Which Method of Suicide Is Fatal 90 of the Time: Firearm Injury, Suffocation, Poisoning, or All of the Above

Firearm injury is the method most often fatal at or above 90% in United States data, though the exact figure depends on wound location and whether the attempt is medically treated. Case-fatality rates for suffocation and poisoning run far lower, and no major method is uniformly 90% fatal across every circumstance. "All of the above" is therefore not accurate. The sections below explain the mechanism behind each gap and why these numbers shift by setting and intent.

Mechanism: Why Firearms Produce the Highest Case Fatality

A firearm injury kills through rapid, irreversible destruction of tissue that the body cannot compensate for in time. High-velocity projectiles create a permanent cavity plus a surrounding shock wave of stretched tissue. When the projectile crosses the heart, aorta, or brainstem, loss of perfusion or respiratory drive can be complete within seconds to a few minutes. There is no window for self-rescue and often no window for bystander first aid beyond compression of an external bleed.

The physics matter. Kinetic energy scales with mass times velocity squared, so a rifle round carries far more destructive energy than a handgun round. This is why long-gun wounds are more lethal than handgun wounds at the same anatomic site. The mode of death differs by location: cardiac and great-vessel injury causes exsanguination or tamponade; brain injury causes immediate loss of consciousness and apnea; liver and spleen injury causes slower hemorrhage that may allow survival if surgery arrives quickly.

Suffocation and poisoning act through slower or more reversible pathways. Hanging and strangulation produce cerebral hypoxia by compressing the airway, the carotid arteries, or both, and loss of consciousness can follow within seconds to a minute of carotid occlusion. But the process is sensitive to body weight, ligature position, and whether a bystander intervenes. Poisoning kills through dose-dependent organ failure, and the interval between ingestion and death is often hours, which creates time for antidotes and supportive care.

The single most important mechanism-level point is treatability. A firearm injury to a non-vital area may still be survivable with rapid surgery, and a hanging victim cut down within minutes may recover. Fatality is not a fixed property of a method; it is a property of method plus injury severity plus time to care.

Quantified Context: Case Fatality by Method

Case fatality is the share of attempts that end in death. It is not the same as the share of deaths attributable to a method, because attempt frequency differs enormously across methods. This distinction is the most common source of confusion in thin explainers.

In United States surveillance data, firearm attempts carry the highest case fatality, commonly cited in the range of roughly 85% to 95% depending on the dataset and whether the denominator includes only medically treated attempts or all attempts. Suffocation, which includes hanging and strangulation, generally falls in a lower band, with estimates varying widely by study design. Poisoning, which includes overdose of medications, pesticides, and gases, tends to have the lowest case fatality of the three because many poisonings are reversible with time and treatment.

These numbers are unstable for a reason. Hospital-based studies capture only people who reached care, which inflates apparent survival for firearms and undercounts deaths at the scene. Vital-statistics studies capture deaths but often miss nonfatal attempts. The two denominators produce different rates. Any single percentage should be read as a range with a stated source, not a fixed constant.

The table is a map of ranges, not a leaderboard of certainty. Treat every cell as approximate and source-dependent.

Edge Cases: Where the 90% Claim Breaks Down

Intent and setting change the math more than most people expect. A self-inflicted firearm injury in a rural county with a long emergency-medical-services response time has a higher fatality than the same wound in a city with a Level I trauma center ten minutes away. Distance to definitive care is a real variable, and it is not captured in method-level averages.

Wound location dominates. A gunshot to the abdomen may be survivable with surgery, while a gunshot to the brainstem is almost uniformly fatal. Aggregating all firearm injuries into one number hides this spread. The same logic applies to hanging: partial suspension with feet on the ground produces different physiology than full suspension, and the presence of a bystander who can cut the ligature changes outcome.

Poisoning has its own edge cases. Some agents are highly lethal at small doses and have no antidote, while others are routinely survived with supportive care. Acetaminophen overdose is a classic example of a poisoning that is often survivable if N-acetylcysteine is given early, but fatal if treatment is delayed past the point of liver failure. Carbon monoxide poisoning depends on concentration, exposure duration, and whether the person is removed to fresh air.

