SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Background Data on relative rates of acquisition of gastrointestinal infections by travelers are incomplete. The objective of this study was to analyze infections associated with oral ingestion of pathogens in international travelers in relation to place of exposure.

Methods We performed a multicenter, retrospective observational analysis of 6,086 travelers ill enough with any gastrointestinal infection to seek medical care at a GeoSentinel clinic after completion of travel during 2000 to 2005. We determined regional and country-specific reporting rate ratios (RRRs) in comparison to risk in northern and western Europe.

Results Travel to sub-Saharan Africa (RRR = 282), South America (RRR = 203), and South Asia (RRR = 890) was associated with the greatest rate of gastrointestinal infections. RRRs were moderate (25–142) for travel to Oceania, the Middle East, North Africa, Central America, the Caribbean, and Southeast Asia. RRRs were least (<28) following travel to southern, central, and eastern Europe; North America; Northeast Asia; and Australasia. Income level of the country visited was inversely proportional to the RRR for gastrointestinal infection. For bacterial and parasitic infections examined separately, the regions group in the same way. RRRs could be estimated for 28 individual countries and together with regional data were used to derive a global RRR map for travel-related gastrointestinal infection.

Conclusions This analysis of morbidity associated with oral ingestion of pathogens abroad determines which parts of the world currently are high-risk destinations.

Gastrointestinal infections are frequent in travelers, as evidenced by the fact that over 60% of individuals visiting tropical and subtropical regions will develop diarrhea.1–7 However, the morbidity associated with gastrointestinal infections in travelers is broader and, together with cases of diarrhea that are more serious or more prolonged, has a spectrum that has been incompletely described.

Studies in a relatively small number of individual destinations or on small numbers of travelers visiting different destinations3,8,9 have described epidemiological and risk factors associated with travelers’ diarrhea.6,7,10–12 The only large-scale global study dates back a quarter of a century.5 These have demonstrated some destination-specific differences in risks of developing diarrhea according to country visited, with South Asia typically among the “most risky” destinations.13–15 Comprehensive reviews11 containing maps of worldwide risk zones for travelers’ diarrhea, of necessity, cite heterogeneous studies from as early as the 1970s. Studies reporting the broader spectrum of infections associated with oral ingestion of pathogens in travelers according to destinations of travel have not been previously performed.

We have used the GeoSentinel database16,17 to provide a standardized, aggregate multinational description of the full spectrum of travelers’ gastrointestinal infections acquired by a broad range of travelers returning from diverse global destinations.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Analyses of the GeoSentinel database are based on where people have traveled to and not the country where travelers are seen. Confirmed diagnoses are made by an indisputable clinical finding or diagnostic test, and probable diagnoses are made by defined clinical and laboratory evidence supporting strong presumption. All sites use best available reference diagnostics in their own country. Codes for a full range of syndromic diagnoses are available to consistently classify patients when a specific etiologic diagnosis is not assignable. Patients included in this analysis visited a GeoSentinel clinic from January 1, 2000, to December 31, 2005, after completion of their travel and with a confirmed or probable infectious gastrointestinal diagnosis. For one analysis, gastrointestinal diagnoses extracted from the GeoSentinel database were grouped as bacterial, parasitic, viral, or syndromic. We did not perform an analysis by specific pathogens according to regions of travel as this was beyond the scope of this study.

Reporting rate ratios

Only patients with exposures restricted to a single one of the 15 regions used by the World Tourism Organization (WTO)18 (Table 1) and a single World Bank income level were included. Annual reports on inbound tourist arrivals for each country for 2000 to 2005 were summed to estimate the denominator of total number of travelers to a region or individual country during the period being analyzed.18

Table 1.  Gastrointestinal infection reporting rate and RRRs by region of travel, GeoSentinel 2000 to 2005
RegionTravelers* (106)Travelers to only one regionStrata
CasesRate/107RRR95% CI
  • RRRs = reporting rate ratios; CI = confidence intervals.

  • *

    Number of travelers to the region in 2000 to 2005 reported to World Tourism Organization.

  • Number of cases of gastrointestinal infections after travel with exposure in the region as reported to GeoSentinel 2000 to 2005.

  • North America not including Mexico.

