Vous êtes sur la page 1sur 6

Policy Forum

The Transit Phase of Migration: Circulation of Malaria

and Its Multidrug-Resistant Forms in Africa
Caroline Lynch, Cally Roper*
London School of Hygiene and Tropical Medicine, London, United Kingdom

This is one article in a six-part where more than 40% of the population Migration, Malaria, and Drug
PLoS Medicine series on Migration have parasites in their blood [5]; the bulk Resistance
& Health. of these infections are asymptomatic.
Transit between different areas exposes International transit of people with
Transit Migration people to changing risks of malaria malaria played a significant role in the
infection, and the burden of infection global dispersal of resistance to both
There are many different types of often falls disproportionately on mobile chloroquine and sulphadoxine-pyrimeth-
migrants and types of movements and no and migrant sectors of the community [6]. amine (also known as SP or Fansidar), two
single commonly agreed definition for Migrants travelling from low to high drugs that were the mainstay of malaria
transit migration. Papadopoulou [1] de- transmission areas are at greater risk of treatment in Africa for 30 years. In both
scribes it as ‘‘the stage between emigration acquiring a malarial infection than those cases, resistance mutations arose in South-
and settlement’’, while in contrast, the travelling in the opposite direction, where east Asia and were subsequently imported
assembly of Inter-Parliamentary Union in also having no acquired immunity means into Africa during the 1970s for chloro-
Geneva [2] states that ‘‘transit migrants they are much more vulnerable to disease. quine, and during the 1980s and 1990s for
are…aliens who stay in the country for pyrimethamine and sulphadoxine, respec-
Evidence from large-scale population re-
some period of time while seeking to tively [11,12]. While there was a 17-year
settlement programs in Ethiopia [7],
migrate permanently to another country’’. lag between the appearance of chloro-
Indonesia, and Brazil [8] show sharp
Although transit is often defined in quine resistance in Southeast Asia and its
increases in malaria morbidity and mor-
terms of international borders, it is highly introduction to Africa, once established in
tality across all age groups in migrants
likely that transit migrants undertake Africa the dispersal of resistance has been
from low to high transmission areas.
similar patterns of movement within their shown to follow the predictable process of
As such, migration has enormous incremental diffusion [12].
countries of origin, and within the transit
significance for patterns of malaria infec- In 2010, Tatem et al. [13] used
and settlement countries. Within a country
tion and disease, and for malaria control. permanent migration data derived from
people can be moving permanently or
When large groups of people move from national census statistics and measures of
multiple times towards urban areas and
high to low transmission areas, the malaria endemicity to describe communi-
back home, especially if they are internally
immediate result, as measurable by par- ties of malaria-endemic African countries
displaced people (IDPs). In practice, tran-
asite prevalence, would be an overall linked by higher levels of infection move-
sitory migration is one component within
a broad framework of movement encom- increase in transmission [9]. But more ment (see Figure 1A). Their analysis
passing permanent, temporary, and circu- common than unidirectional permanent highlights ‘‘natural’’ migration regions
latory migration, all of which can occur in migrations are the regular and cyclical where high levels of malaria infection are
a bidirectional or in a stepwise sequence of movements of migrants or return mi- interchanged. A comparison of malaria
moves [3,4]. grants to areas of low transmission. In migration data (Figure 1A) with the
This article examines types of transit either case, a migrant infected with geographic distribution of parasite drug
migration and their significance for the malaria can serve as a reservoir and seed resistance lineages (Figure 1B) shows there
prevention and treatment of malaria in localised outbreaks or epidemics in those is broad correspondence between migra-
Africa, using a case study of migration in areas, and thus migrants become ‘‘active tion and resistance patterns.
southwest Uganda. transmitters’’ of infection in low transmis- The description of the geographic
sion areas [10]. patterns of drug resistance dispersal comes
People Movement and Malaria
Citation: Lynch C, Roper C (2011) The Transit Phase of Migration: Circulation of Malaria and Its Multidrug-
Plasmodium falciparum malaria—responsi- Resistant Forms in Africa. PLoS Med 8(5): e1001040. doi:10.1371/journal.pmed.1001040
ble for the most severe malaria—occurs Published May 31, 2011
across Africa, although the intensity of
Copyright: ! 2011 Lynch, Roper. This is an open-access article distributed under the terms of the Creative
transmission varies considerably. Half of Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium,
Africans live in areas of high endemicity provided the original author and source are credited.
Funding: No specific funding was received for writing this article.

