04)
Lamotrigine (anticonvulsant) 6 (2%) 13 (5%) 3.10 (1.17, 8.27) 2.44 (0.90, 6.64)
Trazodone (antidepressant) 31 (8%) 20 (7%) 0.89 (0.50, 1.59) 0.67 (0.37, 1.22)
Bupropion (antidepressant) 18 (5%) 23 (8%) 1.84 (0.97, 3.48) 1.82 (0.94, 3.51)
Citalopram (antidepressant) 17 (4%) 13 (5%) 1.07 (0.51, 2.23) 1.16 (0.54, 2.49)
Sertraline (antidepressant) 15 (4%) 13 (5%) 1.21 (0.57, 2.59) 1.31 (0.60, 2.90)
Fluoxetine (antidepressant) 22 (6%) 4 (1%) 0.24 (0.08, 0.71) 0.23 (0.08, 0.69)
Amitriptyline (antidepressant) 2 (1%) 15 (5%) 10.93 (2.48, 48.13) 12.36 (2.74, 55.90)
Quetiapine (antipsychotic) 16 (4%) 30 (11%) 2.79 (1.49, 5.23) 2.54 (1.34, 4.82)
Aripiprazole (antipsychotic) 12 (3%) 6 (2%) 0.69 (0.26, 1.85) 0.55 (0.20, 1.54)
Clonidine (cardiovascular) 5 (1%) 14 (5%) 4.04 (1.44, 11.34) 4.59 (1.57, 13.40)
Oxycodone (opioid) 4 (1%) 15 (5%) 5.44 (1.79, 16.58) 4.21 (1.35, 13.11)
Hydrocodone (opioid) 5 (1%) 10 (4%) 4.21 (1.35, 13.11) 2.06 (0.68, 6.22)
Clonazepam (sedative-hypnotic) 35 (9%) 16 (6%) 0.61 (0.33, 1.13) 0.44 (0.23, 0.82)
Gabapentin (sedative-hypnotic) 15 (4%) 16 (6%) 1.51 (0.73, 3.11) 1.00 (0.48, 2.12)
Cyclobenzaprine (sedative-hypnotic) 9 (2%) 18 (6%) 2.90 (1.28, 6.55) 2.47 (1.07, 5.70)
Alprazolam (sedative-hypnotic) 11 (3%) 14 (5%) 1.81 (0.81, 4.04) 1.55 (0.68, 3.51)
Zolpidem (sedative-hypnotic) 5 (1%) 16 (6%) 4.65 (1.68, 12.85) 3.89 (1.38, 10.98)
Lorazepam (sedative-hypnotic) 10 (3%) 9 (3%) 1.26 (0.50, 3.14) 0.96 (0.38, 2.43)
Cocaine (sympathomimetic) 15 (4%) 19 (7%) 1.81 (0.91, 3.63) 1.10 (0.54, 2.27)
Open in a new tab

ISP, intentional self-poisoning. Column percentages shown. PSS, poisoning severity score. Low severity cases (n = 390) = PSS 0, 1. Moderate–high severity cases (n = 281) = PSS 2, 3, 4. Statistically significant p-values bolded. CI, confidence interval. Adjusted odds ratio with adjustment for age, sex, and co-ingestion of two or more classes of drugs. Reference group = Low severity cases in all analyses. Antichol/antihist = anticholinergic/antihistamine. Sedative-hypnotic = sedative-hypnotic/muscle relaxant.

Sensitivity Analyses

We obtained similar findings using continuous PSS results and linear regression models (results not shown). We also confirmed moderate–severe poisoning cases had a greater likelihood (estimated to be 3½ to eightfold, depending on the type of treatment) of receiving various acute medical treatments including antidote (e.g., physostigmine), pharmacological support (e.g., benzodiazepines), nonpharmacological support (e.g., intravenous fluid resuscitation), and decontamination (e.g., activated charcoal) (data not shown).

