Suicide is considered one of the increasing public health problems. According to the World Health Organization (WHO), more than 700,000 people die by suicide each year. Even more, it is considered the fourth leading cause of death among adolescents aged 15 to 19 years. In 2019, when suicidal cases were analyzed based on the socioeconomic status of different countries, 77% of the total reported suicide cases were among low and middle-income countries. Besides, ingestion of pesticides, hanging, and firearms were the most commonly reported methods. ¹

The incidence of suicide is believed to be the highest among adolescents and increases more among those experiencing stressors of abuse, conflicts, and bullying. Moreover, even though women tend to have more suicidal attempts, the incidence of men dying by suicide is higher due to their tendency to choose more lethal methods and higher lethality acts, which explains why women are more likely to survive each attempt. ^{2 , 3} Regarding the methods of suicide, hanging is the most common method used among women, while firearms are most commonly used by men, with hanging being the most fatal method among both. ^{2 , 3} Concerning the profession, a higher risk of suicide is found to be among physicians, particularly those who are female. ⁴

Insulin is a polypeptide hormone that is secreted by β cells of the pancreas and regulated by glucose level in the blood, some amino acids, autonomic mediators, and other hormones. Artificial insulin is an exogenous insulin given to all patients with T1DM, and some patients with T2DM. Insulin can be divided into rapid-acting (lispro, lispro-aabc, glulisine, aspart, and faster aspart), short-acting (regular insulin), intermediate-acting (neutral protamine hagedorn (NPH)), long-acting (glargine, detemir, and degludec), and premixed insulin. ^{5 , 6} Subcutaneous injection is the most commonly used route of administration of insulin, which can be delivered via various methods, including, vials, syringes, insulin pens, and insulin pumps. ⁷ Since there is no safe medication, artificial insulin carries its side effects, and hypoglycemia is the most common one. ⁶

Few cases of insulin overdose are reported every year, which can be accidental or intentional for the purpose of dying by suicide. Surprisingly, these reported cases include not only individuals with diabetes of both types but also those who do not have diabetes. People with massive insulin overdose might develop neurological complications, such as hypoglycemic coma, which can be reversible or result in death. ^{8 , 9}

Although there are no clear studies stating that diabetes mellitus is a known risk factor for suicide, it is found that individuals with T1DM have 11 times higher risk of attempting suicide than people who do not have diabetes. Moreover, other studies conclude that suicidal ideation in people with diabetes mellitus was associated with these individuals not taking their medication. ¹⁰

As insulin can be used as a method to attempt suicide, we aim in this paper to analyze research published on attempting or dying by suicide with insulin. The review focuses on six dimensions, which are the demographic data of suicidal cases, the relation of psychiatric illness to dying by and attempting suicide using insulin, route of insulin administration and outcome, post-mortem findings in suicide by insulin and collaborative evidence, initial presentation, blood glucose level and complications in attempting suicide cases.

The literature search identified a total of 425 records from PubMed (n = 216), Web of Science (n = 193), and Scopus (n = 16). After removing 196 duplicate records, an initial screening of 229 records was done and led to the exclusion of 187 articles. Further, a full-text review excluded 31 articles due to the following reasons: non-English articles (n = 29) (Derobert L et al., ¹² Domart A et al., ¹³ Lenhardt A, ¹⁴ Richter R, ¹⁵ Szabo Z et al., ¹⁶ Chlud K, ¹⁷ Sachsse B, ¹⁸ Bourgeois M et al., ¹⁹ Vincent V et al., ²⁰ Creutzfeldt W et al., ²¹ Euler-Rolle J et al., ²² Stofer A et al., ²³ Tourniaire J et al., ²⁴ Holldorf L et al., ²⁵ Marigo S et al., ²⁶ Bourgeois M, ²⁷ Stefan J, ²⁸ Zilker T et al., ²⁹ Schneider V et al., ³⁰ Hambrecht M et al., ³¹ Logemann E et al., ³² von Albert H et al., ³³ Szponar J et al., ³⁴ Wehner F et al., ³⁵ Nikolić S et al., ³⁶ Klimaszyk D et al., ³⁷ Uesugi K, ³⁸ Chen L et al., ³⁹ Tong F et al., ⁴⁰ and articles that could not be retrieved (n = 2) (Lionte C et al., ⁴¹ Arbouche N et al. ⁴² ). From the primary search of the databases 11 records met the inclusion criteria. Also, an additional 23 records were identified by screening the references of the included studies from the primary search. Eventually, a total of 34 articles met the inclusion criteria and were considered in the present review. ^{8 , 9 , 43 – 73} A PRISMA flow chart for the literature search is shown in Figure 1.

