Introduction

Puumala hantavirus infection (nephropathia epidemica) causes a mild form of haemorrhagic fever with renal syndrome (1–5), and is prevalent throughout Europe, particularly in Finland and Scandinavia. Usually, onset of symptoms (such as fever, abdominal pain, headache, back pain and/or gastrointestinal symptoms) is sudden (5). In endemic regions it may be the most common cause of acute renal failure and 5% of hospitalised patients may have renal function impaired enough to need dialysis. Diagnosis is based on serology and no specific treatment exists. Some patients can develop severe disease requiring hospitalisation (with an average stay of 11 days) (1), and some may suffer longterm sequelae such as hypertension (3) or impaired hypophyseal function (4). In Finland the reported case fatality ratio of 0.03% (1) may be an underestimate (5), and during epidemic periods the burden on public health services can be considerable.

Puumala virus is transmitted to humans by inhalation of aerosolised excreta of the bank vole Clethrionomys glareolus, which is the reservoir and vector. Activities which bring humans into close contact with bank voles or their excreta, such as living in endemic areas (6,7), farming (2,4,6,8), exposure to the forest (e.g. through forest work (6,8) or other activities which involve exposure to forest dust/earth (9)), raising dust during cleaning (9) or military field exercises (10), are all associated with increased risk of disease. Infection is related to prevalence of the virus in the vole population, as peaks in the incidence of human infection occur every 3-4 years and have correlated with peaks in the vole population (5). In Finland most cases in humans occur during the winter months.

To evaluate trends in incidence of human Puumala virus infections, we analysed national surveillance data from 1995 to 2002. The aim was to shed light on the pattern of infection in Finland, so as to provide information for a more detailed study to model epidemic occurrence.

Methods

Since 1995, all Finnish microbiology laboratories notify to the National Infectious Disease Register (NIDR) at the National Public Health Institute in Helsinki all laboratory-confirmed diagnoses of Puumala virus infection (mostly by antibody detection). Each notification includes information on date of specimen, date of birth, sex and place of treatment. Multiple notifications for persons with the same information received within a 12 month period are combined as one case.

As the incidence of human infections peaks during the winter months, we investigated 12-month periods from March to the following February. This method ensured that the entire winter season for peak infection was included within each period. The periods under study could then be investigated as epidemic or non-epidemic periods.

Surveillance data reported to the NIDR from March 1995 to February 2002 were analysed by age, sex and month of onset as well as by health district (HD) of treatment. There are 21 HDs in Finland, a country of 337 030 km2 with a population of 5.5 million. For each HD, the incidence for each 12-month period was calculated, as well as the mean incidence during non-epidemic and epidemic periods. The HDs were categorised as low, intermediate or high according to the range of their mean incidence for both epidemic and non-epidemic periods. The geographic distribution of mean incidence was examined.

Annual data from the National Population Register (1995–2000) were used as denominators for calculation of age- and sex-specific incidence, as well as incidence by HD. For each 12-month period, the population for the calendar year starting the period was used. Data from 2000 were also used for 2001–02.

Results

Between March 1995 and February 2002, 8184 laboratory-confirmed cases of Puumala infections were notified to the NIDR. Of these, 63% were male and 81% were between 25 and 64 years old. Male tended to be slightly younger than female (median: 42 years vs 46 years old; figure 1). The mean incidence among males was almost twice as high as among females, both for epidemic (43 vs 24 per 100 000/yr) and non-epidemic (19 vs 10 per 100 000/yr) periods.

During the time under study, three epidemic periods were identified (Figure 2), during which the total number of cases was over 1400 (range: 1481–2100). The average number for non-epidemic periods was 756 (range: 624-895). Incidence during non-epidemic periods ranged from 12 to 17 per 100 000 (mean 15 per 100 000), and this was more than twice as high during epidemic periods (mean 33 per 100 000). There was a small overall increase in the number of reported cases over the study period, mostly due to an increase in the eastern part of the country (data not shown).

