Guaroa Virus and Plasmodium vivax Co-Infections, Peruvian Amazon

During April–June 2014 in a malaria-endemic rural community close to the city of Iquitos in Peru, we detected evidence of Guaroa virus (GROV) infection in 14 febrile persons, of whom 6 also had evidence of Plasmodium vivax malaria. Cases were discovered through a long-term febrile illness surveillance network at local participating health facilities. GROV cases were identified by using a combination of seroconversion and virus isolation, and malaria was diagnosed by thick smear and PCR. GROV mono-infections manifested as nonspecific febrile illness and were clinically indistinguishable from GROV and P. vivax co-infections. This cluster of cases highlights the potential for GROV transmission in the rural Peruvian Amazon, particularly in areas where malaria is endemic. Further study of similar areas of the Amazon may provide insights into the extent of GROV transmission in the Amazon basin.

of mosquito species and provide a suitable environment for arbovirus and Plasmodium spp. transmission. Since 1993, at least 13 arboviruses have been detected in this area, of which dengue virus (DENV), Zika virus, Mayaro virus, and Venezuelan equine encephalitis virus are considered to be of public health importance (1).
Guaroa virus (GROV; order Bunyavirales, family Peribunyaviridae, genus Orthobunyavirus) (2) is a known cause of febrile illness in tropical regions of Central and South America (3). It was first isolated from asymptomatic humans in Colombia in 1956 (4) and isolated from symptomatic humans in Brazil in 1964 (5). Recent phylogenetic analysis suggests that a common ancestor of GROV and Wyeomyia virus was introduced into South America in the Brazilian Amazon region ≈250 years ago, with subsequent southward spread of GROV to Peru within the past 60-70 years (6). A serologic survey in 1965 provided evidence of GROV transmission in Peru shortly after this time (7).
More recent antibody prevalence studies on samples collected in Iquitos in 2006 demonstrated an overall GROV seroprevalence of 13% (3). The increasing seroprevalence with age suggests that transmission occurred consistently in the region over several years (3). In addition to seroprevalence, this study also described 15 symptomatic GROV infections in Peru (including 3 in Iquitos) during 1995-2008, providing further evidence of GROV as a cause of symptomatic disease in the region (3).
The only confirmed vector of GROV is Anopheles (Kerteszia) neivai mosquitoes (8). This species is an important vector of human malaria in the Pacific lowlands of Colombia (9). After a successful elimination effort in the 1960s, malaria reemerged in the early 1990s in the Peruvian Amazon, coincident with the reintroduction of An. darlingi mosquitoes and is currently a leading cause of febrile illness (10). Subsequently, a resurgence of P. falciparium peaked in 1997, During April-June 2014 in a malaria-endemic rural community close to the city of Iquitos in Peru, we detected evidence of Guaroa virus (GROV) infection in 14 febrile persons, of whom 6 also had evidence of Plasmodium vivax malaria. Cases were discovered through a longterm febrile illness surveillance network at local participating health facilities. GROV cases were identified by using a combination of seroconversion and virus isolation, and malaria was diagnosed by thick smear and PCR. GROV mono-infections manifested as nonspecific febrile illness and were clinically indistinguishable from GROV and P. vivax co-infections. This cluster of cases highlights the potential for GROV transmission in the rural Peruvian Amazon, particularly in areas where malaria is endemic. Further study of similar areas of the Amazon may provide insights into the extent of GROV transmission in the Amazon basin.
at which point P. falciparium prevalence decreased significantly. The An. darlingi mosquito is the primary vector of malaria in rural areas near Iquitos (11-13). There are numerous reports in the literature of arboviral and malarial co-infections, but reports of GROV and malaria co-infection are scarce, and only co-infection with P. falciparium has been reported (14,15). We describe the clinical and epidemiologic aspects of a GROV outbreak in the rural surroundings of Iquitos and report symptomatic co-infections with GROV and P. vivax.

