Which Drugs Are the Most Addictive?

Heroin ranks highest on dependence potential at 3.00 out of 3.00 in the most widely cited expert assessment, followed by cocaine at 2.39 and tobacco at 2.21, while fentanyl and methamphetamine were never assessed.

Alcohol ranks 6th and benzodiazepines 7th of the 20 substances Nutt and colleagues scored in The Lancet in 2007. Most lists citing that study name substances it never measured.

Key Takeaways

• Heroin scored the maximum 3.00 on dependence in Nutt 2007, taking the top rating on all three dependence subscales at once.
• Amphetamine ranked 8th at 1.67, below alcohol at 1.93 and benzodiazepines at 1.83.
• Dependence potential, acute danger and withdrawal danger are three separate measurements producing three different orders.
• Alcohol and the sedative-hypnotic class produce the withdrawals that kill, and that class includes barbiturates, GHB, GBL and baclofen.
• Per NIDA, genetics and epigenetics account for 40 to 60 percent of a person’s risk of addiction.

What Should You Do in an Overdose?

A suspected overdose requires 911 and naloxone given as quickly as possible, even when opioid involvement is uncertain. Per the FDA label, naloxone is safe to use even when opioids are not present, and CDC confirms naloxone does not harm a person who has taken no opioid.

CDC directs responders not to wait for emergency workers before giving naloxone, to repeat every 2 to 3 minutes without a response, and to stay with the person, because naloxone lasts 30 to 90 minutes while many opioids last longer. Fentanyl-involved overdoses frequently require repeat doses.

Crisis resources. The 988 Suicide and Crisis Lifeline answers three ways: call 988, text 988, or chat at 988lifeline.org/chat. Spanish-language support answers by pressing 2, and the Lifeline’s videophone option serves ASL users.

Most US states operate an overdose Good Samaritan provision, and the scope varies by state. California is one example, and its provision is narrower than the phrase suggests. Health and Safety Code section 11376.5 provides that a person who in good faith seeks medical help for someone overdosing, and who does not obstruct responders, cannot be charged with being under the influence of, or possessing for personal use, a controlled substance or paraphernalia. The same protection covers the person who overdoses. The statute does not cover possession for sale, driving offenses, outstanding warrants, or probation and parole violations, and the statute states that no other immunities or protections from arrest or prosecution may be inferred. California Civil Code section 1714.22 separately shields a person who administers naloxone in good faith and without compensation from civil damages. Anyone relying on a state provision should check their own state’s text.

[LEGAL REVIEW NEEDED: this paragraph summarizes a California statute and requires confirmation by a California-licensed reviewer before publication.]

Which substance produced the overdose matters less to the response than most rankings suggest, and the rankings themselves need explaining.

What Actually Makes a Drug Addictive?

Nutt 2007 scored dependence across three subscales rather than as one property: intensity of pleasure, psychological dependence and physical dependence, each on a 0 to 3 scale. A score of 0 meant no risk and 3 meant extreme risk.

Intensity of pleasure measures how strongly a substance drives repeat use. Drugs producing a large, fast rise in dopamine in the mesolimbic reward pathway reinforce most powerfully, and Nutt 2007 notes that effects on the brain can occur within 30 seconds by the fastest routes.

Psychological dependence covers craving and the compulsion to continue independent of physical symptoms, and heroin, cocaine and tobacco all scored 2.6 or above on that subscale. Physical dependence determines how hard stopping becomes once the body has adapted, and severe withdrawal keeps people using to avoid symptoms rather than to obtain effects. Per the FDA, physical dependence on benzodiazepines can develop in several days to weeks, even when taken as prescribed.

Tolerance is a separate process from dependence, and conflating the two is the most common misunderstanding in this area. Drug tolerance is the loss of effect at a given dose, and dependence is the adaptation that produces withdrawal on stopping. The scores those three subscales produced are published.

What Is the Actual Dependence Ranking?

