Culicoides species involved in the BTV‐3 epidemic, the Netherlands, 2023–2024

Abstract

Bluetongue virus (BTV) is an arthropod‐borne virus that is transmitted between ruminants by Culicoides (Order Diptera, Family Ceratopogonidae) midges. In September 2023, BTV serotype 3 (BTV‐3/NET2023) emerged in the Netherlands, causing a devastating epidemic in sheep and cattle. The aim of this study was to determine which midge species contributed to the spread of BTV‐3 and to what extent the virus is present in local midge populations. Midges were collected using Onderstepoort UV‐light suction traps on BTV‐affected farms in the centre of the Netherlands, from October 2023 till March 2024. Species, sex and parity of the midges were morphologically determined. Pooled female parous and gravid midges were subjected to pan‐BTV and BTV‐3 real‐time polymerase chain reaction (PCR) testing. Pool prevalence, minimum infection rate (MIR) and infection rate (IR) were calculated. In total, 33,093 midges were morphologically identified, all being indigenous Culicoides species. Of these, 10,835 parous or gravid female midges were selected and pooled in 383 pools (mean 28.3 midges per pool, range 1–115). A total of 155 pools (40.5%) tested BTV positive, with a mean MIR/100 of 1.4 and an IR of 2.2 (95% confidence interval: 1.9–2.6). All positive pools were from batches that were collected in October 2023 (week 40 and 41). BTV‐RNA was detected in the Culicoides species C. obsoletus Meigen, C. scoticus Downes and Kettle, C. chiopterus Meigen, C. dewulfi Goetghebuer and C. punctatus Meigen. The high proportion of BTV‐PCR positive midge pools is indicative of a high vector competence for BTV‐3/NET2023 of Dutch indigenous midges present on farms; it could potentially partly explain the rapid spread of the virus throughout the Netherlands.

• A high amount of Culicoides‐midges were captured using Onderstepoort light traps on farms with animals infected with bluetongue serotype 3 (BTV‐3) at the beginning of the outbreak in 2023.

• Species, sex and parity of the captured Culicoides‐midges were morphologically determined and gravid and parous Culicoides‐midges have been selected and pooled for BTV detection by real‐time PCR.

• Genomic material of BTV‐3 have been detected in Culicoides obsoletus, Culicoides scoticus, Culicoides chiopterus, Culicoides dewulffi and Culicoides punctatus species, suggestive of their vector competence.

INTRODUCTION

Bluetongue virus (BTV, Order Reovirales, Family Sedoreoviridae, Genus Orbivirus) is an arthropod‐borne pathogen of veterinary importance. Culicoides biting midges of the subgenera Avaritia Fox and Culicoides Latreille (Order Diptera, Family Ceratopogonidae) are competent vectors. These small insects are able to transmit BTV between susceptible hosts (ruminants and camelids) by subsequent bloodmeals, after the extrinsic incubation period of approx. 3–20 days (Carpenter et al., 2011). In early September 2023, BTV was detected in Dutch sheep located in the middle of the Netherlands, which was identified as serotype 3 (BTV‐3/NET2023) using whole genome sequencing (Holwerda et al., 2024). After introduction, the virus spread rapidly; by the end of November 2023, over 6,000 clinically affected farm locations and several thousand severely sick and dead ruminants (mainly sheep and cattle) were reported (Santman‐Berends et al., 2024; van den Brink et al., 2024).

Considering the situation in Europe, untill the end of the 1990s, BTV occurred only in parts of the Mediterranean Basin where Culicoides imicola Kieffer (subgenus Avaritia) was present. Therefore, this midge species was considered to be the only important BTV vector species in Europe (Mellor et al., 2008). The gradual expansion of BTV across southern Europe in areas where C. imicola appeared to be absent (Savini et al., 2004), together with the first unexpected epidemic of BTV (i.e., BTV‐8) in the northern parts of Europe, led to the identification of other midge species as possible vectors of BTV (Carpenter et al., 2015; Goffredo et al., 2015; Mehlhorn et al., 2007; Meiswinkel, Baldet, et al., 2008; Meiswinkel, van Rijn, et al., 2007; Paslaru et al., 2018; Torina et al., 2004; Vanbinst et al., 2009). The species that are associated with ruminants are among those that were identified as important BTV vectors: Culicoides obsoletus Meigen, Culicoides scoticus Downes and Kettle, Culicoides chiopterus Meigen, Culicoides dewulfi Goetghebuer of the subgenus Avaritia and Culicoides punctatus Meigen and Culicoides pulicaris L. of the subgenus Culicoides. These six Culicoides species can be found near livestock in pastures and farms (Elbers & Gonzales, 2023). The midges feed on ruminants and breed in dung, silage residues and other organic matter present at pastures and farms (Carpenter, Mellor, & Torr, 2008; Werner et al., 2020; Zimmer, Losson, et al., 2013).