The Policy Forum allows health policy makers Competing Interests: The authors have declared that no competing interests exist.
around the world to discuss challenges and Abbreviations: IDP, internally displaced people
opportunities for improving health care in their
societies. * E-mail: Cally.Roper@lshtm.ac.uk
Provenance: Commissioned; externally peer reviewed.

PLoS Medicine | www.plosmedicine.org 1 May 2011 | Volume 8 | Issue 5 | e1001040

Summary Points migration between countries support the
view that parasite circulation through
N Movement of people is the means by which human pathogens are dispersed,
regionally distinct migration streams was
central to the spread of drug resistance
so providing health care to mobile sectors of the community is vital to disease
control interventions. mutations. However, permanent migra-
tion reported in census data as ‘‘place of
N Using malaria as a case study, our article examines types of migrant transit and birth’’ captures only a fraction of the
their significance for prevention and treatment of the disease.
migration and transit taking place. Fur-
N Asymptomatic untreated infections act as reservoirs for malaria transmission. thermore, having been collated at the
Malaria control programmes need to identify those migrant streams with national level, these data lack the spatial
potential to transport malaria and to target prevention and treatment measures resolution to examine migration and
appropriately. mobility within countries. Patterns of
N Transit has also played a role in the dispersal of antimalarial drug resistance, human circulation change according to
with the international transportation of artemisinin-resistant parasites by the dictates of war, trade, and transport
human migration being the greatest threat to the antimalarial treatments used infrastructure. These maps suggest the
in Africa today. existence of networks of high volume
N A geographic framework of human migration at local, national, and migration, whose boundaries are unlikely
international levels is needed so the potential speed and direction of pathogen to be national borders.
dispersal can be predicted, and for health policy and services to respond
appropriately to the needs of migrants and to threats of pandemic.
Case Study: Complex Migration
Patterns and Drug-Resistant
from an analysis [14] that described five 2000–2007. The degree to which parasite
major lineages of sulphadoxine resistance populations share resistance lineages is an Malaria in Southwest Uganda
at the dhps gene in P. falciparum. These indication of the extent of parasite mixing There has emerged a highly resistant
emerged during the 1990s when the SP during the 10–15 years they have been in form of P. falciparum in the Kabale and
drug was widely used for treatment of circulation. The major regional distinc- Rukungiri districts of southwest Uganda
malaria. Each lineage occurred via a single tions between parasite populations on the [15] and in bordering areas of Rwanda
mutation event, and has a distinctive African mainland are between populations [16]. The form carries a mutation in the
geographical distribution that reflects cir- in the east and west, and between the dhfr gene, which confers high level resis-
culation and dispersal from their site of northern and southern populations in the tance to a number of antifolate drugs in
initial emergence [14]. The map in east and the west. general use (including sulfadoxine-pyri-
Figure 1b is a snapshot of the distribution The similarities between the geograph- methamine and chlorproguanil-dapsone).
of these resistance lineages among 20 ical pattern of resistance mutation dispers- To predict the likely path of dispersal of
African populations sampled during al in Africa and the levels of permanent this resistant form of the parasite, we

Figure 1. Regions of malaria parasite exchange strongly resemble the distribution of resistance allele lineages. (A) P. falciparum
migration communities for Africa. The map shows communities connected by comparatively higher levels of malaria migration, with community
membership shown by colour (from Tatem and Smith [13]). (B) The distribution of resistance mutant lineages among 20 African P falciparum
populations (reprinted from Pearce et al. [14]). Most resistance mutations in the dhps gene belong to five major lineages, (indicated by the colours
shown in the key) and each is derived from a single emergence event. Those which do not belong to the major five are shown in grey.

