OBJECTIVETo determine the prevalence of loss to follow-up (PLF) and risk factors among human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) patients from 1989 to 2012 in Baoshan city, Yunnan province.

METHODSThe epidemic and follow-up databases of HIV/AIDS patients by the end of 2012 were downloaded from "the history card downloading site" of HIV/AIDS database in China Information System for Disease Control and Prevention and obtained the related data of patients from 1989 to 2012 who had local residence in Baoshan city. These data included demographic characteristics (genders, age at the time of HIV testing positive, and occupation, marital status, and education levels, et al), transmission routes, and disease staging, antiretroviral therapy (ART), and sources of samples, the first CD4(+)T cell counts, and status of follow-up, et al. Descriptive epidemiological study was used to describe the general characteristics of loss to follow-up. Multivariable Cox regression was used for determining risk factors associated with loss to follow-up.

RESULTSA total of 3 295 eligible HIV/AIDS patients from 1989 to 2012 were included. The accumulative study person-year was 11 416.59 years, 222 HIV/AIDS patients were lost to follow-up, and the PLF was 0.019 4/ person years (py). The highest PLF was 0.052 8/py in 2008, the lowest was 0.006 2/py in 2012. The lost patients included 56.76% (126/222) males and 43.24% (96/222) females, the PLFs were 0.020 4/py, 0.018 3/py, respectively. Baoshan city, other cities in Yunnan province, and other provinces, foreign nationality as the family register reached 53.60% (119/222) , 28.83% (64/222) , and 5.86% (13/222) , 11.71% (26/222) , respectively, and their PLFs were 0.012 5/py, 0.046 3/py, and 0.053 6/py, 0.095 6/py, respectively. Receiving ART and not receiving ART occupied 6.76% (15/222) , 93.24% (207/222) , respectively, and the PLFs were 0.001 9/py, 0.0588/py. AIDS and HIV staging standed at 8.11% (18/222) , 91.89% (204/222) , respectively, and the PLFs were 0.003 3, 0.034 5/py. The first CD4(+)T cell counts < 200, 200-350, and > 350 cells /ml accounted for 4.95% (11/222) , 73.87% (164/222) , 21.17% (47/222) , respectively, and the PLFs were 0.004 8/py, 0.024 0/py, 0.020 3/py. The results of multivariable Cox regression showed the risks of loss to follow-up (RLFs) of family register as other cities in Yunnan province (HR = 3.11, 95%CI:2.28-4.25) , other provinces (HR = 2.55, 95%CI:1.42-4.56) , and foreign nationality (HR = 2.12, 95%CI:1.35-3.33) higher than that of Baoshan city, respectively. The RLFs of not receiving ART (HR = 20.83, 95%CI:11.74-36.96) and HIV staging (HR = 3.61, 95%CI:1.82-7.16) were higher than those of receiving ART and AIDS staging, respectively, moreover, the RFLs of the first CD4(+)T cell counts between 200-350 cells/ml (HR = 0.35, 95%CI:0.15-0.82) and the first CD4(+)T cell counts > 350 cells/ml (HR = 0.29, 95%CI:0.12-0.72) were less than that of first CD4(+)T cell counts < 200 cells /ml, respectively. The RLF of transmission route as injecting drug (HR = 0.60, 95%CI:0.41-0.88) was less than that of heterosexual contact.

CONCLUSIONOverall, the prevalence of loss to follow-up among HIV/AIDS patients shows a downward trend, moreover, patients of outsiders, heterosexual contact, HIV staging, baseline CD4(+)T cell counts < 200 cells/ml are at higher risk of loss to follow-up.