The aim of this study was to get insight into the involvement of midge species in the BTV‐3/NET2023 outbreak and to what extent the virus is present in local midge populations.

METHODS

Farm selection and trapping

Onderstepoort UV‐light suction traps (Goffredo & Meiswinkel, 2004) were set up at 10 farms in the central part of the Netherlands with sheep and/or (dairy) cattle. Farms participated on a voluntary basis. Livestock on all participating farms were recently pan‐BTV‐polymerase chain reaction (PCR) positive (i.e., tested positive with a PCR test that detects all BTV serotypes). BTV had been diagnosed by PCR at each location in at least one animal, but no exact numbers of infected animals could be determined. Traps were placed inside barns, outside in the farmyard or in pasture and operated for at least 12 h (from about 1 h before sunset to and including 1 h after sunrise). Traps operated for variable periods, but all between the beginning of October 2023 (week 40) and the end of March 2024 (week 13). The day after trap operation, samples were collected and transported to the laboratory, stored in 70% ethanol. In addition, at one farm, two CO₂ baited BG‐Pro suction traps (Biogents AG, Regensburg, Germany), a sweep‐net and a mouth aspirator were once applied for insect collection. Samples were processed immediately or stored at −20°C.

Morphological identification

Due to labour and budget restraints, not all 320 collected samples could be processed for morphological identification. A total of 230 samples from six farms were processed: all collected samples from locations Tricht, Vinkeveen, Wilnis and Zoelen and a part of the collected samples from Leuvenum and Uddel (Figure 1, Table 1). Midges were separated from other insects and morphologically identified to species level based upon wing pattern, characteristic features of the head (antennae and palpae) and the abdomen (scutellum and spermathecae) (Delécolle, 1985; Mathieu et al., 2010). Since females of C. obsoletus and C. scoticus cannot easily be distinguished based on morphological characteristics, they were not identified to species level but recorded as C. obsoletus/scoticus. Following morphological identification, female midges were separated according to parity (Larska et al., 2013). Gravid and parous Culicoides were analysed for the presence of BTV, whereas the nulliparous and freshly blood‐fed midges were not tested. Gravid and parous Culicoides were pooled in 70% ethanol with, in principle, a maximum of 50 individuals per species per location per sampling date and processed for BTV‐PCR testing. However, due to convenience reasons, some pools contained more than 50 specimens.

FIGURE 1.

Geographical locations of the farms in the Netherlands where Onderstepoort UV‐light traps were used to collect midges. Each farm is displayed as a green dot. The purple square indicates the area where BTV‐3 was first observed in September 2023. BTV, bluetongue virus.

TABLE 1.

Details of sampling with Onderstepoort Culicoides UV‐light trap on the six farm locations.

Location name	Province	Type of farm	Sample dates	Sample week numbers	Sample frequency	No. of traps inside	No. of traps outside	Total number of traps	No. of samples
Uddel	Gelderland	Sheep	5 October 2023	40	Daily	0	3 a	3	3
Zoelen b	Gelderland	Dairy cattle	5–9 October 2023	40–41	Daily	3	2 c	5	74
			25 October 2023–27 March 2024	43–13	Weekly	1	1 c	2
Leuvenum	Gelderland	Dairy cattle	7–8 October 2023	40	Daily	3	0	3	6
Vinkeveen	Utrecht	Dairy cattle, sheep	11–14 October 2023	41	Daily	4	2 a	6	24
Tricht	Gelderland	Dairy cattle	24–25 October 2023	43	Daily	1	1 c	2	52
			1 November 2023–27 March 2024	44–13	Weekly	1	1 c	2
Wilnis	Utrecht	Sheep, young cattle	1 November 2023–27 March 2024	44–13	Weekly	1	2 (1 c , 1 a )	3	71
Total						16	10	26	230

Note: Traps were placed inside a barn or outside in the farm yard or in pasture.

^a Trap placed in pasture.

^b Sweep‐net, mouth aspirator and BG‐Pro also used for insect collection.

^c Trap placed in yard.

Nucleic acid extraction from midge pools

The ethanol was removed from the midge pools by removing the liquid and evaporation followed by the resuspension of midges in 500 μL of EMEM medium (Gibco) supplemented with 2% foetal calf serum (Capricorn) per pool. Then, the midges were homogenised using a pellet pestle and pellet pestle motor (Sigma) and debris was discarded after centrifugation for 10 min at 6,500× g at room temperature. Then, 200 μL of supernatant and 20 μL of an internal control of phocine distemper virus was subjected to nucleic acid extraction using the Magnapure 96 robotic extraction machine (Roche, Basel, Switzerland) in combination with the Magnapure 96 DNA and viral NA small volume kit. The nucleic acids were eluted in 100 μL.

