Hepatitis A is considered as one of the most common vaccine-preventable travel-related diseases globally.1,2 Despite access to efficient and safe vaccines, the immunization level in travelers to endemic areas is shown to be low in most countries.3–5 Several studies have estimated the true incidence and risk groups of travel-related hepatitis A,1,2,6,7 but they have all been hampered by the lack of a homogeneous denominator or underreporting of cases. Therefore, the aims of this population-based study were to calculate incidence rates and to estimate the risk groups for travel-related hepatitis A infection in travelers returning to Sweden and to relate our findings to immunization coverage in travelers. We have used the national Swedish notification system for communicable diseases, with a very high sensitivity, combined with data from a large ongoing national survey on travel patterns, providing a stable travel denominator.
Background Traveling to highly endemic areas for hepatitis A is increasing while the immunization level in travelers has been shown to be low in the countries studied.
Methods In this population-based study, we have estimated the incidence rate of travel-related hepatitis A during 1997 to 2005 by use of the Swedish notification system of communicable diseases and an ongoing national database on travel patterns. We have also acquired airport-based immunization data from 2007.
Results During the study period, 636 cases of travel-related hepatitis A were notified. Traveling to East Africa was associated with the highest incidence rate (14.1 cases/100,000 person months), followed by the Middle East (5.8/100,000 person months), and India with neighboring countries (5.6/100,000 person months). Visiting Friends and Relatives (VFR) travelers represented 83, 91, and 70% of the cases to these three regions. By age-group, the highest incidence was found in children 0 to 14 years (3.1/100,000 travelers) where 88% of the cases were VFR travelers. Incidence rate in unprotected travelers to East Asia, North Africa, and the Middle East was 2, 12, and 18 cases/100,000 person months, respectively. In 2007, 79% of the travelers were immunized against hepatitis A.
Conclusions We conclude that travelers, and especially children, who are VFR in endemic areas constitute a high-risk group for acquiring hepatitis A infection, while the risk for unprotected tourists to East Asia is low.
Methods and Material
Notification Data on Hepatitis A
Hepatitis A is a notifiable disease in Sweden according to the Swedish Communicable Diseases Act. Notifications are made to the Swedish Institute for Infectious Diseases Control (SMI) and are compulsory both for the clinician and for the laboratory when identifying hepatitis A virus (HAV) IgM antibodies. The notification system is based on the unique personal identification number given to every Swedish citizen. Refugees and individuals who are seeking asylum are given a temporary number very soon after arrival to Sweden. The clinical notification contains information about age, gender, country of infection, time at onset of symptoms, reason for notification, and other epidemiological data when appropriate. The laboratory notification is automatically merged with the clinical notification to avoid duplication of reports. As there is a limited number of laboratories performing HAV diagnostics in Sweden and as most of these laboratories are directly linked to the national notification database, the system is highly sensitive. One study has shown that the notification system in 2002 picked up more than 98% of patients with a microbiologically confirmed diagnosis of a notifiable disease.8 We have collected notification data from all cases of notified hepatitis A in Sweden during the period 1997 to 2005. The travel-related cases have been identified, as well as a subgroup of travelers VFR. The VFR-group consists of persons who have their origin and relatives in a country outside Sweden, and the identification of this group has been made using epidemiological information on the notification form.
Denominator Data on Travel Patterns
As a denominator, we have used data on international travel of Swedish residents, the Swedish Tourist, and Travel Database (TDB) (http://www.resursab.se). The TDB is an ongoing telephone survey based on a fresh randomized selection of 2,000 members of the Swedish population each month. To be randomized for a telephone interview for the TDB, the household must have a nonmobile subscription and one of the household members must be younger than 75 years. The telephone calls are undertaken both at daytime and during evenings, and the first question to a randomized household member is if he or she has been traveling more than one overnight during the recent month. Apart from information on age and gender, the database contains categorized information on destination, length of travel in days, purpose of travel, and more detailed information about their stay.
