To the Editor: In June 2006, the World Health Organization (WHO), the Food and Agriculture Organization of the United Nations (FAO), and the World Organisation for Animal Health (OIE) convened an Expert Consultation to consider the risks to human health represented by the use of antimicrobial drugs in aquaculture. This would, therefore, appear to be an opportune time to reexamine some of the arguments that have been presented with respect to the assessment of these risks.

In their contributions to the debate regarding the risks associated with the use of antimicrobial agents in aquaculture, Angulo (1), Angulo and Griffin (2), Ribot et al. (3), and, more recently, Cabello (4) have argued that the available molecular evidence suggests that the flo gene that encodes chloramphenicol and florfenicol resistance in Salmonella enterica serovar Typhimurium DT104 (DT104) originally emerged in Japanese aquaculture and may have transferred horizontally from this host to DT104. This argument also appears in the report of the WHO/FAO/OIE consultation (ftp://ftp.fao.org/ag/agn/food/aquaculture_rep_13_16june2006.pdf). These authors (1–4) have based their argument on the assertions that florfenicol was first used in Japan and that flo gene–mediated resistance to this agent was first identified in bacteria isolated from Japanese fish farms.

In attempting to identify the date of the emergence of florfenicol resistance in Japanese aquaculture, Angulo and Griffin (2) state that florfenicol had been used in this country since the early 1980s. However, Schering Plough, the manufacturer of florfenicol, reports first marketing this agent for aquacultural use in Japan in 1990 (D. Schofield, pers. comm.), and this date for the introduction of florfenicol is also provided by Kim et al. (5). It should, however, be noted that the flo gene encodes resistance to both florfenicol and chloramphenicol and, therefore, flo-containing bacteria could be selected for by the use of either agent. As a consequence, arguments about the chronology of the first use of florfenicol may have limited relevance.

Florfenicol resistance in Japanese aquaculture (5) was first reported in strains of Pasteurella piscicida (now renamed Vibrio damsela). The sequence of the gene that encoded florfenicol resistance in these strains (6) was demonstrated to have a 97% nucleotide sequence similarity to that found in strains of DT104 resistant to ampicillin, chloramphenicol, streptomycin, sulfonamide, and tetracycline (ACSSuT DT104) (7). The available data (5,8) allow a reasonably accurate estimate of the date when these florfenicol-resistant strains first emerged in Japanese aquaculture. The strains of florfenicol-resistant P. piscicida were first isolated in Japan in 1992 (5). However, a previous study (8) had demonstrated 100% susceptibility to florfenicol of P. piscicida strains isolated from 1989 through 1991. Because this study examined 175 P. piscicida strains isolated from fish farms distributed over a wide geographic area in Japan, the data it generated provide strong support for the conclusions that flo-mediated florfenicol resistance in P. piscicida first emerged in Japanese aquaculture in 1992 (5). However, the presence of the floR gene has been demonstrated in a strain of ACSSuT DT104 isolated in the United States in 1985 (3). Thus, the floR gene was present in DT104 strains isolated in the United States at least 7 years before the first bacteria containing this gene (6) were isolated from an aquaculture setting in Japan (5).

There are, however, data indicating that the flo gene was present in terrestrial bacteria associated with humans long before it was detected in multidrug -resistant DT104. A gene with a 95%–97% nucleotide identity with the flo gene of DT104 was detected in the Inc C plasmid R55 (9). This plasmid was originally identified in a chloramphenicol-resistant strain of Klebsiella pneumoniae isolated from a person in Paris in 1969 (10).

The earliest report of the isolation of a bacterium whose florfenicol resistance was encoded by a flo gene (6) and the earliest accession date for a flo gene sequence in GenBank (www.ncbi.nlm.nih.gov/Genbank) both related to P. damsela isolated from Japanese aquaculture. Publication and accession dates do not, however, constitute evidence of the date of the first isolation of a bacterium containing this gene. Analysis of the available chronological and molecular data presented here indicates that a variant of the floR gene in DT104 (9) was present in a terrestrial bacterium isolated in 1969 (10), 23 years before the first isolation, in 1992, of a bacterium associated with aquaculture that contained this gene (5). It further demonstrates that this gene was present in strains of DT104 isolated in 1985. Thus, these data provide no support for the arguments (1–4) that implicate aquacultural use of florfenicol or the subsequent occurrence of florfenicol-resistant P. damsela in the emergence of the flo gene in DT104.