The low and declining risk of malaria in travellers to Latin America: is there still an indication for chemoprophylaxis?

A comparison was made between local malaria transmission and malaria imported by travellers to identify the utility of national and regional annual parasite index (API) in predicting malaria risk and its value in generating recommendations on malaria prophylaxis for travellers. Regional malaria transmission data was correlated with malaria acquired in Latin America and imported into the USA and nine European countries. Between 2000 and 2004, most countries reported declining malaria transmission. Highest API's in 2003/4 were in Surinam (287.4) Guyana (209.2) and French Guiana (147.4). The major source of travel associated malaria was Honduras, French Guiana, Guatemala, Mexico and Ecuador. During 2004 there were 6.3 million visits from the ten study countries and in 2005, 209 cases of malaria of which 22 (11%) were Plasmodium falciparum. The risk of adverse events are high and the benefit of avoided benign vivax malaria is very low under current policy, which may be causing more harm than benefit.


Background
Many public health bodies base their recommendations for the prevention of malaria in travellers on national surveillance data, which provides information on the inten-sity and risk of malaria in local populations, expressed as the annual parasite index (API), which may reflect regional risks. While this approach appears rational, there is no evidence that patterns of travel-acquired malaria cor-relate with transmission intensity among indigenous populations. Recommendations on prophylaxis for travellers need to balance the threat of malaria, including falciparum malaria, and the risk of a fatal outcome, against the potential toxicity of chemoprophylaxis, a risk which is not relevant to populations living in endemic regions. Rombo called into question the use of API to estimate the risk of travellers acquiring malaria [1]. Highlighting the disparity between native and traveller's vulnerability and exposure to infection he has emphasized the need to consider prophylaxis toxicity when prescribing for travel to low risk malaria regions.
Providing appropriate malaria prophylaxis advice for travellers visiting countries in Central and South America can be complex and challenging, particularly when the journey involves many regions or countries where multiple parasite species are present.
This study was set up to review rates of malaria transmission within Central and South American countries and to compare these with patterns of imported malaria among European and US travellers returned from endemic countries. The aim of the study was to try and identify whether transmission within a country reflects malaria transmission among travellers and to examine the usefulness of API in predicting travellers' risk and its value as a basis for recommendations of malaria prophylaxis.

Methods
The local population risk is based on reports from the Pan American Health Organization (PAHO) Regional Office of the World Health Organization with information on API provided by countries within the region[2,3]. The change in the API over the period 1998 and 2004 has been included to reflect changing trends of malaria transmission. All countries provide regional and district API data, and for this study the highest API in each country is used to represent the maximum risk likely to be faced by travellers. The malaria risk associated with travel is identified through reports from National malaria surveillance bodies, describing malaria imported in returned travellers from Central and South America (Table 1) between the  years 2000 and 2005 (data for 2005 was not available for France) from nine European countries and the USA ( Table  1). Most of this data was provided through and from members of TropNetEurop, a network of clinical sites, which have access to national malaria surveillance reports, and from the literature. Most reports do not contain details on region of travel within countries, and where several countries have been visited, reports do not necessarily reflect the country of acquisition. Cases provided by TropNetEurop were not included in the country malaria analysis as they would duplicate case reports.
Malaria imported from Mexico was analysed separately from the Central America region.
The volume of travel will have a significant bearing on the number of cases of imported malaria and therefore rates, where possible, were calculated. The World Tourism Organization collects data on international arrivals and this data provides an estimate of the number of tourist departures and arrivals by country. This data was used where no national statistics were available to estimate the numbers of visits made from the study countries [4,5] ( Figure 1). Malaria cases recorded in UK travellers were analysed using a denominator, the number of visits made by UK citizens to the countries of malaria acquisition. Data provided by the International Passenger Survey (IPS) is collected through face-to-face interviews of passengers at all major ports within the UK. A quarter of a million passengers are interviewed throughout the year and this sample provides an estimate of the total annual visits to each country, the duration of stay and reason for travel. Malaria cases occurring in United States travellers were extracted from data published by the Centers for Disease Control [6][7][8][9][10][11] and visits made by US citizens to the region were collected by the USA International Air Travel Statistics (or I-92) programme. This provides data on outbound numbers of US citizens travelling, using point-topoint air traffic totals from the USA, on departing flights. Visits to, and malaria from Mexico were analysed separately due to the large travelling population from the USA [12]. Malaria data from France was provided through  Total visits made by international tourists to the study coun-tries adjusted to visits from the reporting countries Figure 1 Total visits made by international tourists to the study countries adjusted to visits from the reporting countries.

