This is the first extensive survey of TBEV prevalences in Ixodes ricinus in Norway, and our study confirms the existence of TBEV endemic foci in Norway. The seven localities tested for TBEV (Figure1) had an estimated overall prevalence rate of 0.53%, which is in accordance with a Swiss study reporting a mean prevalence of 0.46% in endemic areas. The average prevalence in foci of endemic areas in Europe ranges from 0.1–5%. TBEV prevalence point estimates range from 0.11%–1.22% for the various locations in this study and a significantly higher EPP was found at S5 (1.22%) compared to S1 (0.11%) and S7 (0.21%) (Table3). Sites S3 (0.66%), S4 (0.70%) and S5 exhibit the highest point estimates of prevalence, and are located within the municipalities that have the highest number of TBE cases reported in Norway.
Previous studies have proposed that TBEV is distributed in a patchwork pattern, which is probably due to climatic conditions, virus prevalence, vector and host relationship and other factors.
The prevalence of TBEV in ticks is a suitable marker for TBE risk analysis in natural foci[50, 51], but cannot be directly translated into a risk for the population. The risk of human infection is the product of the hazard (number of infected ticks, which is a product of tick abundance and pathogen infection prevalence) and contact rate between the infected ticks and humans. As a zoonotic agent, TBEVs circulates widely in wildlife and the appearance of the virus is likely to be far more widespread than revealed by human cases. The virus primarily circulates in rodents without human cases being reported in the areas until either the hazard or the contact rate increase above a certain variable level. The distribution of TBE cases is therefore shaped by the distribution of the human population and the behaviour that favours human-tick contact. Due to the focal distribution of TBEV, the estimated prevalence cannot be generalized to reflect the true prevalence in the study region, whereas the prevalence of infected ticks within risk areas has also been reported to vary considerably between sites[9, 53]. In Norway the distribution and prevalence of TBE Virus (TBEV) in tick populations is largely unknown. Our findings are important to increase the awareness of TBEV risk areas in Norway, while also being useful for advising the public about TBE infections and its prevention together with vaccine recommendations.
The screening of pooled samples is complicated since it is impossible to determine if a positive result is due to one or more infected ticks. The simplest method to estimate pooled prevalence is the calculation of Minimum Infection Rate (MIR); which assumes that a positive pool only contains a single infected tick and provides an estimate of the average minimum infection prevalence[35, 54]. This method is inaccurate and might underestimate the actual level of infection and gives only point prevalence estimates with no information about the uncertainty of the estimate. The Frequentist methods (EPP) calculate maximum-likelihood estimates of true prevalence and confidence limits, although the difference between the point prevalence estimations of MIR and EPP in our study is negligible due to the sites having a low prevalence level. However, EPP provides a measurement of the uncertainty contained in the confidence interval associated with the apparent prevalence estimations (Table3).
Estimating prevalence is subject to errors influenced by the sampling process. This source of error is influenced by both the number of ticks examined and by the methods used to examine them. More than 730 nymphs were obtained from each site in our study, with the exception of the location S3. This allows for a good estimation of prevalence with a high confidence limit and a high precision (Table3). When prevalence is <10%, pooled testing is comparable to or better than individual testing. This is assuming that the sensitivity of the pool test is equal to the test used for individuals[55, 56]. The pool size will obviously influence the variance of a prevalence estimate, and the effect of reducing the pool size depends primarily on the true prevalence in the population investigated. If the prevalence is approximately 0.1%, the standard deviation will be small, so even up to a pool size of 30, i.e. a smaller pool size would not improve the accuracy of the estimates, while if the prevalence is 0.3 an effect will already be seen at a pool size of 8–10. The effect of having smaller pool sizes is therefore most prominent at the higher prevalence levels, and reducing the pool size would not eliminate the inherent problem with the MIR method. For instance, with a pool size of 10, and assuming a 10% prevalence it is possible that one individual in each pool could be positive, which would result in an estimated prevalence of 100%.
Many papers on TBEV in ticks are based on MIR as a measure of the prevalence of virus infected ticks. A variable number of ticks in each pool, as well as a mix of adults, larvae and nymphs makes it difficult to compare prevalence studies between countries or regions within a given country[17, 57–59]. The current study was performed on a large number of nymphs from each site calculating both EPP and MIR, but using EPP as the method of choice. This study used a fixed number of nymphs in each pool, and only nymphs were included. Using this method ensures that similar amounts of nucleic acid were isolated from each pool, thereby resulting in a reliable EPP when comparing TBEV presence and absence. In our study, high Ct values close to the detection limit of the RT-PCR indicated that only a few nymphs in each pool carried the TBEV.