Methods
We identified GROV cases through a passive febrile surveillance system at 12 health centers in urban, periurban, and rural areas in and around Iquitos, which were described previously (1). Inclusion criteria were age >5 years, oral or tympanic temperature >38°C (or axillary >37.5°C), duration of symptoms <5 days, and no obvious focus of infection. All participants at these sites were initially screened for Plasmodium infection by thick smear and were then invited to undergo screening for arboviruses regardless of their smear results. In addition to serum samples obtained during the acute phase, serum samples were obtained during follow-up evaluations 20 days (± 10 days), 3 months (± 15 days), 6 months (± 15 days), and 12 months (± 30 days) after the initial sample. Serum samples collected on day 20 were used to measure convalescent titers.
Of the 12 clinics, 2 are in rural communities with active malaria transmission (Zungarococha and Quistococha communinities), and 3 urban clinics serve as catchment areas for communities with active malaria transmission (Bella Vista Nanay, 6 de Octubre, and San Juan). The GROV cases reported in this study were captured during April 5-June 26, 2014, from 3 malaria-endemic communities in the district of San Juan (Santo Tomas, Quistococha, and Zungarococha) (Figure).
We obtained approval for the study protocol (NMRCD.2010.0010) from the Institutional Review Board of NAMRU-6 in compliance with all applicable US federal regulations governing the protection of human subjects and from the Institutional Review Board of the Peruvian Ministry of Health. Adults >18 years of age provided written informed consent, and parents or legal guardians of participants <17 years of age provided assent.
We conducted sequencing on a 3130 XL Genetic Analyzer (Applied Biosystems) platform and analyzed sequences using Sequencher software (Gene Codes Corporation); we queried individual sequences by using the nucleotide database with BLAST (17).

Malaria Testing
Microscopy was the primary method of Plasmodium identification conducted by the health centers as part of the Ministry of Health's surveillance. Microscopy was also used as part of our study, although to evaluate Plasmodium spp. infection more accurately in all GROV-positive cases, we also tested blood from acute-phase samples for Plasmodium by PCR regardless of their initial thick-smear results. DNA was extracted from whole blood samples using the DNeasy Blood & Tissue Kit (QIAGEN), following the manufacturer's protocol. Plasmodium DNA was amplified using a nested PCR protocol that targets the small subunit ribosomal RNA 18S gene (18). Both reactions were conducted in a volume of 50 µL containing 1X Taq polymerase buffer (Invitrogen), 2 mmol/L MgCl2, 125 µmol/L dNTPs (Invitrogen), 0.25 µmol/L of each primer, 1 unit of Taq DNA polymerase (Invitrogen), and 5 µL of DNA sample. We ran both PCR reactions on a Verity Thermal Cycler (Applied Biosystems) as previously described (18). We used DNA from P. falciparum 3D7 and P. vivax Sal-I reference strains as positive controls and human DNA from a person from a non-malaria-endemic area as a negative control. We subsequently ran PCR products on a 2% agarose gel; a band of ≈205-bp indicates the presence of P. falciparum DNA, and a PCR product of 120bp indicates the presence of P. vivax DNA.

Results
During the 3-month period when GROV infections were detected, a total of 681 febrile patients were enrolled from all 12 participating clinics around Iquitos. All GROV-infected participants lived in the district of San Juan, in 1 of the 3 malaria-endemic communities: Santo Tomas, Quistococha, or Zungarococha. A total of 121 (18%) of 681 febrile persons resided in 1 of these 3 communities in San Juan. Of those, 14 (12%) tested positive for GROV infection, of whom 6 (43%) were co-infected with P. vivax. Of the 14 persons for whom GROV infection was confirmed, 3 were enrolled in urban health facilities, although these 3 persons reside in the rural San Juan district (Figure). During the 3-month study period, 26 malaria cases were reported in the San Juan district, including the 6 with GROV co-infection. All 14 GROV-infected persons seroconverted and were negative for all other examined arboviruses; 11 (79%) samples collected from these persons yielded GROV isolates in both Vero-76 and C6/36 ( Table 1) All participants reported fever, chills, malaise, body pain, joint pain, and headache (Table 2). Myalgia, anorexia, nausea, and dysgeusia occurred in 8 patients. Only 2 patients reported a rash. Distinct symptom frequencies between persons with GROV mono-infection and co-infected persons were conjunctival injection, dysgeusia, cough, and sore throat. These symptoms occurred more frequently in persons without malaria. These differences in symptoms were not statistically significant (χ 2 with Yates correction significance level 0.05).
Median fever duration for both GROV mono-infection and co-infected persons was 4.5 days. Symptoms with the longest median duration for all participants were body pain, joint pain, headache, and dizziness (all >5 days). Headache was reported during the convalescent phase and for up to 3 months during follow-up in 4 patients with a GROV mono-infection and 2 patients with GROV and malaria co-infection. Duration of symptoms between mono-infected and co-infected persons did not differ significantly. No patients were hospitalized, and all recovered without sequelae.

