A rapidly progressive disease developed in a 2-year-old Standardbred stallion in Finland; clinical signs were fever, scrotal swelling, and ventral edema (Technical Appendix[PDF - 391 KB - 4 pages] Figure); multifocal, hard, nodular skin lesions (Figure 1, panel A) and moderately enlarged lymph nodes were also apparent. The horse was apathetic and lost weight despite a good appetite. The attending clinicians suspected generalized lymphoma. However, a biopsy sample taken from nodular skin lesions showed proliferative dermatitis (Table). The horse had secondary immune-mediated hemolytic anemia 1.5 months after onset of disease; because the prognosis was poor, the horse was euthanized in September 2013. The body was received at the University of Helsinki Faculty of Veterinary Medicine (Helsinki, Finland) for a postmortem examination that month.

In necropsy, the horse was found to be thin and poorly muscled. Multifocal, nodular, dry, hard, proliferative lesions in the skin were mainly on the muzzle, lower forelimbs, and ventral abdomen. Moderate edema was present in the abdomen, scrotum, and all limbs. Thickened and hyperemic mucosa in the small intestine, moderately swollen mesenteric lymph nodes, and ascites were visible.

Histologically, the skin lesions were characterized by severe multifocal lymphohistiocytic dermatitis with intraepidermal vesicles caused by marked ballooning degeneration of the stratum granulosum (Figure 1, panel B). Eosinophilic intracytoplasmic inclusion bodies were seen in keratinocytes. Intestinal tissue, lungs, and mesenteric lymph nodes showed chronic, lymphohistiocytic inflammatory changes (Table). Special stains for mycobacteria were negative.

Because the histological findings of the skin samples suggested poxvirus infection, we collected a frozen plain skin sample and slices from formalin-fixed, paraffin-embedded skin, lung, lymph node, and spleen for virological studies. We attempted virus isolation from the skin sample in green monkey and baby hamster kidney cells and saw negative results. DNA was extracted by using the DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany), but no OPV DNA was detectable by real-time PCR (9) (Technical Appendix[PDF - 391 KB - 4 pages] Table). However, PPV DNA or that of a closely related virus was present in the skin samples: both the Pan-PPV PCR targeting the PPV envelope phospholipase gene (ENV) (11) and the high-GC (guanine-cytosine) pan-pox PCR targeting the large subunit of the poxvirus RNA polymerase gene (RPO147) (10) produced amplicons (Table), although several other primer pairs targeting PPV genes were negative (Technical Appendix[PDF - 391 KB - 4 pages] Table).

Sequencing of the PCR products showed that the ENV (GenBank accession no. KR863114) and RPO147 (GenBank accession no. KR827441) sequences shared 80%–89% nt and aa identity with other PPVs, depending on the virus species. The RPO147 sequence was 99%–100% identical at nt level and 100% identical at aa level to the sequences of the 2 recent poxvirus isolates (2012_37 and 2013_013 RPO147) from humans in the United States (7). In phylogenetic analyses, the sequences from the horse in this study and from these human patients grouped together, forming a different lineage within the PPVs and separate from other related poxviruses, molluscum contagiosum virus and squirrelpox virus (Figure 2). The equine poxvirus was designated F14.1158H.

Although the skin lesions showed poxvirus infection, formalin-fixed samples from internal organs contained no viral inclusion bodies and were negative for PPV by PCR. This finding is in accordance with the fact that PPVs are specialized to replicate in the highly specific immune environment of skin (13). Further investigations are required to show whether the poxvirus caused the generalized infection in addition to dermatitis.

The owner, breeder, and trainers of the horses on the farm where this horse became ill were unaware of any other animal or zoonotic cases in the premises and disclosed no contact between the horse and ruminants. The horse had lived in contact with many horses and several dogs and cats in 3 locations in southern parts of western and eastern Finland before being transferred to the last training stable. A few months before onset of clinical signs, the horse had been trained at a farm where cows had been kept 25 years earlier. During the illness, the horse lived in a stable with 17 horses, shared corrals and equipment, and had muzzle contact with 2 horses in adjacent stalls. Despite the direct and indirect contacts, all other horses, the 3 caretakers, and the trainer remained asymptomatic.

