The present study aimed at finding structural variations in the translated abstracts from Persian into English and comparing them with abstracts originally written in English and published in international journals in the area of medical sciences. To do this, 64 medical article abstracts (32 in Iranian journals, 32 in international journals) were analyzed and compared on the basis of Swales’ model (1990). More detailed analysis was done in the Introduction unit regarding CARS model (Swales, 1990) and also language features of each unit were identified. The IMRC (Introduction, Methods, Results, Conclusion) sequence was considered as the structural conventions for the analysis. The results showed that in terms of structural units, there was a significant difference in using the Methods unit between the two groups of abstracts (p= 0.002). Some variations were observed in the Introduction unit and language features. The data revealed that the translated abstracts from Persian into English in research medical articles meet the determined criteria for scientific writing while the original ones often ignore the criteria, although they are linguistically superior to the original English ones.