Acute Gastroenteritis Caused by GI/2 Sapovirus, Taiwan, 2007

To the Editor: Sapovirus is an etiologic agent of human gastroenteritis. Although many of the previously reported cases were of mild, sporadic infections in young children (1–3), several recent sapovirus-associated gastroenteritis outbreaks have affected adults, which suggests that the virus’s virulence, prevalence, or both, may be increasing (4–6). In this study, we describe a sapovirus-associated outbreak of gastroenteritis that occurred during May 4–8, 2007, and involved college students in northern Taiwan. 
 
A total of 55 students had clinical symptoms of gastroenteritis, including diarrhea (45), vomiting (22), abdominal cramps (17), and fever (2). The clinical symptoms continued for up to 10 days (mean 4.7 days). Stool specimens were collected from 8 of 55 students on May 8 (Table). Initially, the specimens were screened for bacteria, rotavirus, and norovirus, but all specimens were negative for these pathogens. The 8 stool specimens were then examined by electron microscopy (EM), and 1 was positive for calicivirus-like particles. 
 
 
 
Table 
 
Clinical symptoms and laboratory diagnosis results for sapovirus-related outbreak among college students, northern Taiwan, May 2007*† 
 
 
 
To confirm the EM results, we performed reverse transcription–PCR (RT-PCR), real-time RT-PCR, and sequence analysis as previously described (7). Briefly, purified RNA (10 μL) was reverse transcribed by using SuperScript III reverse transcriptase according to the manufacture’s instructions (Invitrogen, Carlsbad, CA, USA). PCR was carried out by using the SV-F11 and SV-R1 primer set directed against the conserved N terminal capsid region (8). The PCR products were analyzed with 2% agarose gel electrophoresis and visualized after ethidium bromide staining. The PCR-generated amplicons (≈780 bp) were excised from the gel and purified by the QIAquick gel extraction kit (QIAGEN, Hilden, Germany). 
 
Nucleotide sequences were prepared with the terminator cycle sequence kit (version 3.1) and determined with the ABI 3130 sequencer (Applied Biosystems, Foster City, CA, USA.). Nucleotide sequences were aligned by using ClustalX (www.clustal.org), and the distances were calculated by using the Kimura 2-parameter method. A phylogenetic tree was generated by the neighbor-joining method as described previously (1,8). 
 
Of the 8 specimens, 7 were positive by RT-PCR and real-time RT-PCR (Table). SaV124F, SaV1F, SaV5F, and SaV1245R primers as well as SaV124TP and SaV5TP minor-groove binding probes were used for real-time RT-PCR diagnosis, which targets the sapovirus RdRp-capsid junction region as described (7). The number of sapovirus cDNA copies ranged from 2.86 × 107 to 1.72 × 1010 copies/g of stool specimen; mean was 2.71 × 109 copies/g of stool specimen (Table). Sequence analysis of the 7 positive specimens showed 100% nucleotide identity (nt 5098–5878), indicating that the outbreak was caused by 1 sapovirus strain. 
 
To better classify the sapovirus, we reamplified the 3′ end of the genome from 1 positive specimen and sequenced ≈2,400 nt (nt 5074-3′) (Hu/SaV/9–5/Taipei/07/TW; GenBank accession no. {"type":"entrez-nucleotide","attrs":{"text":"EU124657","term_id":"158929628"}}EU124657). PCR was performed with SV-F13, SV-F14, and TX30SXN primers as described (1). Database searches found a closely matching sapovirus sequence (99%) that was detected in a patient with gastroenteritis in Japan, in 2004 (Chiba041413 strain; GenBank accession no. {"type":"entrez-nucleotide","attrs":{"text":"AB258427","term_id":"118442701"}}AB258427). The next closely matching sequence was detected in an outbreak of gastroenteritis among adults in the United States in 1994 (Parkville strain; {"type":"entrez-nucleotide","attrs":{"text":"HCU73124","term_id":"1916653"}}HCU73124) (6). Phylogenetic analysis clustered these 3 sapovirus sequences into genogroup I/genotype 2 (GI/2) (Appendix Figure). 
 
