The subjects were obtained from the Lieber Institute for Brain Development (LIBD) brain donation program. Postmortem brain samples were donated to the LIBD from the Office of the Medical Examiner of the Western Michigan University Homer Stryker M.D. School of Medicine, Department of Pathology (WIRB protocol #1126332), with the informed consent of legal next-of-kin at the time of autopsy. An important feature of this brain donation program is the inclusion of a substantial percentage of people who completed suicide. Potential donors are identified each morning by reviewing information received from death scene investigations in the Western Michigan catchment area (13 counties in western Michigan) that took place in the preceding 24 h. Decedents who require forensic autopsy, as determined by a board-certified forensic pathologist, are screened for exclusionary criteria. The exclusionary criteria include: not-at-fault motor vehicle accidents, homicides, anything suspicious otherwise (e.g., elder abuse, child abuse, deaths in custody), specific infectious illnesses (hepatitis, human immunodeficiency virus, meningitis, sepsis, prion disease), structural brain damage (ischemic strokes, brain tumors, brain surgery history, cerebral palsy, or past severe traumatic brain injury), encephalopathy, seizure disorder without a psychiatric history, advanced life support for more than 24 h, last known to be alive more than 72 h prior, and evidence of decomposition.

Next-of-kin of decedents determined acceptable for referral are then contacted by LIBD to request consent for donation. One study documented that more than half (57.0%) of all African-American families and more than two-thirds (74.1%) of all Caucasian families who were contacted by the LIBD at the time of autopsy for consent to donate their deceased loved one's brain for neuropsychiatric research agreed to do so (34). Donor history is documented via medical records, mental health records, and telephone interviews with the next-of-kin (described below). Our subjects were all men aged 50 or greater because the purpose of the study was to examine the association between playing amateur football and death by suicide in older adult men.

All cases were reviewed in a multidisciplinary conference prior to autopsy and/or brain procurement. Manner of death was determined by a board-certified forensic pathologist in each case, according to guidelines from the National Association of Medical Examiners (35). The manner of death is an opinion based on an evaluation of all relevant information, including death scene investigation, information from witnesses, family members, and acquaintances, a review of relevant medical records, and a postmortem examination conducted by the forensic pathologist. Deaths are classified as suicide when an intentional self-inflicted injury or poisoning results in death. The manner of death and cause of death for the decedents is reported in Table 1.

At the time of donation, a 36-item next-of-kin informant telephone interview was conducted to obtain medical, social, demographic, family, and psychiatric history. During the phone interview, next-of-kin were asked whether the decedent had a history of head trauma. This was coded as yes or no and it was accepted as yes if there was a personal history of concussions in sports, head injury in daily life, or what clearly sounded like a severe traumatic brain injury. The family members described a broad range of mechanisms of injury (e.g., sports, car accidents, and falls from a height) and severity injury (e.g., concussion, skull fractures, subdural hematomas, and coma). It is assumed that all levels of brain injury severity, mild, moderate, and severe, are included in this binary classification, but for most cases there was insufficient information to classify brain injury severity. For most subjects, detailed notes on this variable were not available and it was coded as binary. Next-of-kin were also asked whether the decedent had a history of engagement in sports with a high risk of repetitive brain injury, including boxing, football, mixed martial arts, lacrosse, rugby, or auto racing, and this variable was coded as yes or no. We categorized men into groups based on whether they played football, other contact sports, or no contact sports.

A retrospective clinical diagnostic review was conducted on every brain donor, consisting of the telephone screening, autopsy, and forensic investigative data, forensic toxicology data, extensive psychiatric treatment, substance abuse treatment, and/or medical record reviews, and whenever possible, family informant interviews. All data were compiled into a comprehensive psychiatric narrative summary that was reviewed by two board-certified psychiatrists in order to arrive at lifetime DSM-5 psychiatric diagnoses and medical diagnoses. A subgroup of decedents were recruited who were free from psychiatric and substance use diagnoses, and their toxicological data was negative for drugs of abuse. Every brain donor had forensic toxicological analysis, which typically covered ethanol and volatiles, opiates, cocaine/metabolites, amphetamines, and benzodiazepines. Some donors also received supplemental directed toxicological analysis using National Medical Services, Inc., including nicotine/cotinine testing, cannabis testing, and the expanded forensic panel in postmortem blood in order to cover any substances not tested.

