Objectives Traditionally, subjects' migration status has usually been defined on the basis of their registered residency status. We attempted to redefine migration based on the duration of residency in their cities of migration and to explore more precisely the impact of migration on HIV infection risk in men who have sex with men (MSM).

Methods A multisite cross-sectional study was conducted during 2012–2013 in seven Chinese cities. Questionnaire surveys were conducted and blood was drawn to test for antibodies to HIV, syphilis and herpes simplex virus-2 (HSV-2). MSM who were unregistered local residents and had resided in their cities of migration for ≤1 or >1 year were defined as migrant MSM, or transitional MSM, respectively.

Results Compared with transitional MSM and local MSM, migrant MSM had poorer HIV knowledge and higher rates of high-risk behaviour, including earlier sexual debut, multiple sexual partners, participation in commercial sex and recreational drug use. Multivariate logistic regression analysis showed that HIV prevalence among migrant MSM was higher than local MSM (p<0.05). This relationship, however, did not hold for transitional MSM and local MSM (p>0.05). Male sex work, recreational drug use, syphilis infection and HSV-2 infection were independently associated with HIV infection among migrant MSM.

Conclusions Non-local MSM with shorter residence were at greater risk of HIV acquisition. More focus should be placed on HIV behavioural interventions targeting non-local MSM with temporary residence.