A print friendly PDF version is available from this Communicable Diseases Intelligence issue's table of contents.

A report on behalf of the National Arbovirus Advisory Committee of the Communicable Diseases Network Australia

Abstract

Mechanisms for monitoring Murray Valley encephalitis (MVE) virus activity include surveillance of human cases, surveillance for activity in sentinel animals, monitoring of mosquito vectors and monitoring of weather conditions. The monitoring of human cases is only one possible trigger for public health action and the additional surveillance systems are used in concert to signal the risk of human disease, often before the appearance of human cases. Mosquito vector surveillance includes mosquito trapping for speciation and enumeration of mosquitoes to monitor population sizes and relative composition. Virus isolation from mosquitoes can also be undertaken. Monitoring of weather conditions and vector surveillance determines whether there is a potential for MVE activity to occur. Virus isolation from trapped mosquitoes is necessary to define whether MVE is actually present, but is difficult to deliver in a timely fashion in some jurisdictions. Monitoring of sentinel animals indicates whether MVE transmission to vertebrates is actually occurring. Meteorological surveillance can assist in the prediction of potential MVE virus activity by signaling conditions that have been associated with outbreaks of Murray Valley encephalitis in humans in the past. Predictive models of MVE virus activity for south-eastern Australia have been developed, but due to the infrequency of outbreaks, are yet to be demonstrated as useful for the forecasting of major outbreaks. Surveillance mechanisms vary across the jurisdictions. Surveillance of human disease occurs in all States and Territories by reporting of cases to health authorities. Sentinel flocks of chickens are maintained in 4 jurisdictions (Western Australia, the Northern Territory, Victoria and New South Wales) with collaborations between Western Australia and the Northern Territory. Mosquito monitoring complements the surveillance of sentinel animals in these jurisdictions. In addition, other mosquito monitoring programs exist in other States (including South Australia and Queensland). Public health control measures may include advice to the general public and mosquito management programs to reduce the numbers of both mosquito larvae and adult vectors. Strategic plans for public health action in the event of MVE virus activity are currently developed or being developed in New South Wales, the Northern Territory, South Australia, Western Australia and Victoria. A southern tri-State agreement exists between health departments of New South Wales, Victoria and South Australia and the Commonwealth Department of Health and Aged Care. All partners have agreed to co-operate and provide assistance in predicting and combating outbreaks of mosquito-borne disease in south-eastern Australia. The newly formed National Arbovirus Advisory Committee is a working party providing advice to the Communicable Diseases Network Australia on arbovirus surveillance and control. Recommendations for further enhancement of national surveillance for Murray Valley encephalitis are described. Commun Dis Intell 2001;25:33-47.

Top of page

Prologue

Arboviruses (arthropod-borne viruses) of public health importance in Australia include flaviviruses and alphaviruses. Within the flavivirus group the important human pathogens include Murray Valley encephalitis (MVE), Kunjin (KUN), Japanese encephalitis (JE) and Dengue (DEN) viruses. Alphaviruses causing human disease include Ross River (RR) and Barmah Forest (BF) viruses. Other arboviruses including Sindbis (SIN), Alfuy (ALF), Edge Hill (EH), Kokobera (KOK), Gan Gan (GAN), Trubanaman (TRU) and Stratford (STR) virus cause only mild or inapparent infections.¹

Of all the arbovirus infections, Murray Valley encephalitis causes the most severe disease. In the last 2 years MVE virus activity has increased with record levels of cases reported in Western Australia and widespread activity in the Northern Territory. In early 2001, 2 cases of Murray Valley encephalitis acquired in the Alice Springs area of the Northern Territory were reported, and a further case was detected in Mt. Isa, Queensland (see case reports in this issue of Communicable Diseases Intelligence). At the same time sentinel chickens in New South Wales showed seroconversions to MVE virus for the first time since the last national outbreak of Murray Valley encephalitis in 1974 (for sentinel chicken results see report in this issue of Communicable Diseases Intelligence).

