On November 9, 2014, chickens in 1 of 6 poultry houses on a 124,000-bird indoor-layer farm in the Netherlands began dying at an exponentially increasing rate. The dead chickens were submitted for necropsy to the Dutch Animal Health Service (http://www.gdanimalhealth.com) on November 14. RNA was extracted from cloacal and oropharyngeal samples from clinically affected hens with positive results from the screening influenza real-time reverse transcription PCR (7), which detects all avian influenza virus subtypes. As a standard procedure, the swab samples were forwarded to the Central Veterinary Institute, the Netherlands’ national reference laboratory. Positive screening samples were checked for the presence of H5 and H7 influenza subtypes by real-time reverse transcription quantitative PCR as recommended by the European Union reference laboratory. Hemagglutinin (HA) and neuraminidase (NA) sequence analysis was performed by using PCR fragments that were generated according to previously described protocols (8,9). The HA cleavage site showed polybasic properties RNSPLRERRRKR*GLFGAIA, confirming the high pathogenicity of the virus. In addition, HA and NA sequence results showed that the virus subtype was H5N8.

At the start of the outbreak, preliminary sequencing of the cleavage site showed that it shared high similarity with that of the outbreak strain from Germany. However, complete sequencing was necessary for an investigation of the origin and emergence of this virus in Europe, specifically in the Netherlands. Therefore, we amplified all 8 RNA genome segments of the outbreak virus by using universal 8-segment primers and then directly sequenced the segments (10). Purified amplicons were sequenced at high coverage (average >1,000) by using the Nextera library preparation method and MiSeq system (Illumina, San Diego, CA, USA), generating paired-end read lengths of 150 bases. High-quality quality control–passed sequence reads were iteratively mapped by using Bowtie 2 (11) against the genome sequence of the H5N8 virus from South Korea (GenBank accession nos. KJ511809–KJ511816) to generate a majority (>80% evidence) consensus sequence of all segments. The consensus sequences were compared with de novo–assembled sequence reads by using SPAdes version 3 (12); substantial differences were not detected. Most consensus sequences were submitted to the Global Initiative on Sharing Avian Influenza Data (accession no. EPI_ISL_167905). We subsequently performed a molecular phylogenetic analysis on all nucleic acid sequences by using the maximum-likelihood method based on the Tamura-Nei model in MEGA6.0 (13). Accession numbers and sequence providers are listed in the Technical Appendix.

Genetic analysis shows that the H5N8 virus from the Netherlands (A/chicken/Netherlands/14015526/2014) and viruses from poultry and wild birds from Europe and 2 strains from Japan (A/duck/Chiba/26-372-48/2014 and A/duck/Chiba/26-372-61/2014) detected thereafter are closely related. Analysis also showed that the viruses are descendants of 3 strains isolated in early 2014: A/broiler duck/Buan2/2014 and A/Baikal teal/Korea/Donglim3/2014 from South Korea and A/chicken/Kumamoto/1-7/2014 from Japan. HPAI (H5N8) virus isolated from a wigeon (Anas penelope) in Sakha, northeastern Russia, is a precursor phylogenetically located at the node of European and Chiba viruses (Figure); this virus was isolated in September 2014, but the sequence was released in December 2014 (15).

We determined the number of per site base substitutions between the sequences by using the maximum composite likelihood in MEGA6 (13). Sequences of the 8 genome segments in the H5N8 virus from the Netherlands differed from those of strains A/broiler duck/Korea/Buan2/2014, A/Baikal teal/Korea/Donglim3/2014, and A/chicken/Kumamoto/1-7/2014 by a minimum of 0 and a maximum of 0.009 substitutions. On the basis of data in the National Center for Biotechnology Information Influenza Virus Resource (http://www.ncbi.nlm.nih.gov/genomes/FLU/FLU.html) and EpiFlu (http://www.gisaid.org) databases, the H5N8 virus from the Netherlands shares the highest similarity with strain A/Baikal teal/Korea/Donglim3/2014 from South Korea.

Genetic analysis of influenza A(H5N8) virus from the Netherlands indicates that the virus probably was spread by migratory wild birds from Asia, possibly through overlapping flyways and common breeding sites in Siberia. In addition to the outbreak in the Netherlands, several other outbreaks of HPAI (H5N8) virus infections were reported in Europe at the end of 2014 after exponentially increasing deaths occurred in chicken and turkey flocks. Genetic sequences submitted to the EpiFlu database indicated that the viruses from Europe showed a strong similarity to viruses isolated earlier in 2014 in South Korea, China, and Japan. An H5N8 virus isolated from a wigeon in Russia in September 2014 is located in the phylogenetic tree near the node of all sequences for H5N8 viruses from Europe. In regard to time, this location fits the hypothesized route of H5N8 virus introduction into Europe. Furthermore, for several reasons, it is highly likely that the introduction of HPAI (H5N8) virus into the indoor-layer farm in the Netherlands occurred via indirect contact. First, despite intensive monitoring, H5N8 viruses have never been detected in commercial poultry or wild birds in the Netherlands. Second, when the virus was detected, the Netherlands had no direct trade contact with other European countries or Asia that might explain a route of introduction. Third, because of the severity of disease in galliforms, outbreaks of H5N8 in the Netherlands before November 2014 would have been noticed.

Sequences of the H5N8 virus isolates from the Netherlands were compared with sequences obtained from the Global Initiative on Sharing Avian Influenza Data’s EpiFlu Database.