The analyzed studies were published in a time period between 1934 and 2020. Most studies were case reports (n = 22), while few were case series (n = 3), retrospective cohort (n = 2), prospective study (n = 1), hospital based (n = 2), population based (n = 1), case-control (n = 1), case conference (n = 1) and cross-sectional (n = 1) studies. A summary of the important results is presented in Table 1.

Diabetes and psychiatric illnesses can co-exist and have a significant impact on people’s lives. Studies showed that there is a significant association between diabetes and alcohol use disorder, substance use, mood disorders, anxiety disorders, and psychotic disorders, which puts them at an increased risk of suicide. ^{74 , 75} This goes in harmony with this systematic review where 67 out of 179 (37.43%) reported cases were diagnosed with mental illnesses. Furthermore, insulin overdose, alongside its congeners, is not an uncommon method of suicide attempts among individuals with diabetes. ⁷⁶ However, publications regarding the incidence of insulin suicide are scarce. ¹⁰ In the current existing data, those with T1DM showed higher incidence of insulin suicide (n = 102) than in T2DM (n = 21), which represents 63.35% and 13.04%, respectively. This could be explained by two possible reasons. First, T1DM is managed mainly by insulin, which makes it readily accessible to them. ⁷⁶ Second, generally, the incidence of dying by suicide is considered to be higher among newly diagnosed youth with T1DM. The suicidal ideation and attempts are believed to be related to symptomology and initial psychiatric status shortly after being diagnosed. ⁷⁷

Although multiple professions are prone to die by suicide, the medical field is considered a high-risk profession. Multiple studies reported a higher incidence of anxiety and depression among physicians. It may be attributed to the stressful work environment and the emotional burden when facing illness, breaking bad news, and death. ^{4 , 78 , 79} Up to this point, self-poisoning is the most common method of suicide among healthcare providers. ⁸⁰ In addition to that, the thorough medical knowledge of the rapid lethal insulin-induced hypoglycemia can make the available insulin in hospitals a feasible weapon for suicide. ⁸¹ This systematic review showed that 9.93% (n = 16) of victims were healthcare providers and, as anticipated, most of them were suffering from mental illnesses.

In reality, many individuals who are chronically ill, including those with diabetes, attempt suicide using a combination of different prescribed medications. ⁸² A very recent study found that a benzodiazepine overdose can be linked to hypoglycemia. ⁸³ Surprisingly, we found that the most common medication used in combination with insulin is a benzodiazepine, which can be explained by the fact that a lower threshold of insulin will be needed to result in death. In addition to benzodiazepines, the consumption of alcohol was also associated with an increased risk of hypoglycemia the day after the intake. This is most likely due to the inhibitory properties of alcohol to hepatic gluconeogenesis. ^{84 , 85} This systematic review reported eight attempted suicides by individuals with diabetes who were also suffering from alcohol use disorder. This suggests a higher risk for those individuals to develop life-threatening hypoglycemic episodes, even though alcohol was not consumed on the same day of the event.

Despite the fact that hypoglycemia symptoms are usually typical, almost all of them are nonspecific. For this, the only evidence to know that the individual is suffering from hypoglycemia is to measure blood glucose level. Studies have shown that the threshold of developing hypoglycemia symptoms is when plasma glucose level falls below 80 mg/dL, while cognitive dysfunction mainly develops at 45 mg/dL. Moreover, the most commonly reported symptoms were mild, which include a combination of diplopia, diaphoresis, palpitations, and blurred vision. Furthermore, more severe symptoms were also reported in 65% of the cases, which are unconsciousness or amnesia, and generalized convulsions. ⁸⁶ In this systematic review, most of the findings of the reported cases are consistent with the previous studies, excluding a single case who had a glucose level of 25.2 mg/dL without any symptoms except for diaphoresis. For this, it is believed that signs and symptoms of hypoglycemia can vary greatly among people despite the blood glucose level readings.

