To the Editor: Leptospirosis is a major emerging infectious disease with a worldwide distribution (1). It is a systemic disease of humans and domestic animals (2). Regarded globally as a zoonosis because it is acquired by humans from contact with animals or from water contaminated with the urine of infected animals, it is presumed to be the most widespread zoonotic disease in the world (1,2). Species such as mice (Mus spp.) and rats (mainly Rattus norvegicus and R. rattus) serve as reservoirs for their host-related serovars (3).

Human patients usually exhibit a nonspecific self-limiting febrile illness; however, in 5%–10% of cases, severe forms of the disease develop, including Weil disease and severe pulmonary hemorrhagic syndrome. Case-fatality rates for Weil disease and severe pulmonary hemorrhagic syndrome are >10% and >74%, respectively (4,5).

Because leptospirosis has been found in humans in the Canary Islands (J. Alcoba-Flores, pers. comm.), detecting carrier animals is vital to the understanding of enzootic and epizootic leptospirosis in this environment. We examined the possible role of the rodent species found in the Canary Islands in the transmission of this pathogen to determine the risk for humans in these islands.

A total of 149 wild rodents (74 R. rattus and 75 Mus domesticus) were captured during 20009–2010 from 4 of the Canary Islands (Tenerife, El Hierro, La Gomera, Lanzarote). Urinary bladders of the animals were collected and preserved in 100% ethanol. Genomic DNA was extracted by using the Fast DNA Kit (BIO 101 Systems: MP Biomedicals, Santa Ana, CA, USA).

The lipL32 fragment (497 bp), which is present only in pathogenic leptospires, was amplified according to the method of Bomfim et al. (6) by using a MyCycler thermocycler (Bio-Rad, Hercules, CA, USA). L. interrogans serovar Icterohemorragiae (RGA strain) was used as a positive control.

Twenty-two samples were positive for Leptospira spp., indicating a general prevalence of 14.8% in the rodents. Although the prevalence was higher in rats (20.3%) than in mice (9.3%) (Table), the difference was not significant (χ² test). Positive samples were obtained from all the studied islands and for both host species in all of them (Table), without significant differences in the prevalences between host species or between islands.

To confirm the amplified products belonged to pathogenic Leptospira spp., we sequenced some amplicons. Sequencing reactions were performed for both strands at the University of La Laguna Genomic Service. When the sequences were compared, 2 different sequences were obtained. The first sequence, L19 (GenBank accession no. HQ231747), from rats, clustered with L. interrogans serovar Copenhageni (GenBank accession no. AE016823) and different serovars of the same species by BLAST (99% identity). Previous results associate L. interrogans serovar Copenhageni with Rattus spp (7). However, the sequence obtained from mice (GenBank accession no. HQ231748), L47, showed a 100% BLAST identity with L. borgpetersenii (GenBank accession nos. DQ320625.1 and DQ286415.1).

New and published Leptospira sequences were aligned with the multiple alignment program ClustalW in MEGA3.1 (8), and minor corrections were made manually. The alignment for the 497-bp fragment starts at nt position 208, with respect to the complete sequence of the lipL32 (AY609332), and ends at nt position 705.

Phylogenetic relationships were inferred by using the neighbor-joining distance method with MEGA3.1. At least 1,000 bootstrap replicates were used to infer statistical support at branch nodes. The consensus tree yielded 3 monophyletic groups clearly separated by high bootstrap values. The first clade was formed by L. interrogans, L. kirschneri, and L. noguchii (93% bootstrap value). The sequence L19 was included in the L. interrogans node (92% bootstrap value). The second clade included L. borgpetersenii and L. weilii as a monophyletic group (97% bootstrap value). The sequence L47 clustered with L. borgpetersenii DQ286415 with a high bootstrap value (82%). These results are in accordance with those obtained by Haake et al. (9) based on lipL32. Finally, L. santarosai sequences formed the third separate clade (100% bootstrap) (data not shown).

Although the method we used does not enable specific identification, determining the most similar species by BLAST is needed for control programs. L. interrogans serovar Copenhageni is the predominant infecting serovar among patients with severe leptospirosis (7), and L. borgpetersenii is also commonly acquired from mice.

On the basis of these findings, the global distribution of Leptospira spp. must be revised to include the Canary Islands, with rodents as natural hosts. Because pathogenic Leptospira spp. were detected on every island studied and in both analyzed species, R. rattus and M. domesticus, the distribution of this pathogen likely extends to even the islands not studied. The high incidence found suggests that rodents play a role in transmission of human leptospirosis. Further studies are needed to identify other possible reservoir hosts and to determine the risk areas for acquiring pathogenic leptospires in the Canary Islands.