ferred significant risk for suicidality in adolescence and adulthood,
with the strongest and clearest effects for childhood sexual abuse in
particular. Risk of multiple suicide attempts was eight times greater
among those with a sexual abuse history than among others.46
As mentioned in Chapter 2, our study on data collected from the
National Comorbidity Survey—a large project on the occurrence of
mental disorders and associated variables in U.S. adults—told a simi-
lar story. Our analyses showed that some forms of abuse were more
frequently linked to subsequent suicidality than were other forms;
specifically, the effects for childhood physical abuse and sexual abuse
on later suicidal behavior were relatively pronounced and similar to
one another, and exceeded effects for molestation and for verbal abuse.
Various forms of injury and victimization may instill the ability to
lethally harm oneself and increase risk for serious suicidal behavior.
On the one hand, the experience of physical and sexual abuse could
habituate people to self-injury. In fact, regarding childhood sexual
abuse, there is evidence that more painful forms (e.g., severe forced
abuse) are more associated with suicidality than less painful forms.47
On the other hand, the model developed in this book posits that
lethality combines with desire for death to result in serious sui-
cidal behavior, and that desire for death stems from feeling a burden
on loved ones and others, and feeling disconnected and alienated
from others. As noted in Chapter 2, to the degree that any form of
abuse facilitates either lethality (through habituation to pain and
provocation) or desire for death (through increased feelings of bur-
densomeness or disconnection), it should, according to the current
model, constitute a risk for later suicidal behavior. Childhood physi-
cal and sexual abuse may particularly confer risk, because they are
Genetics, Neurobiology, and Mental Disorders ● 191
both painful and imply burdensomeness and disconnection. In fact,
there is evidence that increased alienation (similar to lack of belong-
ingness) is a prevailing psychological link between childhood mal-
treatment and later suicidal behavior.48
What is the main neurobiological mechanism linking early abuse
to later self-injury? There is intriguing evidence that the HPA axis is
involved. Adults who have been abused as children appear to have a
dysregulated HPA system.49 Another effect of an HPA system gone
awry is decreased volume of a brain region called the hippocampus,
which is heavily involved in memory. Too much circulating cortisol
seems to erode hippocampal cells. In one study, depressed women
who had experienced childhood abuse had smaller hippocampal vol-
ume on brain scans as compared to depressed women who had not
been abused.50
Childhood adversity harms the HPA axis and increases risk for
adult suicidal behavior. As was noted earlier in this chapter, HPA
problems may also increase risk for later suicidality. Putting these
facts together, it is plausible that childhood adversity affects later sui-
cidal behavior partly through its effects on the HPA system. This may
be one neurobiological underpinning for the psychological effects of
childhood adversity on suicide. Childhood adversity, especially when
severe, impacts all aspects of my model. It familiarizes people with
pain and provocation, and it makes them feel worthless and alien-
ated—a lethal combination, according to the view developed here.
Mental Disorders
Approximately 95 percent of people who die by suicide experienced a
mental disorder at the time of death.51 As noted in the first chapter, I
believe my dad had bipolar II disorder—serious depressions com-
bined with hypomanic episodes—and this played a role in his death.
Little is known about the other 5 percent, but most if not all of them
likely experienced one or more “subsyndromal” mental disorders—
192 ● WHY PEOPLE DIE BY SUICIDE
that is, they experienced many symptoms of, say, depression, but not
quite enough symptoms to rise to the threshold of formal diagnosis
according to the American Psychiatric Association’s Diagnostic and
Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV).
A brief side note on the DSM is in order. The manual has many
merits, but one of them happens not to be an ultimate monopoly on
truth—it is a work in progress, albeit a very reliable one. Despite the
DSM ’s imperfections, we must discount the views of some scholars
who claim that mental disorders do not exist or represent social
myths. A full discussion of this notion is beyond the scope of this
book, but briefly, one profound problem with this idea is that it is an
affront to people with mental disorders (indeed, people who have
died from them), as well as to their loved ones. Imagine, as your
loved one is dying from cancer, that someone smugly tells you, “Can-
cer doesn’t really exist anyway.” People with major mental disorders
and their loved ones have suffered exactly this offense. Another prob-
lem with this idea is that it is highly implausible in light of current
scientific knowledge. Writing specifically of schizophrenia, Seymour
Kety summed up this issue succinctly: “If schizophrenia is a myth, it
is a myth with a strong genetic component.”52
The DSM has five axes, and the first two are relevant here. Axis I includes the major mental disorders like schizophrenia, mood disorders, anxiety disorders, substance-use disorders, and so on. Axis II
includes the personality disorders. Several of these two categories of
disorders (e.g., schizophrenia, bipolar-spectrum disorders, major de-
pression, some anxiety disorders, some substance-use disorders, and
some personality disorders) appear to play a role in the risk for sui-
cide. In what follows, the relevance of the proposed model is evalu-
ated with respect to suicidality in the context of several different
mental disorders.