PLoS Medicine | www.plosmedicine.org 2 May 2011 | Volume 8 | Issue 5 | e1001040

examined the types and frequency of as well as erosion and soil degradation population mobility, and by the difficulties
population movements in Uganda. The [28], which has led to out-migration from of access to and adequate health care
region of resistance emergence is shown in the region [29]. More recently, improved provision for mobile sectors of the popu-
red in Figure 2. roads and increased vehicle numbers, as lation [3]. A geographical framework of
In 2005, Uganda hosted an international well as market liberalisation, has led to a malaria dispersal is needed for contempo-
migrant pool of 2.25% of its total popula- further increase in temporary mobility rary national planning and regional coor-
tion, with a reported net migration of 5,000 between rural and urban areas [22]. dination of malaria control measures
people [17]. Internal migration is much Another result of high population den- directed towards elimination.
higher: nationally, 19.7% of the population sity has been migration and resettlement, To build such a framework, there is an
of Uganda report to have lived in another both programmed and spontaneous, into urgent need for better data on circulation
village, town, or district for more than 6 neighbouring areas. The first official within countries and short-term circula-
months at any one time in the last 5 years resettlement schemes took place in the tion of populations around porous bor-
[18]. And sub-national or district-level late 1940s, and involved the transfer of ders. Countries that have multiple parasite
studies have shown that only 18% of heads approximately 15,000 people from South populations, such as the Democratic
of households were born in their current Kigezi (now Kabale) into North Kigezi Republic of the Congo, require detailed
areas of residence, and this proportion is (now Rukungiri) [30]. Further resettlement surveys of travel behavior across their
higher in rural areas (29%) [19]. schemes were undertaken to relocate territory, while countries with large pop-
Uganda also has a long history of labour from Kabale and Rukungiri to ulations of IDPs such as Uganda and
hosting refugees from neighbouring coun- Bunyoro in order to clear 500 square miles Sudan (including Darfur) are also key.
tries. In 2008, the country was home to of bush as part of a tsetse barrier scheme Data on permanent internal migration
nearly 250,000 refugees from Sudan, the [31]. can be used to describe migration routes
Democratic Republic of the Congo, and While these schemes ended in the and possible migration streams but cannot
Rwanda in camps containing upwards of 1960s, migration between Kabale, Rukun- inform as to the frequency of travel
3,000 people each [20], and an additional giri, and the Bunyoro region continued between different places. Permanent mi-
915,000 IDPs still remain after being and the original scheme resulted in a gration data from census are informative,
displaced during the ongoing conflict in strong migration route opening between but the national scale of the data is not
northern Uganda [21]. The location of these areas [32]. Around the same time sufficiently high resolution to begin to
refugee and IDP camps is illustrated by (mid to late 1950s) in Kabale, people had disentangle the movement of disease
shaded regions in Figure 2. begun to encroach into the Bwindi forests within national borders. In particular,
The interregional migrant streams in in the district [33]. Migration data, based permanent migration data will not allow
Figure 2 illustrate the central influence of on birthplace information taken during us to capture information on movements
the main city of Kampala. Urbanization is a the 1969 population census in Uganda, between countries with leaky borders, or
major component of permanent migration, clearly showed a large migration stream illegal movement between countries. Some
although it has progressed more slowly in from southwestern districts (Kigezi) into types of circulation in these situations are
East Africa than in other parts of the world. the western Toro and Ankole regions as not captured in surveys because people do
In 2002, just over 12% of the total Ugandan well as into Kampala and its surrounding not wish to disclose travel across border
population was estimated to live in an urban areas. areas. One example is when IDPs may
area [18]. Migration for labour is now These large northward migrant streams make a series of return visits to their areas
accepted as part of a household strategy to from the affected area of southwest Uganda of origin for planting and harvesting but
improve traditional livelihood [19,22] and indicate that the highly drug-resistant report being present in IDP camps in
has been relatively well-documented in malaria parasites have the means to spread order to fully avail themselves of food
Uganda. Migrant labour is characterised quickly within Southern Uganda. The vouchers.
by circulatory rather than permanent maps in Figures 1 and 2 indicate that the Mapping the dispersal of drug resistance
movements, with migrants returning to future dispersal of resistance has the mutations has potential utility in defining
their areas of origin at the end of a formal potential to extend rapidly throughout east regions that encompass significant popu-
contract or wage-paying period [23]. This and southeast Africa provided that selection lation mobility and quantifying population
migration system was disrupted with the in the form of continuing use of SP for connectivity. Such measures summarise
economic collapse in the 1970s and 1980s, treatment is maintained. The relative the circulation of parasites through all
with reports of reverse migration from importance of different types of travel for types of migration and have value in this
major urban areas, such as Kampala, back the dispersal of malaria is difficult to respect, but they are not informative about
to rural areas [24]. This reverse flow quantify individually, but it is clear from which migrant sectors of the population
continued until the mid-1980s [25], when patterns of dispersal of drug resistance that should be prioritized for malaria preven-
urbanisation began to slowly increase again. the most significant migration streams carry tion and treatment intervention.
Apart from labour migration, the huge volumes of traffic and that these
amount of circulatory migration associated should be incorporated into thinking about Need for Migration Data in the
with other activities is difficult to establish. malaria control. Containment of Artemisinin
Return migration appears to be frequent
with an estimated 10% of urban dwellers Resistance
Geographical Framework of
making return visits to their home areas of As we have noted, migration has
Migration for Malaria
origin annually [26,27]. historically played a pivotal role in the
global dissemination of drug resistance.
Migration in Southwest Uganda Programmes for control and eradication Therefore, the recent confirmation of
The southwest of Uganda in particular of malaria in Africa during the 1950s were resistance to artemisinins on the Cambo-
has always suffered from scarcity of land, undermined by failure to take account of dia–Thailand border [34] has triggered a