Molecular detection of BTV using PCR

For molecular detection of BTV‐RNA, a pan‐BTV (Hofmann et al., 2008) and a BTV‐3 specific quantitative real‐time PCR (Lorusso et al., 2018) were used. External pre‐denaturation of the double‐stranded RNA of BTV was performed by warming 8 μL of elution in a plate for 1 min at 98°C, which was immediately cooled on ice for 5 min. Then, 5 μL of the denatured RNA was added to a new 96‐well plate containing 15 μL of the pan‐BTV or serotype 3 specific oligonucleotides (0.6 μM, respectively 1.0 μM final concentration for each primer), probe (0.5 μM final concentration) and 4× TaqMan Fastvirus 1‐step Master mix enzyme (Thermo Fischer, Waltham, USA) per well (Hofmann et al., 2008; Lorusso et al., 2018). The reverse transcription and amplification were performed in Quantstudio 5 PCR system (Thermo Fischer Scientific) using the following real‐time PCR programme: 52°C for 10 min followed by 95°C for 1 min and 45 cycles of 95°C for 15 s and 60°C for 30 s. Analysis was performed using the integrated Quantstudio software. Samples were tested in duplicate. Samples with a mean Ct value of ≤36.0 were considered positive, >36.0 and <40.0 doubtful and Ct values of ≥40.0 were interpreted as real‐time PCR negative. Pools with positive and doubtful results were retested with the pan‐BTV real‐time PCR (Hofmann et al., 2008). Also, these pools were tested with the serotype‐3 specific real‐time PCR (Lorusso et al., 2018) after a new extraction of nucleic acids. Pools with a mean Ct ≤36.0 in both pan‐BTV qPCRs were considered BTV positive.

Estimate of minimal infection rate and infection rate

Minimal infection rate (MIR) provides an estimate of the proportion of infected vectors (midges) in a population. This method assumes that only one vector in a positive pool is infected; hence, it provides a minimum estimate. Calculations were performed in R studio (R core team 2021). MIR was calculated as follows:

$$\mathrm{MIR}=\text{number of positive pools}/\text{total number of midges tested}\times 100$$

Infection rates (IRs) are estimated using a maximum likelihood estimate (MLE) method and take into consideration the pool size to quantify the probability of detection of at least one positive infected vector in a given pool as follows:

If we consider p as the probability a vector is infected, then the probability of one vector being not infected is 1 − p. The probability that all n vectors in pool i are negative is then ${\left(1-p\right)}^{n_i}$. We then estimate the probability that at least one vector is positive (therefore the pool being positive) as

$${\text{Pool}}_{\mathrm{pos}}=1-{\left(1-p\right)}^{n_i}.$$

Linearising this probability we get $\log \left(-\log \left(1-p_i\right)\right)=\log \left(n_i\right)+{\beta }_0$.In this equation $p_i$ is the probability of infection of pool $i$ given pool size $n$ and $\exp \left({\beta }_0\right)$ is the mean IR per vector. We can therefore estimate these rates by fitting a generalised regression model with a binomial distribution and a cloglog transformation (2021). Using this model additional co‐variates were included to compare IR among different species and sample location. For statistical comparisons on IR among locations and species, only samples from week 40 and 41 were considered, as only midge pools collected from those 2 weeks were BTV positive. Adjusted IRs were calculated and used for these comparisons.

Data presentation

The bar graphs and pie diagrams were generated using GraphPad Prism 10, version 10.2.3 (San Diego, USA).

RESULTS

Descriptive data

To investigate the vector role of Culicoides spp. for the circulating BTV‐3, samples that were collected on six farm locations in the Netherlands (Figure 1) between October 5 2023 (week 40) and March 27 2024 (week 13) were further processed. Details per location can be found in Table 1. On three farms, traps operated only in Autumn 2023; on the three other farms from Autumn 2023 to Spring 2024.

In total, 230 samples from the six locations were processed for Culicoides identification (Table 2). The number of collected midges between week 40 of 2023 and week 13 of 2024 is depicted in Figure 2. Less than five Culicoides per trap outside (i.e., three midges in one trap, zero midges in the other two traps) were caught in week 47 of 2023, as an indication of the defined start of the Culicoides free period. In week 51, one midge was caught in an inside trap. No midges were caught from week 52 of 2023 to week 10 of 2024. After a period of 11 weeks without collected midges, six Culicoides sp. were caught in one outside trap in week 11 of 2024, as an indication of the defined end of the Culicoides free period. In the same week, two and five midges were caught in two inside traps. Detailed information about the number of Culicoides spp. per week can be found in Table S1.

