INTRODUCTION
Influenza A viruses (IAVs) are among the most challenging viruses that threaten both human and animal health (1). From a One Health perspective, their capability to transmit from one species to another, causing multiple viral genome reassortments to occur is of major concern. The best studied pandemic IVs, in 1918, 1957, 1968, and 2009, ultimately acquired some or all of their gene segments from the avian IAV gene pool with swine origin genes in a second order (2,3). The last two decades have witnessed a growing list of avian influenza viruses (AIVs) that could infect humans with severe consequences (4), while the role of swine in interspecies transmission of IAVs continues to be of concern (5,6).
There are long-standing evidences that relate presumed equine influenza viruses (EIVs) to human epidemics or pandemics, but lacks confirmation by modern scientific methods (2). Current indications of zoonosis with EIVs mostly come from regions where humans still have close contact with horses (7). Other animals which are in close contact with humans such as dogs were not considered a natural host of IAVs until an H3N8 virus infected dogs through transmission from equine species. This occurrence furthered originated a unique dog stable lineage (8). Dogs can be also be infected by AIVs, which raises concern for public health (9,10).
The unpredictability of an influenza pandemic highlights the need for improved surveillance on animal IVs and for research to understand and manage the risk of spillover between animals and humans. While the world prepared to face a conceivable H5N1 avian-origin influenza Eurasian lineage pandemic, a H1N1 pandemic virus emerged from swine in North America. Information on the geographic variation of IV emergence risk and/or spillover, framed within concurrent surveillance efforts, is essential for the design of future influenza surveillance and pandemic mitigation strategies (11).
There is a unique article summarizing AIVs detections in domestic poultry in the Caribbean and other locations, which covered 2006 to 2008 and mainly focused on case description (12). An updated review of animal IV surveillance programs is needed to assess the capacity for detecting IVs with zoonotic potential. This would assist to determine the global need for increasing surveillance, which targets these viruses in relevant animal species (13,14). However, surveyed countries in this article revealed that in the Caribbean, only four out of 28 nations (14.29 %) provided data. Such disparity with regards to the overall coverage of the study (40,91 %), could be due to the fact that these countries were targeted by the Food and Agriculture Organization of the United Nations (FAO), which excluded small island states and overseas territories. Therefore, information on both research and surveillance of IAVs in the Caribbean was determined by regional capacity to perform the testing.
Several facts underline the importance of such data for the Caribbean countries that may contribute to the global body of knowledge. For this subregion several factors contribute to the risk for avian influenza with high impact (15,16). Moreover, swine, as other species of main concern for IAVs surveillance, account for important populations among some Caribbean countries (17) in which they constitute a main concern of local food security. Therefore, noteworthy levels of contact intensity among animal species and humans for IAVs infections could be present in this subregion. Additionally, either the frequency or intensity in which the Caribbean region is impacted by hurricanes might recurrently enhance the vulnerability of animal and human populations for higher risk of exposure to pathogens, including IAVs. Additionally, island territories face further challenges owing to their isolated geographies (18).
Systematic reviews on health-related issues are essential to determine trends or data collection needs, which provide evidence for policy makers to develop interventions toward risk reduction and management, which will build defensive capacities for the reduction in biological threats. In particular, within the Caribbean there are some regional organizations, including an Animal Health Network (CaribVET) (https://www.caribvet.net/) prone to addressing capacity building for resilience, among other topics that could improve outputs for a systematic review on animal influenza. A systematic review on animal IV research and surveillance was conducted across the Caribbean region based on the assumption that publication trends capture varying and evolving regional capacity on these topics. The aim of this work was to characterize temporal and geographic data and international research collaboration. We addressed the capability for research and surveillance on animal influenza in the Caribbean region.
METHOD
A prior protocol was accomplished, based on the PRISMA statement (19) to specify the search strategy, eligibility criteria, objectives, and the methods of this systematic review. The available scientific publications generated in Caribbean countries were searched, retrieved and analyzed for their content, quality and relevance. Additionally, animal disease web-based applications like WAHIS (http://www.oie.int/wahis_2/public/wahid.php/Wahidhome/Home) from the Organization for Animal Health (OIE), the FAO’s Global Animal Disease Information System (http://empres-i.fao.org/eipws3g/) and GenBank (https://www.ncbi.nlm.nih.gov/genbank/) were used for screening records or reports related to animal influenza in the Caribbean and to supplement the scientific publication data.
