1. Introduction

Malaria has been one of the major causes of devastating epidemics in the past. They occur periodically following excessive rains and floods in arid areas or periods of drought in fertile river valleys, compounding the misery often brought about by such meteorological calamities. Malaria epidemics have obstructed efforts to colonize jungle areas or to irrigate dry lands and have plagued displaced populations and refugees. They erupt following the devastation of war and spread with the demobilization of armies, thereby hampering the quick reconstruction of rural life.

The etymology of the word epidemic, from the Greek arrival or stay in a country (similar to the English visitation), from the verb to come or to reside as a foreigner, shows the essential element of this strange phenomenon. The concept of `epidemic' is itself a relative one. In principle, any sudden increase in disease incidence beyond what is considered normal will constitute an epidemic. As Brès (1986) comments, it would be an error to consider as an epidemic a hitherto unrecognized endemic situation or a mere seasonal increase in the incidence of a disease. It would also be an error to neglect the significance of a single case of a new disease in a country, which might well be the prelude to a further dramatic spread of disease.

An essential prerequisite for the occurrence of an epidemic is the existence of a large enough number of susceptible persons who are likely to become clinically ill when suddenly exposed to infection. There is therefore an inverse relationship between endemicity and epidemicity: malaria epidemics cannot affect the populations of highly endemic areas who develop sufficient immunity early in life.

The actual impact of epidemics depends not only on the increase in specific morbidity, but also on the general health of the affected population. Many malaria epidemics coincide with periods of famine, economic crisis, war or civil disturbances, affecting impoverished or displaced populations that are not only physically weak, but often affected by other diseases and unable to obtain appropriate medication. Malaria epidemics do not always cause dramatic emergencies; instead, they more frequently affect economic development. They may flare up suddenly and subside at the end of one season and not return for several years, or build up over several transmission seasons. The early stages of an epidemic may often pass unnoticed, as malaria is still easily treated and antimalarial drugs are widely available almost everywhere.

The presentation and spread of malaria epidemics varies considerably. At one extreme, they may arise in an apparently explosive way, affecting areas covering hundreds of thousands of square kilometres at practically the same time and then subside after a year or two. At the other extreme, they may progress slowly from locality to locality, or valley to valley, taking several years to spread and showing little tendency to subside.

1.1. Historical background

From historical accounts, it may be difficult to ascribe to malaria some of the major fever epidemics of the past with certainty, but during this century, there have been specifically diagnosed malaria epidemics, which clearly show the magnitude of the impact that they can produce:

• the explosive and disastrous epidemic which affected large parts of India in 1908 was officially estimated to have attacked 100 million people and caused about one million deaths;
• in the aftermath of the First World War and in the midst of civil war, the 1922-1923 epidemic in the Soviet Union caused more than 10 million cases and at least 60 000 deaths;
• the Ceylon (Sri Lanka) epidemic of 1934-1935 caused nearly 3 million cases and 82 000 deaths;
• the epidemic in north-eastern Brazil, following the invasion by A. gambiae, caused over 100 000 cases and at least 14 000 deaths in 193 8;
• in 1942, a similar invasion by A. gambiae of Lower Egypt caused some 160 000 cases and more than 12 000 deaths;
• in 1958, an epidemic in Ethiopia caused more than 3 million cases and 150 000 deaths;
• in 1963, in Haiti, hurricane Flora disrupted the ongoing malaria eradication campaign and caused 75 000 cases;
• the epidemic in Sri Lanka in 1968, which occurred while most of the country was in the consolidation phase of the eradication programme, caused 1.5 million cases during 1968-1970, thus revealing the fragility of this particular phase of the campaign;
• the epidemic of 1976-1977 in the Indian subcontinent caused more than 7 million cases;
• that of south-eastern Turkey in 1977-1978 caused some 270 000 cases (Bruce-Chwatt, 1985).

Malaria epidemics are not a phenomenon of the past. Severe epidemics have recently occurred:

• in Afghanistan since the beginning of the civil war in 1979;
• in northern Iraq and southern Turkey, 1993-1995;
• in Tajikistan, 1993-1994;
• in Azerbaijan, 1993-1994;
• in north-western India, particularly Rajasthan, 1995-1996;
• in southern and east Africa (Zimbabwe, Botswana, Mozambique, Swaziland and South Africa), 1996.

