Ross River Virus Risks in Tropical Australia by Harley - Article Example

Critical Evaluation of a Published Paper on Ross River Virus Risks in Tropical Australia This paper is a critical evaluation of the given research document. The document under evaluation is: Harley, D., Ritchie, S., Bain, C., & Sleigh, A. (2005). Risks for Ross River virus disease in tropical Australia. International Journal of Epidemiology, 34(3), 548-555. This paper is on risks associated with Ross River virus disease in tropical Australia. Introduction Ross River virus came into limelight first in 1928 in New South Wales' Narrandera and Hay region. The first time the virus was isolated was in 1959 from Townsville, Queensland mosquito trapped along the Ross River. Till date almost whole of Australia has seen outbreaks of Ross River virus and the largest outbreak has affected more than 60,000 people between 1979-1980. This outbreak affected the Western Pacific more than any other region. The outbreak was normally referred to as epidemic polyarthritis before the virus was identified (National Notifiable Diseases Surveillance, 2008). This is a mosquito-borne infectious disease and is characteristic of polyarthritis and influenza-like symptoms (Morrison et al, 2007). The most affected regions in Australia include Northern Territory, tropical Western Australia and Queensland, mostly during wet seasons of summer and autumn. The higher population of mosquitoes coincides with higher rainfall. Since in Australia different regions have higher rainfall in different months, Ross River virus gets active in these corresponding months. Lagoons and backwaters are breeding grounds for these mosquitoes. For these outbreaks higher risk areas are wetlands, marshes farms with irrigation systems and waterways. Females and males are equally affected in the age group of 25 to 44 years old. Ross River fever is on the notifiable diseases list of Department of Health and Ageing. Transmission takes place through mosquitoes only and the disease is not contagious. Wallabies, kangaroos, horses, possums and even flying foxes and birds have been identified as reserviour hosts and till date as many as 30 species have been identified as possible vectors. However, Aedes vigilax, Culex annulirostris and Ae. camptorhynchus have been identified as major species responsible for transmission. Severity of the disease determines the symptoms involved, which range from arthritis, arthralgia, rash and fever. Children present with asymptomatic symptoms and the average seven to nine days has been observed as the incubation period (Harley et al, 2001; Harley et al, 2002)). During the twin decades of 1980s and 1990s it was seen that symptoms like fatigue, arthralgia and depression lasted for many years after the virus infected somone. Recent studies have remarked that arthralgias got resolved maximum up to seven months. That way the infection has been equated with Q fever and Epstein-Barr virus. Since no analytical studies were conducted until 2004 on individual risks for RRV, this study was conducted with an intention to explore the same. The study also wanted to study protective factors in high incidence area with an underlying thought of assessing utility of case-control design. In order to reach a conclusion a matched case-control design was employed on new cases that emerged in the community in some specific areas which included Mareeba, Cairns, Douglas and Atherton; all falling within high incidence area of Queensland. The risks were assessed on the statistical parameters. Critical evaluation Background of the paper as enumerated by it is the common occurrence of arbovirus disease in the country and their typical presentation of arthritis and arthralgia. The transmission takes place through mosquitoes from other vertebrates triggering an incubation period of minimum 3 days and maximum 21 days. On an average it is up to 9 days. Among natural reserviour hosts are wallabies and kangaroos, and also possums, birds, horses and flying-foxes. The paper reveals that average number of cases notified for RRV per annum is 5000 and it is held that equal number go unreported or undetected. Majority of cases occur in northern Queensland. Increased transmission occurs when there is a co-occurrence of tides, rainfall, humidity and high temperature (Blumer et al, 2001; Thomson et al, 1998). This study was conducted in the wake of the fact that there was no available data at the time of study which can help predict the RRV disease risk. The study was conducted to first identify the risk factors and then suggest preventive measures. Case-control design was emphasised because the researchers assumed that it could prove as the most effective methodology to identify these risk factors. These risk factors included both environmental and behavioural risks. The most salient feature of this study has been that for studying such risk factors this epidemiological method has not been previously used (Jouan et al, 1989; Murray-Smith et al, 1996; Huan et al, 1999). Thus the methodology used for this study was prospective, community-based case-control in areas as