Travel‐Associated Dengue Infections in the United States, 1996 to 2005

Hamish P. Mohammed PhD, Mary M. Ramos MD, Aidsa Rivera MSc, Michael Johansson PhD, Jorge L. Muñoz‐Jordan PhD, Wellington Sun MD, Kay M. Tomashek MD
DOI: http://dx.doi.org/10.1111/j.1708-8305.2009.00374.x 8-14 First published online: 1 January 2010

Abstract

Background As the incidence of dengue increases globally, US travelers to endemic areas may be at an increased risk of travel‐associated dengue.

Methods Data from the US Centers for Disease Control and Prevention's laboratory‐based Passive Dengue Surveillance System (PDSS) were used to describe trends in travel‐associated dengue reported from January 1, 1996 to December 31, 2005. The PDSS relies on provider‐initiated requests for diagnostic testing of serum samples via state health departments. A case of travel‐associated dengue was defined as a laboratory‐positive dengue infection in a resident of the 50 US states and the District of Columbia who had been in a dengue‐endemic area within 14 days before symptom onset. Dengue infection was confirmed by serologic and virologic techniques.

Results One thousand one hundred and ninety‐six suspected travel‐associated dengue cases were reported—334 (28%) were laboratory‐positive, 597 (50%) were laboratory‐negative, and 265 (22%) were laboratory‐indeterminate. The incidence of laboratory‐positive cases varied from 1996 to 2005, but had an overall increase with no significant trend (53.5 to 121.3 per 108 US travelers, p = 0.36). The most commonly visited regions were the Caribbean, Mexico and Central America, and Asia. The median age of laboratory‐positive cases was 37 years (range: <1 to 75 y) and 166 (50%) were male. Of the 334 laboratory‐positive cases, 41 (12%) were hospitalized, and 2 (1%) died.

Conclusions Residents of the US traveling to dengue‐endemic regions are at risk of dengue infection and need to be instructed on appropriate prevention measures prior to travel. Especially in light of the potential transmissibility of dengue virus via blood transfusion, consistent reporting of travel‐associated dengue infections is essential.

Dengue, the most common arboviral infection in the world, is caused by one of the four dengue viruses (DENV‐1, ‐2, ‐3, and ‐4). Most dengue virus infections are asymptomatic 1–3 ; however, the spectrum of clinical illness includes undifferentiated fever, classic dengue fever (DF), dengue hemorrhagic fever (DHF), and dengue shock syndrome (DSS). The average incubation period for dengue infection is 5 days, followed by an infectious period of viremia lasting for an average of 4.5 days. 4,5 There is no licensed dengue vaccine and the only means of prevention is through avoidance of bites from the mosquito vector. The principal dengue vectors, Aedes aegypti and Aedes albopictus mosquitoes, are common throughout the tropics and subtropics. Roughly, one third of the world's population lives in dengue‐endemic areas in over 100 countries, with an estimated 50 to 100 million dengue cases occurring worldwide annually. 6 Over the past two decades, dengue epidemics with cases of DHF have been occurring with increasing frequency around the globe. 7,8

Although the vast majority of dengue infections occur among residents of dengue‐endemic areas, dengue is being increasingly diagnosed among travelers to these destinations. 9 Recent findings from the GeoSentinel Surveillance Network 9,10 indicate that dengue is the most commonly reported cause of acute febrile illness in travelers returning from the Caribbean, South America, south central Asia, and southeast Asia. Moreover, dengue was found to be the second most common cause of febrile illness (after malaria) in travelers returning from sub‐Saharan Africa and Central America. 9 In the United States, this upward trend will likely continue with the increasing rates of international travel in recent years, 11 and the increasing number of new US immigrants from endemic countries 12 who are likely to visit friends and relatives in their countries of origin. 13,14

Concern over the risk of reintroduction of dengue virus into the United States has been recently expressed. 15 Ae. aegypti mosquitoes exist in a few states in the southeastern United States. 16 However, Ae. albopictus mosquitoes exist in 26 states throughout the southeastern United States and Hawaii. 17 The presence of competent vectors in the continental United States and Hawaii, along with the increasing global incidence of dengue and travel to dengue‐endemic areas, underscores the need for vigilance regarding possible importation of dengue virus via travel‐associated cases.

The Division of Vector‐Borne Infectious Diseases at the US Centers for Disease Control and Prevention (CDC) conducts dengue surveillance in collaboration with the Puerto Rico Department of Health. This laboratory‐based Passive Dengue Surveillance System (PDSS) collects serum samples and epidemiologic information from suspected dengue cases reported by healthcare providers from Puerto Rico, the US Virgin Islands, and the 50 US states and the District of Columbia. This analysis uses PDSS surveillance data to describe the epidemiology of travel‐associated dengue among travelers from the United States during the period of 1996 to 2005.

