Routes of Drug Use Among Drug Overdose Deaths — United States, 2020–2022

Preliminary reports indicate that more than 109,000 drug overdose deaths occurred in the United States in 2022; nearly 70% of these involved synthetic opioids other than methadone, primarily illegally manufactured fentanyl and fentanyl analogs (IMFs). Data from the western United States suggested a transition from injecting heroin to smoking IMFs. CDC analyzed data from the State Unintentional Drug Overdose Reporting System to describe trends in routes of drug use in 27 states and the District of Columbia among overdose deaths that occurred during January 2020-December 2022, overall and by region and drugs detected. From January-June 2020 to July-December 2022, the percentage of overdose deaths with evidence of injection decreased 29.1%, from 22.7% to 16.1%, whereas the percentage with evidence of smoking increased 73.7%, from 13.3% to 23.1%. The number of deaths with evidence of smoking increased 109.1%, from 2,794 to 5,843, and by 2022, smoking was the most commonly documented route of use in overdose deaths. Trends were similar in all U.S. regions. Among deaths with only IMFs detected, the percentage with evidence of injection decreased 41.6%, from 20.9% during January-June 2020 to 12.2% during July-December 2022, whereas the percentage with evidence of smoking increased 78.9%, from 10.9% to 19.5%. Similar trends were observed among deaths with both IMFs and stimulants detected. Strengthening public health and harm reduction services to address overdose risk related to diverse routes of drug use, including smoking and other noninjection routes, might reduce drug overdose deaths.


Introduction
Preliminary data indicate that U.S. drug overdose deaths surpassed 109,000 in 2022; nearly 70% of these deaths involved synthetic opioids other than methadone, primarily illegally manufactured fentanyl and fentanyl analogs (IMFs).*In recent years, deaths co-involving IMFs and stimulants have increased steadily (1).The estimated number of U.S. adults who inject drugs increased from approximately 774,000 in 2011 to nearly 3.7 million in 2018, corresponding to shifts from prescription opioid misuse to the use of heroin and IMFs (2).More recent data suggest transitions from injecting heroin

Methods
Jurisdictions entered data from death certificates, postmortem toxicology testing, and medical examiner or coroner reports on unintentional and undetermined intent drug overdose deaths into CDC's State Unintentional Drug Overdose Reporting System (SUDORS).§ Routes of drug use were identified using information from scene investigations, witness reports, or autopsy data and were categorized into nonmutually exclusive categories of ingestion, ¶ injection,** smoking, † † and snorting § § ; other routes (e.g., transdermal) are not presented because sample sizes were small.Among 28 jurisdictions ¶ ¶ with complete data,*** numbers and percentages of overdose † https://adai.uw.edu/wordpress/wp-content/uploads/dlm_uploads/FentanylQualitativeReport2023.pdf § https://www.cdc.gov/drugoverdose/fatal/sudors.html¶ Evidence of ingestion included witness reports of taking pills or tablets orally or ingesting liquid orally (e.g., liquid methadone), or the discovery of prescription pills, prescription bottles, liquid substances, or vials for containing liquid substances at the scene of the overdose or on the decedent's body.** Evidence of injection included witness reports of injecting drugs, items used to prepare and inject substances found at the scene (e.g., needles, cookers, filters, tourniquets, or alcohol pads), or track marks found on the decedent that appeared to be recent.† † Evidence of smoking included witness reports of smoking drugs or drug paraphernalia at the overdose scene associated with smoking (e.g., pipes, stems, aluminum foil, vape pens, matches, disposable lighters, or gas torches).Fewer than 6.0% of deaths with evidence of smoking had vape pens or e-cigarettes endorsed as evidence; fewer than 3.0% had vape pens or e-cigarettes endorsed with no other evidence of smoking.

