De Granda-Orive et al.1 recently published their thoughts on self-citation in medical journals, advocating the continued penalization or exclusion of self-citations from bibliometric indicators, the basic aim of which, they argue, is to prevent editors from attempting to manipulate these indexes. We believe that some reflections must be made on the legitimate use of self-citation that particularly affects journals such as Archivos de Bronconeumología. Groups that aim not to publish in English have recourse to a very limited number of journals in other languages with a degree of international visibility. This is the case of publications in Spanish in the respiratory system setting: the Web of Science database that devised the impact factor lists 50 respiratory system journals, of which only one is in Spanish (accessed on http://www.accesowok.fecyt.es/), while of the 98 journals listed in the Scopus database, from which the SCImago Journal Rank (SJR) is calculated (accessed on http://www.scimagojr.com/), only 3 are in Spanish. Similarly, research papers from areas in which there are few specialized groups, even those of great scientific significance, necessarily tend to be published in a limited number of journals, making self-citation unavoidable. In this way, limiting self-citation in a journal widens the gap between journals of general interest in English, which receive more citations from different sources, and the more specialized or minority language journals. This is more a reflection of a difference in the interest in the field or knowledge of the journal's language than of the quality of that journal.2

In our opinion, the SJR calculation provides an excellent solution to this problem. This index, recently analyzed by the authors for use in respiratory system journals,3 has some features that make it particularly useful. First, it limits the number of self-citations used in its calculations, thus avoiding potential over-manipulation. Secondly, and perhaps more importantly, the SJR is applied by assigning citations a weighted value according to the influence of the citing journal. Thus, self-citations in a journal do not increase the influence of the journal and have little value in weighting that journal. Another alternative is to include other parameters, such as the h-index, in evaluating journals. This index, initially conceived to evaluate researcher activity, can also be applied to scientific journals4 and its manipulation by self-citation is difficult when h-index values are relatively high.5

To conclude, self-citation may indicate a legitimate body of authors contributing to certain journals for the reasons of specialization or language discussed above. Even this letter contains self-citations of journals and authors that, as we see it, are inevitable.