Western and northern Europe1104.8400.41.0ReferenceLow
North America393.0280.72.01.2–3.2Low
Central/East Europe485.0691.43.92.7–5.8Low
Southern Europe888.22122.46.64.7–9.2Low
Northeast Asia416.11553.710.37.3–14.6Low
Australasia40.2256.217.210.4–28.3Low
Oceania17.02514.740.624.6–66.9Moderate
Middle East183.739921.760.043.3–88.0Moderate
North Africa68.018727.576.054.0–106.9Moderate
Central America151.647531.386.562.7–119.5Moderate
Caribbean104.335634.194.368.0–130.7Moderate
Southeast Asia256.896737.7104.075.8–142.7Moderate
South America90.266373.5203.0147.6–279.3High
Sub-Saharan Africa118.51209102.0281.8205.7–386.1High
South Asia39.61276322.2890.0649.7–1219.2Very high

A GeoSentinel reporting rate for gastrointestinal infection for a region was calculated as the number of travelers reporting to GeoSentinel with an infectious gastrointestinal diagnosis after travel to the region as a proportion of the numbers of tourist arrivals to the region.19 Reporting rate ratios (RRRs) for gastrointestinal illnesses acquired in the destination region were then calculated by using western and northern Europe as a reference region, as these were the regions with the lowest rates of infections associated with oral ingestion of pathogens in travelers. Using the method of Rothman,20 95% confidence intervals (CI) were calculated (estimating the standard error of the natural log of the RRR and using 1.96 times this standard error to derive the bounds of the CI). Regional definitions were the same as the United Nations definitions and have been described previously.19 Mexico was included within Central America and not North America because of the high risk of travelers’ diarrhea associated with travel to this destination and the perceived likelihood of a higher risk of other infections associated with ingestion of pathogens that would artificially elevate the RRR for North America.

For the country analysis, RRRs were calculated with Spain as a reference, as this had the lowest RRR of the destination countries analyzed. Travelers often visit multiple countries on a single trip or make multiple trips in a short time window. Exact country of exposure was attributed only when there was a clear temporal relationship between destination country and disease incubation period. Countries were included if at least 30 patients with any gastrointestinal infection diagnosis had that country as the most likely place of exposure, and tourism data were available for 2000 to 2005.18,21

RRRs by income level of the destination country were compared across the four World Bank groups22 with the high-income group as the reference group. For several countries, per capita income from the Central Intelligence Agency The World Fact Book23 was used to classify the country by World Bank criteria. For countries with missing WTO arrival data for 2004 or 2005, the percentage of all international travel that was to high-, upper-middle-, lower-middle-, and low-income countries was estimated from the percentages from 2003 applied to the total global international visits reported to WTO for 2004 or 2005. Analyses were conducted by using STATA software (Version 6.0 for Windows, StataCorp LP, College Station, TX, USA).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Patients

Of the 22,569 ill-returned travelers seen at GeoSentinel clinics after the completion of their travel during 2000 to 2005, 6,086 (27%) met the criteria for inclusion in the analysis of WTO regions and 5,932 met the criteria for inclusion in the analysis of World Bank income regions. Of the 6,086, 59% (N = 3,591) had diarrhea for less than 2 weeks, 21% (N = 1,278) had chronic diarrhea, and 20% (N = 1,217) had other gastrointestinal infections. The proportion of patients seen in participating clinics was 23% in North America, 62% in Europe, 11% in Asia, and 3% in Australia/New Zealand. Excluded were 373 ill-returned travelers with gastrointestinal final diagnoses who had exposure in multiple WTO regions or had missing travel data as well as 537 who had exposure in multiple World Bank income regions or had missing travel data. An additional 4,608 patients with gastrointestinal infections seen at GeoSentinel sites during the time their travel was ongoing were not analyzed here.

Gastrointestinal diagnoses and regions

The RRRs for infectious gastrointestinal diagnoses among GeoSentinel patients are shown in Table 1. The regions segregate into four RRR strata, which are distinct according to the 95% CI. Among these travelers, the RRRs are least following travel to North America; central, eastern, and southern Europe; Northeast Asia; or Australasia (2–17.2) when compared to travelers visiting western and northern Europe. Risks were moderate, with RRRs ranging from 40.6 to 104, for travel to Oceania, the Middle East, North Africa, Central America, the Caribbean, and Southeast Asia. Travel to South America and sub-Saharan Africa was associated with high RRRs (203–281.8). South Asia was associated with the greatest RRRs (890). For bacterial, parasitic, and viral infections separately, the regions segregate into the same low, moderate, high, and very high relative reporting rate groups, with travel to South America, Africa, and South Asia associated with the highest risks (Table 2). Finally, compared with travelers to high-income countries, travelers to upper middle–income countries had a reporting rate 8.4 times higher (95% CI 7.3–9.7) and travelers to low-income countries had a reporting rate 150.9 times higher (95% CI 134.3–169.6).