Millions
South America All reporting countries 3.4 million All intl. tourist arrivals 12.6 million Central America All reporting countries 1.9 million All intl. tourist arrivals 5.4 million and Brazil were similar, around one case for every 3,000 years exposed, and similar to the risks of UK residents visiting India [17]. Honduras and Guatemala were the highest risk countries with one case for every 103 and 513 years exposed respectively.
Low malaria rates were also noted in UK, French and Dutch visitors to Venezuela, Colombia, Peru and Brazil during 2004. These four countries received a total 2.6 million visitors from the study countries ( Figure 1) and had a maximum incidence of malaria of 2.2 (UK travellers to Colombia, 2004) per 10,000 visits.

Discussion
This study was designed to identify whether local transmission of malaria within countries of Latin America reflected the pattern and trends of malaria acquired by travellers from ten developed countries. During 2004, twenty one PAHO countries with active malaria programs examined 6.7 million slides, of which 13% were positive for malaria, three quarters of them were speciated as P. vivax infections in an "at risk" population of 262 million [3]. The detailed information collected in Central and South American countries and presented by PAHO [3,2,13] provides evidence of a clear trend of declining transmission across most of the countries, most notably in Brazil, which reported a 56% decrease in the high incidence regions, attributed to malaria control programmes initiated in 2000. The total number of tourists visiting Latin America is not known precisely, but the World Tourism Organization [4] estimates there were 16 million international tourist arrivals to South America, with a 16% increase from 2003. Central America, during 2004, received 5.7 million inbound visitors, a 17% growth in arrivals over the previous year. The main country sources of imported malaria were Honduras, French Guiana, Guatemala, Mexico and Ecuador, from where there were 1,066 imported cases over five years, accounting for 64% of all imported cases from Latin America, 75% were non-falciparum malaria. There are a number of important limitations that need to be understood when reflecting on the findings. Local transmission reported to PAHO may be inconsistent and regions not reporting or not diagnosing cases may be interpreted as no malaria transmission. The imported malaria cases collected nationally use different reporting methods and are of varied quality. The denominators used in the analysis are again of different capture methods. The USA, France and the Netherlands record the number of citizens departing to a destination while in the UK samples of departing passengers are interviewed, capturing destination, duration of travel and reason for travel. The pattern of travel through the regions by western trav- ellers is not recorded in the denominator data or through the malaria case reports, and therefore the proportion visiting high transmission regions are unknown.
Although a number of regions within Peru and Brazil have an API above 50/1000 cases/year the actual numbers of malaria cases in returning travellers is low, a total of 145 cases over five years, and in 2005, there were only 30 cases of which three were P. falciparum. Asymptomatic carriage in natives living in the Peruvian Amazon near Iquitos is estimated to be less than 10% and the entomological inoculation rate for the Amazonas region reported as 10-20 annually [18]. The small numbers of travel associated cases from Peru are unlikely to be a result of widespread use of chemoprophylaxis. Currie and colleagues [19] examined prophylaxis use in tourists departing Lima, Peru. Of the 1226 travellers interviewed, 43% were from the USA. Nearly three quarters had visited only Peru and 54% had visited a malarious region (as defined by CDC). Of these around half had taken regular chemoprophylaxis (42% atovaquone/proguanil). During that year (2003) there were 10 (six P vivax) imported malaria cases from Peru. The highest numbers of imported malaria cases, over the 5-year study period, were of P. vivax from Honduras, Guatemala, Ecuador and French Guiana. Despite an increase in local transmission in Honduras, total travel associated cases declined by 20% suggesting that there is no correlation between the two trends. The rates in US and UK travellers to the whole region (excluding Mexico) reveal a similar incidence of 0. Current chemoprophylaxis policies recommend prophylaxis for high risk regions [20][21][22], but many of these regions (as shown in Table 2 have a declining risk for indigenous populations). The inconsistency between focal high transmission areas in countries popular with western travellers and small numbers of travel associated malaria is worth exploring. Significant numbers of travellers may not be using prophylaxis during their travel and the departure lounge suggests approximately 50% of visitors will be using chemoprophylaxis. Other countries visited by significant numbers of tourists as reported by WTO in 2005 -Peru, (1.5 million) Brazil (5.4 million), Guatemala (1.3 million) had small numbers of cases and low rates of malaria. Although these countries have areas of high transmission, the major parts of these countries have no malaria transmission. It would appear that most visitors to these countries are at low or no risk of acquiring infection, whatever their journey and destination within the country.
Protection against P. vivax, disease despite using the most widely available regimens is marginal [23][24][25], as only the primary attack [23] is aborted. Most clinical episodes develop some months after infection when travellers have returned home and are unlikely to be missed through routine reporting systems. Severe adverse events leading to stopping medication during chemoprophylactic drug use were reported in 3-8% of users whilst mild to moderate adverse events were reported by 32%-45% of users [26].
In the 423,416 visitors from reporting countries to Peru in 2003 [4] approximately 25% (105,000 or 50% of those visiting a malarious region) visitors were using chemoprophylaxis as identified by the airport departure lounge study [19]. During that year, 10 (two P. falciparum) cases of malaria were reported in nine study countries after visiting Peru. Using the minimal proportion of users encountering adverse events from the popular prophylaxis regimens [26] an estimated 34,000 travellers would have suffered an adverse event related to chemoprophylaxis use. The risk of adverse events for visitors to Peru and other regions are likely to be significantly higher than avoided infections particularly of benign P. vivax malaria under current policy recommendations. Unless chemoprophylaxis prescribing is significantly reduced, current recommendations are likely to be causing more harm than benefit.