We report a clinical equine infection with a novel poxvirus in Finland. The infection is at least of dermatitic relevance for horses, and veterinary awareness is needed. The sequence analysis based on conserved genes revealed a close relationship between this isolate and recent poxvirus isolates from humans with horse contact in the United States (7). Although sequence data are limited and the geographic distance between this equine case and the recent cases in humans is remote, the close genetic relatedness suggests that horses have a possible role as reservoir or vector of an emerging zoonotic poxvirus, necessitating medical awareness and emphasizing the importance of the One Health approach (https://www.onehealthcommission.org/). The horse as an origin for zoonoses is not uncommon: as many as 58% of emerging zoonotic pathogens infect ungulates (14). As for cowpox virus, horse and human may be infected from a common source, such as rodents, and not necessarily from each other. This case appeared sporadic and not very contagious, and the transmission route remained unresolved. Further studies are needed to elucidate ecology, epidemiology, prevalence, and possible zoonotic transmission.

As our limited sequence analysis suggests, the virus we detected is most closely related to PPVs and may merit being classified as a new Parapoxvirusspecies. However, many of the established PPV primer pairs did not produce PCR product, which suggests that the virus is different from the established PPV species and may represent a new poxvirus genus. More sequence data are needed to validate the taxonomic classification of the equine poxvirus. In conclusion, our results provide further evidence that horses are a possible source of the new poxvirus infection recently observed in humans.

During winter 2014 and into the early spring 2015, we became aware of an unusual number of reported Vibrio infections in northern Europe. Colleagues at the European Centre for Disease Control relayed the initial information to the Centre for Environment, Fisheries and Aquaculture Science (Weymouth, UK) and the University of Bath (Bath, UK). The information suggested that an unprecedented number of Vibrio infections had been observed in Sweden and Finland during summer 2014 and that many cases were reported in high-latitude coastal counties.

To scrutinize cases of infection, we took several approaches. We initially contacted the Public Health Agency of Sweden (Stockholm, Sweden) and the National Institute for Health and Welfare (Helsinki, Finland), as well as other northern Europe reference laboratories, in December 2014. Although vibriosis is not regionally notifiable in Europe, Finland and Sweden maintain national databases of Vibrio infections. In Finland, V. cholerae is a notifiable infection, and isolates from persons with suspected infections are submitted to the reference laboratory for confirmation, serotyping, and PCR testing for the cholera toxin gene (ctx). Also, other Vibrio species (e.g., V. vulnificus, V. parahaemolyticus) may be sent to the reference laboratory for subsequent species-level confirmation. In Sweden, diarrhea with CTX-producing V. cholera O1 or O139 is a notifiable disease, as is infection with other Vibrio species, including V. cholerae not producing CTX that causes wound infections, septicemia, enteritis, and otitis. Isolates of V. cholerae are sent to the Public Health Agency of Sweden for serotyping and confirmation of virulence factors, such as ctx, using appropriate molecular methods, such as PCR.

For cases identified in 2014, the geographic location of each reported infection was established (e.g., town or city where the patient was treated). Where possible, information relevant to disease transmission, such as possible water-associated activities, also was gathered; however, for many cases, this information was not available. Basic epidemiologic data on each case, including patient sex and age, was subsequently collated, as was the site of bacterial isolation (e.g., wound, ear, blood). The date the case was reported to regional authorities was determined, and for a subset of cases, data on the onset of reported symptoms also were established. To assess recent trends regarding infections, we collated Vibrio cases identified in Finland and Sweden from 2005 onward and omitted from analysis cases we suspected of being foreign-acquired.

To assess the possible role of extreme weather events on the emergence and dynamics of Vibrio disease in Finland and Sweden, we analyzed the epidemiologic data alongside long-term sea surface temperature (SST) records (HadISST [Hadley Centre Sea Ice and Sea Surface Temperature dataset] and ERSST [Extended Reconstructed Sea Surface Temperature dataset, v3b from the US National Oceanic and Atmospheric Administration (NOAA)] [4]). We used satellite-derived data to scrutinize temperature conditions and changes in the Baltic Sea area using NOAA’s Optimum Interpolation v2 Daily SST Analysis dataset that integrates satellite SST data retrievals. NOAA data (baseline period of 30 years [1971–2000]) was used to determine anomalies from this dataset. We also scrutinized daily SST and SST anomaly retrieval data from 6 fixed positions in the Baltic Sea area, which included the transitional waters between southern Sweden and Denmark, the southeastern and mideastern Baltic coasts of Sweden, and Bay of Bothnia (northern Baltic) and southern coast of Finland. To assess the significance of climatologic data from summer 2014, we also used long-term oceanographic datasets to analyze SST. In situ SST was provided by the Finnish Meteorological Institute and was downloaded on November 14, 2014. We also used instrumental measurements of SST in coastal areas in the Baltic Sea area. We removed short-term fluctuations from the buoy data by applying a 1-hour wide median filter to the original dataset.