Methods

Surveillance System and Case Definitions

All suspected travel‐associated dengue cases reported from the 50 states or the District of Columbia to the PDSS from January 1, 1996 through December 31, 2005 were considered in this analysis. The PDSS relies on provider‐initiated requests for diagnostic testing of serum specimens via state health departments and collects laboratory, clinical, and epidemiologic data (including travel history) from suspected dengue cases. A suspected dengue case was defined as one with a dengue‐compatible illness (eg, acute febrile illness with rash, myalgia, and arthralgia) and a history of recent travel to a dengue‐endemic area. A case of travel‐associated DF was defined as a laboratory‐positive dengue infection in a resident of one of the 50 states or the District of Columbia who traveled in the 14 days before symptom onset to a dengue‐endemic area. A serum specimen and a CDC Dengue Case Investigation Form (DCIF), which included information on basic demographic data, dates of symptom onset and sample collection, and symptoms, were submitted for all suspected cases. Occasionally, a brief letter summarizing the clinical course, laboratory values, and travel history was also submitted.

Laboratory Testing

All laboratory testing was performed at the Dengue Branch (CDC). Serum specimens taken during the first 5 days after the onset of illness were defined as acute‐phase specimens, whereas those taken six or more days after symptom onset were defined as convalescent specimens. Both acute and convalescent specimens were tested using serologic techniques, whereas virus identification and isolation were attempted only on the acute specimens. Serologic testing was conducted using an IgM capture enzyme‐linked immunosorbent assay (MAC‐ELISA) for detecting anti‐dengue IgM antibodies. 18 Since 2005, viral identification was attempted using a real‐time, reverse transcriptase polymerase chain reaction assay (RT‐PCR, TaqMan Applied Biosystems). 19,20 Prior to that year, viral isolation was attempted by viral culture using C6/36 mosquito cells or tissues from inoculated adult Toxorhynchites amboinensis mosquitoes. 21,22 All cases with positive PCR results or with IgM seroconversion were tested by IgG ELISA 23 to determine primary or secondary status of current infections.

A probable dengue case was defined as a suspected dengue case with a positive IgM MAC‐ELISA result on a single, acute‐ or convalescent‐phase serum specimen, or an IgG‐ELISA antibody titer ≥163,840 on an acute‐ or convalescent‐phase specimen. 23 A confirmed dengue case was defined as a suspected dengue case that had dengue virus identified from an acute‐phase serum specimen or autopsy tissue sample, or one that met at least one of these two criteria: seroconversion from a negative anti‐dengue IgM in the acute‐phase specimen to a positive IgM in a convalescent‐phase specimen, or a fourfold or greater change in IgG or IgM antibody titers in paired serum specimens. The absence or presence of anti‐dengue IgG antibodies in acute‐phase serum samples was used to distinguish primary (first) from secondary (subsequent) dengue infections. All suspected dengue cases with negative acute‐phase specimen results and no convalescent specimens were classified as indeterminate. Suspected cases that did not meet these laboratory criteria were classified as laboratory‐negative. For the purposes of this analysis, both probable and confirmed dengue cases are considered laboratory‐positive.

WHO Criteria for Dengue

The World Health Organization (WHO) defines DF as an acute febrile illness with at least two of the following: headache, retro‐orbital pain, myalgia, arthralgia, rash, hemorrhagic manifestations (such as epistaxis, gingival bleeding, gastrointestinal bleeding, hematuria, or menorrhagia), or leukopenia as well as supportive serology or an epidemiologic link to a confirmed case of DF. 6 DHF is defined as fever or history of fever of 2 to 7 days duration in the presence of thrombocytopenia (≤100,000 cells/mm3), at least one hemorrhagic manifestation, and objective evidence of plasma leakage, including pleural effusion, ascites, low serum albumin or protein, or hemoconcentration. Lastly, DSS is defined as DHF plus a rapid, weak pulse with narrow pulse pressure or hypotension with cold, clammy skin and restlessness.

Statistical Analysis

We performed a univariate analysis to describe the suspected cases by demographic characteristics, state of residence, travel destination, laboratory results, and clinical outcomes such as hospitalization, presence of hemorrhagic manifestations, or those meeting criteria for DF, DHF, and DSS. The number of US resident travelers visiting overseas destinations from 1996 to 2005 was obtained from the Office of Travel and Tourism Industries, 11 and this was used to calculate the incidence of laboratory‐positive dengue in travelers. We used logistic regression to test for significant linear trends in laboratory diagnoses and in the incidence of laboratory‐positive dengue in travelers over the 10‐year period under review. Analyses were performed using SPSS version 12 (SPSS Inc.) and SAS version 9.1 (SAS Institute), and all tests for significance were two‐sided and performed at an alpha error rate of 5%.

