Clinical and virological characteristics of hantavirus infections in a 2014 Croatian outbreak

Introduction: Croatia is endemic for hemorrhagic fever with renal syndrome (HFRS), with both Puumala (PUUV) and Dobrava virus (DOBV) documented. Several large outbreaks were recorded in 1995, 2002, and 2012. We analyzed demographic, clinical, laboratory, and virological characteristics of HFRS cases detected in three geographically close natural foci (Ogulin, Slunj, and the Plitvice Lakes surroundings) during the 2014 outbreak. Methodology: From January to December 2014, 122 patients with suspected HFRS were tested for hantavirus IgM/IgG antibodies using an indirect immunofluorescence assay (IFA). Cross-reactive samples were further tested using a western blot (WB). For hospitalized patients from Ogulin area, clinical and laboratory data were analyzed. Results: Acute infection was documented in 57 (46.7%) patients, of whom 75.4% were hospitalized. Ten (8.2%) patients were found to be IgG seropositive. Patients were 15–69 years of age and predominantly male (74.5%). The outbreak started in winter months, with most cases recorded from May to July (80.7%). The most frequently reported symptoms were fever (96.3%), chills/shivering (62.9%), and lumbar pain (48.1%). Mild clinical form was found in 66.7% patients, moderate in 18.5%, and severe in 14.8% patients (all but one infected with PUUV). One patient died. Using IFA, 48.8% patients showed monotypic antibody response, while in 51.2%, cross-reactive antibodies were found. PUUV was confirmed in 94.7% and DOBV in 5.3% HFRS cases by WB. Conclusions: Central mountainous Croatian regions are still highly endemic areas for HFRS. A higher percentage of severe PUUV infections could be at least partly associated with a patient’s immune status.


Introduction
Hantaviruses are diverse group of rodent-borne viruses that belong to the genus Hantavirus of the family Bunyaviridae.In humans, hantaviruses cause two clinical forms of disease: hemorrhagic fever with renal syndrome (HFRS), which occurs primarily in Europe and Asia, and the hantavirus pulmonary syndrome (HPS), which occurs in the Americas [1,2].Puumala (PUUV), Dobrava (DOBV), and Saaremaa virus (SAAV) are known to cause HFRS in Europe.Hantaan virus (HTNV) is widely distributed in Asia, while Seoul virus (SEOV) is found worldwide [3].The clinical presentation of HFRS varies from subclinical or mild to severe, sometimes fatal infection, depending in part on the causative virus.In general, PUUV and SAAV cause a mild form of disease with mortality below 1%.In contrast, DOBV causes a more severe clinical form, with mortality rates from 5% to 15% [4].
In Croatia, HFRS was described for the first time in 1952 [5].Since then, sporadic cases as well as outbreaks have been continuously reported.Several small outbreaks were reported in the Plitvice Lakes (1967), Velika Gorica (1989), and Novska (1991) [6][7][8].The first large outbreak, which included 125 cases, was recorded in 1995 in several well-known endemic foci (Gorski Kotar, Lika, Slunj, Mala Kapela, and west Slavonia) [9].In addition, new natural focus of HFRS was described on the Dinara mountain [10].During the 2002 Croatian outbreak, 401 cases of HFRS were recorded.Except for the coastal region and islands, the whole country was found to be an area endemic for HFRS [11][12][13].In 2012, 154 patients were described, of whom 69.5% acquired the infection on the Medvednica mountain near the Croatian capital of Zagreb [14].In the interepidemic years, approximately 10-20 cases were diagnosed annually [15].Both PUUV and, to a lesser extent, DOBV have been found to circulate among humans in Croatia [12].Besides PUUV and DOBV, SAAV is also detected in rodents [16][17][18].In the 1990s, 1.6% of forest workers and 5.4% of the general population from the endemic territory of Lika and Gorski Kotar were seropositive for hantaviruses.In addition, 4.9%-5.5% of rodents from the same area were found to be positive for the hantavirus antigen [19].
We analyzed epidemiological, clinical, laboratory, and virological characteristics of HFRS cases detected in three geographically close natural foci during the 2014 Croatian outbreak.The data obtained in this study will contribute to the epidemiology of hantavirus infections.

Methodology
From January to December 2014, a total of 122 patients from three well-known natural foci for HFRS in Croatia (Ogulin and Slunj surroundings, the Plitvice Lakes) were tested at the National Reference Laboratory for Arboviruses, Croatian National Institute of Public Health.All of them presented with clinical symptoms suggestive of hantavirus infection, and the majority reported risk activities, mainly exposure to rodents.The analyzed area is located in the central mountainous part of Croatia, which represents the transition between the continental mainland and the Adriatic coast.The climate in this area is submountainous and mountainous, with cool summers and harsh winters.Summer temperatures range from 22°C to 26°C in the continental and 15°C to 20°C in the mountain region.The mean winter temperature is 4°C to 0°C [20].Serologic tests were performed using a commercial indirect immunofluorescence assay (IFA) based on BioChip Mosaic (Hantavirus mosaic; Euroimmun, Lübeck, Germany) for the simultaneous detection of IgM and IgG antibodies to the most important hantaviruses: PUUV, DOBV, HTNV, SEOV, and SAAV.Cross-reactive samples were further tested using a western blot (WB) method (Euroline Hantavirus; Euroimmun, Lübeck, Germany).For hospitalized patients from the Ogulin area, demographic data and clinical and laboratory parameters obtained from medical reports were analyzed.To classify disease severity, the following parameters were used: shock/hypotension, bleeding, fever ≥ 40°C, visual impairment, urea/creatinine levels, platelet count, aspartate-aminotransferase (AST) and alanine-aminotransferase (ALT) levels, and chest X-ray [11].