Discussion
The circulation of many arboviruses in the Peruvian Amazon is well documented, but most cases are not reported (1), possibly because the asymptomatic and mild self-limiting infections are common and do not usually result in treatment-seeking behavior (14). Other important factors are the absence of diagnostic facilities for detecting arboviruses in areas to which they are endemic, limited access to healthcare for atrisk persons in rural Amazonian populations, and the narrow window of opportunity to collect potentially diagnostic blood samples during the acute phase of the illness (15,19). Another factor limiting arboviral diagnosis and identification of co-infections with other pathogens is that once malaria is diagnosed, secondary diagnoses are rarely pursued (15). For example, in our study GROV infection was diagnosed in 3 patients after they returned to the same health center within 3 days after having tested positive for P. vivax. Because their febrile symptoms persisted despite treatment with antimalarial drugs, they provided additional samples that tested positive for GROV. Although previous exposure to antimalarial drugs might have affected our RT-PCR results for malaria and might have resulted in a false-negative result, only 2 patients were willing to provide blood samples for arbovirus testing during the same visit in which they tested positive for P. vivax by thick smear. The detection of viral and parasitic co-infection demostrates the value of actively looking for viral pathogens in malaria-positive patients at the time they seek medical care during the acute phase, when the probability of identifying a viral infection is highest. Current knowledge about GROV as a human pathogen is limited, and little is known about its reservoir hosts and vectors, particularly in Peru. Anopheles mosquitoes have been implicated elsewhere as vectors for GROV (5,8,20), although few arbovirus are known to be transmitted by Anopheles mosquitoes, such as o'nyong-nyong virus, which is transmitted in Africa by An. funestus and An. gambiae mosquitoes (21). It is possible that a co-infected mosquito could transmit both Plasmodium spp. and GROV; however, vector-competence studies of An. darlingi mosquitoes with GROV have not yet been published (22). The An. darlingi mosquito is the most common and competent malaria vector in the Peruvian Amazon (11-13,23,24). During 2014, the Peruvian Ministry of Health recorded 65,239 malaria cases in the entire country, most of which resulted from P. vivax (84%) and were detected in the Amazon region of Loreto (≈90%) (23). The GROV infection reported in this study coincides temporally with the peak of this P. vivax outbreak observed in 2014 (25), increasing the chances of coinfected mosquitoes, probability of transmission to humans, or both. Furthermore, malaria rates tend to be high along the Nanay River, located only a few kilometers from the southern Iquitos city limits (26). All the GROV cases observed in the present report came from rural communities in the district of San Juan Baustista bordering the Nanay River. Further investigation is needed to understand the nature of these co-infections because they could have occurred solely as a consequence of the elevated levels of circulating P. vivax, a possible peak in Anopheles populations, or subclinical/persistent malaria with fever caused by GROV infection. In addition, identification of nonhuman reservoirs for GROV is critical to clarify the epidemiology of the disease and distinguish GROV from other arboviruses circulating in and near Iquitos (27,28).
Consistent with a previous report (3), the clinical manifestations of GROV infection in this study were highly nonspecific, and persons co-infected with P. vivax were clinically indistinguishable from those with GROV infection alone. It was not possible to determine the relative contribution of malaria and GROV to the symptomatology of co-infected persons; however, 3 of the 5 patients with positive thick smears who were treated with antimalarial drugs returned to the same health center within 3 days after the initial diagnosis because of persistent fever. These symptoms could, at least in part, have been related to the GROV infection consistent with other clinical symptom reports (3). Understanding the origin of symptoms in co-infected persons is further complicated because both asymptomatic malaria and arboviral disease are common (14,29).
The cases of GROV fever we report may underrepresent the true number of total cases in Iquitos during the study period. GROV may be underreported because of overlapping symptoms between pathogens, such as DENV and Plasmodium spp., that are more frequently targeted by diagnostic tests. Our study highlights the importance of the febrile surveillance system and its access to advanced diagnostic facilities that enabled detection of these cases. Continued surveillance is necessary not only to monitor the dynamics of well-studied diseases, such as dengue and malaria, but also to capture the emergence of potentially new pathogens with epidemic potential that pose a public health risk. In particular, examining malaria-endemic communities may provide an opportunity to better quantify the incidence of GROV and other endemic or enzootic pathogens.