OBJECTIVETo analyze factors associated with antiretroviral therapy (ART) initiation and its timeliness among HIV sero-discordant couples in high HIV prevalence regions in China.

METHODSData from the national HIV epidemiology and treatment databases by Dec 31, 2013 were used to collect HIV serodiscordant couples' baseline and follow-up informations in 40 counties which had reported 200 HIV cases cumulatively from Yunnan, Guangxi, Sichuan, Henan, and Xinjiang. Positive couples were divided into ART group and Non-ART group based on their ART initiation status. Sero-discordant couples that were reported between January 1, 2012 and October 31, 2013 and initiated ART treatment by 2013, were classified into two categories: 'having initiated ART within two months of receiving their HIV diagnosis', and not accessing timely treatment as 'having not initiated ART within two months of receiving their HIV diagnosis'. Multivariate logistic regressions were used to analyze potential factors associated with ART initiation and serodiscordant couples' timeliness to treatment. Odds ratios (95% confidence internal) were used to measure the associations.

RESULTSA total of 10 213 HIV-positive individuals were included in this study, among whom 73.9% (7 550/10 213) were males and 26.1% (2 663/10 213) were females, 66.4% (6 780/10 213) had initiated ART and 33.6% (3 433/10 213) had not. There were 1 733 serodiscordant couples who were reported during January 1, 2012 to October 31, 2013 had initiated ART. Among those 64.9% (1 125/1 733) had successfully initiated ART within two months of receiving their HIV diagnosis and 35.1% (608/1 733) had not. Multivariate logistic regression analysis showed that those being male were 0.81(0.71-0.92) less likely to initiate ART, as compared with those being female. Those being Yi ethnicity were 0.29(0.25-0.35) less likely to initiate ART, as compared with those being Han ethnicity. Those being Uygur ethnicity were 1.57 (1.28-1.93) times more likely to initiate ART, as compared with those being Han ethnicity. Those engaging the other jobs were 0.85(0.75-0.96) less likely to initiate ART, as compared with those being peasant. Those being diagnosed in medical institutions were 0.61 (0.53-0.71) less likely to initiate ART, as compared with those being diagnosed in VCT. Those having CD4⁺ T cells of 250-349, 350-550 and > 550 were 0.75 (0.64-0.89), 0.17 (0.14-0.19), 0.10 (0.08-0.11) less likely to initiate ART, respectively, as compared with those having CD4⁺ T cells of < 250. Those having duration of follow-up of 13-36 months and ≥ 37 months were 0.55(0.48-0.63) and 0.32 (0.28-0.37) less likely to initiate ART respectively, as compared with those having duration of follow-up of ≤ 12 months. Multivariate logistic regression analysis showed that those being Yi and Uygur ethnicity were 0.63 (0.44-0.91) and 0.40(0.29-0.56) less likely to initiate ART timely respectively, as compared with those being Han ethnicity. Those being infected through injecting drug use were 0.64 (0.47-0.86) less likely to initiate ART timely, as compared with those being infected through heterosexual intercourse. Those being diagnosed in other institutions were 0.58(0.43-0.78) less likely to initiate ART timely, as compared with those being diagnosed in VCT. Those having CD4⁺T cells of 250-349, 350-550 and > 550 were 0.75(0.56-1.00), 0.44(0.34-0.58), and 0.31(0.22-0.45) less likely to initiate ART timely respectively, as compared with those having CD4⁺ T cells of < 250.

CONCLUSIONBeing males, being Yi ethnicity, engaging other jobs, being diagnosed in medical institutions, having high CD4⁺ T cells and having long duration of follow-up were risk factors for initiating ART among serodiscordant couples.Being Yi and Uygur ethnicity, being infected through injecting drug use, and having CD4⁺ T cells were risk factors for initiating ART timely among serodiscordant couples.

Anti-Retroviral Agents ; Asian Continental Ancestry Group ; CD4 Lymphocyte Count ; China ; Disease Transmission, Infectious ; Family Characteristics ; Female ; HIV Infections ; Humans ; Logistic Models ; Male ; Occupations ; Prevalence ; Risk Factors ; Street Drugs ; Time-to-Treatment