The proportions of men with a manner of death of suicide were computed and compared using χ² tests. To characterize the magnitude of group differences, an odds ratio (OR) was calculated for each analysis as an effect size (36) and interpreted according to widely used criteria (37), i.e., ORs between 1.20 and 1.71 = small, ORs between 1.72 and 2.40 = medium, and ORs >2.40 = large. Independent t-tests were used to compare groups based on age. Exploratory binary logistic regressions were used to model independent contributions of more than one variable. All statistical analyses were conducted using IBM SPSS Statistics 26.

The proportions with suicide as a manner of death, by subgroup, are presented in Figure 1. There was no significant difference in the proportions of suicide between those who played football (25.4%) vs. those who did not (38.8%), or between those who played football and those who did not play any sports (38.8%) (see Table 2). There was no significant difference in the proportions of suicide between those who played contact sports (27.5%) vs. those who did not play contact sports (38.8%). There was also no significant difference in the proportions of suicide between those who had a history of TBI (37.7%) vs. those who did not (36.3%).

The proportions of men with suicide as their manner of death were compared based on the presence or absence of specific clinical characteristics. Those results are provided in Table 2. Those with the following clinical characteristics were significantly more likely to have suicide as their manner of death: (i) history of depression or bipolar disorder, (ii) history of substance use disorder, (iii) history of suicide ideation, (iv) history of suicide attempts, (v) family history of suicide.

The proportions with suicide attempts, by subgroup, are presented in Figure 2. There was a statistically significant difference in the proportions with suicide attempts between those who played football (14.8%) vs. those who did not (34.1%), and between those who played football and those who did not play any sports (33.3%). In both cases, those who played football were less likely to have a lifetime history of a suicide attempt. There was also a significant difference in the proportions with suicide attempts between those who played contact sports (24.0%) vs. those who did not play contact sports (33.3%); those who played contact sports were less likely to have a lifetime history of a suicide attempt. Those with a history of TBI were significantly more likely to have a suicide attempt (44.9%) than those with no reported history of TBI (23.7%).

The proportions of men with a history of suicide attempts were compared based on the presence or absence of specific clinical characteristics. Those results are provided in Table 2. Those with the following clinical characteristics were significantly more likely to have a history of suicide attempts: (i) history of depression or bipolar disorder, (ii) history of suicide ideation, and (iii) family history of suicide.

Exploratory analyses were undertaken to examine further the association between a personal history of TBI and suicidality. As noted above, those with a history of TBI also had a greater history of suicide attempts. There was also a significant difference in the proportions with suicide ideation between those who had a history of TBI vs. those who did not [χ(1, 159)2 = 5.135, p = 0.023, OR = 2.213, 95% CI = 1.106–4.428, medium effect size]. There was not a significant difference in the proportions with documented depression or bipolar disorder between those who had a history of TBI vs. those who did not [χ(1, 171)2 = 1.186, p = 0.276, OR = 1.462, 95% CI = 0.737–2.903, large effect size]. There was a significant difference in the proportions with documented substance abuse disorders between those who had a history of TBI vs. those who did not [χ(1, 173)2 = 4.591, p = 0.032, OR = 2.106, 95% CI = 1.059–4.192, medium effect size].