Using the demographic distribution of respondents in the data set and the demographic distribution of Swedish residents based on official statistics, the total number of travelers can be extrapolated. In the TDB, countries that are less often visited are grouped together into regions (Figure 1). No information about country of birth or medical history is given in this database. As a denominator in the calculation of incidence of travel-related hepatitis A, we have used data from TDB for the same years as the notified cases (1997–2005). For the calculation of time spent traveling, for every specific region, we have multiplied the number of nights spent abroad (1–10, 11–17, 18–24, >24) with the total number of travels. In the groups 11 to 17, 18 to 24, and 25+ nights spent abroad, we used a mean of 14, 21, and 50, respectively.
To estimate the hepatitis A immunity and vaccination coverage in travelers to some destinations of special interest, we have collected information from Swedish travelers to Asia (mainly Thailand and Malaysia), Iran, and North Africa (Egypt and Morocco). The travelers departed from the two main airports of Sweden, Landvetter (Gothenburg) and Arlanda (Stockholm), between April and June 2007. When lining up for check-in, all travelers who were Swedish residents were asked to fill in an anonymous questionnaire about their travel destination, sex, age, and protection against hepatitis A (yes/no/don’t know). They were asked to specify the type of protection with the alternatives previous clinical illness, vaccine or gamma globulin, and if applicable, also number of vaccine doses and when each dose was given. The questionnaire was written in Swedish and collected on-site. All travelers were informed that the results would be published as a part of a research project and that participation was voluntary. The travelers who came very late for check-in and therefore were in a real hurry were not asked at all.
The cumulative incidence of travel-related hepatitis A by region, age, and sex per 100,000 travelers with 95% confidence intervals (CI) was calculated by dividing the notified cases with the estimated number of travelers from the TDB. Incidence rates, as cases per 100,000 person months, with 95% CI, were calculated by dividing notified cases with length of travel to each specific region. The estimated proportions of unprotected travelers were calculated by multiplying the total number of estimated travelers with the percentage of travelers in the vaccination survey who were not protected or had unknown protection against hepatitis A. A p value less than 0.05 was considered significant.
The use of notification data is regulated by the Swedish Communicable Diseases Act. The TDB data are anonymous and aggregated. This study has been approved by the Medical Ethics Committee of the Karolinska Institute (No. 03-667)
Hepatitis A Cases
As shown in Table 1, from 1997 to 2005, 1,885 cases of hepatitis A infection were notified to the SMI. A total of 636 cases were travel related (34%) and 331(52%) of those were categorized as VFR. Children 0 to 14 years accounted for 43% of the travel-related cases, and the majority (57%) were males. The overall trend of notified hepatitis A was decreasing, but the proportion of travel-related cases, as well as the proportion of VFR-cases, was increasing during the study period.
|Category of infection||1997||1998||1999||2000||2001||2002||200 3||2004||2005||1997–2005|
|Travel associated (% of total)||107 (15)||66 (25)||96 (52)||54 (35)||68 (40)||38 (51)||59 (48)||81 (60)||67 (72)||636 (34)|
|VFR (% of travel related)||43 (40)||35 (53)||63 (65)||29 (54)||28 (41)||24 (63)||41 (69)||33 (41)||35 (52)||331 (52)|
During the study period, 22,673 persons were recorded with international travel in the TDB. After statistical weighting and extrapolating, this figure corresponds to a total of 94,443,000 travel episodes by Swedish residents (Table 2).