In response to the increased MVE virus activity, the National Arbovirus Advisory Committee (NAAC), a working party of the Communicable Diseases Network Australia (CDNA) proposed that a scoping study of all current and possible surveillance mechanisms for Murray Valley encephalitis be undertaken. The collation of the information would facilitate, in the event of a national outbreak, rapid identification and co-ordination of these surveillance systems. The call for the review also reflected the perceived need to address cross-border issues that may arise during outbreaks and identify gaps in the current surveillance mechanisms. This document provides a summary of surveillance mechanisms and vectorborne disease control initiatives for MVE virus in Australia. Existing systems are described and other surveillance systems that may be utilised in the event of a widespread outbreak are discussed. Specific recommendations for the improvement of national MVE virus surveillance are proposed.

Top of page

Background

Epidemiology of Murray Valley encephalitis

MVE virus is enzootic in the Kimberley region of Western Australia and the Top End of the Northern Territory. The virus is epizootic in the Pilbara and regions further south in Western Australia and the southern half of the Northern Territory. The situation in Queensland is less well understood due to the dearth of data over the past three decades. However, human cases occur sporadically throughout the State, including southern Queensland.¹ Since 1974, however, nearly all cases of arboviral encephalitis due to MVE virus have been reported from Western Australia and the Northern Territory,^2,3,4 with MVE activity and human disease occurring in most years. Virus activity occurs in the wet season, with human cases being infected between February and July.

Prior to 1974 only 1 case of encephalitis due to MVE virus had been reported from Western Australia, and none from the Northern Territory. Strong circumstantial evidence has indicated that ecological and environmental changes resulting from damming the Ord River and establishing the irrigation area in the north-east Kimberley may have provided conditions conducive to increased MVE virus activity and endemicity.³ Any future changes to the waterways in the north of Australia may further change the ecology of the flaviviruses.

The history of severe epidemics of encephalitis in south-eastern Australia (particularly in the Murray/Darling River system) and the subsequent identification of MVE virus has been previously described.⁴ These outbreaks started in December/January, peaked in February/March and declined in the cooler months. They occurred at irregular intervals, the last being in 1974 and involving approximately 58 cases, 13 of whom died.^5,6,7

The 1974 outbreak spread to all mainland States of Australia and led to the introduction of the term 'Australian encephalitis (AE)'. The term AE has subsequently been used to refer to encephalitis due to either MVE or KUN infection,^8,9 and has led to considerable confusion. It is recommended that this term no longer be used and the terms MVE encephalitis and KUN encephalitis replace this nomenclature. If the infecting flavivirus cannot be differentiated, MVE/KUN encephalitis should be used. While surveillance mechanisms for KUN virus may be similar to that for MVE virus, the focus of this paper will be MVE virus.

Clinical aspects

It has been estimated that 1 in approximately 1,000-2,000 persons infected with MVE virus will develop severe encephalitis.^10,11 A larger proportion will develop a milder illness¹² but the vast majority remain asymptomatic. However, estimations of the case:infection ratios are not based on prospective data and are potentially inaccurate. It is likely that the rates will be higher during epidemics or in those at higher risk of severe disease.¹³

Murray Valley encephalitis is a potentially serious infection, with symptoms that include headache, neck stiffness, fever, tremor, weakness, confusion, fitting, and sometimes coma and death. Burrow and colleagues describe specific clinical features.⁹ These signs may not be immediately associated with Murray Valley encephalitis by physicians when cases occur in non-enzootic areas. Persisting fevers and seizures are common in children. Cerebellar signs, brainstem features (e.g. cranial nerve palsies such as facial palsies and ophthalmoplegias) and spinal cord involvement (pseudo polio) are seen in more severe cases, often with an associated tremor. Computerised tomography scans are usually normal, but abnormalities on magnetic resonance imaging may be dramatic (e.g. thalamic lesions). Examination of the cerebral spinal fluid (CSF) usually shows a lymphocytic pleocytosis and samples should be sent to a reference laboratory for culture, serology and Polymerase Chain Reaction (PCR).