Eligibility criteria
The retrospective study was conducted to review the articles on animal influenza in the Caribbean published from January 1950 to March 2019. This search focused on articles written in English, French, or Spanish. One reviewer screened the retrieved records to exclude those not matching through titles and if necessary, the abstract. In case of doubts, the consensus was reached by discussion between two review authors and if necessary, examining the full text publication of concern. Review articles were excluded, but after a carefully reading to assess if they contained information, which allowed tracing back some relevant articles potentially not identified through the search strategy.
The abstracts or articles were obtained from PubMed Central (https://www.ncbi.nlm.nih.gov/pmc/) and Sciencedirect (https://www.sciencedirect.com/) databases, and Academic Google (https://scholar.google.com.cu/) was also used to judge the first 200 retrieved articles ordered by relevance. The descriptions used in the record retrieval process included the following terms: influenza, Caribbean and animals, combined under the following search strategies: 'influenza' AND 'animals' AND 'Caribbean' and 'influenza' AND 'Caribbean' AND ('swine' OR 'poultry' OR 'bird' OR 'equine'). A search was also carried out using the strategy: 'influenza' AND ('animals' OR 'swine' OR 'poultry' OR 'bird' OR 'equine').
The search strategy replaced the term Caribbean by the single countries/territories denomination such as: Anguilla, Antigua & Barbuda, The Bahamas, Barbados, Bermuda, Bonaire, British Virgin Islands, Cayman Islands, Cuba, Dominica, Dominican Republic, Grenada, Guadeloupe, Haiti, Jamaica, Martinique, Montserrat, Netherland Antilles (includes Aruba, Sint Maarten, Curacao), Puerto Rico, Saint Barthelemy, Saint Eustatius, Saint Kitts & Nevis, Saint Lucia, Saint Martin, Saint Vincent & the Grenadines, Trinidad & Tobago, Turks & Caicos Islands, and the US Virgin Islands. Denomination of group of islands or the vernacular name within a group were also used, namely Hispaniola, Jost Van Dyke, Leeward Islands, Saint Croix, Saint John, Tortola, Virgin Gorda, and Water Island. The remaining CARICOM member states (Belize, Guyana, Suriname, and French Guiana) were also included in the search. Thirty-two mainland or island countries/territories were considered as the Caribbean.
Inclusion and exclusion criteria
Inclusion criteria: Abstracts were included if they described:
data relating to animal influenza in animals or humans in any of the 32 detailed countries in the Caribbean,
data from surveys of animal influenza obtained using serological, virus isolation and nucleic acid detection,
proposal or development of any procedures applicable to the surveillance and control of animal influenza,
development or evaluation of surveillance systems, and
animal influenza vaccine development or evaluation.
Data extraction and synthesis
The database was built in Zotero 5.0 (https://www.zotero.org/). Once the Digital Library was constructed and duplicate records removed, the titles and abstracts were analyzed and revised against the inclusion criteria to determine their relevance and suitability. The following data were extracted from the publications selected for inclusion of the author(s), year of publication, geographic location in which the study was conducted, and the design of the study. Other extracted data included were the subtypes, the animal species and the category tested.
The articles were classified according to their content in topics such as: 1) Diagnosis (virological, molecular or serological); 2) Proposal or development of any procedures applicable to the surveillance and control; 3) Development or evaluation of surveillance systems; 4) Surveillance studies on wild bird or poultry (commercial or backyard), and 5) Vaccine development or evaluation. For analysis of diagnosis, the serological approaches were considered separately, whereas viral isolation and identification, as well the nucleic acid detection were grouped under virological diagnosis. For surveillance studies involving more than one country, descriptive statistics consider the first author´s country/territory. A spreadsheet in Excel® (Microsoft, USA) was structured and the information entered according to the above-mentioned categories to process descriptive indicators of frequency distributions.
Social Network Analysis of collaboration
From the bibliometric database, information was extracted, including institutional names, number of authors, institutions and countries to carry out social network analysis (SNA) of collaboration. For each included paper, one author examined the data based on the full texts with information checked by a second author. A process of standardization was carried out to combine various institution or countries/territories. In the case of departments or branches attached to universities or Ministries of Agriculture, the affiliation was considered as an institution. A binary matrix of relationships between countries and institutions was constructed, considering only contacts (link or edges) involving at least one Caribbean country/territory. Data was imported to Gephi 0.9 program to construct, metric and visualize the network for collaboration to generate pertaining articles.