Table 2.  Gastrointestinal infection reporting rate and RRRs by pathogen type according to region visited (only a single region visited), GeoSentinel 2000 to 2005
RegionTravelers* (106)BacteriaParasiteVirus
CasesRate/107RRR95% CICasesRate/107RRR95% CICasesRate/107RRR95% CI
  • RRRs = reporting rate ratios; CI = confidence intervals.

  • *

    Number of travelers to the region in 2000 to 2005 reported to World Tourism Organization.

  • Number of cases of gastrointestinal infections after travel with exposure in the region reported to GeoSentinel 2000 to 2005.

  • North America not including Mexico.

Western and northern Europe1,104.890.11.0Reference140.11.0Reference20.01.0Reference
North America393.060.21.90.7–5.3120.32.41.1–5.210.01.40.1–15.5
Central/East Europe485.0210.45.32.4–11.6220.53.61.8–7.030.13.40.6–20.4
Southern Europe888.2310.34.32.1–9.0560.65.02.8–8.930.01.90.3–11.2
Northeast Asia416.1320.89.44.5–19.8471.18.94.9–16.240.15.31.0–29.0
Australasia40.230.79.22.5–33.892.217.77.7–40.810.213.71.2–151.5
Oceania17.031.821.75.9–80.0116.551.123.2–112.510.632.52.9–358.4
Middle East183.7683.745.422.7–91.1703.830.116.9–53.490.527.15.8–125.3
North Africa68.0294.352.424.8–110.6324.737.119.8–69.6101.581.217.8–370.8
Central America151.6644.251.825.8–104.117411.590.652.5–156.1140.951.011.6–224.5
Caribbean104.3636.074.136.9–149.11039.977.944.6–136.260.631.86.4–157.4
Southeast Asia256.82329.0110.957.0–215.829611.591.053.2–155.550.210.82.1–55.4
South America90.211612.9157.980.1–311.025328.0221.3129.2–379.1111.267.414.9–303.9
Sub-Saharan Africa118.522418.9232.0119.2–451.845338.2301.7177.2–513.5141.265.314.8–287.2
South Asia39.631579.5976.5503.4–1894.2489123.5974.5572.9–1657.7287.1390.693.0–1639.6

Gastrointestinal diagnoses by country

To estimate country-specific risk where possible and to determine if the risk of presenting with an infection associated with oral ingestion of pathogens was higher for some countries within a WTO region, the RRR for 28 specific countries for which there were data compared to Spain as the reference group are shown in Table 3. Figure 1 provides a profile map of relative rates of acquisition of gastrointestinal infection by destination. South Africa, Malaysia, Singapore, and Chile stand out as having lower RRRs compared to their regional neighbors. Guatemala and Burma have higher RRRs when compared to neighbors. Mexico has a higher RRR than North America. Nepal and India had the highest RRRs of all countries analyzed in the study.

Table 3.  Gastrointestinal infection RRRs for selected countries (only one country visited), GeoSentinel 2000 to 2005
CountryTravelers* (106)CasesRate/107RRR95% CI
  • RRRs = reporting rate ratios; CI = confidence intervals.

  • *

    Number of travelers to the region in 2000 to 2005 reported to World Tourism Organization.

  • Number of cases of gastrointestinal infections after travel with exposure in the region reported to GeoSentinel 2000 to 2005.

Spain312.6301.01.0Reference
Italy231.2381.61.71.1–2.8
Malaysia78.9303.84.02.4–6.6
China224.41004.54.63.1–7.0
Turkey84.0748.89.26.0–14.0
Mexico122.322718.619.313.2–28.3
Tunisia33.27221.722.614.8–34.6
South Africa39.28822.423.415.5–35.4
Morocco28.98128.029.219.2–44.4
Chile10.33029.130.318.3–50.3
Thailand64.233852.654.937.8–79.7
Costa Rica7.75064.967.743.0–106.4
Indonesia30.119865.868.546.7–100.6
Philippines12.58668.871.747.3–108.6
Dominican Republic19.115581.284.657.2–125.0
Colombia3.93282.185.552.0–140.7
Cambodia5.24382.786.254.1–137.3
Cuba11.29584.888.458.6–133.2
Honduras3.73594.698.660.5–160.5
Guatemala5.962105.1109.570.8–169.3
Kenya6.0117195.0203.2136.1–303.5
Peru6.0131218.3227.5153.0–338.3
Ecuador4.4107243.2253.4169.0–379.9
Sri Lanka2.873260.7271.7177.6–415.6
Bolivia2.163300.0312.6202.4–482.8
Myanmar1.348369.2384.7243.8–607.1
India17.6916520.5542.3377.0–780.2
Nepal2.1135642.9669.9451.0–994.9
image