Nutt 2007 assessed 20 substances using a delphic procedure across two independent expert groups, and the published dependence means run from heroin at 3.00 down to alkyl nitrites at 0.87. The top ten in rank order:

Rank	Substance	Dependence mean (0 to 3)
1	Heroin	3.00
2	Cocaine (including crack cocaine)	2.39
3	Tobacco	2.21
4	Street methadone	2.08
5	Barbiturates	2.01
6	Alcohol	1.93
7	Benzodiazepines	1.83
8	Amphetamine	1.67
9	Buprenorphine	1.64
10	Cannabis	1.51

Three features of that table explain most of the confusion around this topic. Heroin scored the ceiling, taking the maximum 3.0 on intensity of pleasure, psychological dependence and physical dependence simultaneously, and no other substance did. Crack sits inside the cocaine row, because Table 2 of the study lists cocaine with the note “includes crack cocaine”; crack received its own score only in the 2010 follow-up.

The third feature is a warning the study issues about a comparison people make constantly. Nutt 2007 states that direct comparison of the scores for tobacco and alcohol with the other drugs is not possible, because their legal status could affect their harms through easier availability. Two substances people expect at the top are absent entirely.

Why Are Fentanyl and Methamphetamine Missing From the Ranking?

Nutt 2007 and its 2010 follow-up assessed neither fentanyl nor methamphetamine, so no dependence score exists for either substance on that scale. Any ranking placing them at the top draws on something else.

The one expert assessment that does include fentanyl places fentanyl below heroin. Van Amsterdam, Pierce and van den Brink’s 2015 ranking of non-medically used prescription opioids gave an overall harm score of 99 for injected heroin, 72 for smoked heroin and 54 for fentanyl. Animal self-administration work using behavioral-economic analysis places fentanyl’s reinforcing efficacy at roughly equal to heroin’s rather than above it.

What fentanyl unambiguously carries is greater potency and faster onset. Potency describes how much drug produces an effect; dependence potential describes how strongly a drug drives repeated use. Treating potency as a dependence ranking is the error that puts fentanyl at the top of lists it does not belong on. Fentanyl’s real distinction is a supply-side one and it is lethal: fentanyl appears in supplies where people do not expect it, so dependence now frequently begins without a decision to use an opioid at all.

Methamphetamine’s nearest proxy in the study, amphetamine, ranked 8th at 1.67, below both alcohol and benzodiazepines. Head-to-head laboratory work gives a sharper picture than a ranking: Shimosato and colleagues measured nucleus accumbens dopamine at 698 percent of basal levels for methamphetamine at 1 mg/kg against 254 percent for cocaine at 10 mg/kg. Methamphetamine both blocks dopamine reuptake and forces dopamine release, and methamphetamine’s half-life of roughly 10 to 12 hours against cocaine’s approximately 60 minutes compounds the effect. Methamphetamine addiction reflects that combination of intensity and duration. The substances the study did score deserve individual accounts.

Which Scored Substances Sit at the Top, and Why?

Five scored substances account for the top of the dependence table: heroin and the opioids, cocaine in both forms, tobacco, alcohol, and the sedative-hypnotics. Each reaches that position by a different route.

Heroin and Other Opioids

Opioids rank first on dependence in Nutt 2007 and first on acute physical harm, where heroin scored 2.8, the highest of the 20. Opioids act directly on receptors the body already uses for pain relief and reward, producing an effect closer to relief than to stimulation. Heroin addiction commonly begins with prescription opioid use rather than with heroin itself.

Cocaine, in Both Forms

Cocaine ranked 2nd on dependence at 2.39, scoring the maximum 3.0 on intensity of pleasure. Crack and powder are the same molecule: per Hatsukami and Fischman, the physiological and psychoactive effects are similar regardless of form. Route accounts for the difference in dependence risk. Smoked cocaine reaches the cerebral circulation in about 6 seconds against 18 seconds intravenously, and controlled dosing gives absorption half-times of 1.1 minutes smoked against 11.7 minutes intranasally.

Tobacco and Nicotine

Tobacco ranked 3rd on dependence at 2.21 while ranking 18th of 20 on acute physical harm at 0.9, and that split illustrates most clearly that dependence and danger are different measurements. Per NIDA, a typical smoker takes about 10 puffs per cigarette, so a pack-a-day smoker gets roughly 200 hits of nicotine to the brain each day. One qualification belongs here: Lachenmeier and Rehm’s margin-of-exposure analysis places nicotine in the highest acute-risk band alongside alcohol, heroin and cocaine, because the gap between a used dose and a lethal dose is narrow.