RESULTS
Search results
During the selection process, PMC yielded 8979 records, Sciendirect1094records, Google Scholar 200 results, and WAHIS 5 reports. A total of 10278 articles, of which 2598 were duplicates, left a total of 7680 reports. The study selection process was summarized in a flow diagram (Fig. 1). After the initial review of all studies by the first and last author and subsequent discussion, and the consensus of the remaining author, 6115 articles/records were discarded after screening the title and if necessary the abstract, while 1383articles/records were discarded as they did not meet the eligibility criteria. One hundred eighty-two publication articles from the electronic databases mentioned were eventually deemed potentially relevant studies for inclusion in this review and 32 were found to be pertinent and included after a full-text screening.
Characterization of included studies or reports
During the retrieval of pertaining studies (2007- March 2019), 32 articles or reports generated in the Caribbean were identified (≈2.66/year). There was a ratio of 5.4 articles per disease report. The temporal distribution of studies (Fig. 2) showed a clear overall trend, which increased over time although it varied among topics. The largest amount of studies per year was six in 2018, the last entirely concluded year.
The 32 studies (articles/reports) occurred in six sections (Table 1). This distribution showed that the serological and virological diagnosis paired with each other (7, 21.87 %) and represented the largest number (14, 43.75 %) of studies within a particular topic followed by vaccine development (5, 15.62 %). Thereafter, the amount of studies on development of nucleic acid detection approaches (4, 12.5 %) equated to the OIE disease reports and were followed by the development or evaluation of surveillance systems (3, 9.37 %). Cuba was the only country that published studies on AIV vaccine development. Among these included the diagnostic tests used for vaccine evaluation.
Thematic | No. of articles | References |
---|---|---|
Development of molecular approaches | 4 | 13,20,21,22 |
Diseaseprioritization | 1 | 23 |
Disease reports | 5 | 24,25,26,27,28 |
Development of vaccines | 5 | 29,30,31,32,33 |
Serological diagnosis | 7 | 34,35,36,37,38,39,40 |
Surveillance systems | 3 | 41,42,43 |
Virological diagnosis | 7 | 15,16,44,45,46,47,48 |
The spatial distribution of data for swine and horses was narrow, with only 3 articles concerning swine in one country, whereas the other in horses was carried out in 3 Leeward Islands (St. Kit, Nevis, and Sint Eustatius). In contrast, the OIE animal influenza reports listed by species other than avian were absent, and no reports were retrieved through alternative disease information sources (http://empres-i.fao.org/eipws3g/).
The diagnosis, the most represented data, when scattered per species (Fig. 3) showed a vast predominance (almost two thirds) of studies targeted avian (wild bird or poultry) compared to swine and horses. However, within the studies on avian (15,16,24,25,26,27,28,35,36,45,49), the majority addressed poultry (commercial or backyard) from which 8 out of 10 were on commercially reared poultry.
The studies listed by the first author's affiliation were distributed in nine countries/territories, highlighting the fact that two of them were found in OIE disease reports (25,26,27,28). The frequency of studies per country revealed that the greatest number corresponded to Cuba (15, 46.87 %), which coincided with a broadest thematic diversity. Cuba produced almost a fourfold number compared to Trinidad and Tobago, which was the second highest country (Fig. 4). Cuba was the most productive, even subtracting studies on vaccine development.
Excluding OIE diseases reports, only 7 out of 32 countries/territories provided articles. Among such countries/territories, Cuba, Trinidad and Tobago, Grenada, and Guadeloupe provided the largest number of articles, which accounted for 22 out of 27 (81.48 %).
Social Network Analysis of collaboration
The network of collaboration to generate articles on animal influenza in the Caribbean (Fig. 5) had a size of 23 nodes (countries/territories) and 129 contacts or edges. Slightly more than 30 % of countries (Cuba, Dominican Republic, Guadeloupe, Haiti, Barbados, France, and Belize) represented 58.89 % of data. Eighty-three of the overall contacts (64.34 %) occurred among Caribbean countries, whereas the remaining 46 edges connected Caribbean countries with European or American nations at a ratio of 32 and 14 respectively. Region interactions were largely represented by France with 55.81 % of relationships, which was six-fold higher than the US as the following most collaborative. Noteworthy, interactions with France were not restricted and were not part of French overseas territories.
The network of collaboration to generate articles on animal influenza at institutional level (Fig. 6) had a size of 51 nodes (institutions) and 350 interactions (edges) distributed mainly into three (3) main communities of nodes with dense connections within each group and sparser connections between different groups. Most of contacts occurred between Caribbean institutions (61.45 %) followed by other European (24.02 %) and American organizations (14.53 %).