TABLE 2.

Number of collected Culicoides species per location.

Culicoides sp.	Leuvenum		Tricht		Uddel		Vinkeveen		Wilnis		Zoelen		Total
	F	M	F	M	F	M	F	M	F	M	F	M
C. achrayi	6	6											12
C. alazanicus												1	1
C. chiopterus	681	122	41	6	2		5196	459	6	1	2244	163	8921
C. circumscriptus		1					22	1	1				25
C. dewulfi	191	273	3	2	4		176	5	1		711	16	2582
C. duddingstoni											1		1
C. fascipennis							1						1
C. festivipennis	1	1			3		1				1	1	8
C. gornostaevae								1					1
C. impunctatus			1								2		3
C. lupicaris							2						2
C. newsteadi							1						1
C. nubeculosus s.l.											1		1
C. obsoletus		146		29		1		152		13		563	904
C. obsoletus/scoticus a	2575		600		23		2431		43		10,691	34	16,397
C. pallidicornis							1						1
C. pulicaris	4	1	4				136	1	2		3		151
C. punctatus	113	170	173	27	2		2661	166	179	23	444	35	3993
C. riethi											1	3	4
C. scoticus		24				1		4				31	60
Culicoides sp.		17										7	24
Total	4771	761	822	64	34	2	10,628	789	232	37	14,154	854	33,093

Abbreviations: F, female; M, male.

^a Not all males could be identified to species level.

FIGURE 2.

Distribution of collected midges per sampling week in 2023 and 2024. No sampling effort was performed in week 42. No midges were collected in the Onderstepoort traps between week 52 and week 10. See Additional file S1: Table S2 for detailed information.

In the 230 processed samples, 33,093 midges were identified, of which 2,507 males (7.6%) and 30,586 females (92.4%) (Table 2). Overall, C. obsoletus/scoticus were most abundant (52.5%), followed by C. chiopterus (27.0%), C. punctatus (12.1%) and C. dewulfi (7.8%). Culicoides pulicaris, who is also associated with sheep and cattle, was far less abundant (0.5%).

Parity determination

To assess which midge species could contribute to the spread of BTV, parity of female midges was determined. In total, parity of 16,481 female midges was determined, of which 8,145 (49.4%) were parous, 3,856 (23.4%) gravid, 2,464 (15.0%) nulliparous and 2016 (12.2%) bloodfed (Figure 3, Table S2). Considering the most abundant species, 84.7% of C. chiopterus were parous or gravid, 78.5% of C. dewulfi and 58.4% of C. obsoletus/scoticus.

FIGURE 3.

Pie charts of the most abundant Culicoides species of our study including the percentages of parity for each. Each colour represents each stage of parity: gravid (purple), parous (orange), bloodfed (blue) and nulliparous (green).

BTV pool prevalence

In total, 383 pools with a total of 10,835 parous and gravid Culicoides (range: 1–115 midges per pool; mean: 28.3 midges per pool) were tested with pan‐BTV real‐time PCR. Of those 383 pools, 209 pools showed positive or doubtful results (Figure 4), with 22.63 being the lowest mean Ct value and 39.89 the highest. These 209 pools were retested with the same pan‐BTV qPCR, leading to a total of 155 BTV positive pools. Genomic material of BTV could be detected in pools of C. obsoletus/scoticus, C. chiopterus, C. dewulfi and C. punctatus. Also, the 209 pools were tested with a BTV‐3 specific qPCR to confirm the present serotype, which resulted in the detection of BTV‐3 RNA in 185 pools (Ct values range: 22.71–40.99). In the pools where BTV‐3 could not be detected, a high mean Ct value was observed in the pan‐BTV, which could be due to the lower sensitivity of the BTV‐3 PCR in comparison to the pan‐BTV‐PCR.

FIGURE 4.

The distribution of Ct values of the midge pools after the first test for presence of BTV‐RNA with pan‐BTV qPCR (the grey bar 36–40 represents doubtful results which were excluded from further analysis). BTV, bluetongue virus.

Overall, mean pool prevalence was 40.5% (155/383), mean MIR per 100 was 1.4 and mean IR was 2.2 (95% confidence interval [CI]: 1.9–2.6). Detailed pool prevalences, MIR and IR per Culicoides species tested are shown in Table 3, and per farm location in Table 4. All positive pools were collected on three farms, and only in week 40 and 41 of 2023.