DISCUSSION

Antidepressants and sedative-hypnotic/muscle relaxants were the most commonly ingested classes of drugs in ISP events, and co-ingestion was commonplace, consistent with prior research of ISP (Bhaksaran et al., 2015). Opiates were the only class of medications to be associated with moderate–severe poisoning in adjusted analyses, with an estimate that ingestion of an opiate approximately triples the risk of moderate–severe ISP. The implication is that, along with the emphasis placed on the role of opiate medications in unintentional overdose morbidity and mortality (Rudd, Seth, David, & Scholl, 2016), opiates may also play an important but less recognized role in medically serious intentional overdose.

There were numerous statistically significant results when considering specific agents, suggesting the relative calm at the class level belies more dynamic patterns of the agents within. In antidepressants, fluoxetine, a selective serotonin reuptake inhibiter, was associated with lower risk for moderate–severe ISP, and amitriptyline, a TCA, was associated with pronounced risk for moderate–severe ISP, unsurprising in light of the known risks associated with TCAs (Shah et al., 2001; Woolf et al., 2007). In opiates, ingestion of oxycodone was associated with moderate–severe ISP, while hydrocodone did not show an association after statistical adjustment. The differing results may be attributable to oxycodone formulations potentially containing greater amounts of opioid, on a morphine-equivalent basis, than hydrocodone products. Hydrocodone products typically contain 5 to 10 mg hydrocodone with acetaminophen. This coformulation, while introducing another drug, potentially limits the overall amount of hydrocodone ingested. Oxycodone- and acetaminophen-containing products are also available in the same dosage ranges, along with products that contain significantly greater amounts of oxycodone. Oxycodone single ingredient immediate release formulations can contain up to 30 mg oxycodone, and an extended-release form comes in several different strengths with up to 80 mg oxycodone available.

In sedative-hypnotics/muscle relaxants, ingestion of clonazepam was associated with lower risk for moderate–severe ISP, whereas zolpidem was associated with higher risk. While both of these agents are sedatives with similar mechanisms involving the inhibitory neurotransmitter gamma aminobutyric acid (GABA), clonazepam is a long-acting benzodiazepine typically prescribed in scheduled doses for anxiety, characteristics that may lead to tolerance, while zolpidem is a short-acting sedative more commonly prescribed for insomnia on an as needed basis that, conceptually, has lower potential for the development of tolerance. Tolerance may reduce acute risk for overdose, for example as demonstrated in studies of opiate overdose (Tagliaro, De Battisti, Smith, & Marigo, 1998), potentially explaining a protective effect for clonazepam. This idea is preliminary as we do not have data to formally test such tolerance as a protective mechanism. Moreover, analysis of a large cohort of military Veterans prescribed an opiate analgesic did not show clonazepam was associated with lower risk of overdose deaths compared to other benzodiazepines (Park, Saitz, Ganoczy, Ilgen, & Bohnert, 2015), underscoring the need for further analysis of the association between use of clonazepam and overdose mortality and morbidity.

Several drugs that may produce anticholinergic effects in overdose were associated with greater likelihood of moderate–severe ISP in the current analyses. These drugs include diphenhydramine, amitriptyline, cyclobenzaprine, and quetiapine. With the exception of diphenhydramine, ToxIC lists these drugs under different categories rather than as an anticholinergic agent per se; for example, quetiapine is listed as an antipsychotic, consistent with its primary use. However, in our experience, the anticholinergic effects of these medications in overdose including tachycardia, sedation, coma, and delirium are main driver of their toxicity.

Clonidine, a cardiovascular medication, is used to treat high blood pressure and may be used adjunctively for attention deficit hyperactivity disorder, along with off-label uses for anxiety and other symptoms. In the current analyses, clonidine was associated with higher risk for moderate–severe ISP. Explanations for the result include sedative and hemodynamic effects of clonidine overdose. Moreover, large ingestions can result in coma and bradycardia with hypotension due to its effect on presynaptic alpha-2 receptors which results in decreased outflow of norepinephrine and subsequent sympatholytic effect. Prior research of overdose events broadly (including unintentional, undetermined, and intentional cases) has reported increased risk of sedation and bradycardia with clonidine ingestion but not more severe poisonings per se (Isbister, Heppell, Page, & Ryan, 2017), indicating the need for further study.