Alcohol

Alcohol ranked 6th on dependence at 1.93, which reads as mid-table and understates alcohol’s position badly. The 2010 follow-up study, using multicriteria decision analysis, ranked alcohol the most harmful drug overall at 72, above heroin at 55 and crack at 54. Alcohol’s withdrawal is among the most dangerous of any substance, and availability produces far more cue exposure than a drug requiring deliberate acquisition. Lachenmeier and Rehm give alcohol the lowest margin of exposure of any substance assessed.

Barbiturates and Benzodiazepines

Barbiturates ranked 5th at 2.01 and benzodiazepines 7th at 1.83. Their dependence risk attracts frequent underestimation because both arrive by prescription, which reads as safety. The FDA’s 2020 class boxed warning states that stopping abruptly or reducing too quickly can result in withdrawal reactions, including seizures, which can be life-threatening.

Dependence ranking and withdrawal danger diverge sharply at this point in the table.

Which Withdrawals Can Kill?

Alcohol and the sedative-hypnotic class produce the withdrawals that kill, and that class runs wider than most lists suggest: benzodiazepines, barbiturates, gamma-hydroxybutyric acid, gamma-butyrolactone, baclofen and ethanol.

Abrupt cessation or a sharp reduction in any of those substances can precipitate a life-threatening withdrawal syndrome. Alcohol carries the best-quantified risk: per StatPearls, delirium tremens carries a mortality of up to 37 percent without appropriate treatment, falling below 5 percent when treated early, and 3 to 5 percent of people in alcohol withdrawal develop it. GHB withdrawal has produced documented deaths, including one on hospital day 13 in an eight-patient case series.

Opioid withdrawal sits differently rather than lower. Per SAMHSA TIP 45, uncomplicated opioid withdrawal is not life-threatening, and the same chapter groups opioids with alcohol and the sedative-hypnotics in preferring hospitalization or some form of 24-hour medical care. Deaths from untreated opioid withdrawal are documented, through dehydration and electrolyte disturbance from persistent vomiting and diarrhea. Neonatal opioid withdrawal is a separate matter: per the FDA naloxone label, neonatal opioid withdrawal syndrome, unlike opioid withdrawal in adults, may be life-threatening. Stimulant withdrawal carries a risk of a different kind.

Does Stimulant Withdrawal Need Medical Care?

Stimulant withdrawal requires close monitoring for suicide risk rather than for physical danger, and SAMHSA TIP 45 names that risk as a potentially lethal medical danger. The reason is not the one most people expect.

Stimulant withdrawal produces no seizures or autonomic collapse comparable to alcohol, and per TIP 45 stimulant withdrawal usually does not involve medical danger. The same section then names the exception directly: TIP 45 describes an often overlooked but potentially lethal medical danger during stimulant withdrawal, a profound dysphoria that may include suicidal ideas or attempts. TIP 45 instructs that amphetamine users in particular receive close monitoring during detoxification for signs of suicidality, notes that seizures may also complicate the picture, and directs that chest pain and persistent headache receive evaluation.

Medication has a role. TIP 45 names diphenhydramine, trazodone or hydroxyzine for insomnia, and SSRIs where depression accompanies withdrawal. The crash itself runs long: per TIP 33, methamphetamine withdrawal symptoms begin 2 to 4 days after stopping and may persist 2 to 4 weeks, with anhedonia and dysphoria lasting for months. Combining substances changes every one of these pictures.

How Often Does More Than One Substance Contribute to an Overdose?

In CDC’s State Unintentional Drug Overdose Reporting System data covering 28 states and DC, 34.9 percent of 38,048 deaths involved opioids together with stimulants. A further 48.1 percent involved opioids without stimulants and 12.6 percent stimulants without opioids.

CDC’s MMWR Vital Signs analysis of 24 states and DC found illicitly manufactured fentanyl, heroin, cocaine or methamphetamine involved in 83.8 percent of deaths, and of those, 49.8 percent involved two or more of the four. Among synthetic-opioid-involved deaths in 2016, almost 80 percent involved another drug or alcohol.