Tick-borne flaviviruses are known to occur at a relatively consistent, low prevalence in tick populations. The mechanism that leads to a stable perpetuation of tick-borne flavivirus in nature has not been fully researched and needs further study. The key to TBEV maintenance was proposed by Labuda and Randolph (1999) to be the non-viraemic transmission of TBEV between infected nymphs and uninfected larvae when co-feeding on rodents. Co-feeding has been shown to be favoured in the spring, when temperatures rise rapidly and exceed 10°C, with the rate of spring warming being used to explain the distribution of TBEV foci. The rate of spring warming together with the mean January minimum temperature, estimated in our study, is in line with Randolph and Sumilo’s (2007) results for TBEV-presence sites. Lindgren and Gustafson found that an increase in TBEV prevalence was significantly related to a temperature range between 5 and 8°C. Randolph and Sumilo used the slope of the 10-day mean Land Surface Temperature (LST) obtained from satellites to calculate the number of days corresponding to a daily mean temperature between 7 and 10°C in spring. LST is regarded as the Earth’s skin temperature; however, the land surface is far from being a skin or a homogenous two-dimensional entity: it is composed of different materials with various geometries both of which complicate LST estimation. In our study, LST data were not suited due to the great variability in surface types within 1 × 1 km in the study areas, corresponding to the pixel resolution in the LST dataset. In addition, there are only minor differences between the slopes of the increase in maximum surface temperature and the increase in a 10-day running mean of daily mean temperatures for the actual sites during spring.
The meteorological data used in our study (i.e. the grid data and observations from the weather stations) should be considered as a fairly rough estimate for the local micro climate found near or at the ground surface in the habitat of the ticks. Tick development depends on the combination of RH and temperature, and these combined factors of RH and temperature had a highly significant effect on the moulting of engorged larvae into nymphs and engorged nymphs into adults. Previous studies have shown that SD affects tick dynamics and thus tick numbers, while RH on its own influences the virus replication level in the ticks. Gilbert found that there was a strong negative effect of altitude on questing tick abundance, and deduced that the effect of altitude was probably due to temperature and humidity, e.g. with SD playing an important role.
It is crucial to study the microclimatic conditions in particular RH and SD (Figures2 and3) because the tick phenology, and therefore the degree of co-feeding transmission could be influenced by SD, as well as TBEV replication within the tick body possibly being influenced by RH[45, 64, 66, 67]. Studies have shown that RH has an influence on the maintenance of the TBEV within ticks[64, 68]. Experimental studies show that low humidity promotes the disappearance of the virus from the tick body, and that higher humidity affects the multiplication of the virus in ticks and influences the degree of overall virus infection rate in the tick population. An increased SD causes a decreased tick density (or decreased availability) and limits questing duration, and that SD is a key factor that affects not only questing nymph abundance, but also the seasonal pattern of nymph questing activity[15, 66, 69]. Larvae are more sensitive to even smaller changes in SD than nymphs[15, 70], and once SD becomes so high that it limits tick questing, the TBEV transmission will be lower. High SD also affects tick survival since if high SD prevail, ticks will eventually use up their fixed energy reserves and die. Thus long-lasting high SD might limit the development of tick populations and affect tick abundance. It is not primarily the density of ticks (ticks per square meter or ticks per volume) that is reduced when the density of water molecules in the air at ground level decreases. Instead, the activity of the ticks, which to the collector gives the impression that the tick density has diminished – in reality it is the availability (the product of tick density and tick activity) that is reduced by the increasing “deficit” of water in the air. In our study, we found that the site with the lowest mean RH also had the lowest EPP, whereas the site with the highest mean RH had the correspondingly highest EPP, which is in accordance with Naumov and Danielova’s findings[64, 68]. The literature indicates that SD acts primarily on the TBEV transmission pathway by reducing the degree of co-feeding, while RH seems to directly affect the level of TBEV replication within the tick population. Our results suggest that microclimatic conditions, particularly RH and SD, are of major importance locally, and are in line with previous findings[13, 15, 45]. In order to improve the representativeness of the climate data, it is necessary to perform measurements of temperature and humidity at the actual site and at sites where I. ricinus is abundant but where TBEV seems to be absent, which will be considered in future studies.