LIMITATIONS

There were limitations of the study. The database did not contain a validated assessment of suicide attempt (e.g., Posner, Oquendo, Gould, Stanley, & Davies, 2007). The study was a secondary analysis of cross-sectional data with multiple tests run, increasing the probability of chance findings, and power was limited by small numbers of observations for some medications. More refined data on dosage (e.g., pill counts) of the various medications ingestion were unavailable. The results were based on ISP patients who ingested a pharmaceutical agent, with unclear generalizability to other population. We did not have reliability data on PSS, for example by comparing two independent raters. PSS has not typically been dichotomized, although sensitivity analyses using continuous scores were consistent with primary analyses.

CONCLUSION

The current results indicate moderate–severe ISP cases were more likely to require a range of acute interventions (antidote, etc.) during hospitalization compared to ISP cases of low severity, illustrating the importance of research on moderate–severe ISP events. The primary finding of the study suggests that ingestion of opiates is associated with increased risk for ISP. The result indicates the need for caution in prescribing opiates to patients known to be at risk for suicidal behavior (e.g., based on a prior history of attempt) or suspected of being at increased risk (e.g., based on depressive symptoms or other indicators). Several specific agents were also associated with greater ISP severity, and we have offered preliminary explanations for these results based on current knowledge.

Acknowledgments

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. An earlier version was presented in May 2017 at the annual meeting of the European Association of Poisons Centres and Clinical Toxicologists (EAPCCT), Basel, Switzerland.

