Malaria transmission and climate change in Australia

Although endemic malaria was eradicated from Australia by 1981, the vectors remain and transmission from imported cases still occurs. Climate modelling shows that global warming will enlarge the potential range of the main vector, Anopheles farauti sensu stricto; by the year 2030 it could extend along the Queensland coast to Gladstone, 800 km south of its present limit. Vigilance and a dispassionate assessment of risk are needed to meet this challenge.

Introduction

Australia has long been under threat from malaria and this has led many to predict, from early times, that without vigilance there will be outbreaks.¹ Climate change due to the enhanced greenhouse effect is expected to extend vector ranges²and potentially expand the threat. It is timely to review the epidemiology of malaria in Australia and explore its future with climate change.

Although malaria transmission has occurred widely in Australia,³ the disease was eradicated by 1981.⁴ Imported cases (acquired outside Australia) and introduced cases (derived from imported cases), which still occur, do not invalidate this "free from endemic malaria" status; malaria is considered endemic only if parasitaemia arises from introduced cases or other endemic infections.⁵ However, Australia remains vulnerable to malaria transmission, as the vectors have not been eradicated and hundreds of cases are imported annually.

Natural infections have been found in Anopheles hilli and An. farauti⁶sensu lato (s.l.) (a group of three morphologically indistinguishable species,^7,8 found in northern Australia,⁶ which have not yet been given formal names). In southern Australia, An. annulipes s.l. is the putative vector.³An. amictus, An. bancroftii and An. stigmaticus are susceptible to malaria in the laboratory⁶ and may play a role in the field. Nevertheless, experience since the 1900s strongly supports the hypothesis that An. farauti s.l. is the most important vector.

Receptivity to malaria

Ford⁹ delineated the potentially malarious area of Australia, within which "implantation" may occur, as extending south to latitude 19^oS in Queensland, to about 17^oS in the Northern Territory and to about 19^oS in Western Australia, except for a non-receptive coastal strip including Broome at 18^oS. Although transmission had occurred widely outside this area, Ford thought conditions elsewhere did not favour maintenance of transmission. The area north of 19^oS, largely coinciding with the range of An. farauti s.l.,⁶ is now regarded as receptive.³

Australia's receptivity was tested with the return of large numbers of infected servicemen after the world wars: few malaria infections were transmitted outside the receptive zone.³ Even within the receptive zone, local transmission is rare despite imported cases. Possible reasons include:

• Australian mosquitoes may not be susceptible to imported malaria parasites. Not all vectors and malaria strains are compatible; the European vector, An. atroparvus, cannot be infected with an African strain of Plasmodium falciparum.¹⁰ The susceptibility of Australian mosquitoes to parasites from different geographic regions is largely unknown.
• Many cases of falciparum malaria do not produce gametocytes to infect vectors, as over half the notified cases are diagnosed (and presumably treated) within three days of onset of symptoms; ¹¹P. falciparum gametocytes mature about 10 days after the appearance of asexual parasites.⁵ Only 25% of P. falciparum-infected patients from 1989 to 1995 in New South Wales were gametocytaemic at diagnosis (Dr J Walker, Centre for Infectious Diseases and Microbiology, Westmead Hospital, NSW, personal communication).
• Mosquito-to-human density is low. Anopheles mosquitoes usually have a restricted flight range of 1-2 km. The most prolific larval habitats are permanent or semipermanent natural pools, which are uncommon in urban areas.
• A low proportion of blood meals are taken from humans. Anopheles mosquitoes feed at night and in Australia are both reluctant to enter houses and restricted by the widespread use of screens (some introduced cases have occurred in soldiers in camps and prospectors who were not so protected³). Feeding on humans is reduced further by the use of repellents and the presence of alternative hosts; An. farauti s.l. in Darwin fed preferentially from dogs rather than humans.¹²
• Vector longevity is short. In Darwin only 3%-4.3% of An. farauti s.l. mosquitoes, 0.1%-2.6% of An. bancroftii and 0.3% of An. hilli lived long enough to allow sporozoite development.¹³

The future

The worsening malaria situation in some neighbouring countries and increasing travel between them and Australia led the incidence of imported malaria to nearly double in Australia between 1981 and 1991.¹¹ Combined with the possibility of expanded vector distribution due to climate change, this has led to calls for reassessment of Australia's receptiveness to malaria.¹⁴

Climate change is expected to lead to:

• More frequent cyclones and floods, which will increase vector density and the risk of malaria. Past epidemics were often associated with above-average rainfall.³
• Inundation of low-lying areas, due to raised sea levels. Within the Asia-Pacific Region, many such areas are malarious and refugees from them could provide a large reservoir of infection. Emergency relocation of refugees, particularly if aircraft are used, will increase the possibility of introducing exotic vectors into Australia. Drug-resistant malaria parasites will add to the difficulties of treatment.

Changes in vector distribution

As An. farauti sensu stricto (s.s.) (or No. 1 of the three species belonging to An. farauti sensu lato)¹⁵ is an important vector,¹⁶ we investigated its present and future distribution with the CLIMEX climate-matching model¹⁷ and the method of Hutchinson to infer meteorological data.¹⁸ An annual ecoclimatic index was derived for each location to describe its climatic suitability for the persistence and growth of the mosquito population on a scale of 0-100. Parameter values for climate matching were derived from field studies of the distribution and phenology of An. farauti s.s. and temperature-controlled experiments with the mosquito.^16,19,20

The potential distribution of An. farauti s.s. was estimated under a recent climate-change scenario for the year 2030, with increases of 1.5^oC in temperature and 10% in summer rainfall in northern Australia²¹(Figure 3).

The present modelled distribution of An. farauti s.s. is largely coastal, which agrees with its known distribution on the Cape York Peninsula and in the Northern Territory. An. farauti s.l. occurs rarely in northern Western Australia (it has been recorded only twice, at Kununurra¹⁶), but this was not predicted, probably because it requires exceptionally rainy years or irrigation to survive.

Under the climate-change scenario for the year 2030, the potential distribution of An. farauti s.s. extends a further 800 km south in coastal Queensland, to Gladstone (23^o50'S). This would encompass the Whitsunday Passage with its tourist population. However, the ecoclimatic indices from Townsville south are low, indicating a limited vector density and hence a low receptivity for malaria.

Conclusions

Malaria has long been a threat in Australia, but recently the threat has been sensationalised.¹⁴ It is timely to recall Mapleton's advice, given in 1922 in the MJA,²²to cease making "tropical Australia appear a hot-bed of dread diseases". In the last 70 years, our understanding of malaria and our ability to control it have increased. Our study indicates that conditions under climate change will become more suitable for the primary vector. To meet this challenge we require vigilance, an understanding of malaria epidemiology and a dispassionate assessment of the risks.

References

Author details

University of Queensland, Brisbane, QLD.