International travelers were at risk of acquiring influenza A(H1N1)pdm09 (H1N1pdm09) virus infection during travel and importing the virus to their home or other countries.
International travelers were at risk of acquiring influenza A(H1N1)pdm09 (H1N1pdm09) virus infection during travel and importing the virus to their home or other countries.
Characteristics of travelers reported to the GeoSentinel Surveillance Network who carried H1N1pdm09 influenza virus across international borders into a receiving country from April 1, 2009, through October 24, 2009, are described. The relationship between the detection of H1N1pdm09 in travelers and the level of H1N1pdm09 transmission in the exposure country as defined by pandemic intervals was examined using analysis of variance (anova).
Among the 203 (189 confirmed; 14 probable) H1N1pdm09 case-travelers identified, 56% were male; a majority, 60%, traveled for tourism; and 20% traveled for business. Paralleling age profiles in population-based studies only 13% of H1N1pdm09 case-travelers were older than 45 years. H1N1pdm09 case-travelers sought pre-travel medical advice less often (8%) than travelers with non-H1N1pdm09 unspecified respiratory illnesses (24%), and less often than travelers with nonrespiratory illnesses (43%; p < 0.0001). The number of days from first official H1N1pdm09 case reported by a country to WHO and the first GeoSentinel site report of a H1N1pdm09-exported case in a traveler originated from that country was inversely associated with each country's assigned pandemic interval, or local level of transmission intensity.
Detection of travel-related cases appeared to be a reliable indicator of sustained influenza transmission within the exposure country and may aid planning for targeted surveillance, interventions, and quarantine protocols.
International travelers were at risk of acquiring influenza A(H1N1)pdm09 (H1N1pdm09) virus infection during travel to affected areas and importing the virus to their home or other countries.[1, 2] For example, a positive correlation was found between the volume of airline travelers departing from Mexico and confirmed H1N1pdm09 imported cases identified in various countries during the early stage of the 2009 influenza pandemic.[3, 4] We investigated more broadly whether travelers can function as sentinels for sustained transmission in an affected country and could complement traditional surveillance systems and aid public health planning for targeted surveillance, interventions, and quarantine protocols at international borders.
We describe the profile of travelers who carried H1N1pdm09 virus across international borders throughout the world and explore the relationship between detection of H1N1pdm09 in travelers and the level of H1N1pdm09 transmission in the exposure country during the first phase of the H1N1pdm09 pandemic.
The 49 GeoSentinel sites in six continents contributing data are specialized travel and tropical medicine clinics that systematically provide clinical information on all ill returning travelers, as described elsewhere (www.geosentinel.org). Intake at the sites reflects a mixed population of patients requiring tertiary care and self-referred patients. Some sites are restricted to outpatient care, and at no site is practice limited to the care of ill travelers. The GeoSentinel data-collection protocol was reviewed by the institutional review board officer at the National Center for Emerging and Zoonotic Infectious Diseases at the Centers for Disease Control and Prevention and classified as public health surveillance and not as human-subjects research requiring submission to institutional review boards.
Cases were defined as travelers with confirmed or probable diagnoses of H1N1pdm09 reported to GeoSentinel from April 1, 2009, through October 24, 2009, when H1N1pdm09 virus transmission was well established worldwide. Confirmed H1N1pdm09 cases required positive results by real-time reverse transcriptase-polymerase chain reaction (rRT-PCR). At GeoSentinel sites, testing uses the best national reference laboratories in the site country. In the context of the 2009 pandemic, testing was done by public health authorities in all countries. Probable cases were defined as those with positive rapid tests for influenza A in acquisition countries where the predominant circulating strain was H1N1pdm09 or those with acute respiratory illness with an epidemiologic link to an rRT-PCR-confirmed H1N1pdm09 case.
Two separate groups of travelers who carried the infection across an international border are described. The first group, the majority of cases in this analysis (159; 78%), included travelers who were diagnosed with H1N1pdm09 after returning to their countries of residence and whose place of exposure was definitively attributed by GeoSentinel clinicians as their last exposure country. The second group included travelers diagnosed with H1N1pdm09 while in an exposure country and whose exposures were attributed to either their country of residence before travel or to a prior exposure country on the same trip. No differences between the groups were observed for any analysis, so pooled data is shown. Cases with uncertain country of exposure were excluded from some analyses.
