Seoul Virus Infection in Humans, France, 2014–2016

We report detection of Seoul virus in 3 patients in France over a 2-year period. These patients accounted for 3 of the 4 Seoul virus infections among 434 hantavirus infections (1.7%) reported during this time. More attention should be given to this virus in Europe where surveillance has been focused mostly on Puumala and Dobrava-Belgrade hantaviruses.

likely source of human infection. Organs from 10 rats sampled at this unit were positive for SEOV by nested reverse transcription PCR (7). A partial large RNA sequence obtained from a rat was identical to that obtained from the case-patient. Only 1 wild rat caught near the hen house was sampled; results were negative for SEOV.
We obtained complete small RNA coding domain sequences (GenBank accession nos. KX064270-KX064275) from specimens from case-patients 2 and 3 (samples were not available for case-patient 1); specimens from the pet brown rat and raised brown rats suspected to be sources of infection for case-patients 1 and 2; and specimens from 2 wild brown rats suspected to sources of infection for a serologically confirmed human infection with SEOV detected in 2014 (6). Using MEGA version 5.1 (8) and a generalized time-reversible model with a gamma distribution and 5 rate categories (according to the best fit substitution model proposed), we performed phylogenetic analysis of the small RNA coding sequence. This analysis confirmed that all virus strains detected were SEOV ( Figure 2).

Conclusions
Four hantaviruses (Puumala [PUUV], SEOV, Tula, and Nova viruses), have been detected in France; the first 3 viruses were associated with humans, and most human infections were with PUUV (9). Human SEOV infection in Europe was initially reported in 2013 (5); however, human *ALT, alanine aminotransferase; AST, aspartate aminotransferase; D, day after symptom onset; GGT -glutamyl transferase; LDH, lactate dehydrogenase; NA, not available; †Dysphagia, diarrhea, jaundice, sore throat, rash, bleeding manifestation, and loss of consciousness were not observed for these patients. ‡Patient 1 also had posttraumatic rhabdomyolysis: myoglobin level 869.5 g/mL at D3 (reference range 16.0-116.0 µg/mL) and creatine kinase level 9,444 g/mL at D5 (reference range 39-308 g/mL).  SEOV infections were previously suspected by serologic analysis in France (6 cases were reported during 1977-1996). SEOV was also detected in rodents from several areas in France ( Figure 1) (10)(11)(12)(13). Detection of SEOV in 3 case-patients in our study and a recent report of a human infected with SEOV (6) within a 2-year period indicate that SEOV infections in humans are not uncommon in France. These infections accounted for 4 (1.7%) of 234 cases of hantavirus infection, mostly with PUUV, detected serologically or virologically during the same period in France. Detection of such cases might be caused by improvements in detection of SEOV infections, rather than by emergence of SEOV infections (SEOV antigen in serologic assays and molecular detection have been used at NRC since 2012).
Some infections with SEOV are probably missed in France. Commercials kits are used for hantavirus serologic diagnosis by 15 public hospital or private clinical laboratories in France. Eleven of these hospitals used a POC Puumala IgM rapid test (Reagena, Toivala, Finland). Four other hospitals used IgM and IgG ELISA kits with mixtures of recombinant antigens, including those from SEOV or Hantaan virus strains.
Most (83.4%) diagnostic tests in 2014 were requested for patients in the area of France to which PUUV is endemic. Consequently, some SEOV infections might have been be missed because hantavirus infection is rarely suspected outside this area. Samples with positive results for SEOV are sent to NRC for diagnostic confirmation and surveillance purposes. The 3 case-patients we report initially showed positive results by ELISA at local laboratories but then showed negative results by a ReaScan Puumala IgM test (Reagena) at NRC (Table 2). Therefore, SEOV infections were probably not detected initially by the 11 local laboratories who used the PUUV IgM Rapid Test, which led to underestimation of SEOV infections in humans in France.
This negative result could have also occurred in other countries in Europe that used the same test. Consequently, use of a pan hantavirus serologic assay is preferred. Furthermore, PUUV was detected by molecular techniques for most PUUV-infected patients during the acute phase of the disease (14). Thus, although there are no similar data for SEOV, to avoid misdiagnosis, we suggest using molecular diagnostic tests to more specifically detect hantavirus infections.  Ragnaud et al. (10), Le Guenno (11), and Bour et al. (6); solid blue circles indicate virologically confirmed human infections with SEOV reported in this study (patients 1, 2, and 3); solid yellow square indicates virologically confirmed human infection with SEOV reported by Macé al. (5); solid red diamonds indicate virologically confirmed SEOV infections in brown rats reported in this study; and open blue diamonds indicate virologically confirmed SEOV infections in brown rats reported by Heyman et al. (12) and Dupinay et al. (13).

DISPATCHES
Furthermore, epidemiologic investigations showed that several animal traders in Europe sold pet rats or food rats to case-patients 2 and 3. These investigations also identified weak traceability of rat batches, absence of health-monitoring data for rats, and no information for the 3 case-patients on zoonotic risks for infection. Awareness of rat owners, traceability, and health monitoring of these animals, as performed for laboratory rats (15), should be improved. Triangles indicate sequences of strains detected in wild brown rats (this study) associated with a serologically confirmed human infection reported by Bour A et al. (6); squares indicate sequences of strains detected in case-patient 2 and in his pet brown rat; and circles indicate sequences of strains detected in case-patient 3 and 1 of his farmed brown rats. Bootstrap percentages >70% (from 500 resamplings) are indicated at each node; GenBank accession numbers are indicated for reference strains.Scale bar indicates nucleotide substitutions per site.