TABLE 3.

Pool prevalence, minimum infection rate (MIR) and infection rate (IR) for the Culicoides species.

Culicoides sp.	No. of specimens	No. of pools	Mean pool size	No. of pools BTV positive	Pool prevalence [%]	MIR/100	IR (95% CI)
C. chiopterus	6327	165	38.4	97	58.8	1.5	2.8 (2.2–3.4)
C. circumscriptus	1	1	1.0	0	0	0	NA
C. dewulfi	817	48	17.0	7	14.6	0.9	1.0 (0.4–2.0)
C. newsteadi	1	1	1.0	0	0	0	NA
C. obs/scot	2699	86	31.4	24	27.9	0.9	1.2 (0.8–1.7)
C. pulicaris	38	10	3.8	0	0	0	NA
C. punctatus	952	72	13.2	27	37.5	2.8	5.3 (3.4–7.9)
Total	10,835	383	28.3	155	40.5	1.4	2.2 (1.9–2.6)

Abbreviations: BTV, Bluetongue virus; CI, confidence interval; NA, not applicable.

TABLE 4.

Pool prevalence, minimum infection rate (MIR) and infection rate (IR) per location.

Location	No. of specimens	No. of pools	Mean pool size	No. of pools BTV positive	Pool prevalence [%]	MIR/100	IR (95% CI)
Leuvenum	1856	44	42.2	34	77.3	1.8	4.0 (2.7–5.8)
Tricht	128	24	5.3	0	0	0	NA
Uddel	4	3	1.3	0	0	0	NA
Vinkeveen	5428	173	31.4	106	61.3	2.0	4.2 (3.3–5.2)
Wilnis	47	13	3.6	0	0	0	NA
Zoelen	3372	126	26.8	15	11.9	0.4	0.5 (0.3–0.8)
Total	10,835	383	28.3	155	40.4	1.4	2.2 (1.9–2.6)

Abbreviations: BTV, Bluetongue virus; CI, confidence interval; NA, not applicable.

For statistical comparisons, only positive pools (which were collected in week 40 and 41) were considered, for which adjusted IRs were calculated. Using C. chiopterus (adjusted IR 2.8, 95% CI: 2.3–3.5) as a reference, the adjusted IR for C. dewulfi (1.1, 95% CI: 0.5–2.1) and C. obsoletus/scoticus (1.3, 95% CI: 0.8–1.9) were significantly (respectively p = 0.015 and p = 0.001) lower, and the adjusted IR of C. punctatus (6.7, 95% CI: 4.3–10.2) was significantly (p < 0.001) higher. Additionally, the adjusted IR at location Zoelen (0.5, 95% CI: 0.3–0.8) was 7.7 times significantly (p < 0.001) lower than the adjusted IR at the location Leuvenum (4.0, 95% CI: 2.7–5.8) and 8.1 times significantly (p < 0.001) lower than at location Vinkeveen (4.2, 95% CI: 3.3–5.2). No significant difference in adjusted IR between locations Vinkeveen and Leuvenum was observed.

DISCUSSION

Midges were collected at different time points and lengths at BTV‐3‐affected sheep and/or cattle farms in the central part of the Netherlands and processed for investigation of their possible vector role in the BTV‐3 outbreak. The results show that a high proportion of midge pools were BTV‐PCR positive and that five midge species are likely vectors for BTV‐3/NET2023: C. obsoletus, C. scoticus, C. chiopterus, C. dewulfi and C. punctatus.

Sampling for midges was performed on 10 BTV positive farms, but due to the time‐consuming nature of morphological identification, unfortunately not all samples could be processed. We aimed to at least complete analyses of the farms in the region where the outbreak started, and the farms at which sampling was continued till the next Spring.

The six species of midges that were collected most were the species that are associated with livestock farms (Cuéllar et al., 2018; Elbers et al., 2018; Elbers & Meiswinkel, 2015; Griffioen et al., 2011; Takken et al., 2008). Of the total of collected midges, the females of C. obsoletus/scoticus were the most abundant. Morphological characteristics can be used to distinguish females of C. obsoletus/scoticus (Foxi et al., 2016), which would need morphometric measurements on dissected abdomens. Due to time restraints, we did not identify females of C. obsoletus/scoticus to species level. In view of the distribution of these two species in the male specimens (Table 2) and considering the lesser attraction of the Onderstepoort UV‐light trap for C. scoticus (Carpenter, Szmaragd, et al., 2008), it is likely that the larger part of the collected females were C. obsoletus and the lesser part C. scoticus. This is in accordance with other findings (González et al., 2013; Hoffmann et al., 2009; Sanders et al., 2012). Even though these two species share many ecological and behavioural traits such as host and breeding site preferences, C. obsoletus is more likely to breed in manure and C. scoticus more likely to breed in plant material such as rotting leaf litter (González et al., 2013). Two other prevalent species in our study, C. chiopterus and C. dewulfi, also show coprophilic breeding behaviour (Zimmer et al., 2014). The two members of the subgenus Culicoides (C. punctatus and C. pulicaris) that were also collected in our study preferably breed in plant material such as silage residues (Zimmer et al., 2014). In short, findings of these six species as the most abundant Culicoides species were expected, as they breed in habitats and feed on hosts that are widely available on cattle and sheep farms.

