Chinese men who have sex with men (MSM) have a higher prevalence of depression than the general population. Long-term depression symptoms could increase the risk of having physical harm, self-harm and suicide. Therefore, depression issues among MSM need to be solved urgently.

To explore the latent classes of depression, and the influence of HIV-related knowledge and anticipated stigma on depression among MSM.

Usingthe General Information Questionnaire, HIV Knowledge Questionnaire, Anticipated HIV Stigma Scale and 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10) , an online questionnaire survey was conducted from December 2020 to March 2021 via the same-sex social network of Blued7.5. A total of 1396 MSM with HIVinfection or unclear status of HIVinfection were sampled by use of convenience sampling to attend the survey. The latent classes of depressive symptoms were analyzed by latent class analysis. Multinomial Logistic regression analysis was used to explore the correlation of the latent class of depression with HIV-related knowledge or anticipated HIV stigma.

Finally, 1 394 cases (99.9%.) who returned responsive questionnaires were included. The 10 items of CES-D-10 were taken as observed indicators, and exploratory latent class analysis was employed to identify the latent classes of depression based on model from 1 to 5 classes. The 4-class model was finally selected, and the probabilities of each class were 40.1% (no obvious depression) , 21.6% (possible risk of depression) , 28.0% (possible mild depression) , and 10.3% (possible moderate to severe depression) . The accuracy of the classification was assessed by test for homogeneity according to the cut-off value of CES-D-10 (0=no depression, 1=depression) and the results of latent class analysis (0= no obvious depression and possible risk of depression, 1= possible mild depression and possible moderate to severe depression) , and the result showed that Kappa=0.735 (P<0.001) . Multinomial Logistic regression analysis indicated that a high level of HIV-related knowledge (OR=0.926, P=0.001) , undergraduate attainment or above (OR=0.642, P=0.003) were related to possible risk of depression. The anticipated HIV stigma (OR=1.594, P<0.001) and being married (OR=0.593, P=0.026) were related to possible mild depression. A high level of HIV-related knowledge (OR=0.935, P=0.026) , anticipated HIV stigma (OR=2.239, P<0.001) and no long-term employment (OR=1.518, P=0.045) were related to possible moderate to severe depression.

The depression in MSM population presented obvious category characteristics, which may be a new idea for tailoring interventions to depression in MSM. Relevant authorities may cooperate with social media to deliver interventions incorporated with HIV-related knowledge and anticipated HIV stigma to prevent and control the development of depression in MSM.

中国男男性行为（MSM）者的抑郁发生率高于一般人群，长期的抑郁症状不仅会对身体造成损害，还会增加自残、自杀等行为的发生，因此MSM者的抑郁问题亟待解决。

探索MSM者抑郁的潜在类别，并进一步探究HIV相关知识及预期污名化对抑郁的影响。

本研究采用便利抽样的方法，于2020-12-16至2021-03-02通过同性社交软件Blued7.5平台选取1 394例未感染HIV或感染状态未知的MSM者进行线上问卷调查，包括一般资料调查表、HIV知识问卷、HIV预期污名量表、简版流调中心抑郁量表（CES-D10），对抑郁进行潜在类别分析，并采用多因素Logistic回归分析探讨不同抑郁类别与HIV相关知识及预期污名化的相关性。

最终收回问卷1 396份，有效问卷1 394份，有效率为99.9%。以CES-D10的10个条目为外显指标，依次选取1~5个潜在类别模型对MSM者抑郁状况进行探索性潜在类别分析，最终选择含有4个类别的模型，各组的类别概率："无明显抑郁"占40.1%，"可能存在抑郁风险"占21.6%，"可能存在轻微抑郁"占28.0%，"可能存在中至重度抑郁"占10.3%。为检验该分类的准确性，将根据CES-D10的截断值所判断的抑郁结果（0=无抑郁，1=存在抑郁）与潜在类别分析所得结果进行一致性检验（0=无明显抑郁+可能存在抑郁风险，1=可能存在轻微抑郁+可能存在中至重度抑郁），结果显示Kappa=0.735（P<0.001）。多因素Logistic回归分析结果显示，高水平HIV相关知识（OR=0.926，P=0.001）、本科及以上受教育程度（OR=0.642，P=0.003）是可能存在抑郁风险的影响因素；HIV预期污名化（OR=1.594，P<0.001）、已婚（OR=0.593，P=0.026）是可能存在轻微抑郁的影响因素；高水平HIV相关知识（OR=0.935，P=0.026）、HIV预期污名化（OR=2.239，P<0.001）、无固定工作（OR=1.518，P=0.045）是可能存在中至重度抑郁的影响因素。

MSM者的抑郁具有明显的类别特征，本研究可为针对MSM者的抑郁状况施行精准干预提供新思路。相关部门可考虑结合社交媒体针对MSM者的HIV相关知识及预期污名化进行干预，从而预防及控制MSM者抑郁的发生和发展。