U.S. Heroin Statistics: Use, Disorder, Overdose Deaths, and Treatment

SAMHSA estimated that 556,000 people aged 12 or older used heroin in 2024, while CDC final mortality data recorded 2,743 U.S. overdose deaths involving heroin that year. SAMHSA also estimated that 496,000 people aged 12 or older met survey criteria for heroin use disorder in 2024. These figures measure different outcomes and should not be combined.

CDC reported that the age-adjusted heroin-involved overdose death rate fell 33.3 percent from 2023 to 2024, declining from 1.2 to 0.8 deaths per 100,000 people. The decline is substantial, but heroin use, opioid use disorder, and overdose still require prevention, treatment, and recovery support.

If a person cannot wake or is breathing slowly, irregularly, or not at all, call 911, give naloxone, and support breathing when trained. Place a breathing person on their side when appropriate and remain with the person until emergency help arrives.

How many people used heroin in the United States in 2024?

SAMHSA estimated that 556,000 people aged 12 or older used heroin in the past year in 2024, representing 0.2 percent of the U.S. civilian, noninstitutionalized population in that age range. The National Survey on Drug Use and Health, or NSDUH, produces a weighted population estimate rather than a direct count of every person who used heroin.

SAMHSA divided the 2024 estimate into 495,000 adults aged 26 or older, 51,000 young adults aged 18 to 25, and 10,000 adolescents aged 12 to 17. The corresponding prevalence estimates were 0.2 percent for adults aged 26 or older, 0.1 percent for young adults, and less than 0.1 percent for adolescents.

NSDUH covers people living in households and noninstitutional group settings, including shelters, college dormitories, and halfway houses. It does not represent people in institutional settings, and self-reported heroin use may be affected by recall or willingness to disclose use.

How many people had heroin use disorder in 2024?

SAMHSA estimated that 496,000 people aged 12 or older met NSDUH criteria for past-year heroin use disorder in 2024. The detailed tables estimated 486,000 adults aged 18 or older, including 33,000 adults aged 18 to 25 and 453,000 adults aged 26 or older, plus 9,000 adolescents aged 12 to 17. SAMHSA rounds weighted estimates independently, so displayed age-group values may not sum exactly to the displayed total.

Heroin use and heroin use disorder are not interchangeable measures. Use disorder reflects a clinically significant pattern of impaired control, risky use, and continued use despite harm; the health effects and clinical signs of heroin use require individual assessment rather than inference from population estimates.

How many U.S. overdose deaths involved heroin in 2024?

CDC final mortality data recorded 2,743 U.S. drug overdose deaths involving heroin in 2024, with an age-adjusted rate of 0.8 deaths per 100,000 people. NCHS identifies a heroin-involved death when drug overdose is the underlying cause and the death certificate includes the ICD-10 multiple-cause code T40.1 for heroin.

NCHS reported that 96 percent of drug overdose death records identified at least one specific drug in both 2023 and 2024. A heroin-involved death does not mean heroin acted alone, because one death certificate can identify heroin, fentanyl, cocaine, alcohol, benzodiazepines, or other substances together.

How did heroin-involved overdose mortality change from 2023 to 2024?

CDC recorded 3,984 heroin-involved overdose deaths in 2023 and 2,743 in 2024. Over the same period, NCHS reported that the age-adjusted rate decreased 33.3 percent, falling from 1.2 to 0.8 deaths per 100,000 people.

The final mortality data establish a national decline in heroin-involved deaths, but they do not establish why the decline occurred. Changes in the illicit drug supply, toxicology, treatment access, naloxone use, population behavior, and death-certificate coding require separate evidence before any factor is credited with causing the trend.

How do heroin deaths compare with synthetic-opioid deaths?

CDC final mortality data recorded 47,735 overdose deaths involving synthetic opioids other than methadone and 2,743 deaths involving heroin in 2024. The synthetic-opioid category includes fentanyl, fentanyl analogs, and tramadol, while the heroin category is coded separately.

A death involving both heroin and fentanyl appears in both applicable categories, so U.S. drug overdose mortality by substance must be interpreted without adding drug-specific totals. The larger synthetic-opioid count does not make heroin exposure predictable, and the smaller heroin count does not measure nonfatal overdose or current use.

What do national treatment-admission statistics show about heroin?

SAMHSA's 2023 TEDS quick statistics recorded 178,944 admissions with heroin as the primary substance, representing 10.8 percent of 1,651,788 reported substance use treatment admissions. The same table recorded 67.0 percent of heroin-primary admissions among male clients and 33.0 percent among female clients in that reporting year.

A TEDS admission is a treatment episode, not a unique person. One person may have more than one admission, facilities differ in whether they report publicly or privately funded care, and state administrative coverage affects the national total. Treatment admissions therefore describe service use within reporting systems rather than the prevalence of heroin use disorder in the entire population.

Why do federal heroin statistics differ?

Federal heroin statistics differ because NSDUH estimates people, NVSS counts death records, and TEDS counts treatment admissions. Each system uses a different population, collection method, time period, and definition, so apparently similar numbers can answer different questions.

Federal systems used to measure national heroin use, disorder, mortality, and treatment demand

Federal system	What it measures	Primary interpretation limit
2024 NSDUH	Weighted estimates of past-year heroin use and heroin use disorder	Represents the civilian, noninstitutionalized population aged 12 or older
2024 NVSS mortality	Final death-certificate counts and rates for overdose deaths involving heroin	Counts drug involvement, not heroin acting alone
2023 TEDS-A	Admissions with heroin reported as the primary substance	Counts treatment episodes in state reporting systems, not unique people

Counts and rates are also different measures. Counts describe the number of recorded events, crude rates relate events to the population, and age-adjusted rates support comparisons across populations or years with different age structures.

What do heroin statistics mean for treatment and recovery?

National heroin statistics show that treatment and overdose prevention remain necessary even as heroin-involved mortality declines. Population data cannot determine one person's diagnosis, withdrawal needs, overdose risk, or appropriate level of care; those decisions require an individual clinical assessment.

CDC guidance supports methadone, buprenorphine, or naltrexone for opioid use disorder and warns that detoxification alone is not recommended. A structured heroin addiction treatment program can combine medication, counseling, overdose prevention, relapse planning, and continuing recovery support.

Do lower heroin death rates mean heroin is safe?

Lower heroin-involved death rates do not make heroin safe. Mortality is one outcome, while nonfatal overdose, opioid use disorder, infection, withdrawal, and impaired functioning remain clinically important harms that death statistics do not count.

Can heroin and fentanyl death counts be added together?

Heroin and fentanyl death counts cannot be added because a single overdose death may involve both drugs and appear in both categories. NCHS reports drug involvement through multiple-cause codes, so adding category totals would count some deaths more than once.

Are treatment admissions the same as people in treatment?

Treatment admissions are not the same as unique people in treatment. TEDS records episodes reported by state administrative systems, and one person can enter treatment more than once during a reporting period.