Policy change
Despite its limitations, this study suggests that the risk of adverse events from chemoprophylaxis is likely to be significantly higher than the risk of acquiring malaria in the most popular tourist destinations in Central and South America. Although current national and international policy focuses on chemoprophylaxis for focal, highly endemic malaria transmission regions in countries which have overall low API's, this strategy appears to provide limited benefit as travellers appear to have a low malaria attack rate and will acquire P. vivax rather than P. falciparum infection. The benefit of chemoprophylaxis in preventing the former is unclear. An alternate strategy adopted by a number of European countries, for example Switzerland [27], is to provide travellers with emergency standby treatment in case of malaria symptoms during travel. This has the benefit of dealing with a life threatening attack of falciparum malaria, but avoiding adverse events associated with excessive chemoprophylaxis. It has the disadvantage of cost, as all travellers will have to purchase therapy. Two of the highest risk countries reported by PAHO -French Guiana, and Surinam, correlated to countries where visitors were at high risk of malaria and chemoprophylaxis would be appropriate for travel to risk areas in these countries. There appears to be no clear benefit and significant potential for toxicity in recommending chemoprophylaxis for visitors to Mexico, where the highest API is less than 0.07 for local residents and 20 imported cases annually. Despite the low or falling risk of malaria, the continued use of bite prevention measures remains important as these are effective, safe and have the added benefit of reducing other vector borne diseases.

Authors' contributions
RHB and BC designed the study, collated the data and prepared the first draft.
JB, OB, UH, CH, TJ, FL, NM, BM, HS AND LV obtained and analysed national data. All authors contributed to the interpretation of the data and agreed the final draft.