In this study, we expected two other species to be collected in higher numbers than we actually did: Culicoides newsteadi Austen and C. impunctatus Goetghebuer. Higher numbers of the first species, C. newsteadi, were expected due to previous findings on farms in the Netherlands (Meiswinkel, Goffredo, et al., 2008) and in Italy (Foxi et al., 2016; Goffredo et al., 2015), even though our findings are in accordance with another study on farms in the Netherlands (Takken et al., 2008). In recent years, taxonomic changes have been made within the Pulicaris complex: what used to be called Culicoides species near newsteadi, which was the species that was found in Meiswinkel, Goffredo, et al. (2008), is now considered (part of) C. punctatus. Also, Culicoides halophilus is not considered a separate species anymore, but is now synonymised with C. newsteadi (Borkent & Dominiak, 2020), which is a species associated with brackish water habitats. So, given these taxonomic reclassifications, our findings of only one specimen of C. newsteadi are appropriate for the farm locations and habitats where the traps were placed.

Additionally, we expected to collect higher numbers of C. impunctatus than the three specimens that were found. Two farms (Vinkeveen and Wilnis) were located in a low‐lying boggy marshland area in the centre of the Netherlands (het Groene Hart) and C. impunctatus breeds in such habitat (Takken et al., 2008; Zimmer, Smeets, et al., 2013). Controversially, the three specimens of C. impunctatus that were detected were found on two farms outside the Groene Hart area (Tricht and Zoelen). The period of midge collection, which started in October and ended the following April, may partly explain why we found so few specimens, because C. impunctatus peaks in late Spring/early Summer (Takken et al., 2008; Verdonschot & van Noord, 2021). Also, to catch C. impunctatus, traps with CO₂ and bait rather than with UV‐light are preferred (Bray et al., 2020), so our use of Onderstepoort UV‐light traps may not represent the population present. This species is of interest because BTV vector competence has been shown for C. impunctatus during laboratory experiments (Carpenter et al., 2006; Jennings & Mellor, 1988). The first batches of eggs of C. impunctatus in Spring are deposited before a bloodmeal has been taken (autogeny) (Blackwell et al., 1992), which means that, besides that high initial population densities can be reached without hosts being present, females of C. impunctatus need to survive at least three reproductive cycles before transmission of an arbovirus such as BTV can occur (Carpenter et al., 2013). Even though C. impunctatus will feed on ruminants, it is a general feeder that feeds on many host species (Blackwell et al., 1994). This ‘dilution effect’ and the autogenous egg‐laying behaviour make involvement of C. impunctatus in BTV outbreaks less likely. On the other hand, population densities of this species can be huge, resulting in high biting rates, which would favour pathogen transmission. In conclusion, our results do not provide clarity on the possible involvement of C. impunctatus in the BTV‐3 outbreak.

In our study, no C. imicola Kieffer specimens were found. This specific midge species is within Europe only circulating in the Mediterranean Basin and is a well‐known vector in Africa for different orbiviruses like BTV, epizootic haemorrhagic disease virus (EHDV) and African horse sickness virus (AHSV). Our findings are in accordance with other findings that found no C. imicola in northern European countries (Brugger & Rubel, 2013; Griffioen et al., 2011; Hoffmann et al., 2009; Meiswinkel, Baldet, et al., 2008; Meiswinkel, Goffredo, et al., 2008).