Eight out of 51 (15.69 %) institutions represented over 49 % of the contacts of collaboration, with individual values ranging from 8 % to 4 %. In decreasing order of the degree of contact, these institutions are listed as their acronyms 1, CENSA, MA Barbados, BAHA, DIGEGA DR, MA Haiti, CIRAD Guadeloupe, MA Belize, and MA Guadeloupe.
There was a relative high rate of collaboration in the Caribbean (Fig. 7), which averaged 3.25 participating institution per study. Noteworthy, the 62.50 % of the studies generated more than one institution based on different countries/territories.
DISCUSSION
The first animal influenza study of wild bird surveillance in Barbados yielded 2 H4N3 AIVs with an isolation rate of 5.0 % (2/40) in hunter-killed Blue-winged teals (45). This study entered the public domain in 2007, but it was carried out between 2003 and 2004, in coincidence with the period of increased awareness of AI. Globally research and surveillance for AIVs increased with the spread of the H5N1 Goose/Guangdong lineage in late 2003s and early 2004s.
However, AI was reported for the first time in the Caribbean in December 2007, when the Dominican Republic determined the presence of an H5N2 LPAIV in poultry (25), as specified by the OIE Terrestrial Animal Health Code (50). Six months later, a similar situation was also confirmed in Haiti (28). In both reports, diagnosis was conducted outside of the affected countries by an OIE Reference Laboratory (National Veterinary Services Laboratories USDA, APHIS, Ames, Iowa, United States).
A subsequent AI event occurred within the Caribbean in 2015, when Belize reported an H5N2 LPAI (24). In this case, confirmation was conducted by an OIE reference laboratory, but affected flocks were first detected under routine serological monitoring implemented as part of a national Avian Influenza Program. This emphasized the importance of having implemented surveillance programs for early alert and appropriate response.
An H5N2 LPAI was reported by the Dominican Republic again in 2017 (26) as well more recently with an unresolved event resulting in seven (7) outbreaks (27). The continuation of Dominican AI reports could be due to the fact that they are reporting on the national infrastructure for diagnosis. Since 2017, the Dominican National laboratory conducted either serology (agar-gel immunodiffusion) or real-time reverse transcriptase/polymerase chain reaction (RRT-PCR) in disease outbreak investigations although they continued to submit samples to an OIE reference laboratory.
Interestingly, all AI reports have been caused by the same AIV subtype and pathotype (H5N2 LPAI) (12), since the H5N2 viruses in the Dominican Republic and Haiti, which are closely related to the Mexican lineage of H5N2, which has been circulating in Mexico since early 1994. However, despite of the recurrence of H5N2 LPAI in the Caribbean, detailed phylogenetic analysis of these viruses has not yet been published and virus genetic data are limited to first outbreak in the Dominican Republic (https://www.ncbi.nlm.nih.gov/genbank/). Timely sharing of virological and epidemiological information between the animal and the human health sectors and with other key partners is crucial in developing a better understanding of influenza viruses and their risks, as well as for providing an early warning of emerging threats (51). Six sequences from Dominican outbreaks were submitted to the network of expertise on animal influenza in 2018(52).
The cleavage site within the hemagglutinin precursor protein is the first target of sequencing for both molecular pathotyping and subtyping of AIV(53), (54). However, a larger genome sequencing may also assist to optimize diagnosis through the assessment of mismatches in the primer/probe recognition region as it has occurred (21,55). In addition, it is also important to determine either the genetic evidence of antiviral resistance or molecular markers of interspecies transmission (3,56,57,58), which might have public health implications.
The small number of Caribbean studies on animal influenza in other than avian species could be related to the fact that detections of swine influenza viruses (SIVs) are not officially notifiable to the OIE. Hence, information on its occurrence could go unnoticed when no local arrangements for surveillance are made. While SIV is not officially reportable, it can significantly impact swine production and can have public health implications. It is well acknowledged that swine are often present worldwide at the human-animal interface and maintain a reservoir for antigenically divergent hemagglutinin and neuraminidase proteins that pose a risk for pandemic emergence(52).
Availability of funds could be another reason for few studies on SIV. Noteworthy, all studies reported in the Caribbean (47,46,48) occurred later than the emergence of the swine-origin H1N1 pandemic virus, which could both have attracted attention and mobilize funds to survey SIV. In some studies (46,47), there was financial support provided by a FAO technical cooperation project (TCP/RLA/3206).
Equine influenza is not a listed OIE disease, but it is ostensibly regulated in countries where the breeding and racing of horses are a major industry (10). Equine populations in the Caribbean are not high, as only three (3) countries, namely Cuba, Haiti and the Dominican Republic, have populations between half a million and one million (17).