The case fatality rate for Murray Valley encephalitis is 20 per cent and approximately 40 per cent of survivors will be left with permanent neurological damage.^6,12 Young Aboriginal children in Western Australia have a particularly poor outcome.¹³ The incubation period has not been well defined due to the difficulty in defining exact exposure episodes. No primary sources of data provide information regarding the MVE virus-specific incubation period, other than from a single case report from the 1974 outbreak, with an incubation period of 28 days.⁵ This exceeds the range reported for other arboviruses (5 to 15 days)¹⁴ and at the moment the best estimate of the incubation period is 5-28 days.

Top of page

MVE virus life cycle

There is a complex relationship between humans, vertebrate hosts, mosquito vectors and the environment involved in the ecology of most arboviruses. A range of factors may be involved in establishing and maintaining the MVE virus life cycle (Table 1).

MVE virus is transmitted to humans by mosquitoes and there is no direct transmission from person to person. The most common vector is the fresh water mosquito Culex annulirostris,^1,7 although other possible vectors have been identified.^1,11 The possibility of survival of MVE virus in arid areas via desiccation resistant Aedes tremulus eggs has been described.¹⁵

Table 1. Factors that may affect establishment and maintenance of MVE virus life cycle

Virus	Enzootic presence or reintroduction into the environment
	Inter-epidemic survival
Vector	Density, fecundity and longevity
	Feeding patterns and feeding preferences
	Oviposition and over-wintering
	Distribution of vectors
	Natural predators of vectors
	Control mechanisms by humans
Vertebrate host	Range of potential hosts species
	Viral titre and duration of viraemia
	Host species density and breeding
	Prior exposure to the MVE virus
	Movement and migration
	Mosquito avoidance mechanisms
Environment	Climate and weather, particularly temperature, rainfall and humidity
	Physical landscape, such as presence of waterways
	Human interventions on the environment, such as irrigation, drainage of swamps etc
Human	Prior exposure to the MVE virus
	Population distribution
	Lifestyle factors
	Use of preventative measures to avoid being bitten by mosquitoes

Both field and experimental infection studies have been used to investigate a number of vertebrate species as potential hosts for MVE virus. A comprehensive review of studies has been previously published.⁷ Investigation of wild and domestic animals around the time of the 1974 outbreak revealed infections in domestic fowls,¹⁶ wild birds and horses.¹⁷ While serological field studies do confirm that particular species can be infected with MVE virus, they do not indicate what viral titres are achieved during infection and how long these infections are maintained. Both of these factors influence whether a particular vertebrate species is likely to be a major host in the MVE virus life cycle. Experimental studies have been undertaken to address these issues. Following experimental infection, wild birds, including herons and egrets, have been shown to develop viraemias of 3 to 5 days duration. Maximal titres were obtained in younger birds.¹⁸ Studies have shown that domesticated animals such as fowl, pigs, cattle and horses may also be experimentally infected¹⁹ but the role of these species in natural transmission cycles is not believed to be important. On the basis of laboratory and field experiments, wild birds, particularly wading water birds are thought to be important in the life cycle of MVE virus. The rufous night heron (Nycticorax caledonicus, also known as the nankeen night heron) is recognised as a major vertebrate host of MVE virus.⁷