CDC attributes the increase primarily to the supply of illicitly manufactured fentanyls and methamphetamine, alongside illicit co-use of opioids and stimulants, so polysubstance use is a major contributor rather than the single cause. Polysubstance use treatment addresses that pattern directly. The drug explains only part of who develops a disorder.

Why Does the Same Drug Affect People Differently?

Four factors shape vulnerability to a substance use disorder: genetics, age of first use, trauma history and co-occurring mental health conditions. The drug sets a ceiling on risk, and the person and their circumstances determine where within that range they land.

• Genetics. Per NIDA, genes and epigenetics account for 40 to 60 percent of a person’s risk of addiction. Heritability estimates by substance range from about 0.39 for hallucinogens to 0.72 for cocaine, and substance use disorder is polygenic, so no single gene determines outcome.
• Age of first use. Grant and Dawson’s analysis of 27,616 people found lifetime DSM-IV alcohol dependence falling from more than 40 percent among those who started drinking at 14 or younger to roughly 10 percent among those who started at 20 or older, with the odds decreasing 14 percent for each additional year of onset age.
• Trauma history. In the CDC-Kaiser adverse childhood experiences work, people reporting five or more adverse childhood experiences were 7 to 10 times more likely to report illicit drug use problems, addiction to illicit drugs and parenteral drug use than those reporting none.
• Co-occurring mental health conditions. Anxiety, depression, PTSD and bipolar disorder each raise risk, partly through self-medication.

The US Surgeon General’s report states that the frontal cortex does not fully develop until the early to mid-20s, which explains the age effect. Those four factors explain how two people using the same substance the same number of times arrive at completely different places. Two beliefs about this ranking cause the most trouble.

Which Beliefs About the Most Addictive Drugs Are Wrong?

Two widely repeated claims about drug dependence rankings fail against the primary sources, and both name substances the cited study never measured.

Common misconception	What the evidence shows
“Fentanyl is the most addictive drug, because it is the most potent.”	Nutt 2007 and its 2010 follow-up never assessed fentanyl. Van Amsterdam, Pierce and van den Brink’s 2015 expert ranking gave fentanyl an overall harm score of 54 against 99 for injected heroin. Potency describes how much drug produces an effect; dependence potential describes how strongly a drug drives repeated use.
“Nicotine is as addictive as heroin.”	Nutt 2007 scored heroin’s dependence at 3.00 against tobacco’s 2.21, and the study states that direct comparison of tobacco and alcohol scores with the other drugs is not possible because legal status affects their harms. Population capture-rate data point the other way, suggesting smoked tobacco is the most addictive commonly used drug. The two metrics disagree and neither settles it.

The ranking that matters clinically is the withdrawal one rather than the dependence one, and treatment follows from that.

Does the Ranking Change How Addiction Is Treated?

The dependence ranking changes the first phase of treatment and much less of the rest, because withdrawal management differs by substance while the work after withdrawal does not.

Per SAMHSA’s TIP 45 consensus panel, alcohol and sedative-hypnotic withdrawal require medical monitoring because of seizure risk, and the failure to treat incipient convulsions is a deviation from the established general standard of care. Opioid withdrawal requires symptom management and usually medication, with hospitalization or 24-hour care preferred. Stimulant withdrawal requires close monitoring for suicidality alongside symptomatic medication.

Beyond that phase the treatment is substantially similar regardless of substance, because what maintains a substance use disorder is not the drug’s pharmacology. The function the substance serves, the cues that trigger use, and whatever sits underneath all persist after the drug has cleared, which is why programs treat mixed cohorts rather than sorting people by substance.

Where Can You Get Treatment in Los Angeles?

New Spirit Recovery treats alcohol, opioid, benzodiazepine, stimulant and prescription medication use disorders at its Tarzana, Northridge and Encino locations, with medically monitored detox followed by residential treatment.

Clinicians treat co-occurring mental health conditions alongside the substance use rather than after it. New Spirit Recovery admissions screens for those co-occurring conditions at intake.

“People ask which drug is worst, and the ranking matters far less than they expect. What gets treated is why somebody needed it, and that part looks remarkably similar whichever substance brought them in.” Dr. Patrick Lockwood, Psy.D., Clinical Consultant at New Spirit Recovery.

Frequently Asked Questions

What is the single most addictive drug?