Statistical tests used to infer the relationship between maximum SST and annual Vibrio case occurrence were investigated by using a generalized linear model that assumed a quasi-Poisson error distribution (log link function) in R version 3.1.3 (http://www.R-project.org). We analyzed daily long-term SST and anomaly data (1981–2015) using a Welch t test (which enables analysis of the unbalanced size of the 2 datasets).

A total of 89 Vibrio infections were reported in Sweden and Finland during the summer and autumn 2014, the largest yearly total number of cases, to our knowledge, identified in these countries. Infections were apparent across most Baltic coastal counties of Sweden and Finland. Numerous cases were reported at extreme subarctic regions, and as far north as >65°N, <100 miles (160 km) from the Arctic Circle. Reported infections began in July 2014 and peaked in August, before decreasing significantly in September (Figure 1). Infections were spread across persons of widely varying ages (range 3–93 years; median 36.2 years). In general, those infected were more commonly male (61 [67%] cases). One known fatality was noted: a V. cholerae non–O1/O139 infection reported in August 2014 from southern Sweden. Data on possible transmission was largely absent from the dataset from Finland; however, most cases in Sweden during 2014 occurred among persons who reported recreational exposure to seawater (e.g., the Baltic Sea) or lake water before infection (33 [78%] cases). Most (70 [77%]) infections were attributed to V. cholerae non–O1/O139; in 1 case, a ctx-negative O1 strain was reported. Other species reported were V. alginolyticus (3 cases), V. parahaemolyticus (4 cases), V. vulnificus (2 cases), V. mimicus (1 case), and unspecified Vibrio species (8 cases) (Table). Thirty-three (37%) infections were associated with ear or ear secretion isolations; however, for 17 (19%) of the 89 reported cases, Vibrio organisms were isolated directly from blood, suggesting more serious systemic disease progression.

The temporal and spatial distribution of reported cases corresponded closely with a highly anomalous heat wave in northern Finland and Sweden during July and August 2014, where SSTs in the northern Baltic exceeded all known long-term climatic and oceanographic records. A persistent and long-lasting period of high pressure occurred in northern Finland and Sweden beginning in May 2014, and this weather pattern persisted until mid-August. Concomitantly, SST in the Baltic Sea area was highly anomalous during July and August 2014; temperatures peaked toward the end of July. In some coastal regions, SSTs were ?10°C higher than the long-term average, indicating the extreme severity of this anomaly (Figure 2). Across the northern Baltic Sea area, SSTs were several degrees Celsius warmer than had been reported since the early 1980s. SSTs across large swathes of the Baltic and the Gulf of Bothnia area, in particular, had SSTs >18°C for several weeks beginning in mid-July and ceasing in mid-August (Figure 3). SST’s reported in the Gulf of Bothnia at the end of July were the most extreme reported during 1981–2016, exceeding 21.7°C on July 29, 2014, and with several days of temperatures >20°C. The observed SST anomaly during this period was also the largest ever seen in this dataset, encompassing almost 13,000 data points, with an anomaly of 9.79°C on July 29, 2014.

A statistical analysis between maximum SST and annual Vibrio cases using a generalized linear model showed that maximum SST explained a significant amount of the variability in cases (as determined by a significant reduction in the residual deviance from 120.55 to 42.16). The model predicted that, as the maximum SST increases, the number of annual number of cases also will increase significantly (? = 0.33002, SE = 0.08045,t = 4.102, p = 0.00343).