Statistical analysis
The data were analyzed by the percentage of patients with symptoms/clinical signs or abnormal laboratory findings.Chi-square (χ 2 ) and analysis of variance (ANOVA) tests were used to compare differences between groups (MedCalc for Windows version 7.0).P value < 0.05 was considered to be statistically significant.
All patients received symptomatic and supportive therapy (antipyretics and rehydration).Antibiotics were administered to patients with a strong inflammatory response.One patient was treated with hemodialysis to manage renal failure.The mean duration of hospitalization was 9.6 (range, 4-19) days.Twenty-six   patients (96.3%) recovered, and one PUUV-infected patient (3.7%) died.Three patients with a severe form of the disease were infected with PUUV (cases 1, 3, 4) and one with DOBV (case 2).Comorbidity was noted in three patients.Case 1 (M, 52 years) reported hypertension; during hospitalization, he developed massive polyuria.Case 2 (M, 20 years) reported diabetes and cardiomyopathy; he developed polyuria and prolonged lumbar pain.Case 3 (M, 35 years) was previously healthy; during hospitalization, he developed pericarditis and cardiac failure.Patients were hospitalized for 19, 16, and 14 days, respectively, and fully recovered.Case 4 (M, 48 years) was admitted to the hospital with high fever, severe intraorbital pain, and visual impairment.His previous medical history revealed myocardial infarction that occurred 7 years ago, when he was resuscitated and required hemodialysis thereafter for one month.During hospitalization, he developed multiorgan failure and died two months after disease onset.

