
Methods: The prospective observational SHCS, founded in 1988, introduced a comprehensive questionnaire on drug use in April 2007. We studied men who have sex with men (MSM), heterosexual persons, current and former injection drug users (IDU), registered prior to April 2007, who had at least one follow-up visit with the drug use questionnaire until December 2012. Robust multivariable logistic regression models were used to assess the association between the different outcomes and co-factors, including drug use behavior.
Results: 6493 participants (69% men, 31% women; median age 44 years; 41% heterosexuals, 39% MSM, 20% IDU) had 58,641 visits during 30,002 years of follow-up. At baseline, 10% were ART-naïve, 81% on ART, and 9% off ART. 5272 (81%) participants reported no drug use (NODU), 1017 (16%) non-injecting drug use without injection drug use (NIDU), and 204 (3%) current injecting drug use (CIDU). Among the NODU, NIDU, and CIDU group, the drop-out rate from the cohort was 8.1%, 11.0%, and 16.7% (p<0.001), respectively. Among persons on ART, adherence to therapy was 92.7%, 89.9%, and 77.1% (p<0.001); the proportion of HIV RNA <50 copies/mL was 87.6%, 85.9%, and 87.1% (n.s.). Compared with NODU, the probability of drop-out, with death as competing risk, in NIDU was 1.34 (95% confidence intervals: 1.08, 1.65; p=0.007), and in CIDU 2.00 (1.39, 2.90; p<0.001). Time to start ART among ART-naïve participants was similar in the three groups, but the probability of interrupting ART >14 days was higher among NIDU and CIDU (p<0.001). The median percent of follow-up time with HIV-RNA <50 copies/mL, irrespective of ART status, in NODU, NIDU, and CIDU, was 95.4 (interquartile range: 77.0-100), 94.0 (71.5-100; p [NIDU vs NODU] =0.005), and 90.3 (64.6-100; p [CIDU vs NODU] =0.001). Compared with NODU, the probability of death, with drop-out as competing risk, in NIDU was 2.46 (1.81, 3.36; p<0.001), and in CIDU 4.31 (2.79, 6.66; p<0.001).
Conclusion: Not only injection drug use, but also non-injection drug use was associated with a lower retention rate in the cohort, more problems to maintain successful ART, and higher mortality.

R. Weber,
None
M. Battegay, None
H. Furrer, None
E. Castro, None
A. Calmy, None
P. Vernazza, None
E. Bernasconi, None
B. Ledergerber, None