Re-emergence of HIV related to injecting drug use despite a comprehensive harm reduction environment: a cross-sectional analysis.

Background: In 2015, an outbreak of HIV was identified among people who inject drugs (PWID) in the Greater Glasgow and Clyde (GGC) area of Scotland, an area which distributes more than 1 million needles and syringes per year. This is the largest such incident in the UK for 30 years. Here, we provide an epidemiological analysis of the impact of the outbreak on HIV prevalence trends in the population and the individual and environmental risk factors associated with infection.

Methods: Four cross-sectional, anonymous, bio-behavioural surveys of almost 4000 PWID attending services providing injecting equipment across GGC between 2011 and 2018 were analysed. Participants were recruited by trained independent interviewers and eligible if they had a history of injecting drug use, either current (within the past 6 months) or historical. Interviewers asked participants questions about demographics, behaviours, and service use and to give a dried blood spot sample that was tested anonymously for the presence of blood-borne viruses. Our primary outcome measure was HIV infection status, as determined by the dried blood spot sample. We removed duplicates and participants with missing data and used all remaining participants to examine trends in prevalence of HIV infection, risk behaviours, and intervention coverage. We then did multivariate analysis with adjusted and unadjusted logistic regression to determine individual and environmental factors associated with HIV infection.

Findings: The overall GGC sample comprised 3641 PWID; data from 2712 PWID were available for multivariate analysis after further removal of duplicate participants and missing data. Between 2011 and 2018, HIV prevalence in GGC rose from 0·1% (95% CI 0·0-0·6) to 4·8% (3·4-6·2) overall, and from 1·1% (0·2-6·2) to 10·8% (7·4-15·5) in Glasgow city centre. Over the same period, the prevalence of cocaine injecting in all individuals in GGC in our sample rose from 16% (129/805) to 50% (291/583) overall, and from 37% (26/70) to 77% (117/153) in Glasgow city centre. HIV infection was more likely among PWID who had participated in surveys after the start of the outbreak in 2014 (adjusted odds ratio 3·4, 95% CI 1·7-6·7; p=0·00052), been homeless in the past 6 months (3·0, 1·7-5·0; p<0·0001), had had more than five incarcerations since they first began injecting (2·1, 1·2-3·7; p=0·0098); and had injected cocaine within the past 6 months (6·7, 3·8-12·1; p<0·0001). Age (per 1-year increase) was also a significant factor (1·1, 1·0-1·1; p=0·0016) but sex was not (1·7, 0·9-3·2; p=0·083).

Interpretation: Despite high coverage of harm reduction interventions, Glasgow has experienced a rapid rise in prevalence of HIV among its PWID population, associated with homelessness, incarceration, and a major shift to injection of cocaine. Robust surveillance through regular HIV testing of high-risk populations is crucial to ensure outbreaks are detected and rapid responses are informed by the best available evidence.

Funding: Health Protection Scotland.