Figure 1. A profile map of relative rates of acquisition of gastrointestinal infection by destination. Global distribution of reporting rate ratios for all gastrointestinal infections in ill-returned travelers presenting to 30 GeoSentinel clinics on six continents. Twenty-eight countries with available country-specific data are outlined. Where country-specific data are lacking, a country assumes the characteristics of its region as per data shown in Table 1 (approach as per13,24).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Gastrointestinal infections are a common problem for travelers whether they visit Paris,25 Mombasa, or Goa.9 This study focuses on physician-verified diagnoses in a large cohort of returned travelers with a broad range of post-travel gastrointestinal illness.

Similar to our findings for all gastrointestinal infections (Figure 1) occurring after travel, in studies of acute travelers’ diarrhea occurring during the actual travel, South Asia8,14–16 and Africa8,13,14 are consistently areas of high risk. Central and South America,8,14,15 Southeast Asia,8,14 the Pacific, and the Caribbean8,16 are areas of moderate to high risk. North America, northern and central Europe, Australia, and New Zealand8 are areas of low risk.

In contrast to the diarrhea-only studies of Hill14 and Steffen,8 we found Northeast Asia to be areas of lower risk. Steffen8 suggests resorts of the northern Mediterranean as an area of moderate risk, while our analysis found southern and Mediterranean Europe to be an area of only slightly higher risk than western and northern Europe. These differences may relate to the 15-year time difference in the two studies and improvement in tourism infrastructure. The Middle East was an area of highest risk in the study by Cobelens and colleagues15 but only an area of moderate risk in our analysis.

While we focused only on travelers who sought medical care after travel had finished, our results are generally consistent with those of single-center travelers’ diarrhea studies that often are limited by local or national differences in the makeup of the traveling population and the destinations visited. One important component of this study is that the multicenter GeoSentinel data set focuses on the destination of travel and not on the clinical site where the patient presented. GeoSentinel clinics capture a sentinel sample of the travelers ill enough to seek medical care. Most travelers’ diarrhea, a frequent manifestation of gastrointestinal infections in travelers, is mild and resolves rapidly without requiring medical attention.13–15 Thus, GeoSentinel data may underestimate the full spectrum of travel-related gastrointestinal risks of travel.

Estimation of the risk of infection associated with travel to a particular destination would ideally be expressed as the rate of illness in all travelers to that destination, but the exact ascertainment of either numerator or denominator is very difficult for most countries. A number of approaches are possible that provide complementary data.16 In the present analysis, we have used WTO data to estimate the denominator to provide a measure of the relative rate of gastrointestinal infection in all travelers whether they become ill or not. However, the use of WTO data has important limitations. Using denominators, counting all travelers may prevent a country from appearing high risk simply because many tourists visit it. Tourist information is not available for all countries for every year, so differing methods are used by each country in counting arrivals, the inclusion of day visitors with some countries, and extrapolation for missing values that affect the estimated distribution of tourists. Some countries have a large number of “same-day visitors” who are not “tourists” and are not included in estimations of the travelers at risk. Cruise ship passengers represent a significant proportion of visitors in some areas. Not including these visitors in denominators may inflate the apparent risk of such areas.

Other factors may also inflate the apparent risks associated with certain regions and countries perceived by patients as having exotic risk. Unwell returned travelers may be more likely to seek medical advice if they have been to an area they perceive as potentially “risky.” Further, they may be more likely to seek care at a specialized travel medicine clinic (as the GeoSentinel clinics are) than their general practitioner if they have been to an area seen as associated with exotic health risks. This may account for the relatively low RRR we describe for Mexico, a destination with known high risk for self-limited travelers’ diarrhea10,26 but that is not publicly perceived as risky for serious illness.

