Objective: To construct a multi-state Markov model and analyze the disease outcomes and its influencing factors in HIV infected individuals receiving antiretroviral therapy. Methods: A retrospective cohort analysis was conducted in HIV infected individuals receiving antiretroviral therapy in Luzhou of Sichuan province from 2010 to 2021. The disease status was divided into CD4+T lymphocytes (CD4) counts >500 cells/μl, 350-500 cells/μl, 200-349 cells/μl, ≤199 cells/μl and death indicated by S1-S5 in turn. A reversible continuous-time discrete-state multi-state Markov model was constructed for the analysis of disease progression features. Results: A total of 7 542 HIV infected individuals receiving antiretroviral therapy were included, and the median age (Q1, Q3) was 53.4 (41.2, 64.5) years old. The transition intensity of S3→S2 was higher. During follow-up, the transition probability of S4→S5 increased gradually. Influencing factors analysis of disease outcomes in HIV infected individuals receiving antiretroviral therapy showed that compared with individuals 15-24 years old, the transition intensities of S2→S1, S3→S2 and S4→S3 were lower and the transition intensity of S3→S4 was higher in individuals ≥45 years old. Compared with single individuals, the transition intensities of S3→S2 and S4→S3 were higher and the transition intensities of S3→S4 and S4→S5 were lower in married individuals. The transition intensity of S1→S2 was higher in individuals with baseline CD4 counts ≤500 cells/μl than in individuals with baseline CD4 counts >500 cells/μl. The transition intensity of S3→S4 in individuals diagnosed during 2011-2015 was lower than that in individuals diagnosed in 2010 and before. Conclusions: HIV infected individuals receiving antiretroviral therapy tended to shift to the previous disease status, suggesting that antiretroviral therapy was conducive to immune reconstitution. Older age (≥45 years old), being married, low baseline CD4 counts and being diagnosed in 2010 and before were the risk factors for disease progression.
目的： 构建多状态马尔科夫模型分析抗病毒治疗HIV感染者的疾病转归及其影响因素。 方法： 对四川省泸州市2010-2021年抗病毒治疗HIV感染者进行回顾性队列分析，将疾病状态划分为CD4+T淋巴细胞（CD4）计数>500、350~、200~、≤199个/μl和死亡（S1~S5依次表示），构建连续时间离散状态的可逆多状态马尔科夫模型分析疾病进展规律。 结果： 共纳入7 542例抗病毒治疗HIV感染者，年龄M（Q1，Q3）为53.4（41.2，64.5）岁。感染者S3→S2的转移强度较大。随访期间，感染者S4→S5的转移概率逐渐升高。抗病毒治疗HIV感染者疾病转归的影响因素分析结果显示，与15~24岁者相比，≥45岁者S2→S1、S3→S2和S4→S3的转移强度较低，S3→S4的转移强度较高；与单身者相比，已婚者S3→S2和S4→S3的转移强度较高，S3→S4和S4→S5的转移强度较低；基线CD4计数≤500个/μl者S1→S2的转移强度高于>500个/μl者；2011-2015年被诊断者S3→S4的转移强度低于2010年及以前被诊断者。 结论： 抗病毒治疗HIV感染者倾向于向上一疾病状态转移，提示抗病毒治疗有利于免疫重建；较大年龄（≥45岁）、已婚、基线CD4计数较低和2010年及以前被诊断均是其疾病进展的危险因素。.