mentioned above. In order to accomplish the task the researchers David Harley, Scott Ritchie, Chris Bain and Adrian C Sleigh, who came from different institutions, namely Queensland Centre for Intellectual and Developmental Disability, School of Population Health, University of Queensland, Mater Hospital, Brisbane; Tropical Public Health Unit, Queensland Health, Cairns; School of Public Health and Tropical Medicine, James Cook University, Cairns; School of Population Health, University of Queensland, Herston, Queensland; Queensland Institute of Medical Research, Herston; and National Centre for Epidemiology and Population Health, Australian National University, Canberra, organised a surveillance system the job of which was to ascertain new RRV infection cases as and when Queensland Health get notified with the same. The notifications were issued either by public or private pathology laboratories in these areas. On account of the privacy laws governing the areas, most crucial private data of infected cases was not accessed by data including date of birth, sec, and treating doctor were obtained. Persons who had acquired RRV disease locally were defined as eligible cases and the data was collated from January 01 to May 31 in the year 1998; period corresponding to the wet season in the area. PanBio, the only test kits available then were used for indirect ELISA assessment. Serological diagnosis was conducted using either IgG seroconversion or single positive IgM test. The latter comprised of 40 cases and the former 15 cases. It is notable that all these persons had lived in the area where they contracted the infection for either one month or more. These cases were clinically examined and the examination took place in their homes. At the same time the researchers completed a household environmental proforma and a risk assessment questionnaire, following which measurement of 108 exposures was done. Some major ones of these were grouped according to leisure and sporting activities like golfing, bushwalking and fishing. Besides this peri-domestic activities that were measured included smoking in the evenings (outside of homes) and gardening. Since kangaroos and wallabies were presumed as hosts, people's exposures to the same were as well measured. The measurement of risk factors was not only limited to the same but also to other major or minor activities that were suspected of either directly or indirectly in touch with vectors (Russell & Doggett, 2008). These included low set houses, exposure to mosquito bites, work conditions, domestic reserviours like cats and dogs (that were presumed as potential vectors) and even use of deodorant or perfume. Incubation periods were selected for measuring exposures since that was the high time when spread is widespread. GPs' help was sought to select controls from their practice. Controls were needed to be willing to provide blood samples so that absence of RRV IgM or IgG was ascertained. One or two controls were provided by GPs even as up to 4 controls were required for each case. Different methods were used to select controls and GPs used their own judgment on how to select them. Some relied on recall and some accessed their datasets to provide sex and age-matched controls. Some even asked their staff to recall any potential controls. One of the investigators conducted all inspections and interviews for which he used a standard questionnaire. Three different ways were used to assess risk and protective factors. These included those across the incubation period which was inferred, over the year before there was an onset of symptoms, and with no reference to time. Incubation period was referred to as the three-week period ending four days before there was an onset of symptoms. This was considered as the period in which chances of infection occurring was the maximum. The researchers were calculative when they decided to chose these three methods. It is because they wanted to avoid exclusive dependence since a past exposure was being recorded. The controls and the matches were selected in the same time period in order to create greater validity in assessing the risks and protective measures. However, in order to assess environmental risks related to the housing inspections were carried out at the time for conducting the study. SPSS was used to conduct analysis of the data collected and confidence intervals (CIs) and odds rations (ORs) were used by following Cox regression methodology (SPSS, 2002). The best feature of such analysis was that they could be matched with any number of controls per case. ORs were calculated for all measured exposures and the results suggested that the cases presented with an epidemiological effect. The study was approved by the ethics committee of the University of Queensland Medical Research and the Cairns Base Hospital. The major risk factors for which ORs and CIs were calculated included camping, bromeliads, banana trees, working outside, kangaroos and wallabies in the gardens. Protective exposures which were analysed for CIs and ORs included personal repellents, mosquito coils, citronella candles, air-conditioning