|Sex/age/region||Estimated no. of travelers||Cases||Cases per 100,000||95% CI||100,000 months||Cases per 100,000 months||95% CI||%VFR|
|Eastern Europe (including Baltic Republics)||5,150,000||15||0.3||0.2–0.5||50||0.3||0.3–0.5||40|
|Russia and former Soviet union||381,000||13||3.4||1.9–6.1||5||2.6||2.4–4.6||38|
|India with neighbors||215,000||40||18.6||12.4–28.0||7||5.6||5.4–8.40||70|
Hepatitis A Risk Groups
In Table 2, we show that the overall risk of travel-associated hepatitis A was 0.67 per 100,000 travelers (0.72 among males and 0.62 among females). By age-group, the highest risk was seen in children 0 to 14 years (3.1/100,000 travelers). The overall incidence rate was 0.61 per 100,000 travel months, and travel to East Africa was associated with the highest risk (14.1 cases/100,000 travel months), followed by the Middle East (5.8/100,000 travel months), and the Indian subcontinent (5.6/100,000 travel months).The lowest incidence rate in a high-risk hepatitis A area (as defined by the World Health Organization) was found in East Asia where the incidence rate was 0.3 cases per 100,000 travel months, which is comparable with the incidence rate of Europe and North America (Figure 1). Fifty-two percent of the travel-related cases were coming from the VFR group. In the age group 0 to 14 years, the corresponding figure was 88%. The VFR-travelers consisted of 91, 83, and 70% of the hepatitis A cases to the respective regions of the Middle East, East Africa, and the Indian subcontinent.
We collected information from 399 travelers. Fifteen percent (n= 60) were children and 12% (n= 47) were classified as VFR. The median age was 26 years (range 0.5–73), 47% were males, 47% females, and 6% did not answer the question about their sex. In total, 79% (n= 314) were immunized against hepatitis A. Travelers to Asia were significantly more often immunized than travelers to North Africa (p < 0.0001). The group of VFR travelers aged 0 to 14 years were immunized to less extent than non-VFR children, but the difference was not significant. In total, 12% of the travelers were not immunized, including two children <14 years of age. Nine percent of the travelers did not know if they were protected against hepatitis A or not and 0.7% (n= 3) stated previous clinical illness (Table 3). Approximately 10% of the travelers who were asked at the airports declined participation. The reasons were mostly either unwillingness to fill in a questionnaire or a sense of lack of time.
|Protection against hepatitis A||All travelers (n= 399) (%)||Asia (n= 245) (%)||North Africa (n= 120) (%)||Iran (n= 34, VFR) (%)||Non-VFR children (n= 45) (%)||VFR children (n= 15) (%)|
|Clinical disease||3 (0.7)||2 (1)||0||1 (3)||0||0|
|Immunized||314 (79)||213 (87)||80 (66)||22 (64)||43 (96)||11 (73)|
|Not protected||46 (12)||11 (5)||31 (26)||4 (12)||2 (2)||1 (7)|
|Unknown protection||36 (9)||19 (7)||9 (8)||7 (21)||3 (2)||3 (20)|
Hepatitis A Incidence in Unprotected Travelers
Considering the immunization rates and previous clinical illness of travelers to Iran, East Asia, and North Africa, we calculated the incidence rates for the estimated number of unprotected travelers to these regions. The incidence rates were 2, 12, and 18 cases per 100,000 travel months for East Asia, North Africa, and the Middle East, respectively (Table 4).
|Region||Estimated no. of unprotected travelers||Cases||Cases per 100,000 unprotected travelers||95% CI||100,000 person months||Cases per 100,000 person months||95% CI||VFR%|
|Middle East (Iran)||129,690||103||79||59.4–106.1||5.9||18.0||13.1–23.4||91.0|
The main advantage with our study is the high sensitivity of the Swedish notification system in combination with a reliable travel denominator from a large ongoing national survey on travel patterns (TDB). The high accuracy of the data from the TDB survey has been shown in a previous article.9 There is a possible selection bias of not so frequent travelers being more at home and nonmobile telephone owners only. This might slightly overestimate the incidence of travel-related hepatitis A. The Swedish system on notification of communicable diseases is very effective,8 and the vast majority of clinical cases of hepatitis A were therefore notified during the study period. A big proportion of unknown source of infection in notified cases in 1997 coincides with the introduction of a new electronic surveillance system at SMI as well as a large domestic outbreak within a community of intravenous drug users. This might have underestimated the travel-related hepatitis cases in that year leading to a slight underestimation of incidence. Another contribution to a possible underestimation of true incidence is the fact that subclinical cases are not diagnosed and thus not reported. However, this would be the same for all comparative studies.