Meteorological events such as rainfall, temperature and humidity also play a major role in the transmission of MVE virus.²⁰ Mosquito abundance is affected by the availability of aquatic breeding habitats. Other factors such as temperature, wind speed and wind direction affect their distribution and life cycle. Outbreaks of Murray Valley encephalitis may occur after unusually heavy and persistent rainfall and subsequent flooding. Abnormal rainfall may increase the numbers of mosquitoes and lead to movement of infected birds from enzootic regions to epizootic regions.^5,21 The mechanisms by which outbreaks in south-eastern Australia commence are unclear. One possibility is that MVE virus may be enzootic in south-eastern Australia in cryptic foci that are not detected by vector and vertebrate surveillance mechanisms in the intervening periods between outbreaks, but this seems to be an unlikely explanation given the extensive surveillance efforts. A more plausible explanation is that the virus is reintroduced by birds from the northern latitudes following periods of extreme rainfall and flooding. Indeed, genetic evidence demonstrates a lack of independent divergence of Australian MVE lineages, which strongly supports the re-introduction of MVE virus rather than the presence of cryptic foci.²²

Top of page

Surveillance mechanisms

Mechanisms for monitoring MVE virus activity include surveillance of human cases, surveillance of MVE virus activity in vertebrate hosts, monitoring of mosquito vectors for abundance, virus isolation from mosquitoes and climate surveillance.²³ In some jurisdictions monitoring of human cases alone is insufficient for public health action, particularly when there are alternative surveillance mechanisms which may trigger action prior to the detection of human cases, and when the outcomes of infection can be so severe. If viewed in purely economic terms, the financial costs of these additional surveillance systems must be weighed against the cost of preventing human disease. The prevention of 1 human case with permanent neurological damage would make these systems cost effective. In the United States of America (US) it has been estimated that the community cost of a patient with permanent neurological damage is $US3 million. No equivalent costings are available in Australia, although it is likely that the value may be less than that estimated for the US.

The additional surveillance systems are used in concert to signal the risk of human disease. Monitoring sentinel animals for MVE virus activity provides information regarding whether MVE transmission to vertebrates is actually occurring. Sentinel chickens flocks maintained in a number of states and territories provide information on MVE virus activity.²⁴ Vector surveillance includes mosquito trapping for speciation and enumeration of mosquitoes to monitor population sizes and composition. Monitoring of weather conditions and vector surveillance determines whether there is potential for MVE activity to occur. Virus isolation from trapped mosquitoes is necessary to define whether MVE is actually present, but is difficult to deliver in a timely fashion in some jurisdictions.

The trapping of live vector collections does require the use of CDC/EVS CO₂ baited traps, which may have logistical constraints for remote regions of Australia, such as the Kimberley region of Western Australia. The development of PCR assays for detection of MVE virus in pools of mosquitoes would allow more timely reporting of the detection of virus in mosquitoes and may obviate the need for live mosquito collections. It would be advantageous if these methods could be applied to mosquitoes that have been collected in traps for up to a week. These methods remain in the developmental stage in a number of laboratories across Australia until technical problems are overcome. These problems include the storage of trapped mosquitoes, prevention of fungal growth, viral RNA degradation and the presence of PCR inhibitors in large pools of mosquitoes. Optimal pooling sizes must be determined before PCR is a cost-effective and timely replacement for virus isolation using current cell culture techniques.

Meteorological surveillance is used in the prediction of MVE virus activity by signalling conditions that have been associated with outbreaks of Murray Valley encephalitis in humans in the past. Examination of climate meteorologic information including rainfall, temperature and the Southern Oscillation (SO) may assist the prediction of risk situations. The SO is an inter-annual oscillation in tropical sea level pressure between eastern and western regions of the Pacific Ocean. The Southern Oscillation Index (SOI) is calculated from the monthly or seasonal fluctuations in the air pressure difference between Tahiti and Darwin. Positive SOI values suggest that rainfall will be above average across eastern Australia, while negative values suggests that rainfall will be below average. The SOI predicts rainfall to a lesser extent in central and western Australian states.