Apart from cyclical meteorological phenomena, new factors increasing the risk of severe epidemics include the economic crises and the widespread civil unrest affecting many countries of the so-called third world.

The disastrous epidemic of 1908 in the Punjab gave rise to a sustained programme for the study of malaria epidemiology, with particular emphasis on meteorological and physiographical conditions (Gill, 1928). The postwar epidemics of 1920-1921, which affected many countries of Europe, particularly the USSR, and pushed the limit of malaria transmission beyond the Arctic Circle, spurred governments to support malaria control and to pay special attention to the prevention and control of epidemics. In the Punjab, Gill (1923) was asked to elaborate a system of epidemic forecasting and control, which was progressively developed and adapted to other epidemic-prone areas and which, after the severe epidemic of 1934-1935 in Ceylon, received the full support of the Malaria Commission of the League of Nations (League of Nations, 1938).

The search for methods of epidemic forecasting was nevertheless discontinued as during the 1950s and 1960s, it was believed that malaria could be eradicated. As a result, all the efforts of malariologists were focused on the management of mass campaigns of insecticide spraying or drug distribution, the epidemiology of declining malaria, the detection of residual foci and the prevention or elimination of resurgences of transmission.

The return to a strategy of malaria control in the 1970s occurred at a time of economic crisis, the unprecedented development of transport facilities and the globalization of the economy. These factors created population pressures and migration for economic reasons, which resulted in new and sometimes disastrous focal outbreaks of malaria. At the same time, the reduction or discontinuation of control efforts in many areas produced resurgences of sometimes dramatic proportions, while the periodic epidemic waves returned to the well-defined epidemic-prone areas. The rise in temperature over recent cars may have been an important contributory factor in the occurrence of the numerous local epidemics in highland areas in many parts of the world.

1.2 Major determinants of malaria epidemics

An epidemic is the result of the disturbance of a previously existing equilibrium of the ecological system comprising human, parasite and vector populations in a particular environmental niche. Depending on the resilience of the system, and whether or not the disturbance has changed some of its essential components, it will either return to its previous state of equilibrium after the end of the disturbance, or tend to find a new equilibrium, with or without going through a period of oscillation. The epidemiological history of the area and a study of the current situation will indicate the stability of the possible states of equilibrium which may be reached. In this sense it is useful to distinguish epidemics which are the result of:

• temporary disturbances of a stable hypoendemic equilibrium, such as those resulting from abnormal meteorological conditions; these epidemics, if left to themselves, will return to the previous hypo- or mesoendemic situation;
• major changes in the eco-epidemiological system, shifting towards a new equilibrium of higher endemicity, such as those resulting from major environmental changes, e.g., the introduction of irrigation or the colonization of sparsely populated areas. These changes create the conditions for a higher level of endemicity, so that the epidemics are part of the process of finding a new state of equilibrium. In the absence of intervention, therefore, the higher endemicity will be established and maintained. Environmental modifications for purposes of economic development should not cause increased malaria transmission and, if properly designed, should actually contribute to malaria control. If this has not been the case and an epidemic has occurred, control should be aimed at correcting the defects in the design or implementation of the development project, taking into consideration the sustainability of the controlled situation expected. Somewhat similar situations result from the invasion of an area by an exotic and highly efficient vector which finds a permanently suitable environment, as happened with A. gambiae in Brazil and Egypt. In such cases, it may be possible to eliminate the invading vector before it becomes fully established, as happened with the above-mentioned A. gambiae invasions;
• interruptions of antimalarial measures which have kept malaria under control, but in an unstable equilibrium, in an area with all the epidemiological characteristics of high endemicity. The resulting epidemics are the true resurgences or failures of control, the real disturbance of the epidemiological equilibrium being the introduction of unsustainable control measures. The magnitude and impact of the outbreak will depend on the length of time that the unstable state o controlled transmission was maintained. Left to itself, the original endemic situation will be re-established after one epidemic-type wave or more over a few transmission seasons. Attempts t( reintroduce all the previous control measures should be weighed against the risk of repeated unsustainability, as the impact in terms of human suffering and socioeconomic disturbances of repeated epidemic-type resurgences will generally be greater than that of undisturbed endemic malaria. In most cases, it will be preferable to reduce the peak of transmission so as to arrive at the new endemic situation as painlessly as possible and later to adopt a control strategy based on case management and selective vector control.