Louisa Degenhardt, Paige Webb, Samantha Colledge-Frisby, Jeremy Ireland, Alice Wheeler, Sophie Ottaviano, Alex Willing, Abe Kairouz, Evan Cunningham, Behzad Hajarizadeh, Janni Leung, Lucy T Tran, Olivia Price, Amy Peacock, Peter Vickerman, Michael Farrell, Gregory J Dore, Matthew Hickman, Jason Grebely
8 min
This systematic review provides a comprehensive global update on the prevalence of injecting drug use (IDU) and the health harms faced by people who inject drugs (PWID). By analyzing data from 2017 to 2022, the authors identified 14.8 million individuals who inject drugs across 190 countries, representing over 99% of the global population aged 15-64. The study goes beyond mere prevalence counts to characterize the social and environmental risks that exacerbate health crises among this population.
Researchers conducted a multi-stage systematic review, adhering to GATHER and PRISMA guidelines. They screened over 40,000 reports, incorporating peer-reviewed literature, grey literature, and data from international agencies. Key variables assessed included demographic characteristics (gender, age, sexuality), patterns of drug use, and exposure to structural risks such as homelessness, incarceration, and sex work. The team used meta-analyses to pool estimates for HIV, HBV, and HCV, as well as non-fatal overdose and other injecting-related diseases, providing a global and regional snapshot of the burden of disease.
The review reveals that HIV and hepatitis C remain highly prevalent among PWID, with an estimated 15.2% living with HIV and 38.8% having current HCV infection. Beyond these viruses, the findings highlight that 18.5% of PWID have experienced a recent non-fatal overdose, and 31.7% have suffered from recent skin or soft tissue infections. The data also illuminate the 'risk environment,' noting that nearly 60% of PWID have a history of incarceration and approximately 25% have experienced recent homelessness. These findings demonstrate that drug-related harm is deeply intertwined with broader social and economic vulnerabilities.
This study underscores that despite international goals to reduce blood-borne viruses and drug-related deaths, there have been no substantial improvements in the global burden of harm. The findings highlight an urgent need for scaled-up, evidence-based interventions like needle-syringe programs and opioid agonist treatment. Furthermore, the researchers argue that policy responses must move beyond individual behavioral change and instead address the structural, legal, and environmental factors—such as homelessness and the impacts of incarceration—that perpetuate these risks.
BACKGROUND: People who inject drugs are exposed to various and changing risk environments and are at risk of multiple harms related to injecting drug use (IDU). We aimed to undertake a global systematic review of the prevalence of IDU, key IDU-related harms (including HIV, hepatitis C virus [HCV], and hepatitis B virus [HBV] infection and overdose), and key sociodemographic characteristics and risk exposures for people who inject drugs. METHODS: We systematically searched for data published between Jan 1, 2017, and March 31, 2022, in databases of peer-reviewed literature (MEDLINE, Embase, and PsycINFO) and grey literature as well as various agency or organisational websites, and disseminated data requests to international experts and agencies. We searched for data on the prevalence, characteristics, and risks of people who inject drugs, including gender, age, sexuality, drug-use patterns, HIV, HCV, and HBV infections, non-fatal overdose, depression, anxiety, and injecting-related disease. Additional data were extracted from studies identified in our previous review. Meta-analyses were used to pool the data where multiple estimates were available for a country. We present country, regional, and global estimates for each variable examined. FINDINGS: We screened 40 427 reports published between 2017 and 2022, and the 871 eligible reports identified were added to the 1147 documents from the previous review. Evidence of IDU was documented in 190 of 207 countries and territories, and 14·8 million people (95% uncertainty interval [UI] 10·0-21·7) aged 15-64 years globally were estimated to inject drugs. Existing evidence suggests that there might be 2·8 million (95% UI 2·4-3·2) women and 12·1 million (95% UI 11·0-13·3) men who inject drugs globally, and that 0·4% (95% CI 0·3-1·3) of people who inject drugs identify as transgender. The amount of available data on key health and social risks among people who inject drugs varied widely across countries and regions. We estimated that 24·8% (95% CI 19·5-31·6) of people who inject drugs globally had experienced recent homelessness or unstable housing, 58·4% (95% CI 52·0-64·8) had a lifetime history of incarceration, and 14·9% (95% CI 8·1-24·3) had recently engaged in sex work, with substantial geographical variation. Injecting and sexual risk behaviour varied considerably geographically, as did risks of harms. Globally, we estimated that 15·2% (95% CI 10·3-20·9) of people who inject drugs are living with HIV, 38·8% (95% CI 31·4-46·9) have current HCV infection, 18·5% (95% CI 13·9-24·1) have recently overdosed, and 31·7% (95% CI 23·6-40·5) have had a recent skin or soft tissue infection. INTERPRETATION: IDU is being identified in a growing number of countries and territories that comprise more than 99% of the global population. IDU-related health harms are common, and people who inject drugs continue to be exposed to multiple adverse risk environments. However, quantification of many of these exposure and harms is inadequate and must be improved to allow for better targeting of harm-reduction interventions for these risks. FUNDING: Australian National Health and Medical Research Council.