To investigate the association between transmission intensity in a country and the time of H1N1pdm09 exportation from that country, we classified the 22 countries into three different pandemic intervals by using the classification scheme available from the US Department of Health and Human Services (Figure 1) as described in the following text.
Definitions of the observed three pandemic intervals are given as follows: Initiation Interval, this interval begins with the identification and laboratory confirmation of the first human case due to pandemic influenza virus in the [Country]; Acceleration Interval, this interval begins in a [Country] when public health officials have identified that containment efforts have not succeeded, onward transmission is occurring, or there are two or more laboratory-confirmed cases in the [Country] that are not epidemiologically linked to any previous case; and Peak Transmission Interval, this interval encompasses the time when there is extensive transmission in the community and the [Country] has reached its greatest number of newly identified cases.
We used available official country-specific surveillance data and web-based reports to define the pandemic interval (transmission intensity) for each country (see text below). For most countries, the pandemic interval was assessed at the time of exportation (defined as the clinic visit date of the first GeoSentinel case for that country). For countries whose clinic visit date of the first GeoSentinel case was after June 30, 2009, the transmission intensity on June 30, 2009, was used to assess the pandemic interval in each country. By June 30, 2009, the pandemic strain had been circulating for nearly 2 months and the WHO had officially reported a case in each of the 22 countries of interest, so that transmission intensity on that date is an indicator of overall country status in the face of the fully established worldwide pandemic. Even if the first exported case from a country was much later during a subsequent wave, that country would still be counted as an initiation phase country for the purpose of the statistical analysis performed.
Countries were classified into the following pandemic intervals: initiation (low-transmission intensity), acceleration (moderate-transmission intensity), and peak transmission (high-transmission intensity). During the study period, deceleration and resolution of the pandemic were not observed. In this analysis, eight countries were classified at the initiation interval (Brazil, China, Cuba, Hungary, India, Ireland, Norway, and Philippines); eight countries at the acceleration interval (Argentina, Chile, Greece, New Zealand, Panama, Spain, Thailand, and UK); and six countries at the peak-transmission interval (Australia, Canada, Dominican Republic,[25, 26] Indonesia, Mexico, and the United States).
Chi-square or Fisher's exact test was used as appropriate (SAS v9.2). Analysis of variance (anova) was used to assess the association between pandemic interval in the exposure country and the identification of sentinel travelers with H1N1pdm09. A p value of <0.05 was considered statistically significant.
An increase in the number of unspecified respiratory illnesses reported in GeoSentinel was observed during the early 2009 pandemic compared with data on respiratory illness reported from the same period in 2008 (Figure 2). Distribution of our laboratory-confirmed H1N1pdm09 cases coincided with the peak of respiratory illnesses documented from the week of April 26, 2009, through the end of June 2009.
Among the 203 (189 confirmed; 14 probable) H1N1pdm09 case-travelers identified, 56% were male; a majority, 60%, traveled for tourism; 20% traveled for business; and 86% were 10 to 44 years of age (Table 1). We compared H1N1pdm09 case-travelers with travelers in the GeoSentinel database with non-H1N1pdm09 unspecified respiratory illnesses or with nonrespiratory illnesses during the same period. Overall, the age profile of the three groups was significantly different (p < 0.0001; χ2). Paralleling age profiles in population-based studies only 13% of our H1N1pdm09 case-travelers were older than 45 years, while 32% of our travelers with non-H1N1pdm09 unspecified respiratory illnesses and 29% of our travelers with nonrespiratory illnesses were in the above 45 years cohort. A higher proportion of H1N1pdm09 case-travelers were hospitalized (75%) compared with those with non-H1N1pdm09 unspecified respiratory illnesses (40%) and those with non-respiratory illnesses (13%) (p < 0.0001; χ2). H1N1pdm09 case-travelers self-declared having sought pre-travel medical advice from a medical provider less often (8%) than travelers with non-H1N1pdm09 unspecified respiratory illnesses (24%), and less often than travelers with nonrespiratory illnesses (43%) (p < 0.0001; χ2).