Sapovirus was reported in Japan in water samples (untreated wastewater, treated wastewater, and a river) and in clam samples intended for human consumption (1). Apart from these 2 environmental studies, little is known about reservoir of sapovirus or its route of infection in the natural environment. The source of contamination in this current outbreak was not determined; however, none of the food handlers associated with the college reported symptoms of gastroenteritis. However, in a recent molecular epidemiologic study in Japan, a large number of symptomatic and asymptomatic food handlers were found to be infected with noroviruses (9). Several seroprevalence studies also indicated high prevalence rates of antibodies to sapovirus in adults and children (10). All of these findings highlight the need to collect stool specimens from asymptomatic persons and indicate possible “silent” transmission through an asymptomatic route. Symptoms of sapovirus infection are thought to be milder than symptoms of norovirus infections. However, in this study approximately one third (17) of the 55 students reported symptoms of abdominal pain and 22 (40%) reported symptoms of vomiting. Many of the earlier sapovirus studies described sapovirus GI/1 infections in young Japanese children (1), which indicated that infecting virus had a different genotype than the virus detected in this study (GI/2). 
 
In addition, the viral load in this study appeared to be comparatively high. These results suggest that some sapovirus genotypes are more virulent than others. Similar findings were obtained with norovirus infections around the world; strains belonging to norovirus GII/4 were the most prevalent in many countries. Although several recombinant sapovirus strains have been identified and found to be the cause of increased numbers of infections in some countries (1,5), they were not observed in this study. Increased sapovirus surveillance and reporting are needed to shed some more light on this poorly understood virus.

Antibodies to B. melitensis were detected by agglutination by using the Rose Bengal and Brucella Wright tests (both from BioRad, Hercules, CA, USA). Of the 63 patients, 9 had a positive antibody response against a tested antigen (Table): 1 to phase I C. burnetii and 8 to Bartonella spp. (IgG >200). Of these, 7 had a 1-fold dilution higher titer to B. quintana than to B. henselae, including 1 with a lowlevel cross-reaction with C. burnetii, and 1 with identical titers to both. For all 8 patients, Western blot results were consistent with Bartonella endocarditis. For 7, cross-adsorption identifi ed B. quintana as the causative species; for the other, the infecting Bartonella species remained undetermined because adsorption with B. quintana and B. henselae antigens removed all antibodies. Serologic results for B. melitensis were negative for all patients.
B. quintana is mostly associated with human body lice but has also been found in fl eas (9). The predisposing factors for B. quintana endocarditis are homelessness, alcoholism, and exposure to body lice (10). For our patients, the common predisposing factors were poor hygiene and low socioeconomic status, which may expose them to ectoparasites including lice and fl eas. In contrast with previous study fi ndings, B. quintana infectious endocarditis developed on preexisting valvular lesions in all patients (10). This fi nding may refl ect a different clinical evolution than in Europe, where studies have suggested that B. quintana infectious endocarditis followed chronic bacteremia in patients who did not have previous valvular defects (10).
In summary, prevalence of negative blood culture among patients with infectious endocarditis was high (72%). The most commonly associated risk factor was rheumatic heart disease (Table). C. burnetii and Bartonella spp. were responsible for 8% of all infectious endocarditis cases and 14% of blood culture-negative cases. No case of infectious endocarditis caused by B. melitensis was identifi ed.
Our preliminary study suggests that zoonotic agents, especially Bartonella spp., are prevalent causative organisms of blood culture-negative endocarditis in India. We recommend serologic screening for antibodies to zoonotic microorganisms as diagnostic tools for this disease in India. To the Editor: Sapovirus is an etiologic agent of human gastroenteritis. Although many of the previously reported cases were of mild, sporadic infections in young children (1-3), several recent sapovirus-associated gastroenteritis outbreaks have affected adults, which suggests that the virus's virulence, prevalence, or both, may be increasing (4)(5)(6). In this study, we describe a sapovirus-associated outbreak of gastroenteritis that occurred during May 4-8, 2007, and involved college students in northern Taiwan.
A total of 55 students had clinical symptoms of gastroenteritis, including diarrhea (45), vomiting (22), abdominal cramps (17), and fever (2). The clinical symptoms continued for up to 10 days (mean 4.7 days). Stool specimens were collected from 8 of 55 students on May 8 (Table). Initially, the specimens were screened for bacteria, rotavirus, and norovirus, but all specimens were negative for these pathogens. The 8 stool specimens were then examined by electron microscopy (EM), and 1 was positive for calicivirus-like particles.
To confi rm the EM results, we performed reverse transcription-PCR (RT-PCR), real-time RT-PCR, and sequence analysis as previously described (7). Briefl y, purifi ed RNA (10 μL) was reverse transcribed by using SuperScript III reverse transcriptase according to the manufacture's instructions (Invitrogen, Carlsbad, CA, USA). PCR was carried out by using the SV-F11 and SV-R1 primer set directed against the conserved N terminal capsid region (8). The PCR products were analyzed with 2% agarose gel electrophoresis and visualized after ethidium bromide staining. The PCR-generated amplicons (≈780 bp) were excised from the gel and purifi ed by the QIAquick gel extraction kit (QIAGEN, Hilden, Germany).
Nucleotide sequences were prepared with the terminator cycle sequence kit (version 3.1) and determined with the ABI 3130 sequencer (Applied Biosystems, Foster City, CA, USA.). Nucleotide sequences were aligned by using ClustalX (www.clustal.org), and the distances were calculated by using the Kimura 2-parameter method. A phylogenetic tree was generated by the neighbor-joining method as described previously (1,8).
Of the 8 specimens, 7 were positive by RT-PCR and real-time RT-PCR (Table). SaV124F, SaV1F, SaV5F, and SaV1245R primers as well as SaV124TP and SaV5TP minor-groove binding probes were used for real-time RT-PCR diagnosis, which targets the sapovirus RdRp-capsid junction region as described (7). The number of sapovirus cDNA copies ranged from 2.86 × 10 7 to 1.72 × 10 10 copies/g of stool specimen; mean was 2.71 × 10 9 copies/g of stool specimen (Table). Sequence analysis of the 7 positive specimens showed 100% nucleotide identity (nt 5098-5878), indicating that the outbreak was caused by 1 sapovirus strain.
Sapovirus was reported in Japan in water samples (untreated wastewater, treated wastewater, and a river) and in clam samples intended for human consumption (1). Apart from these 2 environmental studies, little is known about reservoir of sapovirus or its route of infection in the natural environment. The source of contamination in this current outbreak was not determined; however, none of the food handlers associated with the college reported symptoms of gastroenteritis. However, in a recent molecular epidemiologic study in Japan, a large number of symptomatic and asymptomatic food handlers were found to be infected with noroviruses (9). Several seroprevalence studies also indicated high prevalence rates of antibodies to sapovirus in adults and children (10). All of these fi ndings highlight the need to collect stool specimens from asymptomatic persons and indicate possible "silent" transmission through an asymptomatic route. Symptoms of sapovirus infection are thought to be milder than symptoms of norovirus infections. However, in this study approximately one third (17) of the 55 students reported symptoms of abdominal pain and 22 (40%) reported symptoms of vomiting. Many of the earlier sapovirus studies described sapovirus GI/1 infections in young Japanese children (1), which indicated that infecting virus had a different genotype than the virus detected in this study (GI/2).