Alex: Pooling can hide differences between studies. How did their approach handle that?
Sam: They used random-effects meta-analysis, which combines estimates while allowing underlying differences between studies. They built country estimates first, then weighted regional and global estimates using population information. They also carried uncertainty through calculations, rather than treating every input as a known value.
Alex: Can you give a concrete example where the choice of measurement changes what the estimate means?
Sam: Hepatitis C is a useful example. Antibodies indicate previous exposure, but they don’t establish that infection is still present. Hepatitis C RNA, the virus’s genetic material, indicates current infection. The authors used direct RNA estimates where available; elsewhere, they inferred current infection from antibody prevalence using an assumed clearance rate.
Alex: So “current infection” isn’t measured directly everywhere. That assumption matters when interpreting the global estimate.
Sam: It does. The authors also flag limited data on hepatitis C treatment uptake, which complicates estimating active infection. They argue the impact should be relatively small because countries where treatment likely made a sizable difference commonly had direct RNA measurements. That’s their assessment, not a reason to ignore the limitation.
Alex: With that distinction in place, what does the review say about the scale of injecting and its harms?
Sam: It estimates roughly fifteen million people globally injected drugs, among adults of working age. The uncertainty around that total is wide. Evidence of injecting exists across countries containing almost the entire global population, but far fewer countries have an estimate of how common it is.
Alex: Documenting that injecting occurs somewhere is different from measuring the population affected. What does the review estimate for infections among people who inject?
Sam: About one in six were estimated to be living with HIV. The burden varied considerably between regions and countries. For current hepatitis C infection, the global estimate was about four in ten.
Alex: Those averages convey a large burden. But if I’m planning services in one country, how misleading could the global profile be?
Sam: Quite misleading if treated as a local description. Opioids were generally the main drugs injected globally, but stimulants predominated in some countries. The authors highlight the Czech Republic and Puerto Rico as examples. That matters because evidence for effective interventions addressing dependent stimulant use is more limited than for opioids.
Alex: The drug profile changes the intervention options. And different injecting frequencies could change how much sterile equipment services need to provide.
Sam: The authors make that connection too. They also find substantial variation in syringe sharing and reuse. Their results argue against designing services around a single global model of injecting behaviour.
Alex: What does the wider lens add beyond infections and drug-use patterns?
Sam: Around six in ten people who inject drugs were estimated to have experienced incarceration during their lives. That’s lifetime exposure, not recent imprisonment. The review also documents substantial recent homelessness or unstable housing, with especially high estimated exposure in North America.
Alex: Those are associated risks, though. This review doesn’t isolate how much harm housing instability or incarceration causes, does it?
Sam: No. This is a synthesis of prevalence and exposure data, not a test of a structural intervention. The authors draw on evidence linking these environments with greater infection risk and poorer health. Their argument is that prevention planning must consider those conditions alongside individual behaviour.
Alex: And the health outcomes that usually receive less attention—did the review find enough evidence to say anything useful?
Sam: Available studies suggest substantial non-fatal overdose and mental health burdens. Around one in three were estimated to have had a recent skin or soft tissue infection. But coverage was uneven: depression and anxiety were assessed in only a small group of countries. High estimated burden and limited measurement coexist here.
Alex: Could we at least compare this review with its predecessor and judge whether prevention has improved?
Sam: Not cleanly. The authors report no substantial overall reductions in injecting prevalence or the major viral harms compared with the earlier review. But they repeatedly warn that changing study methods obstruct direct comparisons. A changed estimate can reflect changed measurement rather than a changed population.
Alex: So this is stronger as a map of burden and evidence gaps than as a clean time trend.
Sam: That’s my reading. Its breadth and systematic searching are strengths. But injecting is stigmatised and poorly captured by ordinary population surveys, so estimates often depend on indirect methods or targeted surveys. Those can be biased, and many countries lack regular updates. The review cannot make uneven source evidence uniformly reliable.
Alex: Given those limits, who should read the full document, and where should they start?
Sam: Researchers in addiction, infectious disease, surveillance, and harm reduction should read it. Start with the methods for selecting and pooling country estimates, then the country-level appendices relevant to your setting. Read the implications section for its argument about environmental risks. For everyone else, remember this: preventing injecting-related harm means addressing the conditions people live in, not only how they inject.
Alex: Thanks for listening.