There are 2 models for the prediction of Murray Valley encephalitis activity for south-eastern Australia using climatic information. The Forbes model² utilises rainfall patterns and provides a quantitative approach to predicting outbreaks of Murray Valley encephalitis. The model relies on rainfall patterns in the preceding and current season of MVE virus activity. The model predicts MVE amplification where there has been above average rainfall in the current and preceding summers, with the underlying hypothesis that abnormal rainfall enhances breeding of both wading birds and mosquitoes. The Nicholls model²⁵ suggests a qualitative association between Darwin atmospheric pressure (a measure of SO) in autumn, winter and spring of the preceding season and Murray Valley encephalitis activity.

A mathematical model based on host and vector factors, has been developed²⁶ for the rural amplification of MVE virus in southern Australia during the 1951 and 1974 outbreaks. This model predicted the likely duration of the rural amplification phase, estimated to have commenced in October of the year prior to an outbreak. Thus it appears that seeding of the south-eastern areas of Australia in the previous year is important for the establishment of an outbreak in the following season.

Top of page

Evidence supporting the use of animal, vector and climate surveillance mechanisms to predict disease in humans

The complex MVE virus life cycle means that a number of surveillance mechanisms can be utilised to predict MVE virus activity. Whether this activity heralds human disease is enhanced by drawing together of data from a number of surveillance systems and evaluating their predictive ability in light of human cases. The appropriate surveillance tools for monitoring MVE virus activity may vary from jurisdiction to jurisdiction. Factors such as whether the virus is enzootic or epizootic, the frequency of human disease, the geography and climate, the availability of laboratory facilities and other infrastructure, competing public health concerns and the availability of public health resources will affect what surveillance mechanisms are appropriate. The ability to evaluate the use of animal, vector and climate surveillance is affected by the frequency of human disease, indeed, it is difficult to evaluate the use of surveillance mechanisms in the Murray Valley region given the last human cases occurred in 1974.

Some of the best evidence for the use of non-human surveillance mechanisms for predicting MVE activity comes from Western Australia. The State has marked climate variability and encompasses enzootic and epizootic regions, as well as regions where no MVE virus activity has ever been detected. Due to the logistical difficulties associated with mosquito monitoring, sentinel chicken and climate surveillance are the key elements for predicting MVE virus activity in Western Australia. Large outbreaks (e.g. in 1993 and 2000) have been associated with abnormal weather patterns in Western Australia. Data collected over the last 10 years of the sentinel chicken program indicate that seroconversions in sentinel chickens have preceded likely dates of exposure of the first human cases by 2 to 18 weeks in all but one situation during that period.²⁷

While it has been shown that large outbreaks of Murray Valley encephalitis are associated with abnormal weather patterns in Western Australia, the use of climate surveillance to predict outbreaks in south-eastern Australia remains controversial, and should be used in conjunction with other surveillance data. The Nicholls model provides more timely prediction of the risk of Murray Valley encephalitis activity compared with the Forbes model as it does not rely on collection of data during the current season. The Forbes model, however, has proved more accurate in recent years. This model suggested there would be MVE virus activity in south-eastern Australia during the 1999/2000 season when there were a number of cases of Murray Valley encephalitis in the Alice Springs area of the Northern Territory and a single case in the north of South Australia. The Forbes model again predicted activity for the 2000/2001 season, which did occur. In comparison, the Nicholls model suggested activity was unlikely in both seasons (personal communication, S Doggett).

One of the difficulties with both the Forbes and the Nicholls models is that they were based on major outbreaks of arboviral encephalitis in south-eastern Australia, which have not occurred since 1974. Both models were developed using very small data sets and neither incorporates observed activity in sentinel animals or addresses subclinical infections. Neither scheme takes into account the impact of other factors such as the breeding and movements of vertebrate hosts and vectors, or the influence of human activities on the natural landscape, such as irrigation and land development. Similarly, public health messages regarding the risk of arbovirus infection and widespread vector management programs may have reduced the incidence of human disease despite the presence of weather conditions that have been associated with outbreaks in the past.

