Classical content-based recommender systems (CB) help users to find preferred items in overloaded search spaces, comparing items descriptions with user profiles. However, classical CBs do not take into account that user preferences may change over time influenced by the user context. This paper propounds to consider context-awareness (CA) in order to improve the quality of recommendations, using contextual information obtained from streams of status updates in microblogging platforms. A novel CA-CB approach is proposed, which provides context awareness recommendations based on topic detection within the current trend interest in Twitter. Finally, some guidelines for the implementation, using the Map Reduce paradigm, are given.