PLoS Medicine | www.plosmedicine.org 3 May 2011 | Volume 8 | Issue 5 | e1001040

Figure 2. The complex patterns of migration flow in Uganda. The area of highly drug-resistant malaria is shown in red, major population
migration flows are shown by arrows, major refugee camps are hatched, and the major urban population in Kampala is indicated in pink. The
permanent migration streams are estimated from birthplace data from 1969 population census from Goddard [37].

PLoS Medicine | www.plosmedicine.org 4 May 2011 | Volume 8 | Issue 5 | e1001040

concerted effort to contain it. In the policy respond appropriately to the needs of misinin resistance and African destina-
document ‘‘Global Planning Artemisinin migrants and at the same time sustain tions.
Resistance Containment’’ [35], the World the gains made in malaria control initia- 6.
2. Establish what the rate of malaria
Health Organization identifies three tiers tives. For example, we need to: infection among such travelers is and
of risk, and implies that migration data are what proportion will travel to areas
central to their definition. Tier I areas 1. Identify networks of high volume
where transmission can potentially
have credible evidence of artemisinin transit and migration within Africa.
This should include non-permanent occur. Once the risks of importation
resistance; Tier II areas have significant are quantified ‘‘a speedy, scientifically
inflows of people from Tier I; and Tier III and cyclical transit both within and
between countries. sound, and coordinated response from
areas have no evidence of artemisinin affected countries, donors, and inter-
resistance and limited contact with Tier I. 2. Apply malaria control and elimination
national organisations’’ is needed [36].
Although most of Africa can be com- interventions across areas that encom-
fortably classified as Tier III, the risk pass significant volumes of migration. 7.
3. When artemisinin resistance eventually
posed by incoming migrants cannot be emerges in Africa, apply migration
3. Identify mobile communities and mo-
assessed without knowing the number and frameworks to guide strategies for the
bile sectors of the population who need
final destination of travelers, together with containment or management of resis-
to be prioritized in the provision of
their likelihood of carrying a malaria malaria treatment and prevention mea- tance through use of alternative treat-
infection. There is currently no policy for sures. ment combinations.
screening travelers to Africa and no 4. Provide targeted health care to these Molecular markers for resistance to
interventions to prevent the importation communities. drugs of the past provide insight into the
of artemisinin-resistant malaria into
Africa. Gathering data on malaria infec- drivers of resistance dispersal. If the rules
Transit at the inter-continental, nation-
tion rates among incoming travelers governing the spread of resistance can be
al, and local level have all played a role in
should be a priority for African national elucidated now, then policy interventions
the dispersal of antimalarial drug resis-
malaria control programmes. to protect the efficacy of artesimin combi-
tance in the past. Arguably, the interna-
nation therapies can be designed proac-
tional transportation of artemisinin-resis-
tively rather than reactively and the threat
Policy Needs and tant parasites by human migration is the
greatest threat to the antimalarial treat- of resistance to new treatments managed
Recommendations with foresight.
ments used in Africa today. There are
Clearly the current data on transit in currently no policy measures in place to
Africa are limited, and more migration prevent the importation of artemisinin- Author Contributions
data are needed. A geographic framework resistant parasites to Africa. As a result, we Wrote the first draft: CR CL. Wrote the paper:
of human migration at local, national, and also need to: CR and CL contributed equally. ICMJE
international levels is particularly essential. criteria for authorship read and met: CR CL.
It is only through the establishment of such 5.
1. Gather data on the volume of migra- Agree with the manuscript’s results and conclu-
a framework that policy and services can tion between areas of confirmed arte- sions: CR CL.