Reported pool prevalences (Table 5) for the previous largest BTV epidemic in northwestern Europe (BTV‐8 in 2006–2008) were between 1.2% and 2.5% (Dijkstra et al., 2008; Hoffmann et al., 2009; Meiswinkel, Baldet, et al., 2007), even though the study of Mehlhorn et al. (2007) found a higher prevalence of 14.6% (Mehlhorn et al., 2007). The latter may be explained by the fact that they tested both unfed and fed midges, even though Hoffmann et al. (2009) also did not differentiate between parity of the tested midges and found a pool prevalence of 2.4%. Voigt et al. (2024) collected midges on German farms at the end of September 2023 and reported a BTV‐3 pool prevalence of 0.06%. An explanation for the different pool prevalence between our and their study might be that the epidemiological situation in Germany at the time and location of trapping was different than in our study, as the first BTV‐3 related clinical cases in Germany were only reported in October 2023. An Italian study that used Embryonating Chicken Eggs (ECE) test for BTV detection, which is a less sensitive test for BTV detection than PCR, reports pool prevalences of 5.4% (BTV‐2) and 1.8% (BTV‐9) (Savini et al., 2004). Two other studies from Italy report BTV pool prevalences of ~38% (Foxi et al., 2016; Goffredo et al., 2015). The BTV pool prevalence found in our study (40.5%) is more in accordance with entomological results from southern Europe (Italy) than from northwestern European countries (Table 5).

TABLE 5.

Bluetongue virus (BTV) pool prevalence (tested with polymerase chain reaction, unless stated otherwise), minimum infection rate (MIR/100, unless stated otherwise), infection rate (IR) in Culicoides sp. and reported specifics about sampling, in this and other studies.

Country	Outbreak BTV serotype	Sampling period	Sampling location specifics	Pool prevalence [%] (positive pools/tested pools)	MIR	IR (95% CI)	Reference
The Netherlands	BTV‐3	October 2023–March 2024	6 BTV‐affected sheep and/or cattle farms	40.5 (155/383)	1.4	2.2 (1.9–2.6)	This study
The Netherlands	BTV‐8	October 2007	5 sheep and/or cattle farms	1.2 (1/81)			Dijkstra et al. (2008)
The Netherlands	BTV‐8	September 2006	1 BTV‐affected dairy cattle farm	2.5 (1/40)			Meiswinkel, van Rijn, et al. (2007), Meiswinkel, Baldet, et al. (2007)
Germany	BTV‐8	April 2007–May 2008	85 cattle farms, mostly in BT restriction zone	2.4 (585/24,513 batches)			Hoffmann et al. (2009) a
Germany	BTV‐8	August 2006–January 2007	2 farms with seropositive cattle	14.6 (39/267)			Mehlhorn et al. (2007) a
Italy	BTV‐2 BTV‐9	July + September 2002	3 BTV‐affected sheep and goat farms	5.4 (3/56) 1.8 (1/56)			Savini et al. (2004) b
Italy	BTV‐1 and BTV‐4	2012–2014	339 affected farms	37.8 (1107/2925)	1.3		Goffredo et al. (2015)
Italy	BTV‐1	April–July 2013	1 BTV‐affected sheep farm	38.7 (63/163)	3.1		Foxi et al. (2016) c
	BTV‐2	2001	1 BTV‐affected sheep farm	Not clear (?/68)	Not reported
Morocco	Not clear	2018–2021	Not clear	82 (45/55)	2.2	4.9 (3.3–6.4)	Daif et al. (2024) c

Note: Parous and gravid midges were tested, unless stated otherwise.

Abbreviation: CI, confidence interval.

^a All female Culicoides parities tested for BTV.

^b Tested with Embryonating Chicken Eggs (ECE) test.

^c Only parous female Culicoides tested for BTV.

Besides pool prevalence, MIR and IR were calculated. MIR is the ratio of the number of positive pools to the total number of midges tested, with the underlying assumption of only one infected midge in a positive pool. This assumption may become problematic when IRs are high and/or pool sizes are large. IR is calculated as a MLE and is estimated using a binomial model to approximate the IR based on the available data. MIR estimates the lower bound of the IRs, and IR estimates the IR within the vector population itself. IR can be considered the estimated proportion of infected vectors that are also infectious in a population at a given timepoint and is one of the factors to estimate transmission risk (Bustamante & Lord, 2010). When IRs are low and pool sizes are small, MIR and IR are alike, but can be different in other conditions. In our study, overall MIR/100 was 1.4 and IR was 2.2 (95% CI: 1.9–2.6). The difference between these two can be explained by relatively high IRs and pool sizes that were not small. The significantly higher adjusted IR for C. punctatus shows that this species, even though it was not the most abundant species, represents a relatively high transmission risk. The significant differences in the adjusted IR between the three locations where BTV positive midge pools were found may be explained by possible differences in BTV circulation among the ruminants (e.g., epidemiological situation) and/or differences in farm management.

MIR and especially IR are less often reported in entomological studies for BTV (Table 5). The combination of the pool prevalence and MIR found in our study is in accordance with the abovementioned Italian study (Goffredo et al., 2015) but not with the other (Foxi et al., 2016). This could be due to differences in pool size (50 Culicoides/pool vs. 10/pool). To our knowledge, the only BTV entomological study in or near Europe that reports IR is a study performed in Morocco with an IR of 4.9, a MIR of 2.2 and pool prevalence of 81.8% (Daif et al., 2024).