The anxiety disorders include panic disorder, social phobia, gener-
alized anxiety disorder, post-traumatic stress disorder, obsessive-
Genetics, Neurobiology, and Mental Disorders ● 193
compulsive disorder, and specific phobia. Symptoms of some of the
anxiety disorders have been repeatedly implicated in serious suicid-
ality. For example, an analysis of the Food and Drug Administration
(FDA) database of treatment outcome studies found a significant as-
sociation between anxiety disorders and suicide.53 Jan Fawcett and
colleagues have repeatedly shown that severe anxiety is an important
sign of acute suicide risk.54
Of all the anxiety disorders, panic disorder probably has received
the most attention with regard to associations with suicidality. Panic
disorder involves repeated experiences of severe panic attacks that
often come “out of the blue”; people going through a panic attack of-
ten believe they are having a heart attack or that some other cata-
strophic thing is happening. Indeed, there does seem to be a sig-
>
nificant association between panic disorder and suicidal symptoms,55
although the connection may be explained mostly by the fact that
panic disorder often co-occurs with mood disorders,56 and of course
suicidality often emerges in the context of mood disorders.
From the standpoint of the model developed in this book, it is in-
teresting to note that there is a form of panic disorder, panic disorder
with agoraphobia, that particularly affects the need to belong. People
who experience this form of panic disorder are so concerned about
experiencing panic attacks that they rarely leave their house and thus
experience extreme reductions in social contact. By the logic of the
current model, those who experience this form of panic disorder
should be more prone to suicidality than those who experience panic
disorder without agoraphobia. In fact, there is some evidence to this
effect.57
Substance-use disorders confer risk for suicidality.58 As has already
been noted, according to the perspective proposed here, this associa-
tion is mainly a result of substance abuse facilitating provocative ex-
periences and thus the acquisition of the ability to enact lethal self-
injury. In Chapter 2, for example, it was noted that heroin users are
194 ● WHY PEOPLE DIE BY SUICIDE
fourteen times more likely than peers to die from suicide, and that
the prevalence of attempted suicide is also many orders of magnitude
greater than that of community samples.59
Some people report feeling more courageous while intoxicated;60
this sense of courage can be misdirected toward self-injury in some
people. In a relevant study on this point, some participants were
given alcohol to a blood alcohol level of .10 percent, and some were
given a placebo drink. Then, all were provided the opportunity to
self-administer shock during a task disguised as a reaction-time
game, with self-aggression defined by the intensity of shock chosen.
Men who had consumed alcohol self-inflicted more shock than those
who did not.61
A recent example shows how substance abuse can facilitate pain-
ful and provocative experiences, including self-injury. On July 13,
2004, the Associated Press reported that in March of that year a man
in England drank fifteen pints of beer, then got in an argument with
a friend about whose turn it was to buy the next beer. Apparently
the argument was unresolved, so the man went home to retrieve a
sawed-off shotgun. He stuffed the gun in his pants. On the way back
to the pub, the gun discharged. His lawyer stated, “He still feels quite
severe pain” and added that some shotgun pellets remained lodged in
the man’s groin area, potentially rendering him infertile. To make it
worse, the man was jailed for illegal possession of a firearm. In this
example, substance use clearly led to a painful, self-injurious experi-
ence.
Prolonged substance abuse can certainly deteriorate social capital
(leading to low belongingness) and diminish feelings of overall effec-
tiveness (producing feelings of perceived burdensomeness). Indeed, a
review of the literature on alcohol abuse and loneliness revealed
that alcohol abusers experience more loneliness than do members of
most other groups.62 A series of phone interviews with a predomi-
nantly crack-cocaine-using sample found that those continuing to
Genetics, Neurobiology, and Mental Disorders ● 195
use the drug after treatment report lower self-efficacy.63 Findings like
these suggest additional points of consilience between parameters of
substance-use disorders and the current model’s emphasis on per-
ceived burdensomeness and failed belongingness.