Binary logistic regression was used to examine simultaneously TBI history and substance abuse history as predictors of suicidality. For suicide attempts, neither variable significantly contributed to the model when both were considered simultaneously [TBI B = 0.710, SE = 0.373, Wald = 3.621, p = 0.057, Exp (B) = 2.034, 95% CI = 0.979–4.226; substance use disorder B = 0.640, SE = 0.363, Wald = 3.106, p = 0.078, Exp (B) = 1.896, 95% CI = 0.031–3.862]. For suicide ideation, substance use disorder, but not TBI, contributed to the model when both were considered simultaneously [TBI B = 0.522, SE = 0.396, Wald = 1.742, p = 0.187, Exp (B) = 1.686, 95% CI = 0.776–3.663; substance use disorder B = 1.743, SE = 0.391, Wald = 19.883, p < 0.001, Exp (B) = 5.715, 95% CI = 2.656–12.297].

The results of this study are consistent with six prior studies relating to participation in high school football, depression, suicidality, and suicide. Three large-scale studies, using the same longitudinal database, have found that men who played high school football, compared to those who did not play sports, did not have greater rates of depression or suicidality in their early or late 20s (20–22). A survey study of middle-aged men who played high school football reported that they did not have higher lifetime rates of being prescribed medication for anxiety or depression or receiving treatment from a mental health professional, and they did not endorse greater depression, anxiety, or anger over the past year (39). A large-scale longitudinal study of older adult men, with an average age of 65, revealed no association between playing football in high school and current problems with depression (23).

Very little is known about suicide in middle-aged or older adult men who played high school football. A recently published study (40) provides information relating to the long-term mental health of former amateur football players in the online Supplementary Material. The study was focused on CTE neuropathologic change, but the unpublished online Supplementary Material provided extensive information about the mental health of these men. These researchers used autopsy cases from the Mayo Clinic Tissue Registry. They relied on primary historical obituary and yearbook records to identify and compare former athletes to people who did not participate in sports, and a medical records-linkage system to examine billing codes relating to various psychiatric and neurological diagnoses. Using the frequency data presented in their online Supplementary Table 3, we statistically compared the 140 men who played football to the 245 men who did not play sports. The former football players had significantly lower rates of depression [25.0% vs. 37.1%, χ(1)2 = 5.97, p = 0.015, odds ratio (OR) = 1.77, 95% confidence interval (CI) = 1.12–2.81], and there was no statistically significant difference in their rate of suicide [3.6% in those who played football vs. 8.2% in those who did not play sports; χ(1)2 = 3.094, p = 0.079].

It is important to note that there are large scale studies illustrating that former NFL players are at lower risk for suicide than men from the general population (12, 13). Lehman and colleagues (13) examined suicide mortality in a cohort of 3,439 NFL players who played between 1960 and 2013 and identified 12 who had died by suicide (0.3% of the cohort). A total of 537 had died during the observation period, so 2.2% had suicide as their manner of death. The expected number of deaths by suicide from men in the general population was more than double the number in former players (i.e., 25.6), resulting in a standardized mortality ratio of 0.47 (95% CI = 0.24–0.82) (13). Relatedly, there was no significant difference in the rate of suicide as a manner of death or the age at which a person completed suicide between former professional soccer players and their matched controls (41). In that study, suicide was recorded as the manner of death in 19 (0.25%) of 7,676 former professional soccer players and 93 (0.40%) of 23,028 population control subjects. Of those who had died during the 18-year observation period, the rate of death by suicide was 1.6% (19/1,180 deaths) for the former soccer players and 2.4% (93/3,807 deaths) for the matched control subjects.

Suicide is a leading cause of death in adolescents and young adults, and men are three times more likely to complete suicide than women (27). In 2016, the incidence of suicide in men between the ages of 15 and 19 was estimated to be 14.8/100,000, and for those between 20 and 24 it was 26.8/100,000. For men between the ages of 55 and 64, the incidence was 28.5/100,000, and for those between the ages of 75 and 84 it was 33.6/100,000 (27). The causes of suicide are usually complex, multifactorial, and difficult to predict or understand in individual cases.