Even though to prove vector competency individual heads including the salivatory glands need to be tested (Koenraadt et al., 2014), in our study we pooled and tested parous and gravid female Culicoides. This was done according to the reasoning that if a female had already laid eggs (parous) or contained eggs (gravid), a blood meal had already been digested (assuming anautogenous egg laying behaviour, which the ruminant‐associated Culicoides spp. show), and in case the bloodmeal contained BTV and the Culicoides species is not vector competent, testing of parous or gravid midges results in negative PCR or with high Ct values. However, if the midge is vector competent, BTV could already have replicated and disseminated into the body, including the salivary glands. In that case, BTV will be detected by PCR testing, with low Ct values. However, the Ct value of most positive pools ranged from 30 to 36, making it unknown at what stage of spread the virus is in the midge at the time of capture. In a study performed by Veronesi et al. (2013) lab‐adapted midges, after digestion of an infected blood meal, showed a Ct value of ±32–33 during 2–3 days after start of the infection, which could be representative of a baseline at the start of the viral dissemination.

So, finding positive pools of parous and gravid midges is merely suggestive of vector competence, even though not probative. For a midge species to contribute as a vector to a BTV outbreak, the species must show a certain degree of vector capacity. Vector capacity includes vector competency and ecological factors such as seasonality, longevity, population size, biting rate and extrinsic incubation period (Mullens et al., 2004). Our study showed varying BTV IRs of five native Culicoides species: C. obsoletus, C. scoticus, C. chiopterus, C. dewulfi and C. punctatus. It can be argued that these species show more or less vector capacity due to their host preferences, seasonality and possible vector competency.

A recent study performed by (Boender et al., 2024) showed with a kernel analysis that two of the three important kernel parameters describing the spread of the epidemic were comparable to the BTV‐8 outbreak in 2006–2008 in the Netherlands. The amplitude of the transmission kernel was significantly higher for the BTV‐3 epidemic compared to the 2006–2008 BTV‐8 epidemic. This might be explained by a much higher average temperature during the first year of the BTV‐3 epidemic and may also, according to our results, be partly explained by a higher proportion of infected midges. The high proportion of positive midge pools and estimated IR may partly explain the rapid spread of BTV‐affected farms in the beginning of the outbreak in 2023 and shows that the vector capacity of BTV‐3 seems very high in the Dutch indigenous midge population.

CONCLUSIONS

Culicoides species that show preference for feeding on and breeding near ruminants were the most abundant species that were collected at the six sheep and cattle farms. Genomic material of BTV was quantitatively detected in varying amounts in five of these Culicoides species: in C. obsoletus/scoticus, C. chiopterus, C. dewulfi and C. punctatus. Pool prevalences were high compared to earlier described BTV epidemics in Europe, indicating an important role of one or more of these Dutch indigenous midges in the spread of BTV‐3/NET2023. Estimated IRs were at least twice as high as the estimated MIR and probably reflect better the true level of infection of midges. In conclusion, these ruminant‐associated Culicoides species, commonly present in northwestern Europe, are likely to be competent vectors that are involved in the transmission of BTV‐3.

AUTHOR CONTRIBUTIONS

Mathilde Uiterwijk: Conceptualization; data curation; investigation; project administration; methodology; supervision; writing – original draft; writing – review and editing. Frans Jacobs: Investigation; project administration. Karst de Boer: Investigation. Arno‐Jan Feddema: Investigation. Rianka P. M. Vloet: Investigation. Marian Dik: Investigation. José L. Gonzales: Formal analysis; writing – original draft. Piet A. van Rijn: Writing – original draft. Armin R. W. Elbers: Conceptualization; methodology; writing – original draft. Melle Holwerda: Conceptualization; data curation; funding acquisition; methodology; writing – original draft; writing – review and editing; project administration; supervision; visualization.

FUNDING INFORMATION

For this study, WBVR was financially supported by the Ministry of Agriculture, Fisheries, Food security and Nature (project number: KB‐37‐003‐062 WOT‐KB BTV in knutten).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Table S1. Number of midges per sampling week.

Table S2. Distribution of parity per Culicoides species.

Uiterwijk, M. , Jacobs, F. , de Boer, K. , Feddema, A.‐J. , Vloet, R.P.M. , Dik, M. et al. (2026) Culicoides species involved in the BTV‐3 epidemic, the Netherlands, 2023–2024. Medical and Veterinary Entomology, 40(1), 111–122. Available from: 10.1111/mve.70007

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.