|Variables||H1N1pdm09 (n = 203)||Non-H1N1pdm09 unspecified respiratory illnesses (n = 470)||Nonrespiratory illnesses (n = 3,470)|
|n||%||n||%||p Valuea||n||%||p Valuea|
|Age group, yearsb|
|Visiting friends and relatives||16||8||54||11||577||17|
Month-by-month clinic visit dates for 187 case-travelers were ascertained for 22 exposure countries (Table 2); 92% occurred from May to July 2009. The United States was the most frequently identified exposure countries (starting in May 2009), followed by Australia, the Philippines, UK, and Thailand.
|Country||Number per month||Number per country|
|Number per month||3||13||121||39||7||3||1||187|
The number of days between date of the first official H1N1pdm09 case reported by a country to WHO and the date of the first GeoSentinel site report of a H1N1pdm09-exported case in a traveler originated from that country varied greatly for each of the 22 exposure countries (Figure 3), the mean being 50 days (median: 46 days, interquartile range: 24–75 days). Two GeoSentinel sites reported cases with exposure in Mexico just 4 days after the first official report of H1N1pdm09 cases by Mexico to WHO on April 24, 2009, coinciding with the time at which peak transmission had already been reached in that country. There was an association between H1N1pdm09-exported cases and the level of H1N1pdm09 transmission in the case-traveler's country of exposure (p = 0.0001). We used the CDC pandemic intervals to represent influenza transmission intensity in the country of exposure (see Methods for definition of pandemic intervals): initiation (n = 8 countries), acceleration (n = 8 countries), and peak transmission (n = 6 countries). Countries with high H1N1pdm09 transmission (peak transmission pandemic level) had shorter interval (mean days) between official report to WHO of in-country first H1N1pdm09 case and export of cases identified in sentinel-travelers by GeoSentinel; mean days by pandemic interval were: initiation (84 days), acceleration (42 days), and peak (15 days).
Although travelers with respiratory illness may present in settings other than sites comprising GeoSentinel, the network was robust enough to distinguish travelers with confirmed H1N1pdm09 from those travelers seeking medical care because of other travel-related illnesses (Figure 2). Respiratory illnesses caused by influenza virus infection are difficult to distinguish from illnesses caused by other respiratory pathogens on the basis of signs and symptoms alone. That the majority of the cases of unspecified respiratory illness in travelers during the 2009 pandemic were due to other respiratory pathogens has been previously shown.
The increased number of reported respiratory illness in 2009 could reflect heightened awareness of the new influenza virus circulating as well as a real increase in disease frequency among travelers. The World Health Organization (WHO) declaration of a pandemic on June 11, 2009, followed documented spread of H1N1pdm09 virus in more than 70 countries. Thus, sentinel travelers detected by GeoSentinel clinics effectively mirrored the increasing global circulation of H1N1pdm09 virus during these early months of the first pandemic wave. From the beginning of July 2009, the guidelines in most countries were to seek medical care only for very severe illness. In addition, physicians were instructed not to send specimens for testing for nonhospitalized patients. This is reflected in Figure 2 as well.
As in population-based studies, case-travelers were mostly young (Table 1) and not traveling to “exotic” destinations. Travelers with H1N1pdm09 sought pre-travel medical advice significantly less often (8%) than those with non-H1N1pdm09 unspecified respiratory illnesses (24%), and those with nonrespiratory illnesses (43%). The rate of hospitalization in H1N1pdm09 reported in this study was much higher than those reported elsewhere[33, 34] for H1N1pdm09 cases and may not represent severity of illness in this population. This has more likely resulted from some countries' (eg, Singapore, Italy, France) policies to hospitalize all H1N1pdm09 cases identified during the initial pandemic phase, regardless of severity.
The mean days from first official H1N1pdm09 case reported by a country to WHO and the first GeoSentinel site report of a H1N1pdm09-exported case in a traveler originated from that country was inversely associated with each country's assigned pandemic interval, or local level of transmission intensity. This might indicate that a certain threshold of influenza transmission needs to be present locally before there is sufficient probability that a traveler can export the virus across international borders. In this context, the detection of travel-related pandemic influenza cases by a sentinel system such as GeoSentinel could be a reliable indicator of the onset of sustained transmission within the exposure country as infected travelers captured in the system function as sentinels for sustained influenza transmission.
The first cases of H1N1pdm09 in GeoSentinel acquired infection in Mexico in April 2009, but overall few cases from Mexico were identified. This could reflect lack of widely available diagnostics in most countries during the major wave of exportation from Mexico in the early days of the pandemic.