The airport-based vaccination survey gives a fair estimate of the current hepatitis A immunization rate in 2007, but unfortunately, we lack knowledge about this during the years of the study period when the vaccination coverage probably was lower. The fact that some travelers did not want to answer the questionnaire might indicate that these persons did not want to report that they were unprotected, thus leading to a possible overestimation of immunization rate. Furthermore, the immunization figures are based on personal information without supporting serological data.
According to our study, the proportion of travel-related hepatitis A seems to be bigger in Sweden than in other countries,2,6,7,10 which we believe reflects the high sensitivity of the national reporting system. Compared with the study by Mutsch and colleagues2 where the overall incidence was 3 to 11 and the incidence in nonimmune travelers was 6 to 28/100,000 person months abroad, our findings indicate that the incidence in Swedish travelers is lower. This might be explained both by a more reliable denominator but also by a higher immunization rate in Swedish travelers. The hepatitis A immunization rate in 588 Swedish travelers in 2002 was 40%,4 which is already a high percentage compared to international figures3,11,12 and perhaps a more adequate figure when relating to our study period. The most striking finding is that the risk for hepatitis A in a non-VFR tourist in Asia is low, even for an unprotected tourist. Unfortunately, we do not have specific demographic analysis of the region except for the years 2000 to 2005 when specific figures for Thailand alone are presented and reveal that the majority of the travels go there. Thailand has gone through an epidemiological shift in HAV seroprevalence during the past years,13 and our figures confirm the low risk also in travelers. Another contributing factor explaining the low risk to East Asia could be a lower proportion of VFR travelers and a bigger proportion of tourist staying at better hotels. On the other hand, the majority of the travelers in the Middle East region are VFR, which might explain the higher incidence. This region is also heterogeneous and a shift in seroprevalence has been documented in Qatar, United Arab Emirates, Saudi Arabia, and Lebanon.14,15 However, Swedish immigrants would rather originate from Iraq and Iran. Generally, the impact of the HAV seroepidemiological shift in Asia, the Middle East, and Latin America16,17 is important and must be addressed in pretravel counseling to immigrants regardless of age and whom were previously considered to be naturally immune.
The fact that tourists to North Africa are vaccinated against hepatitis A to less extent than tourists to other high-risk destinations has been highlighted before.4,18 Popular tourist resorts of Egypt and Morocco are probably perceived as having the same risk profile as the rest of the countries around the Mediterranean. HAV vaccination is strongly recommended to this group.18
The most important finding is that VFR travelers, and especially the children, accounted for a majority of the travel-related hepatitis A episodes diagnosed. This is also consistent with earlier findings,2,19–22 and a recent study showing that this group tend to travel more to HAV high endemic countries and stay longer than non-VFR travelers.23 Infected children constitute a risk to their playmates as well as to nonimmune household members. Secondary cases of older age can be critically ill when infected with HAV.10,24,25 We conclude that the group of VFR travelers, and especially the youngest travelers, constitute a high-risk group for hepatitis A. On the other hand, the risk for a non-VFR tourist to Asia is low. These results can probably be generalized to other European countries. Information of the need for hepatitis A vaccine should be better targeted to the main risk group of young VFR travelers.
We would like to thank the passengers and Airport Duty officers at Landvetter and Arlanda airports in Sweden. This study was funded in part by the SMI and by the Centre for Clinical Research, Sormland County Council, Eskilstuna, Sweden.
Declaration of Interests
The authors state that they have no conflicts of interest.