The absence of autopsy and histology findings makes the diagnosis of hypoglycemia very challenging. ⁸⁷ Although biochemistry analysis is the only current way to estimate hypoglycemia, post-mortem blood glucose levels have no diagnostic value in determining hypoglycemia. After death, the persisting cells still have the ability to metabolize blood glucose, which results in a rapid variable decrease in blood glucose levels, making the diagnosis of insulin-induced hypoglycemia difficult. ⁸⁸ Thus, in order to estimate ante-mortem glucose levels, according to Traube’s formula, a measurement of post-mortem glucose and lactate in both cerebrospinal fluid (CSF) and vitreous humor, and also hemoglobin A1c should be obtained. All these measurements are combined in a score, where a score of less than 50 is almost diagnostic of hypoglycemia. ⁵⁸ However, some studies suggested that the combined measurements of glucose and lactate in vitreous humor and CSF have no value in estimating blood glucose concentration before death. ⁸⁸ Yet, this cannot be evaluated by this systematic review since no data was reported in this regard. We found that the mean glucose levels in vitreous humor is 22.37, while 5.9 in CSF. These findings were limited to the few studies that had performed the biochemistry analysis in vitreous humor and CSF. Even though post-mortem analysis of insulin is challenging, fortunately, a new promising validated quantitative method has been introduced. It quantifies human insulin and its various analogues in samples obtained from post-mortem blood and tissues (kidney, liver, and skeletal muscles) with an accuracy reaching 70% up to 130%, using liquid chromatography-mass spectrometry with high-resolution mass spectrometry (LC-MS/HRMS) analysis. This method provides quantitative measures of both therapeutic and toxic levels of insulin. It was tested on several forensic cases, and it identified the insulin prescribed to the victims even though many types of insulin were undetermined before death. However, this assay has five steps for sample preparation (tissue sample size reduction, homogenization, extraction, concentration, and immunopurification), which is a very time-consuming process. ^{89 , 90} Moreover, insulin is considered a large molecular weight substance, which makes it challenging to reach certainty of post-mortem insulin toxicity. For this, routine determination of insulin in post-mortem specimens is not offered frequently. ⁹¹ This can explain the variation among the cases included in this systematic review where the results of insulin levels were ranging from negative to 1,848.8 mU/ml.

Our review was subject to several limitations. First, evaluation of the accuracy of post-mortem glucose measurements could not be done since the included articles were mainly focusing on the characteristics of victims, interventions, and complications rather than detection of insulin overdose. Second, as post-mortem insulin analysis is a new method, data in case reports were limited in this regard. Thus, it is recommended that further research and assessment of the accuracy and feasibility of postmortem insulin analysis as it is a new promising method to detect insulin in suicide cases.

Insulin overdose as a method of dying by suicide is not uncommon among those with diabetes, however, it can lead to unfavorable neurological sequelae and death. As the prevalence of diabetes mellitus is increasing worldwide and insulin is one of the medications used in the management, it is expected that self-poisoning with insulin will be also increased in both individuals with diabetes and those without diabetes. It is crucial that physicians suspect insulin self-poisoning as one of the differential diagnoses when an individual, with diabetes or without, presents with a coma or unexplained hypoglycemia especially if it is prolonged or recurrent. Evaluation and early management of psychiatric illnesses especially among those with diabetes, mainly depression, are some of the important measures that may prevent suicidality. In addition, easy access to insulin in those without diabetes in the case of individuals with relatives or friends who have diabetes, or among healthcare providers should be considered. Since post-mortem evaluation of insulin toxicity victims is challenging, attention needs to be given to developing a more accurate approach to establish and confirm the diagnosis of insulin poisoning. At the end, managing an individual who presents with intentional insulin poisoning should include a multidisciplinary team and be treated as a whole, from all aspects rather than managing only the individual’s presentation.