Overall, more than 50% of global travel is from the developed world,18 so the data set mostly offers insight into the relative risks faced by residents of these Western nations. Risks differ even within single countries according to the type, duration, and season of travel as well as to individual factors and even individual accommodation and travel habits.8,9,13,15,27,28 Further, the relationship between country income level and risk of illness may be confounded by climate.10 Many of these factors are difficult to either classify or ascertain and were not included in this analysis.

The current data (Figure 1, Table 3) provide a reference of RRRs for clinicians advising and caring for travelers who are at risk of infections associated with oral ingestion of pathogens severe enough to require post-travel medical care. Finally, these data provide travelers with estimates of the relative risks of returning with a gastrointestinal infection for different global destinations.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

The GeoSentinel data collection protocol was reviewed by the institutional review board officer at the National Center for 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.

Financial support

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

The Global Surveillance Network of the International Society of Travel Medicine is supported by Cooperative Agreement U50/CCU412347 from the Centers for Disease Control and Prevention, which had a role in the design, conduct, and reporting of this analysis. The Centre for Clinical Research Excellence, Royal Melbourne Hospital, Victoria, is funded by the National Health and Medical Research Council of Australia, which had no direct role in the design, conduct, and reporting of this analysis.

Declaration of interests

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

D. O. F. has accepted honoraria from Salix Pharmaceuticals, GlaxoSmithKline, and sanofi pasteur. B. A. C. is a consultant to and has accepted grants, honoraria, and stock from Salix Pharmaceuticals; he is also a consultant and has accepted honoraria from GlaxoSmithKline, AVANT Immunotherapeutics, and Acambis. L. W. owns shares of Amgen. Z. G., J. B., D. O., K. L., F. v. S., P. P., E. S., G. B., and J. T. report no conflicts of interest. This manuscript contains surveillance data only, without any interventions or discussion of therapeutics or products. To be inclusive, we disclose these relationships as these entities produce gastrointestinal-related products whose sale could be perceived as benefiting from a statement of the risk of associated disease entities.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Appendix

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. Financial support
  8. Declaration of interests
  9. References
  10. Appendix

Appendix. GeoSentinel Surveillance Network Members

In addition to the authors, GeoSentinel Surveillance Network members contributing data for this analysis include: Elizabeth Barnett, Boston University, Boston, MA, USA; Giampiero Carosi and Francesco Castelli, University of Brescia, Brescia, Italy; Lin Chen and Mary Wilson, Harvard University, Cambridge, MA, USA; Jean Delmont and Philippe Parola, Hôpital Nord, Marseille, France; Alejandra Gurtman, Mount Sinai Medical Center, New York City, NY, USA; Devon Hale and Stephanie Gelman, University of Utah, Salt Lake City, UT, USA; Nancy Piper Jenks, Hudson River Health Care, Peekskill, NY, USA; Elaine Jong and Jean Haulman, University of Washington, Seattle, WA, USA; Jay Keystone and Kevin Kain, University of Toronto, Toronto, Ontario, Canada; Phyllis Kozarsky and Carlos Franco-Paredes, Emory University, Atlanta, GA, USA; Carmelo Licitra, Orlando Regional Health Center, Orlando, FL, USA; Louis Loutan, University of Geneva, Geneva, Switzerland; Michael Lynch, Fresno International Travel Medical Center, Fresno, CA, USA; Susan McDonald, Beijing United Family Hospital and Clinics, Beijing, China; Susan McLellan, Tulane University, New Orleans, LA, USA; Robert Müller, Travel Clinic Services, Johannesburg, South Africa (June 2003 to June 2005 only); Thomas Nutman and Amy Klion, National Institutes of Health, Bethesda, MD, USA; Cecelia Perret Pérez, Catholic University of Chile, Santiago, Chile; Bradley Sack and Robin McKenzie, Johns Hopkins University, Baltimore, MD, USA; Hiroko Sagara, Yokohama Municipal Citizen’s Hospital, Yokohama, Japan; Patricia Schlagenhauf and Robert Steffen, University of Zurich, Zurich, Switzerland; Marc Shaw, Travellers Health and Vaccination Centre, Auckland, New Zealand; William Stauffer and Patricia Walker, University of Minnesota, St Paul, MN, USA; Annelies Wilder-Smith and Po Lian Lim, Tan Tock Seng Hospital, Singapore; Murry Wittner, Albert Einstein School of Medicine, Bronx, NY, USA.