Human Subjects Review

This analysis of routinely collected, de‐identified, and confidential dengue surveillance data was determined to be a non‐research activity and did not require institutional review by the CDC Human Subjects Review Committee.

Results

During 1996 to 2005, 1,196 suspected travel‐associated dengue cases from 49 states and the District of Columbia were reported to the PDSS. Of the 1,196 suspected cases, 334 (28%) were laboratory‐positive, 597 (50%) were laboratory‐negative, and 265 (22%) were laboratory‐indeterminate. Those with positive, negative, and indeterminate results did not vary significantly by age or sex. Suspected travel‐associated dengue cases by laboratory diagnosis are shown in Figure 1. The proportion of laboratory‐positive cases varied by year, with an overall increase over the period under review (25% to 39% laboratory‐positive cases from 1996 to 2005). However, there was no significant linear trend in proportions (p = 0.31). Similarly, despite an overall increase in the incidence of laboratory‐positive cases per 108 US travelers from 53.5 to 121.3 from 1996 to 2005, there was no significant linear trend (p = 0.36) (Figure 2). Dengue virus serotype was successfully identified in 36 (9%) of the 393 acute samples submitted; 5 were positive by RT‐PCR, 27 by viral culture, and 4 by both. Of these 36 samples, 10 cases of DENV‐1, 11 cases of DENV‐2, 7 cases of DENV‐3, and 8 cases of DENV‐4 were identified.

Figure 1

Number of suspected travel‐associated dengue cases by laboratory result, United States, 1996 to 2005.

Figure 2

Incidence of laboratory‐positive cases per 108 US travelers, United States, 1996 to 2005.

Just over half (52%) of the 334 laboratory‐positive cases were reported from four states: New York, Massachusetts, Texas, and Hawaii (Figure 3). Of all laboratory‐positive cases, travel destinations were documented for 240 (72%). The most commonly visited regions were the Caribbean (23%), Mexico and Central America (20%), and southeast Asia (17%) (Table 1). The most commonly visited destinations within each region were Puerto Rico (n = 25), Mexico (n = 36), and Thailand (n = 20), respectively. The median age of all laboratory‐positive cases was 37 years (range: <1 to 75 y); 166 (50%) were male.

Figure 3

Proportion of laboratory‐positive cases by state of residence, United States, 1996 to 2005.

View this table:
Table 1

Laboratory‐positive cases by year and region of travel—United States, 1996 to 2005

Region of travel
YearAfrica (%)Caribbean (%)Central America* (%)South America (%)Southeast Asia (%)South Pacific (%)Unknown (%)Total (n)
19960360425729 28
1997031179171411 35
19983392062408 66
199902429035012 17
20000388038015 13
2001514510142924 21
20020216927927 33
20030261304948 23
2004014181823027 22
2005034300055 77
Total12320517528335
  • * Includes Mexico

  • One traveler is counted twice, having reported a travel history to both “Africa and India.”

Among the 334 laboratory‐positive patients, 30 (9%) had primary infections and 55 (16%) had secondary infections. The most commonly reported symptoms were fever (55%), headache (35%), myalgia (30%), and rash (28%). Other reported symptoms included chills (26%), nausea or vomiting (17%), arthralgia (14%), diarrhea (14%), and retro‐orbital pain (10%). Some travelers had severe illness: 41 (12%) were hospitalized, 41 (12%) had at least one hemorrhagic manifestation (most common: petechiae, n = 25), 31 (9%) had platelet counts ≤100,000/mm3, and 4 (1%) had evidence of capillary leakage. Of the laboratory‐positive cases, 119 (36%) met WHO criteria for DF, 2 (1%) met criteria for DHF, and none met criteria for DSS. Two (1%) fatal cases occurred in previously healthy young adults who had traveled to Mexico and acquired secondary dengue infections.

Discussion

This review of 10 years of dengue surveillance data among travelers from the 50 US states and the District of Columbia provides an important measure of the frequency and severity of travel‐associated dengue illness. An average of 120 suspected travel‐associated dengue infections were reported annually to the PDSS, and there was no significant increase in the incidence of laboratory‐positive cases in travelers. Most reported infections were mild; relatively few cases were hospitalized. However, the data underscore the risk of dengue infection for travelers to dengue‐endemic areas.