Domestically acquired Vibrio infections are rare in northern Europe, and the spike in recorded cases of vibriosis reported in this region is particularly noteworthy. The cases in 2014 are the largest yearly total of reportedVibrio infections in Sweden and Finland, more than double the number of reported cases than in other recent years (Figure 4). In Sweden, 2014 was the warmest year on record since recordkeeping began in 1860; in Finland, 2014 was the second-warmest year on record (10,11). Across Finland, 50 days of hot summer weather (temperatures >25°C) were recorded during May–August, which is 14 days more than the long-term average (10). The large number of reported infections corresponded closely with an intense and northerly SST anomaly, suggesting that these unusual oceanographic and climatic conditions drove this episode of waterborne disease. A subsequent quantitative and statistical analysis of SST data from this region revealed 3 further observations: 1) the peak SSTs in late July 2014 were the most intense observed in the Bay of Bothnia; 2) the anomaly is the most intense in almost 35 years of climate data (1981–2015); and 3) the likelihood of such an event occurring based on recent climate data (1981–2015) is highly unlikely—the 2014 maximum observed temperature was significantly higher than the maximum expected based on the data for other years, and based on the distribution of maximum temperatures observed, a temperature this much higher than the mean would be expected only in 0.78% of years (once every 128 years).

Vibrio species such as V. cholerae grow preferentially in low-salinity warm water, and recreational exposure to water, which appears to have been responsible for a sizeable proportion of these reported infections, also increases substantially during heat waves. That 2014 followed several other recent heat wave years (e.g., 1994, 1997, 2003, 2006, and 2010), during which recorded domestically acquired Vibrio cases spiked in northern Europe (4,5), is particularly noteworthy. Previous epidemiologic analysis regarding the emergence of Vibrioinfections in the region (5) indicated that sustained SSTs >18°C were a notable risk factor, significantly increasing reported cases. The relation between maximum SST and annual Vibrio case occurrence analyzed by using generalized linear model–based methods demonstrated similarly to previous studies in the region (4) that maximum temperature correlates highly with risk, and cooler years (e.g., 2005, 2007, and 2012) indicate lower levels of reported infections than heat wave years (e.g., 2006, 2010, and 2014). In our study, the observation that a sizeable proportion of described cases were reported in subarctic latitudes (>65°N) and within 100 miles (160 km) of the Arctic Circle is striking. Ten V. cholerae infections were reported above 63°N, of which 6 cases were identified in the Oulo area (?65°N). The cases recorded here are, to our knowledge, the most northerly reported instances of vibriosis documented, exceeding previous studies where cases have been reported at high latitudes, such as Alaska (9) and previously in northern Europe (5).

Disease data, such as those reported here, often are sporadic and usually grossly underreported. Likewise, a major limitation of our investigation was the lack of detailed trace-back epidemiologic data, which limits the assessment of exposure and subsequent risk. For many reported cases, data about prior exposure (e.g., specific information about the timing and location of recreational exposure to water) and subsequent routes of transmission were absent. However, almost without exception, cases from Finland and Sweden were reported in coastal rather than inland medical centers. Second, when prior transmission information was available from confirmed cases, most patients reported exposure to seawater in the days before symptom onset. These 2 factors, coupled with the striking climatic and oceanographic conditions during summer 2014, suggest that exposure to seawater was largely responsible for these episodes of disease emergence. The limitations underscore the need for a centralized system of surveillance and reporting. In the United States, the Centers for Disease Control and Prevention’s COVIS (Cholera and Other Vibrio Illness Surveillance) maintains a national database of vibriosis that contained detailed epidemiologic and transmission route information (12). A similar centralized reporting, monitoring, and surveillance system would greatly enhance risk assessment and risk management of vibriosis in Europe. Across the region, and with the exception of toxigenic V. cholerae infection, vibriosis is not a notifiable disease (5). Given that these rare waterborne infections appear to have emerged and increased in northern Europe recently (13) (e.g., 1994, 2006, 2014), this event underlies the need for clinicians to identify possible exposure to seawater. This event is particularly relevant for patients who have a history of conditions where progression of vibriosis to systemic infection is more likely, including diabetes, immune disorders, and liver dysfunction.

Climatic anomalies, such as the heat wave conditions during summer 2014 in northern Europe, appear to be responsible for restructuring the geographic distribution of waterborne infectious diseases and resulted in major and far reaching consequences for the identification, treatment, and management of these pathogens. The greater number and intensity of large heat wave events in northern Europe during the past 20 years or so (1994, 1997, 2003, 2006, 2010, 2014) further highlights the need for improved epidemiology and reporting, coupled with enhanced diagnostic capability in clinical settings to manage and ameliorate risk.