Heroin. In Nutt 2007 heroin scored the maximum 3.00 on dependence, taking the top rating on intensity of pleasure, psychological dependence and physical dependence simultaneously. Cocaine followed at 2.39 and tobacco at 2.21.

Where does fentanyl rank for addictiveness?

Nutt 2007 and its 2010 follow-up never assessed fentanyl, so fentanyl holds no score on that scale. The one expert ranking including fentanyl gave it an overall harm score of 54 against 99 for injected heroin. Fentanyl’s distinction is potency and speed of onset rather than dependence potential.

Which withdrawal is the most dangerous?

Alcohol and the sedative-hypnotic class, which includes benzodiazepines, barbiturates, GHB, GBL and baclofen. Delirium tremens carries a mortality of up to 37 percent without appropriate treatment. Opioid withdrawal produces more discomfort and generally proves survivable when managed.

Is methamphetamine more addictive than alcohol?

Not on the Nutt 2007 scale, which never assessed methamphetamine. Amphetamine, the nearest substance in the study, ranked 8th at 1.67, below alcohol at 1.93. Laboratory work does show methamphetamine producing a substantially larger dopamine response than cocaine.

Why is crack considered more addictive than powder cocaine?

Chiefly delivery speed. Crack and powder are the same molecule with similar effects per Hatsukami and Fischman. Smoked cocaine reaches the cerebral circulation in about 6 seconds with an absorption half-time of 1.1 minutes, against 11.7 minutes intranasally. Nutt 2007 scored both forms in a single combined row.

Can you get addicted to a drug after one use?

A substance use disorder does not form from a single use, because the DSM-5-TR diagnosis requires at least 2 of 11 criteria within a rolling 12-month period. A single exposure to a strongly reinforcing drug establishes a powerful memory of the reward, which is a different and real risk.

Does a drug’s legal status reflect how addictive it is?

Not in the jurisdiction Nutt 2007 tested. The study compared its overall harm ratings against classification under the UK Misuse of Drugs Act 1971 and found the correlation not significant, reporting a Kendall’s rank correlation of -0.18 at p=0.25. A legal schedule is not a harm ranking.

References

1. Nutt, D., King, L. A., Saulsbury, W., Blakemore, C. (2007). Development of a rational scale to assess the harm of drugs of potential misuse. The Lancet, 369(9566), 1047-1053.
2. Nutt, D. J., King, L. A., Phillips, L. D.; Independent Scientific Committee on Drugs. (2010). Drug harms in the UK: a multicriteria decision analysis. The Lancet, 376(9752), 1558-1565.
3. van Amsterdam, J., Pierce, M., van den Brink, W. (2015). Ranking the harm of non-medically used prescription opioids in the UK. Regulatory Toxicology and Pharmacology, 73(3), 999-1004.
4. Lachenmeier, D. W., Rehm, J. (2015). Comparative risk assessment of alcohol, tobacco, cannabis and other illicit drugs using the margin of exposure approach. Scientific Reports, 5, 8126.
5. Hatsukami, D. K., Fischman, M. W. (1996). Crack cocaine and cocaine hydrochloride: are the differences myth or reality? JAMA, 276(19), 1580-1588.
6. Grant, B. F., Dawson, D. A. (1997). Age at onset of alcohol use and its association with DSM-IV alcohol abuse and dependence. Journal of Substance Abuse, 9, 103-110.
7. Dube, S. R., et al. (2003). Childhood abuse, neglect, and household dysfunction and the risk of illicit drug use. Pediatrics, 111(3), 564-572.
8. Substance Abuse and Mental Health Services Administration. (2006). Detoxification and Substance Abuse Treatment. TIP Series, No. 45; and (2021) Treatment for Stimulant Use Disorders. TIP Series, No. 33.
9. Centers for Disease Control and Prevention. Drug Overdose Deaths Involving Opioids and Stimulants. SUDORS Data Brief; and MMWR Vital Signs, 69(35).
10. US Food and Drug Administration. (2020). Boxed Warning Updated to Improve Safe Use of Benzodiazepine Drug Class.
11. National Institute on Drug Abuse. Drugs, Brains, and Behavior: The Science of Addiction.
12. American Psychiatric Association. (2022). Diagnostic and Statistical Manual of Mental Disorders (5th ed., text rev.).