As it was shown herein, animal influenza studies had an upward trend from 2007 to March 2019. However, publications were concentrated in only a few countries/territories, which constrained their spatial distribution. A decreasing effect on the spatial distribution metric could be present, since studies were consigned according to the first author's country/territory. It is worth noting that 4 studies had regional coverage (15,16,23,43) with another study, on EIV, included 3 Leeward Island (34). Regardless, their importance could be negligible, because they represented a minimal proportion of the reported data (7.14 %).
The high number of articles from Cuba (almost fourfold over the second highest country) could be associated to the thematic diversity and the amount of local institutions working on the subject. At country level, thematic diversification is considered a driver of the productivity or efficiency of the research system (59). Additionally, either the overall or particular differences by topic between and among countries/territories could also be influenced by the success of the scientific technical innovation policy regarding certain topics, expressed in institutions, associated scientific community, presence of the subject in scholarship, and articulation of the science and decision makers (Nuñez, 2019 pers com).
Considering the huge dissimilarities among the Caribbean countries/territories in terms of surface, economies and animal populations (FAOSTAT), it is hard to ascertain the accurate scattering of research and surveillance. However, according to the shape of the retrieved studies, a broader spatial distribution of surveillance efforts on animal influenza is advisable. The scarce surveillance for AIVs in Latin America and the Caribbean is related to the lack of adequate resources and local infrastructure, as well as more urgent animal health priorities (60). Noteworthy, a coordinated surveillance system with a global or a regional perspective does not necessarily require participation from every country (61).
In case of a suspicious disease, reference laboratories support diagnosis, but sustainable surveillance requires local capacities. Alternatively, there are cheap and easy-to-implement serological approaches, such as agar gel precipitation (AGP), that are effective for surveillance, as demonstrated by the detection in Belize (24). Since AIVs exist commonly in wild free flying ducks as LPAI viruses (5,6), subclinical infections in poultry are more difficult to detect. Consequently, LPAI surveillance often relies on diagnostic tests, particularly serology, and molecular and virological tests are used to follow up positive cases (51).
A better understanding of AIV research and surveillance will aid in risk assessment. Progress has been made in surveillance reporting in this region (42); however, ongoing work for avian influenza needs to be strengthened (https://www.caribvet.net/).
The interaction between animals and humans is a major source of emerging infectious diseases, some of which have the potential to cause global pandemics in the future (62). From the One Health perspective, one of the major concerns for IAVs are intermediate hosts that have dual specificity sialic acid receptors on their cells (e.g. swine and domestic poultry (quails and turkeys)), which may be involved in the interspecies transfer of influenza viruses (3). Under circumstances of interaction among such species with humans, there are also opportunities for reassortment of the genes from these viruses, which may result in a mammalian-adapted influenza virus. In contrast, influenza surveillance in production animals is largely focused on facilitating the international trade (14).
Some global spatial frameworks have been developed to quantify the geographic variation in outbreak emergence based on potential species transmission to humans (11,63), but they have been scarcely implemented at regional or national level with lower spatial resolution.
Network analysis from a social science perspective was developed to characterize a group of authors and the nature and extent of the connections or interactions between and among them. In the current work, the SNA program allowed to identify the most productive countries/territories and institutions, as well as the identification of groups of them with more intense collaboration, which have published studies on animal influenza through the period 2007-March 2019. It was confirmed the ‘same country or institution phenomenon’, which might be because scientists and researchers in the same country or institution have the quickest and easiest ways to communicate, and further form collaborative relationship (64). However, it was remarkable that 62.50 % of the studies were generated from institutions based on more than one country/territory. Research collaborations contribute to the advancement of knowledge by exploiting the results of scientific efforts more cost-effectively and also enhance the sharing of resources among nations and disciplines (65).
CONCLUSIONS
Reported studies of animal influenza in the Caribbean have tended to increase since 2007, but they have been concentrated in a few countries/territories. Such constrained spatial distribution may constitute gaps in the early alert capacity of the existence of these viruses, which in turn may threaten both human and animal health for optimizing pandemic mitigation strategies. A better balance of surveillance efforts in the short term may increase country/territory capacities to implement feasible and low-cost approaches to constrain or eliminate the threat. The spatial risk assessment of animal influenza and further design of risk-based sampling can optimize resources for animal influenza surveillance in this region. This systematic review suggested the need to further expand the scientific studies on animal influenza generated in the Caribbean and the opportunities to improve risk assessment, management and communication through international collaboration.