Figure 2 shows the distribution of cases during the study period by month of diagnosis. Except for 2000/2001, when numbers were exceptionally low, non-epidemic periods contained two peak months (August and December). All three epidemic periods had December as the peak month. Nonetheless, there were at least three other months during epidemic periods in which the number of cases was greater than even the peak December month in any non-epidemic period.

The mean number of notified cases by HD was variable, from 0 to 96 during non-epidemic periods and from 3 to 206 during epidemic periods (data not shown). In addition, some HDs had epidemic periods at times that were different from those observed overall.

Helsinki was one of the five lowest HDs for mean incidence of disease. Other HDs with low incidence were also situated in southern Finland (Table 1 and map). The HDs with the highest incidence were located in the east central area of Finland, and the remaining (intermediate incidence) HDs were spread throughout the country.

Table 1
Health districts by mean incidence of reported Puumala virus infection per 100 000 population, Finland, 1995–2002

Discussion

There is a clear geographical variation in the incidence of reported Puumala virus infections in Finland. The rates are highest in the east central and northern areas of the country and lowest in the south. The lower incidence in the primarily urban and less forested south could be because this area has a smaller vole population, and therefore less likelihood of exposure (1). Cases reported in persons residing in these areas could have occurred after travel to endemic areas in the north or east. Typically, many Finns holiday in forest and lakeside cottages during the summer and autumn months, and forest trekking is common, especially in Finnish Lapland in the north.

The small increase in reported cases overall was partly due to increases in the eastern part of the country. This may, however, also be an artefact of the time under study. Epidemics of Puumala virus infections occur every few years, and the three epidemic periods happened to occur at the latter end of the timeframe. More years of data will be needed to more accurately estimate the trend.

The bank vole population in rural Finland peaks during the autumn months. As the temperature drops, and the ground freezes, rodents (including voles) seek shelter in or near human habitats (11), increasing the potential for exposure in humans. Hence the main peak in incidence of reported human cases in early winter, which then drops to a low point by the following spring (as the vole population density declines) (12,13). A previous study has suggested that, in humans, the smaller August peak is in the urban population (1).

The relation between vole population density and incidence of human infection is complex and poorly understood (14,15). However, since the early 1980s, bank vole populations have been reported to follow a 3-4-year cycle (1,15,16), often with two consecutive high density years, giving potential for epidemics in humans. Different parts of Finland have been in different phases of the cycle. These variations in seasonality and geographical distribution of the vole population in Finland are believed to have led to the irregular pattern of Puumala virus infection in humans.

As only microbiologically confirmed cases are reported nationally, the true extent of this disease is difficult to estimate. Both under-diagnosis and under-notification exist. In some HDs, the disease is so rare that it may not be recognised; and in all areas there may be persons with mild disease, who never visit a physician. In epidemic HDs, the disease may be so well recognised that some cases will be treated on clinical suspicion without microbiological confirmation, and district-to-district variation in the ratio of serological to clinical diagnoses has been reported (1). In addition, although for specialist physicians, clinical diagnosis may correlate well with serodiagnosis (17), this may not be the case in general. For patients with severe disease, differential diagnoses can include systemic inflammatory response syndrome or sepsis (18).

Puumala virus infection is an emerging disease, with a potentially increasing trend in Finland, for which there are no current official prevention guidelines. The main general recommendation, to avoid contact with bank voles or their excreta, is difficult both to implement and to monitor. In order to develop effective guidelines, it might be helpful if knowledge of high bank vole populations (preceding potential human epidemic periods) could be predicted. Further study of longer-term data, including investigation of the correlation between human incidence of disease and bank vole population cycles and/or seasonal weather patterns, might enable creation of a model to predict future epidemics. In addition, it is possible that any increase in number of cases is due to increasing bank vole population density, changes in the pattern of vole epizootics and/or changing human behaviour patterns. Identification of changes in risk factors will help in the design of interventions to reduce risk of infection.

Acknowledgements

We thank Pekka Holmström and Jaana Heino for their assistance during the investigation.