In 2007, the U.S. National Library of Medicine (NLM), which created and maintains the MEDLINE database, raised Nutrition to the category of thematic area and consequently generated a new Medical Subject Heading (Mesh): ‘‘Nutritional Sciences.’’ This situation brought about a change in the way scientific documentation for this area of knowledge is indexed, including the disappearance of the Mesh descriptor ‘‘Nutrition.’’ At the same time, new more sensitive descriptors appeared for indexing all the literature that had previously included the word nutrition. Therefore, the need arose for a thematic search filter that would guarantee efficient and correct access to this scientific literature in the most appropriate manner (1). The former nutrition search filter consisted of 30 Mesh term until 31 December 2008. The Mesh descriptors were structured by means of the Boolean union of the Mesh terms associated with the connector ‘‘OR.’’ On 1 January 2009, the nutritional filter included 26 Mesh terms for Human Nutrition (see Annex). This new configuration was the result of the annual review NLM carried out on its Thesaurus. The sensitivity of this filter can increase if the Descriptors (Mesh) are used as a Major Topic (Majr). The architecture of the filter and its evaluation were based on a previously tried and tested methodology (2,3) and took into account possible application and valuation errors (4). Due to its modular structure (Boolean equation), the filter can be easily modified by adding or subtracting any of its elements and can also be updated at any time. Through its use, the filter can be adapted to different areas of health science, and possible improvements can thus be evaluated. It must also be updated according to the new versions of the NLM Thesaurus (5). In conclusion, a simple (copy and paste) thematic filter has been created to enable the recovery of scientific production on nutrition contained in MEDLINE through PubMed, with a high degree of sensitivity and specificity. Furthermore, slight modifications to the filter will enable its use, either total or partial, in other Bibliographical Data Bases that share the NLM Thesaurus (EMBASE, The Cochrane Library Plus, LILACS, etc).