1. Papadopoulou A (2005) Exploring the asylum- 12. Naidoo I, Roper C (2010) Following the path of 19. Bryceson DF, Maunder DAC, Mbara TC,
migration nexus: A case study of transit migrants most resistance: dhps K540E dispersal in African Kibombo R, Davis ASC, et al. (2003) Sustainable
in Europe. Geneva: Global Commission on Plasmodium falciparum. Trends Parasitol 26: livelihoods, mobility and access needs. Crowthorne,
International Migration. 447–456. UK: Transport Research Laboratory.
2. IPU (2005) Migration and Development. Geneva: 13. Tatem AJ, Smith DL (2010) International popu- 20. UNHCR (2009) United Nations High Commis-
Inter-Parliamentary Union (IPU). lation movements and regional Plasmodium sioner for Refugees Global Appeal Report 2008–
3. Prothero RM (1977) Disease and mobility: a falciparum malaria elimination strategies. Pro- 2009. http://www.unhcr.org/474ac8d00.html.
neglected factor in epidemiology. Int J Epidemiol ceedings of the National Academy of Sciences Kampala.
6: 259–267. 107: 12222–12227. 21. IDMC (2010) Internal Displacement Monitoring
4. Zimmerman C, Kiss L, Hossain M (2011) 14. Pearce RJ, Pota H, Evehe MS, Ba el H, Mombo- Centre-Country Profile: Uganda. Internal Displa-
Migration and health: A framework for 21st Ngoma G, et al. (2009) Multiple origins and cement Monitoring Centre. http://www.internal-
century policy-making. PLoS Medicine In press. regional dispersal of resistant dhps in African displacement.org/idmc/website/countries.nsf/
5. Hay SI, Guerra CA, Gething PW, Patil AP, Plasmodium falciparum malaria. PLoS Med 6: (httpEnvelopes)/2439C2AC21E16365C125719C
Tatem AJ, et al. (2009) A world malaria map: e1000055. 004177C7?OpenDocument.
15. Lynch C, Pearce R, Pota H, Cox J, Abeku TA, 22. Deshingkar P, Grimm S (2004) Voluntary
Plasmodium falciparum endemicity in 2007.
et al. (2008) Emergence of a dhfr mutation Internal Migration: An Update. Paper commis-
PLoS Medicine 6: e1000048.
conferring high-level drug resistance in Plasmo- sioned jointly by Urban Rural Change Team and
6. Najera JA (1996) Malaria Control among Refu-
dium falciparum populations from southwest the Migration Team DFID Overseas Develop-
gees and Displaced persons. Geneva.
Uganda. J Infect Dis 197: 1598–1604. ment Institute: London.
7. Prothero RM (2001) Migration and malaria risk. 16. Corine K, Imwong M, Fanello CI, Stepniewska K, 23. Gould WTS (1995) Regional Labour Migration
Health, Risk & Society 3: 19–38. Uwimana A, et al. (2009) Molecular correlates of Systems in East Africa: Continuity and Change.
8. Sawyer D (1993) Economic and social conse- high level antifolate resistance in Rwandan children The Cambridge Survey of World Migration. 183 p.
quences of malaria in new colonization projects in with Plasmodium falciparum malaria. Antimicrob 24. Potts D (1997) Urban lives: Adopting new
Brazil. Social Science and Medicine 37: Agents Chemother: AAC.00498-00409. strategies and adapting rural links. In: Rakodi C,
1131–1131. 17. Parsons C, Skeldon R, Walmsley T, Winters L ed. The urban challenge in Africa: Growth and
9. Kazmi JH, Pandit K (2001) Disease and disloca- (2007) A database of bilateral migrant stocks. management of its large cities. Tokyo- New York
tion: the impact of refugee movements on the Available at http://www.migrationdrc.org/re- - Paris: United Nations University Press.
geography of malaria in NWFP, Pakistan. Social search/typesofmigration/global_migrant_ori- 25. Jamal V, Weeks J (1993) Africa misunderstood:
Science & Medicine 52: 1043–1055. gin_database.html. Or whatever happened to the rural-urban gap?
10. Prothero RM (1987) Populations on the Move. 18. UBOS (2007) Uganda Demographic and Health London: Macmillan.
Third World Quarterly 9: 1282–1310. Survey 2006. Kampala, Uganda and Calverton, 26. Oucho JO (1996) Urban migrants and rural
11. Anderson TJ, Roper C (2005) The origins and Maryland, USA. Kampala, Uganda: Uganda development in Kenya Nairobi: University Press.
spread of antimalarial drug resistance: lessons for Bureau of Statistics (UBOS) and Macro Interna- 27. Tacoli C (2001) Urbanisation and migration in
policy makers. Acta Trop 94: 269–280. tional Inc. Sub-Saharan Africa: Changing patterns and