A counterintuitive professional caveat: case-fatality rates are not a measure of method "seriousness" or of intent. People who survive a highly lethal attempt are not necessarily less intent on dying; they may simply have had a more treatable injury. Clinicians are trained not to rank intent by method, because doing so leads to missed risk.

Any prior suicide attempt raises future risk, regardless of method lethality. Access to a firearm in the home is a strong independent risk factor for death by suicide. Sudden withdrawal or agitation during a depressive episode can signal elevated risk. Giving away possessions or saying goodbye can precede an attempt. A "calm" demeanor after a crisis is not necessarily improvement; it can reflect a decision made. Mixing alcohol or sedatives with emotional crisis increases impulsivity and overdose risk.

These are warning signs, not diagnostic criteria. They belong in the hands of a clinician, not a search engine.

Misconceptions: What Thin Explainers Get Wrong

The most common error is treating case fatality as a property of the method alone. In reality it is a property of the method, the injury, the setting, and the time to care. A method with a high average fatality can still produce survivors, and a method with a low average fatality can still kill when the dose or location is severe.

A second error is confusing deaths with attempts. Firearms account for a majority of suicide deaths in the United States while representing a smaller share of attempts, because firearm attempts are more often fatal. This is why method-level death counts and method-level attempt counts tell opposite stories. Neither is wrong; they answer different questions.

A third error is assuming "all of the above" is a safe answer. It is not, and it is not supported by data. The three methods do not share a common fatality rate, and claiming they do misleads readers who are trying to understand relative risk.

A fourth error is ignoring the survival side of the ledger. Many people who survive a serious attempt do not go on to die by suicide, and the period after an attempt is a window for intervention. Framing these methods only as death statistics erases the clinical reality that survival is common and treatable.

Practical Actions: What to Do With This Information

The practical takeaway is about reducing access and buying time, not about ranking methods. Means restriction is one of the few interventions with strong evidence for reducing suicide deaths. If a firearm is in the home, storing it locked, unloaded, and separate from ammunition reduces the chance that an impulsive crisis becomes a fatal injury. Off-site storage with a trusted person during a crisis is even stronger.

For medications, the goal is to limit the quantity available during a high-risk period. Clinicians sometimes prescribe smaller supplies or involve a family member in dispensing. For people at risk of overdose, removing high-lethality agents from the home and having naloxone available for opioid exposure are concrete steps.

For hanging and strangulation risk, the home environment matters: removing anchor points, ligatures, and other means during a crisis is a recognized prevention step. These actions are not about surveillance; they are about inserting friction between an impulse and a lethal outcome.

Concrete thresholds to keep in mind: any talk of suicide warrants same-day contact with a crisis line or clinician; any attempt warrants emergency evaluation; and any access to a firearm during a crisis warrants immediate means restriction. If you are unsure whether a situation is urgent, treat it as urgent and call.

Limitations: What These Numbers Cannot Tell You

Case-fatality estimates are subject to selection bias, reporting differences, and definitional inconsistency across studies. Some datasets count only hospital-treated attempts, which excludes deaths at the scene. Others count deaths from vital records, which excludes nonfatal attempts. The result is that the same method can appear to have different fatality rates in different studies without any real-world change.

The numbers also vary by country and by year, so a United States estimate should not be generalized globally. Age, sex, rurality, and access to trauma care all shift the figures. A single percentage presented without a denominator, a setting, and a date is not a reliable claim, and this article deliberately avoids inventing precision that the underlying data do not support.

Finally, method-level statistics describe populations, not individuals. They cannot tell you what will happen in a specific case, and they should never be used to reassure or alarm someone about a particular act. Clinical judgment, not a table, governs individual risk.

If you or someone you know is in immediate danger, call or text 988 in the United States to reach the Suicide and Crisis Lifeline, or contact local emergency services. This article is general educational information about injury mechanisms and population statistics. It is not a diagnosis, treatment plan, or clinical recommendation, and it does not promise any outcome; for personal risk or care decisions, consult a qualified local clinician or crisis service.