Overall Trends
During January 2020-December 2022, a total of 139,740 overdose deaths occurred in 28 jurisdictions; deaths increased 20.2%, from 21,046 during January-June 2020 to 25,301 during July-December 2022.The percentage of deaths with IMFs detected increased 8.4% from 71.4% during January-June 2020 to 77.4% during July-December 2022.Evidence of at least one route of drug use was documented in 71,480 (51.2%) overdose deaths.From January-June 2020 to July-December 2022, the number and percentage of overdose deaths with evidence of smoking increased 109.1% (from 2,794 to 5,843) and 73.7% (from 13.3% to 23.1%), respectively (Figure 1).The number and percentage of deaths with evidence of snorting increased 43.1% (from 2,858 to 4,090) and 19.1% (from 13.6% to 16.2%), respectively.In contrast, the number and percentage of overdose deaths with evidence of injection decreased 14.6% (from 4,780 to 4,080) and 29.1% (from 22.7% to 16.1%), respectively, from January-June 2020 to July-December 2022.Although the number of deaths with evidence of ingestion increased 14.6%, from 3,189 to 3,656, the percentage of such deaths declined 4.6%, from 15.2% to 14.5%.† † † U.S. Census Bureau regions were used to stratify jurisdictions into geographic regions (https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf ).Region analysis included eight of nine jurisdictions in the Northeast Region, five of 12 jurisdictions in the Midwest Region, nine of 17 jurisdictions in the South Region, and six of 13 jurisdictions in the West Region.§ § § Fentanyl was classified as likely illegally manufactured using toxicology, scene, and witness evidence.For the 8.1% of deaths involving fentanyl that had insufficient evidence for classification as illegal or prescription, fentanyl was classified as illegal because the majority of fentanyl overdose deaths involve illegal fentanyl.All fentanyl analogs except alfentanil, remifentanil, and sufentanil, which have legitimate human medical use, were included as IMFs.The leading route of use in drug overdose deaths changed from injection during January-June 2020 (22.7% of deaths) compared with ingestion (15.2%), snorting (13.6%), and smoking (13.3%) to smoking during July-December 2022 (23.1% of deaths) compared with snorting (16.2%), injection (16.1%), and ingestion (14.5%).During July-December 2022, most deaths with evidence of smoking (79.7%), snorting (84.5%), or ingestion (86.5%) had no evidence of injection; among deaths with information on route of use, 81.9% had evidence of a noninjection route.

Trends by Drugs Detected
Among overdose deaths with only IMFs detected (13,107; 9.6%), deaths with both IMFs and stimulants detected (58,754; 43.1%), and deaths with only stimulants detected (8,525; 6.2%), the percentage with evidence of smoking increased, and the percentage with evidence of injection decreased from January-June 2020 to July-December 2022 (Figure 3).For IMFs only, the percentage of overdose deaths with evidence of smoking increased 78.9%, from 10.9% to 19.5%, whereas the percentage with evidence of injection decreased 41.6%, from 20.9% to 12.2%.Among deaths with both IMFs and stimulants detected, the percentage with evidence of smoking increased 65.4%, from 17.9% to 29.6%, whereas the percentage with evidence of injection decreased 25.5%, from 28.6% to 21.3%.A similar pattern was observed among deaths with only stimulants detected (smoking: 29.7% increase, from 15.5% to 20.1%; injection: 22.5% decrease, from 10.2% to 7.9%).Among deaths with neither IMFs nor stimulants detected (10,628; 7.8%), the percentage with * Percentages with evidence of other routes (i.e., buccal, sublingual, suppository, or transdermal) (583; 0.4%) are not presented because of small sample sizes; decedents with drug use via these routes are included in the denominators.In addition, percentages of decedents with no information on route (68,260; 48.8%) are not shown; these decedents are also included in the denominators.† Routes of drug use are not mutually exclusive; decedents might have used multiple routes.§ Alaska, Arizona, Colorado, Connecticut, Delaware, District of Columbia, Georgia, Illinois, Kansas, Kentucky, Maine, Maryland, Massachusetts, Minnesota, Nebraska, New Hampshire, New Jersey, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, Utah, Vermont, Virginia, Washington, and West Virginia.Illinois and Washington reported deaths from counties that accounted for ≥75% of drug overdose deaths in the respective state in 2017, per SUDORS funding requirements; all other jurisdictions reported deaths from the full jurisdiction.¶ Jurisdictions were included if medical examiner or coroner reports and toxicology reports were available for ≥75% of deaths during January 2020-December 2022.
Analysis was restricted to deaths with an available medical examiner or coroner report (39,740; 95.8% of all deaths).
evidence of smoking did not change, and the percentage with evidence of injection decreased 42.2% (11.6% to 6.7%); ingestion was the most common route during July-December 2022 (39.4% of deaths) and throughout the study period.