Although 12% of laboratory‐positive dengue cases were hospitalized, cases of severe dengue illness were uncommon among US travelers. Over the 10‐year analysis period, few cases were reported as having hemorrhagic manifestations, and even fewer met WHO criteria for DHF. These findings are consistent with previous research on travel‐associated dengue. 24–27

Secondary dengue infections are associated with severe illness including DHF and DSS, 28 and are more likely to occur in residents of endemic areas (who are repeatedly exposed) than in visitors. However, international travel among US residents is increasing, and frequent travelers may be at risk of secondary dengue infection and thus, more severe dengue illness. The volume of US residents traveling abroad hit a record high of 64 million in 2007, reflecting an increase of roughly 15% since 1998. 11 Moreover, increased travel to Central America, South America, Africa, and Asia, all regions with dengue‐endemic countries, contributed to the new record for US outbound travel. 11

There is potential for limited secondary transmission of dengue upon return of an infected traveler to the United States as competent vectors exist throughout much of the southeastern region. With incubation and viremic periods of roughly 5 days each, 4,5 travelers may be infectious for several days upon return to their state of residence. In 2001, Hawaii experienced its first dengue outbreak in over 50 years, an outbreak likely caused by importation of dengue virus from an infected traveler. 29 Sporadic outbreaks of DF have occurred in the past two decades in southern Texas along the US‐Mexico border. 30–32 US healthcare providers are often unfamiliar with DF, which can delay accurate diagnosis in symptomatic travelers, thereby increasing the risk of secondary transmission. Despite the risk of secondary dengue transmission in the southeastern United States, infrastructural factors such as the widespread usage of air‐conditioning in homes in the United States may prevent the establishment of autochthonous transmission. 30,32 Lastly, asymptomatic dengue infections may also potentially pose a risk to others via blood donations, 33–36 as current screening practices do not defer persons from donating blood solely on the basis of recent travel to the tropics.

Future work is needed to more accurately determine the burden of dengue infection and the risk of infection among US travelers. Mathematical modeling techniques may be employed to determine this risk. 37 Data captured by the PDSS could be supplemented by reports of suspected and confirmed dengue cases from the major commercial reference laboratories throughout the United States which perform dengue diagnostic testing. We recommend making dengue a nationally reportable disease and strongly encourage reporting from all state and local health departments to the CDC. A timely and sensitive surveillance system with more complete data is essential for detecting introductions of dengue virus, preventing secondary transmission within the households and communities of returning travelers, and guiding prevention efforts.

Persons traveling from the United States should be given pre‐travel advice on lowering their risk of dengue infection while overseas. Preventive measures include wearing long‐sleeve shirts and long pants, using insect repellent with 30% to 50% N,N‐diethyl‐m‐toluamide (DEET), and staying in accommodations with air‐conditioning and screened windows. There is also a need to raise awareness among continental US physicians so that dengue is considered in the differential diagnosis of febrile patients with recent travel histories to the tropics and subtropics.

Our study has several limitations. Considerable underreporting of dengue is likely because the PDSS is a passive system, dengue is not nationally reportable in the United States, and reporting of cases to the CDC is voluntary. Given the diagnostic challenges 38 and lack of awareness among US physicians, dengue in travelers may be often misdiagnosed. Furthermore, dengue infections are often asymptomatic or present only with mild undifferentiated febrile illness. 39 Serological testing of paired acute‐ and convalescent‐phase specimens has been the foundation of dengue diagnostics, but this approach generally confirms dengue cases only after patients recover and sensitivity varies between tests. 40 Detection of viral RNA is being increasingly used for dengue diagnosis with promisingly high sensitivity and specificity; however, costs associated with these tests are still prohibitive in many endemic areas. 41 The development and improvement of sensitive, fast, and inexpensive tests for early diagnosis of dengue is crucial to timely dengue surveillance. In this study, 22% of all suspected cases had an indeterminate laboratory diagnosis, indicating the lack of paired samples.

Further underreporting of dengue is possible as, given the 5‐day incubation period, many travelers may become ill and seek care in the country of travel. Lastly, many physicians who reported suspected cases inadequately completed the DCIF. A missing date of onset of illness, in particular, limits the interpretability of the laboratory results.

Given the global dengue pandemic, increasing travel among US residents, and the presence of dengue vector mosquitoes in much of the continental United States, strong consideration should be given to making dengue a nationally reportable disease. US residents traveling to dengue‐endemic regions need to be instructed on appropriate prevention measures prior to travel. Physicians practicing in the continental United States should be alerted to the possibility of dengue infection among travelers to the tropics and subtropics. Repeat travelers to dengue‐endemic areas are at a higher risk of secondary dengue infection and, as a consequence, more severe illness. Surveillance of dengue in US travelers is essential for the early detection of any introductions of dengue virus into the continental United States.

Declaration of interests

The authors state they have no conflicts of interest to declare.

Acknowledgments

We acknowledge the assistance of the state and local health departments of the United States, as well as the staff of the Dengue Branch, and Jennifer Lehman (DVBID).

Footnotes

  • Data have been previously presented at the Late Breakers in Clinical Tropical Medicine Session at the American Society of Tropical Medicine and Hygiene Annual Meeting in 2006 (Atlanta, GA).

References

View Abstract