PLoS Medicine | www.plosmedicine.org 5 May 2011 | Volume 8 | Issue 5 | e1001040

trends. Mobile Africa: changing patterns of 32. Belshaw DGR (1984) Planning and agrarian 34. Dondorp AM, Nosten F, Poravuth Y, D D, P. P.A,
movement in Africa and beyond. 141 p. change in East Africa: appropriate and inappro- et al. (2009) Artemisinin Resistance in Plasmodium
28. Carswell G (2002) Farmers and Following: priate models for land settlement schemes. falciparum malaria. New England Journal of
Agricultural Change in Kigezi District, Uganda. Understanding Green Revolutions: Agrarian Medicine 361: 455–467.
Geographical Journal. pp 130–140. Change and Development Planning in South 35. WHO (2011) Global Planning for Artemisinin
29. Olson J, Berry L (2003) Land degradation in Asia. 270 p. Resistance Containment (GPARC). Geneva.
Uganda: its extent and impact. available at lada 33. Jelliffe DB, Bennett FJ, Stroud CE, Novotny ME, 36. White NJ (2010) Artemisinin resistance–the clock
virtualcentre org/eims/download asp. Karrach HA, et al. (1961) Field Survey of the is ticking. Lancet 376: 2051–2052.
30. Swanzy H (1951) Quarterly Notes. Afr Aff (Lond) Health of Bachiga Children in the Kayonza 37. Goddard AD, Gould WTS, Masser FI (1975)
50: 89–123. District of Kigezi, Uganda. The American Census data and migration analysis in tropical
31. Swanzy H (1947) Quarterly Notes. Afr Aff (Lond) Journal of Tropical Medicine and Hygiene 10: Africa. Geografiska Annaler Series B Human
46: 181–197. 435. Geography. pp 26–41.

PLoS Medicine | www.plosmedicine.org 6 May 2011 | Volume 8 | Issue 5 | e1001040