Laboratory testing

Testing for MVE virus in humans

Virus isolation from blood is only possible in the very early acute phase of the illness prior to the appearance of antibodies. MVE virus has only been isolated from a small number of human cases, and none since 1974. While detection of viral RNA in CSF⁸ or blood²⁸ using PCR has a higher yield, most infections are diagnosed serologically. Due to high levels of background flavivirus infection in endemic areas, and the long-term persistence of IgM, it is important to demonstrate rising titres of IgG or to have a positive viral detection test (culture or PCR) to confirm acute infection. If confirmatory laboratory evidence is unavailable or inconclusive, then a detailed exposure and clinical assessment is required to determine the likelihood of recent infection. As there is broad cross-reactivity in antibodies to the flaviviruses, assigning a particular virus as the cause based on serology requires a test that is sufficiently specific. Diagnostic and reporting guidelines for MVE (and other arboviral diseases) have been developed²⁹ and have subsequently been refined in the Public Health Laboratory Network case definitions.³⁰

Vertebrates and vectors

There is a network of laboratories across Australia that provides a range of testing for MVE virus activity in vertebrates and vectors, including serological testing, mosquito identification and viral isolation from mosquitoes (Table 2). In addition, there are laboratories that provide such services, but are not currently contributing to surveillance systems. These laboratories may provide services on an ad hoc basis for research purposes, for example, for the opportunistic testing of domestic animals. Local councils in some jurisdictions may also undertake mosquito identification. Top of page

Table 2. Arbovirus research laboratories in Australia providing testing for vertebrate and vector surveillance systems for MVE virus*

Location	Laboratory	Institute	Testing provided
Western Australia	Western Australian Arbovirus Surveillance and Research Laboratory	University of Western Australia	Serological testing of vertebrate hosts Mosquito collection and identification Virus isolation from mosquitoes
Northern Territory	AL Rose Virology Laboratory	Department of Primary Industry and Fisheries	Serological testing of vertebrate hosts Mosquito collection and identification
	Medical Entomology Branch	Territory Health Services	Mosquito collection and identification
New South Wales	NSW Arbovirus Laboratory	Institute of Clinical Pathology and Medical Research, Westmead	Serological testing of vertebrate host Mosquito collection and identification Virus isolation from mosquitoes
Victoria	Victorian Institute of Animal Sciences	Department of Natural Resources and Environment	Serological testing of vertebrate hosts Mosquito collection and identification Virus isolation from mosquitoes
	Australian Animal Health Laboratory	Commonwealth Scientific and Industrial Research Organisation	Serological testing of vertebrate hosts
Queensland	Queensland Health Scientific Services	Queensland Health	Mosquito collection and identification
	Arbovirus and Emerging Diseases Laboratory	University of Queensland	Serological testing of vertebrate hosts Mosquito collection and identification Virus isolation from mosquitoes
	Tropical Public Health Unit	Queensland Health	Mosquito collection and identification
South Australia	Mosquito Research Laboratory	University of South Australia	Mosquito collection and identification

* excluding opportunistic testing

Public health action

Public health action is determined by assessing data from the various surveillance mechanisms. It is impossible to fully eliminate mosquito breeding, therefore, it is important to warn the general public of the risk of Murray Valley encephalitis once conditions are optimal for virus transmission. Advice on personal protection and reducing risk behaviour are the major public health messages. Such warnings can be developed specifically to target the lifestyles and literacy levels of at-risk communities. Mosquito control programs may reduce the numbers of both mosquito larvae and adult vectors in certain circumstances. However, as there is no specific treatment for Murray Valley encephalitis, prevention remains the most important strategy for averting disease.

While it has not been possible to formally evaluate their effectiveness, targeted public health campaigns, drawing on evidence from animal, vector and climate surveillance, are believed to be more effective than general warnings. Data from these additional surveillance mechanisms can be used to stimulate public awareness prior to the detection of human cases.

Continue to Part Two

This article was published in Communicable Diseases Intelligence Volume 25, No 2, April 2001.