REFERENCES

  1. Agresti A (2002). Categorical data analysis. Hoboken, NJ: Wiley & Sons. [Google Scholar]
  2. Akdur O, Durukan P, Ozkan S, Avsarogullari L, Vardar A, Kavalci C, et al. (2010). Poisoning severity score, Glasgow coma scale, corrected QT interval in acute organophosphate poisoning. Human and Experimental Toxicolology, 29, 419–425. [DOI] [PubMed] [Google Scholar]
  3. Berman AL, Shepherd G, & Silverman MM (2003). The LSARS-II: Lethality of suicide attempt rating scale – Updated. Suicide and Life-Threatening Behavior, 33, 261–276. [DOI] [PubMed] [Google Scholar]
  4. Bhaksaran J, Johnson E, Bolton JM, Randall JR, Mota N, Katz C, et al. (2015). Population trends in substances used in deliberate self-poisoning leading to intensive care unit admissions from 2000 to 2010. Journal of Clinical Psychiatry, 76, e1583–e1589. [DOI] [PubMed] [Google Scholar]
  5. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control. (2016). Web-based injury statistics query and reporting system (WISQARS). RetrievedApril 10, 2016, from http://www.cdc.gov/injury/wisqars.
  6. Chappell PB, Stewart M, Alphs L, DiCesare F, DuBrava S, Harkavy-Friedman J, et al. (2017). Assessment of suicidal ideation and behavior: Report of the International Society for CNS Clinical Trials and Methodology Consensus Meeting. Journal of Clinical Psychiatry, 78, e638–e647. 10.4088/jcp.16cs11417. Downloaded 7/15/2017. [DOI] [PubMed] [Google Scholar]
  7. Churi S, Ramesh M, Bhakta K, & Chris J (2012). Prospective assessment of patterns, severity and clinical outcome of Indian poisoning incidents. Chemical & Pharmaceutical Bulletin (Tokyo), 60, 859–864. [DOI] [PubMed] [Google Scholar]
  8. Cohen J, Cohen P, West SG, & Aiken LS (2003). Applied multiple regression/correlation analysis for the behavioral sciences, 3rd ed. Hillsdale, NJ: Lawrence Erlbaum Associates. [Google Scholar]
  9. Finkelstein Y, Macdonald EM, Hollands S, Sivilotti ML, Hutson JR, Mamdani MM, et al. (2015). Risk of suicide following deliberate self-poisoning. JAMA Psychiatry, 72, 570–575. [DOI] [PubMed] [Google Scholar]
  10. Hawton K, Bergen H, Simkin S, Dodd S, Pocock P, Bernal W, et al. (2013). Long term effect of reduced pack sizes of paracetamol on poisoning deaths and liver transplant activity in England and Wales: Interrupted time series analyses. BMJ, 346, f403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hosmer DW, & Lemeshow S (2000). Applied logistic regression. New York, NY: Wiley & Sons. [Google Scholar]
  12. Isbister GK, Heppell SP, Page CB, & Ryan NM (2017). Adult clonidine overdose: Prolonged bradycardia and central nervous system depression, but not severe toxicity. Clinical Toxicology, 55, 187–192. [DOI] [PubMed] [Google Scholar]
  13. Oh SH, Kim HJ, Kim SH, Kim YM, & Park KN (2015). Which deliberate self-poisoning patients are most likely to make high-lethality suicide attempts? International Journal of Mental Health Systems, 9, 35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Park TW, Saitz R, Ganoczy D, Ilgen MA, & Bohnert AS (2015). Benzodiazepine prescribing patterns and deaths from drug overdose among US veterans receiving opioid analgesics: Case-cohort study. BMJ, 350, h2698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Parks SE, Johnson LL, McDaniel DD, & Gladden M (2014). Surveillance for violent deaths – National Violent Death Reporting System, 16 states, 2010. MMWR, 63(ss01), 1–33. Available at: http://www.cdc.gov/mmwr/preview/mmwrhtml/ss6301a1.htm. Downloaded 6/11/2016. [PubMed] [Google Scholar]
  16. Persson HE, Sjöberg GK, Haines JA, & Pronczuk de Garbino J (1998). Poisoning severity score. Grading of acute poisoning. Journal of Toxicology: Clinical Toxicology, 36, 205–213. [DOI] [PubMed] [Google Scholar]
  17. Peter JV, Thomas L, & Graham PL (2013). Performance of clinical scoring systems in acute organophosphate poisoning. Clinical Toxicology, 51, 850–854. [DOI] [PubMed] [Google Scholar]
  18. Posner K, Oquendo MA, Gould M, Stanley B, & Davies M (2007). Algorithm of Suicide Assessment (C-CASA): Classification of suicidal events in the FDA’s pediatric suicidal risk analysis of antidepressants. American Journal of Psychiatry, 164, 1035–1043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Rhyee SH, Farrugia L, Campleman SL, Wax PM, & Brent J; Toxicology Investigators Consortium (2015). The Toxicology Investigators Consortium Case Registry – The 2014 experience. Journal of Medical Toxicology, 11, 388–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Rudd RA, Seth P, David F, & Scholl L (2016). Increases in drug and opioid-involved overdose deaths – United States, 2010–2015. MMWR, 65, 1445–1452. 10.15585/mmwr.mm655051e1. [DOI] [PubMed] [Google Scholar]
  21. Shah R, Uren Z, Baker A, & Majeed A (2001). Deaths from antidepressants in England and Wales 1993–1997: Analysis of a new national database. Psychological Medicine, 31, 1203–1210. [DOI] [PubMed] [Google Scholar]
  22. Tagliaro F, De Battisti Z, Smith FP, & Marigo M (1998). Death from heroin overdose: Findings from hair analysis. Lancet, 351, 1923–1925. [DOI] [PubMed] [Google Scholar]
  23. Thanacoody RH, Gilfillan C, Bradberry SM, Davies J, Jackson G, Vale AJ, et al. (2016). Management of poisoning with ethylene glycol and methanol in the UK: A prospective study conducted by the National Poisons Information Service (NPIS). Clinical Toxicology, 54, 134–140. [DOI] [PubMed] [Google Scholar]
  24. Woolf AD, Erdman AR, Nelson LS, Caravati EM, Cobaugh DJ, Booze LL, et al. (2007). Tricyclic antidepressant poisoning: An evidence-based consensus guideline for outof-hospital management. Clinical Toxicology, 45, 203–233. [DOI] [PubMed] [Google Scholar]
  25. World Health Organization (2014). Preventing suicide: A global imperative. Geneva, Switzerland: World Health Organization. [Google Scholar]

RESOURCES