Discussion
During the 2014 outbreak, a total of 209 HFRS cases were reported in Croatia (data from the Reference Epidemiology Centre, Croatian National Institute of Public Health).More than one-third of cases were from the central mountainous Croatian regions (Ogulin, Slunj, Plitvice Lakes).
Although cases were reported throughout the year, this study detected the majority of HFRS cases from May to July, with a peak in June (40.3%).During the 2013/2014 winter season, the average monthly air temperature was above the multi-annual average, with corresponding temperature anomalies within the range of 2.4°C to 6°C [20], which could explain the earlier occurrence of an HFRS outbreak.A similar seasonal distribution was observed in 2002 [11], while in 1995, outbreak peak incidence shifted two months earlier (i.e., in April [46%]) [9].In addition to climate factors, environmental conditions favor the occurrence of HFRS.Several hantavirus outbreaks following years of beech mast tree, which promotes reproduction of bank vole, were recently reported [22,23].Similarly, in autumn 2013, beech trees produced large quantities of seed (masts) in Croatia [24].
The most common clinical symptoms reported in our patients (fever, chills, malaise, headache, and back and abdominal pain) were similar to those previously reported in Slovenia, Germany, and Croatia [12,25,26].However, there are some differences in the frequency of symptoms as well as in disease severity compared to previous Croatian outbreaks.Lumbar pain was reported less frequently (48.1%) compared to 50%-91.7% in previous outbreaks.Although respiratory symptoms appear frequently, especially in PUUV infections [4], in 2014, respiratory symptoms were present less frequently (11.1%) compared to 2002 and 1995 outbreaks (35.4% and 30%, respectively).Based on disease severity, 66.7% patients presented with a mild form of the disease in 2014, which is similar to the 2002 outbreak (65%) [11], while in 2012, a mild form was observed in 76.5% patients [14].The highest frequency of severe disease was noted in 1995 (50%) [9], compared to 5.3% in 2002 [12], 4.7% in 2012 [14], and 14.8% in 2014.
Several studies showed that the severity of clinical symptoms depends on the patient's characteristics.In Sweden and Finland, gender-specific differences in symptoms as well as mortality rates were identified for epidemic nephropathy [27,28].A similar observation was reported for HFRS in China [29].Our results showed no association between gender and severity of disease (p = 0.945), which is similar to results reported in German patients [30].In addition, the disease severity in Croatian patients differed among outbreaks and the interepidemic period (86.2% patients had a mild clinical form) [15].A recently published German study showed that polyuria in HFRS reflected disease severity and is associated with prolonged hospital stay [31].The duration of hospital stay for two patients with severe HFRS who developed polyuria was 19 and 16 days, respectively, compared to the average of 9.6 days.
Although a higher frequency of the severe form was observed in DOBV-infected patients [12], severe cases may also occur during PUUV infection [9].In this study, three of four hospitalized patients who presented with severe hantavirus infection were infected with PUUV, of whom one died.In other Balkan countries, severe forms and fatal cases were associated mostly with DOBV infection [4,32].In 1995, 20.5% HFRS cases were caused by HTNV, which could explain the high frequency of the severe form of the disease.Since the proportion of PUUV-infected Croatian patients was similar in 2012 and 2014 (95.3% and 94.7%, respectively), a higher frequency of severe infections in 2014 could be at least partly related to the patient's immune status, including comorbidities.
Due to the hazardous nature of hantaviruses and a short-term viremia in hantavirus-infected individuals, diagnosis of HFRS is usually based on serology [33].The main limitation of serology is cross-reactivity, which often complicates interpretation of results, especially in areas where multiple hantaviruses cocirculate [34].The results of this study indicated broadly serological cross-reactivity among hantaviruses.Using IFA, 17.5% IgM and 51.2% IgG positive samples cross-reacted with different hantavirus antigens.Cross-reactivity was found to be much higher for DOBV compared to PUUV.A high/very high crossreactivity of DOBV IgG antibodies was observed with SEOV, HTNV, and SAAV antigens (80.0%-100%).These results are in accordance with the phylogenetic relatedness of hantaviruses.However, a high crossreactivity (80.0%) was also found with PUUV, although PUUV is phylogenetically distantly from DOBV.A recently published multicenter study on the simultaneous of hantaviruses using IFA showed a high cross-reactivity of sera from DOBVinfected patients with SAAV, HTNV, and SEOV (60%-100%) and a moderate cross-reactivity with PUUV (up to 43%) [35].
In this study, PUUV IgG antibodies reacted highly/moderately with HTNV and DOBV (70.0% and 60.0%, respectively) and lower with SEOV and SAAV (30.0% and 26.6%, respectively).In the study of Lederer et al. [35], cross-reactivity between PUUV and HTNV IgG was somewhat higher (79%) compared to our results, while the reactivity to other tested hantaviruses was similar.Using WB, serotype determination was successful in most cross-reactive samples.In two samples that cross-reacted in both IFA and WB, a monotypic IgM response indicated the type of the causative virus.
In addition to acute HFRS cases, this study showed that 8.2% patients were IgG seropositive to hantaviruses.A study from 1991 conducted in the same geographic area showed a wide range of seroprevalence rates of 1.85%-3.26%among forest workers and 4.39%-8.89% in the general population [19].A limited number of tested participants might be the cause of differences in seropositivity in these groups.
In Croatia, the bank vole (Myodes glareolus) is considered to be the principal vector for PUUV, with yellow-necked field mice (Apodemus flavicolis), longtailed field mice (Apodemus sylvaticus), and striped field mice (Apodemus agrarius) playing a minor role in hantavirus transmission [12,19].Several studies addressed the prevalence of hantaviruses in rodent reservoirs in Croatia.Between the 1980s and 2000s, hantavirus antigen was detected in 5.4% of rodents (Clethrionomys glareolus, Apodemus flavicolis) captured in the Plitvice region [19] and in 11.9% rodents captured on the Dinara mountain [10].A study conducted in 2007-2008 revealed specific antibodies against hantaviruses in 25.5% of rodents from the mountainous area, while 29% of rodents from the lowland area were positive for either PUUV or DOBV RNA [36].In addition, SAAV and DOBV-Kurkino genome sequences were also found in rodents, but so far, no human cases have been associated with these viruses in Croatia [18,37].Moreover, a high infection rate of bank voles with PUUV (77.4%) was documented on the Medvednica mountain near Zagreb during the 2012 outbreak [17].

Conclusions
Our results indicate that central mountainous Croatian regions are still highly endemic areas for hantavirus infections, with co-circulation of PUUV and DOBV.Although the number of patients included in this study is limited, our results suggest that PUUV also caused a severe form of HFRS, with one fatal case.The respiratory symptoms that are frequently reported in PUUV infection were reported less frequently in our study.Since PUUV infection is generally mild to moderately severe, a higher percentage of the severe form recorded in this study indicates that disease severity could be at least partly associated with the patient's immune status.

Figure 1 .
Figure 1.Seasonal distribution of hemorrhagic fever with renal syndrome cases in 2014 Croatian outbreak.

Figure 2 .
Figure 2. Geographical distribution of hemorrhagic fever with renal syndrome cases and IgG seropositive subjects.The size of circles/triangles is proportional to a number of positive samples (1-8).

Figure 3 .
Figure 3.Western blot showing signal intensity of hantavirus indirect immunofluorescence assay cross-reactive samples.

Table 1 .
Demographic characteristics, clinical symptoms, and laboratory parameters of 27 hospitalized patients with hemorrhagic fever with renal syndrome.

Table 2 .
IgG cross-reactivity patterns among hantaviruses by IFA and western blot.