As was noted in Chapter 2, two mental disorders, borderline per-
sonality disorder and anorexia nervosa, are of particular interest, be-
cause they are among the most lethal of all psychiatric disorders (de-
spite being more common in women than men), with the usual
mechanism of death (including for anorexia nervosa) being suicide.64
Borderline personality disorder is characterized by a longstanding
pattern of stormy interpersonal relationships, self-destructive behav-
iors such as self-cutting or -burning, marked emotional lability and
impulsivity, and an empty or diffuse sense of identity. Unfortunately,
in some clinical settings, patients with the disorder have the reputa-
tion for manipulation, including manipulating others through self-
destructive behaviors (e.g., “gesturing suicide”), as well as for “split-
ting” (e.g., pitting people, including clinicians, against one another);
some people roll their eyes about such patients, take a subtly or
overtly demeaning tone about them, and make disparaging com-
ments.
In some clinical settings, mental health professionals harbor de-
meaning attitudes toward people with borderline personality disor-
der. I recently read a hospital progress note for a person with border-
line personality disorder that stated, “This patient is certainly not
getting treatment from me. ” One reason for sentiments like these is the belief that many such patients merely “gesture” suicide. In other
words, they engage in suicidal behaviors, such as cutting themselves,
but do not really intend to kill themselves; instead, they only intend
to provoke or manipulate others.
If only this were true. Those with borderline personality disorder
have a 10 percent lifetime rate of death by suicide; at least 50 percent
of people with borderline personality disorder have made a mini-
196 ● WHY PEOPLE DIE BY SUICIDE
mum of one very severe suicide attempt;65 and among those with this
syndrome, an average of over three lifetime suicide attempts has been
reported.66 Further, history of previous attempt among people with
borderline personality disorder is a stronger predictor of completed
suicides than for any other diagnostic group (e.g., 65 percent of sui-
cides among those with borderline personality disorder have made a
prior attempt; 33 percent of suicides among those with major de-
pression have made a prior attempt).67 Through repeated self-injury,
people with borderline personality disorder become practiced re-
garding suicidal behavior and may thus become courageous and
competent about suicide. Moreover, a common and pervasive sense
of self-doubt and feelings of alienation and abandonment are very
likely to instill perceptions of being a burden and create difficulty in
belonging. As a consequence, suicide risk is usually elevated in pa-
tients with borderline personality disorder.
Women with anorexia, too, put themselves through a physical or-
deal—namely, self-starvation. In addition, those with the binge-
purge subtype of anorexia also endure various compensatory efforts
like self-induced vomiting, ingesting agents like ipecac syrup that in-
duce vomiting, repeated enemas, and so on. Through these provoca-
tive experiences, the theory proposed here suggests that women with
anorexia may acquire the ability to enact lethal self-injury. Indeed,
there is evidence that anorexic women have elevated pain thresh-
olds,68 as has also been shown in suicide attempters,69 and suicide
rates among women with anorexia are quite elevated. Over the course
of a ten-year follow-up interval, the rate of death by suicide among
246 women with eating disorders was fifty-eight times the expected
rate.70 All of those who died were anorexic; no women with bulimia
died (but see below).
Shneidman’s case example of Beatrice represents an example of
the co-occurrence of anorexia and suicidality.71 Beatrice said of her
suicide attempt by self-cutting, “The evening dragged on with me
busy reopening the stubborn veins that insisted upon clotting up. I
Genetics, Neurobiology, and Mental Disorders ● 197
was patient and persistent, and cut away at myself for over an hour.
The battle with my body to die was unexpected, and after waging a
good fight, I passed out.” She also said, “For the next two years . . . ev-
ery night, before fading off to sleep, I imagined committing suicide. I
became obsessed with death. I rehearsed my own funeral over and
over, adding careful details each time.” Beatrice later planned her sui-
cide for three months and tried again to die by self-cutting; she
survived.
Shneidman wrote something about Beatrice that is very revealing,
I think. In her voice, he wrote, “I can try to control myself (and oth-
ers) through controlling my body. My body is my only practical han-
dle on the world, a rheostat (that I can turn up and down)—gain or
lose the same 15 pounds—I can control my life by controlling my
body. And if life gets too painful—I can turn it off completely.”
Through control of eating and body, she comes to the ability to beat
down the self-preservation instinct. She has developed that “little
switch”—the ability to turn life off—through the painful and pro-
vocative experiences of suicidality and anorexia. In her own words,
Beatrice makes it plain: “For me, restricting my food intake is not
about being fashionably thin, it’s about my death wish.”
As compared to women with anorexia, who may combine self-
starvation with intermittent binges and severe compensatory behav-
iors, women with bulimia may not be at as high a risk, because their
experiences are relatively less provocative (e.g., self-starvation is not
Why People Die By Suicide Page 25