In the paper by Velez-Montoya et al. [1], the authors reported the first description of a case of intraocular infection in humans caused by an antibiotic-resistant strain of Streptococcus uberis. We would like to point out that the absence of any description of the method used for bacteria identification in this paper raises some concerns related to the possibility of a misidentification of this bacterium as a pathogen affecting the human eye. S. uberis is an environmental pathogen responsible for a high proportion of cases of clinical and subclinical mastitis in ruminant and nonruminant species [2]. The nutritional flexibility associated with an assortment of metabolic options allows S. uberis to occupy a discrete ecological niche [3]. Some studies have hypothesized that the flexibility of this bacterium under various environments and conditions might possibly favour infection also in humans [4, 5]. However, the evidence and putative role of S. uberis as a human pathogen are very limited and the methods used for the identification are frequently questionable [6]. In fact, phenotypic bacterial identification by commonly used systems such as Vitek, Facklam scheme, and similar conventional methods has been generally employed. However, in most cases of supposed human infections by S. uberis these techniques showed a low level of accuracy [6, 7]. Facklam described a case of human infection where all the isolates, previously classified as S. uberis, have been subsequently identified as Globicatella sanguinis [7] and a consistent body of evidence supports the notion that one of the most recurrent mistakes in the identification of gram-positive cocci, using phenotypic bacterial identification methods, is represented by the lack of distinction between S. uberis and Enterococcus spp. [8, 9]. A conventional scheme for the identification of S. uberis strains isolated from bovine milk samples and based on 11 biochemical tests also showed 6% frequency of misidentifications between S. uberis and Enterococcus faecalis [10]. On the other hand, infections caused by E. faecalis are largely described in the literature [11–17]. E. faecalis is known to represent a virulent pathogen frequently associated with endophthalmitis with very poor clinical prognosis [14, 18]. Endophthalmitis caused by E. faecalis has been described in a diabetic patient after biliary surgery [19], while other reports described ocular infections after cataract extractions [20–22]. Recently Bains et al. and Tang et al. also reported the emergence of endophthalmitis caused by E. faecium vancomycin-resistant strains [23, 24]. Indeed, the intraocular infections caused by E. faecium previously described in the literature are not in contrast with the image reported in Figure 1(b) of the paper by Velez-Montoya et al. [1]. In conclusion the phenotypic bacterial identification systems have been repeatedly found to fail the classification of E. faecalis on behalf of S. uberis. Thus, in our opinion the absence of any detailed description of the technique used for the bacterium identification in the paper by Velez-Montoya et al. [1] raises some concern since the method of identification may affect the validity and reliability of the diagnosis. Therefore we consider some information from the authors necessary regarding the description of the methods used for the identification, particularly considering that this might represent the first case of human intraocular infection caused by S. uberis and also in consideration that the pathogenic potential of this bacterium in humans is still under debate.
Read full abstract