1170
Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 14, No. 7, July 2008 In addition, the viral load in this study appeared to be comparatively high. These results suggest that some sapovirus genotypes are more virulent than others. Similar fi ndings were obtained with norovirus infections around the world; strains belonging to norovirus GII/4 were the most prevalent in many countries. Although several recombinant sapovirus strains have been identifi ed and found to be the cause of increased numbers of infections in some countries (1,5), they were not observed in this study. Increased sapovirus surveillance and reporting are needed to shed some more light on this poorly understood virus.

Importation of West Nile Virus Infection from Nicaragua to Spain
To the Editor: We report the case of a 51-year-old Spanish missionary who had lived Nicaragua (Managua) from 2004 to 2006. He had no other notable travel history during that period. In June 2006, he noticed malaise and nausea, followed by abrupt onset of fever (39°C), headache, cervical pain, and right hemiparesis. He was admitted to a local hospital in Nicaragua, at which time routine results of hematologic and biochemistry tests were within normal limits, except for mild neutrophilia. After cerebral magnetic resonance imaging (MRI), a diagnosis of ischemic cerebrovascular accident was made. He was treated with aspirin and ceftriaxone for an oropharyngeal infection.
Because neurologic symptoms persisted, 13 days later he was transferred to a hospital in Madrid, Spain. At that time, physical examination showed neck stiffness, a diminished level of consciousness, right fl accid hemiparesis, and facial weakness. Peripheral blood examination showed only mild neutrophilia. Cerebrospinal fl uid (CSF) analysis showed a 65 mg/dL glucose level (blood glucose 140), proteins 136 g/dL, and 18 cells/ mm 3 (mainly lymphocytes). Serologic test results for HIV, hepatitis B virus, hepatitis C virus, syphilis, Toxoplasma spp., and Brucella spp., and CSF cultures for mycobacterial, bacterial, and fungal infections were all negative. Results of a computed tomographic scan of the brain were within normal limits. MRI showed nonspecifi c abnormal intensity of white matter signal. Electrophysiologic studies showed severe axonal motor neuropathy and moderate sensitive axonal neuropathy in the right upper limb. Gammaglobulin was administered intravenously for 5 days; the patient improved slightly. At