Discussion
The percentage of drug overdose deaths with evidence of smoking increased sharply in all U.S. regions from 2020 to 2022, indicating the importance of an updated response.By late 2022, among decedents with information on route of drug use, more than three fourths had evidence of a noninjection route, highlighting the diversification of methods through which they used drugs.
From January-June 2020 to July-December 2022, the number of overdose deaths with evidence of smoking doubled, and the percentage of deaths with evidence of smoking increased across all geographic regions.By late 2022, smoking was the predominant route of use among drug overdose deaths overall and in the Midwest and West regions.Increases were most pronounced when IMFs were detected, with or without stimulants.Increases in the number and percentage of deaths with evidence of smoking, and the corresponding decrease in those with evidence of injection, might be partially driven by 1) the transition from injecting heroin to smoking IMFs (3,4), 2) increases in deaths co-involving IMFs and stimulants that might be smoked † † † † (1), and 3) increases in the use of ); these decedents are also included in the denominators.† Analysis included some, but not all, of the jurisdictions in each U.S. Census Bureau region.Northeast: Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, Pennsylvania, Rhode Island, and Vermont; Midwest: Illinois, Kansas, Minnesota, Nebraska, and Ohio; South: Delaware, District of Columbia, Georgia, Kentucky, Maryland, North Carolina, Oklahoma, Virginia, and West Virginia; West: Alaska, Arizona, Colorado, Oregon, Utah, and Washington.§ Jurisdictions were included if medical examiner or coroner reports and toxicology reports were available for ≥75% of deaths during January 2020-December 2022.
Analysis was restricted to deaths with an available medical examiner or coroner report (139,740; 95.8% of all deaths).
counterfeit pills, which frequently contain IMFs and are often smoked (7).Motivations for transitioning from injection to smoking include fewer adverse health effects (e.g., fewer abscesses), reduced cost and stigma, sense of more control over drug quantity consumed per use (e.g., smoking small amounts during a period versus a single injection bolus), and a perception of reduced overdose risk among persons who use drugs (3,5,8).These motivations might also signify lower barriers for initiating drug use by smoking, or for transitioning from ingestion to smoking; compared with ingestion, smoking can intensify drug effects and increase overdose risk (9).Despite some risk reduction associated with smoking compared with injection (e.g., fewer bloodborne infections), smoking carries substantial overdose risk because of rapid drug absorption (5,9).Nearly 80% of overdose deaths with evidence of smoking had no evidence of injection; persons who use drugs by smoking but do not inject drugs might not use traditional syringe services programs where harm reduction messaging and supplies are often provided.In response, some jurisdictions have adapted harm reduction services to provide safer smoking supplies or  ); these decedents are also included in the denominators.† Data on drugs detected come from postmortem toxicology reports; among decedents with a medical examiner or coroner report, analysis was further restricted to decedents with a toxicology report (136,466; 97.7% of decedents with a medical examiner or coroner report).§ Ethanol and other selected drugs (e.g., naloxone and cotinine) were not considered a drug for this analysis; deaths categorized as IMFs only (Panel A) or stimulant only (Panel B) might have also had ethanol or these other selected drugs detected.¶ Deaths with IMFs and stimulants detected (Panel C) could also have other drugs detected (e.g., prescription opioids).** Deaths with neither IMFs nor stimulants detected primarily had prescription opioids (65.3%) or benzodiazepines (37.3%) detected.
† † Drug categories are not comprehensive; some deaths are excluded because they contain drug combinations that are not presented in the panels (e.g., deaths with only IMFs and prescription opioids detected).§ § Alaska, Arizona, Colorado, Connecticut, Delaware, District of Columbia, Georgia, Illinois, Kansas, Kentucky, Maine, Maryland, Massachusetts, Minnesota, Nebraska, New Hampshire, New Jersey, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, Utah, Vermont, Virginia, Washington, and West Virginia.Illinois and Washington reported deaths from counties that accounted for ≥75% of drug overdose deaths in the respective state in 2017, per SUDORS funding requirements; all other jurisdictions reported deaths from the full jurisdiction.
established health hubs to expand reach to persons using drugs through noninjection routes.§ § § § In addition, harm reduction

FIGURE 1 .
FIGURE 1. Number and percentage of drug overdose deaths with evidence of selected routes of drug use,* , † by 6-month period of death (N = 139,740) -State Unintentional Drug Overdose Reporting System, 28 jurisdictions, §, ¶ January 2020-December 2022

FIGURE 3 .
FIGURE 3. Percentage of drug overdose deaths with evidence of selected routes of drug use,* by drugs detected †, §, ¶, ** , † † and 6-month period of death -State Unintentional Drug Overdose Reporting System, 28 jurisdictions, § § January 2020-December 2022 A. IMFs only (n = 13,107) B. Stimulants only (n = 8,525) § § Evidence of snorting included witness reports of snorting drugs, drug paraphernalia at the overdose scene associated with snorting (e.g., razor blades or credit cards used to chop and separate powder; straws, rolled paper, dollar bills, or tubes for nasal inhalation; or powder visible on a table or mirror), or powder on the decedent's nose.
(e.g., peer outreach and provision of fentanyl test strips for testing drug products and naloxone to reverse opioid overdoses), messaging specific to smoking drugs, and linkage to treatment for substance use disorders can be integrated into other health care delivery (e.g., emergency departments) and public safety (e.g., drug diversion) settings.