This report contains a number of important observations on an opportunistic, multinational, and sentinel sample of travelers using data gathered at existing surveillance sites that happened to be in a position to capture these travelers in the face of a sudden pandemic. This validation of ongoing international efforts by consortia like GeoSentinel in setting up surveillance for travelers in key countries all over the world is the strength of this article. The design however would have been different if data capture could have been planned in advance, but this was an unexpected pandemic with an unexpected origin and it is not possible now to go back and ascertain new data that was not part of our standard data collection form. It is also not possible to obtain reports from network sites with normal referral patterns that would exclude travelers with acute respiratory illness in the face of an influenza pandemic. This is not a comprehensive worldwide study of every border in each country. And therefore, the results are not reflective of broad national data. The observations are on the travelers enrolled and sampled. Thus, some biases in spectrum of severity or epidemiologic exposure cannot be ruled out.
Differences between surveillance systems in different countries could lead to misclassification bias in determining the pandemic interval if there were detection delays. The wide and rapid dissemination of test reagents by WHO during this pandemic and many years of pandemic planning in most countries make this unlikely to be a significant issue, and no reports to this effect have appeared in the literature. Distinction between “initiation,” “acceleration,” and “peak” pandemic intervals was made by application of enduring definitions (since 2008) to best available information emanating from each country. Differentiation of acceleration from peak intervals would be most affected by limitations in interpretation of available information.
In summary, we found that ill travelers with known countries of exposure can mirror significant transmission intensity within the source country and serve as a separate and important indicator from initial case detection and reporting within that country. Other sensitive mechanisms for initial case detection otherwise exist in most countries. That travelers are important vectors of novel respiratory pathogens may be thought intuitive, however, our specific and detailed descriptive findings have not been documented elsewhere for H1N1pdm09 or other emerging respiratory pathogens. For future novel respiratory events in which an age profile or predominance emerges early, travelers can function as sentinels for sustained transmission and could complement traditional surveillance systems and aid public health planning for targeted surveillance, interventions, and quarantine protocols at international borders. Additionally, these sentinel systems might fill the gaps in epidemiologically “silent” surveillance zones.
This work was supported by the GeoSentinel Surveillance Network through a cooperative agreement with the Centers for Disease Control and Prevention (CDC; grant 5U50CI000359), by a tender from the European Centre for Disease Prevention and Control (ECDC; tender OJ/2008/07/08-PROC/2008/019), and by funding from the International Society of Travel Medicine (ISTM).
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Payments from the CDC funding grant were made to authors or their institutions (D. A. P., P. L. L., E. C., F. C., P. E. K., and D. O. F). Consulting fees were paid by Baxter (to E. C.) and Crucell (to P. E. K.). Payment for development of educational presentations was paid by Sanofi (to P. E. K.). All other authors report no potential conflicts.
Additional members of the GeoSentinel Surveillance Network who contributed data (in descending order) are: Alice Pérignon, Hôpital Pitié-Salpêtrière, Paris, France; Giampiero Carosi, University of Brescia, Brescia, Italy; Philippe Parola and Fabrice Simon, Hôpital Nord and Hôpital Lavaran, Marseille, France; Gerd-Dieter Burchard, Bernhard-Nocht-Institute for Tropical Medicine, Hamburg, Germany; Natsuo Tachikawa and Hanako Kurai, Yokohama Municipal Citizen's Hospital, Yokohama, Japan; Frank von Sonnenburg, University of Munich, Munich, Germany; Patrick W. Doyle and Wayne G. Ghesquiere, Vancouver General Hospital and Vancouver Island Health Authority, Vancouver and Victoria, British Colombia, Canada; John D. Cahill and George McKinley, St. Luke's-Roosevelt Hospital Center, New York, New York, USA; Mogens Jensenius, Oslo University Hospital, Oslo, Norway; Andy Wang and Jane Eason, Beijing United Family Hospital and Clinics, Beijing, People's Republic of China; Watcharapong Piyaphanee and Udomsak Silachamroon, Mahidol University, Bangkok, Thailand; Marc Mendelson and Peter Vincent, University of Cape Town and Tokai Medicross Travel Clinic, Cape Town, South Africa; and Rogelio López-Vélez and Jose Antonio Perez Molina, Hospital Ramon y Cajal, Madrid, Spain.