Suicidality is a clinical feature of depression. Rates of suicidality and suicide also are higher in people with anxiety disorders (42). People who experience major depressive disorders are at considerable risk for attempting suicide (24), and people with psychotic or delusional depression are at even greater risk for attempting suicide (43). For people who are experiencing depression, some risk factors for completing suicide include male gender, family history of a psychiatric disorder, previous attempted suicide, more severe depression, hopelessness, comorbid anxiety, and misuse of alcohol or drugs (44). People who misuse alcohol (28), drugs (29), benzodiazepines (45), and opioids (46, 47) are at increased risk for suicidal ideation. Psychological autopsy studies indicate that drug use disorders are strong proximate risk factors for completed suicide (30).

In adults, all different types of childhood maltreatment including sexual abuse, physical abuse, and emotional abuse are associated with increased risk for depression and suicidality (48, 49), and there is a statistically significant association between lifetime history of sexual abuse and a lifetime diagnosis of anxiety disorder, depression, post-traumatic stress disorder, sleep disorders, and suicide attempts (50). In older adults, suicidal behavior is associated with functional disability and specific medical problems such as malignant diseases, neurological disorders, pain, arthritis, chronic obstructive pulmonary disease, and liver disease (51). In older adults, moderate to severe pain is associated with increased risk for suicidal ideation and attempts; arthritis, back/neck problems, and headaches are associated with higher risks of suicidal behavior; and pain is a much stronger predictor for suicide in men than in women (52). Alzheimer's disease also is associated with a risk of suicide (53).

A CDC report examining suicides identified common precipitating events as depression, substance abuse, a life crisis, intimate partner problems, physical health problems, occupational or financial problems, and criminal or legal proceedings (27). These circumstances, in a sample of more than 18,000 men who completed suicide, are summarized in Table 3. High risk patterns of thinking, or cognitive states, associated with suicidality include hopelessness (54, 55), problem-solving deficits, such as difficulty conceptualizing adaptive solutions to life problems (56), and perceived burdensomeness and thwarted belongingness (a belief that one does not have meaningful interpersonal relationships) (57). Mental pain, considered distinct from depressogenic negativism, is characterized by a person's perception of strong negative feelings and changes in one's sense of self and functioning. It is conceptually related to grief. Mental pain is considered a strong vulnerability factor for suicidal ideation (58).

There are four important methodological issues to consider when interpreting the primary results of this study, the null findings relating to suicidality and participation in football or contact sports. First, information about participation in contact sports was provided during an interview of the next-of-kin. It was not verified using other sources, such as school yearbooks. Second, we had no information relating to how much exposure to football the decedents had, or how many concussions they might have sustained while playing football or over the course of their lives. Third, by design, we studied men aged 50 and over—so the results do not generalize to men under the age of 50. Finally, the sample we used, without question, is not representative of men from the general population who played football, including the fact that there were virtually no men who were Black or African American in the sample. This sample has major ascertainment bias inherent in the nature of the brain donation program—individuals being referred for autopsy to investigate cause of death, many of whom have a known history of psychiatric illness and/or substance abuse. However, it is the very nature of this program that allowed us to have a sufficient sample size of men with suicide as their manner of death to complete this study.

Personal history of TBI was examined in this study as an exploratory aim. There are major limitations associated with this variable. The variable is based on information from a family member, and the accuracy of this was not independently verified through medical records. The severity of injury could sometimes be inferred, such as descriptions involving several concussions in sports, multiple skull fractures and intracranial bleeding after a fall from a height, a period of coma following a motorcycle accident, or need for neurosurgery to evacuate a subdural hematoma after a fall—but for most cases we could not determine severity of injury. These results, therefore, should be interpreted with caution and replication with a larger sample would increase confidence in these findings.

In the present study, suicide attempts and completed suicide were not related to participation in amateur football or contact sports. Suicide was, however, related to well-established risk factors such as a personal history of a mood disorder, substance abuse disorder, prior suicide ideation, suicide attempts, and a family history of suicide attempts. In conclusion, this study adds to a steadily growing body of evidence suggesting that playing high school football is not associated with increased risk for suicidality or suicide during adulthood.