and preference for clothing which was light-coloured so as to attract lesser mosquitoes (Edman & Spielman, 1998). The study comprised of 85 controls and 55 cases. Up to four controls were matched with each case. In the given period, which is mentioned above, 55 cases represented almost half of all cases (129 in all) that infected the population in the given area. The controls selected for participation in the study had tested negative for RRV IgG and IgM ELISA. Forty eight of all controls had not been to their GPs at least one month before they were inducted into the study. Of the remaining controls - 37 in all - 21 presented with various injuries, infections and undifferentiated symptoms like heart palpitations and lethargy and 16 had consulted their GPs for non-acute causes. Controls and cases were selected on the basis of same age group with 39 and 40 as mean ages for corresponding ranges of 17-63 and 16-69. There were 51 percent cases that were females matched by 55 percent controls. Controls were qualified people and were either managers, professionals or administrators. However, the study summarised that these things were not so significant for the study. Ninety four percent of controls and 93 percent of cases lived in houses, while others lived either in caravan parks or flats and so on. Both cases and controls consumed alcohol on a similar scale; some didn’t drink at all, some weekly and some monthly. The paper reveals that both cases and controls presented with similar co-morbidities and interview of controls took place long after the inferred exposures. One hundred and five and 36 days was the mean period between onset and interview. Risk factor analysis was conducted by using unadjusted but matched odds ration to assess exposures, which is termed as crude ORs. It was seen that ever vs. never camping in the year before there was onset of symptoms did actually double the risk for acquiring RRV disease. With never camping as the reference category, camping 1–3 times was associated with an OR of 1.87 (95% CI 0.76–4.60) and 4 or more times increased the risk 2.5-fold (OR 2.41, 95% CI 1.04–5.56; Wald chi-squared for linear trend 5.0, P = 0.08). None of the 15 other peri-domestic or sporting activities studied increased the crude risk significantly, although some had reasonably strong associations. According to the paper no statistically significant association has been shown between presence of wild and domestic animals assumed as potential RRV infection reserviours. However an association has been revealed between wallabies and kangaroos if they have been present around during the incubation period. Since eight controls and four cases couldn't recall their presence during the mentioned period, researchers presented a sensitivity analysis. If all controls and cases were missing it was assumed that exposure data were exposed i.e., the OR (95% CI) was 1.91 (95% CI 0.72–5.08); if they were assumed unexposed the OR (95% CI) was 3.50 (95% CI 0.87–14.1). This association nevertheless has come under the suspicion of the researchers because cases are more likely to recall their exposure to vectors than controls. This was one reason why researchers assumed that this association does not need to be explored further as it actually disappears. In other words the association does not exist at all. The presence of banana trees and bromeliads was shown to have significant association with the risk. A three-fold decrease in exposure was on account of light-coloured clothing and one stood at lesser risk for having RRV by using citronella candles, insect repellents and mosquito coils. More protective measured used meant lesser risk of getting infected. Peri-domestic environment was not seen as having any major association with the RRV disease risk. It did not matter did a house have a patio, porch or a veranda or not. Risk did not even increase when the dwellings were screened for vectors or breeding grounds. During the early onset period insect repellents were found as reducing the RRV risk. In the latter period there was attenuation in the risk factors if one wore protective light-coloured clothing. Immediately before the RRV disease risk increased camping was seen to boost the risk. This risk was more pronounced in the coastal than inland areas. Similarly presence of ornamental plants in the coastal areas increased the risk but did not do so in the tablelands stratum areas. This study has not been able to implicate any specific reserviour or vector but shows consistency with risk factors known otherwise. Presence of wallabies and kangaroos in gardens has been mentioned as not increasing the risk but since it has not been substantiated with relevant data, a bias can be predicted. However one important point that emerges is that the same kangaroos and wallabies must not be mistaken as not being potential risk factors when found in forest areas. Humans venturing in to the same are at greater risk than what they are back home. To sumamrise, this study could be termed as ecologically (Rothman & Greenland, 1998) and biologically coherent and cannot be termed as 'chance findings' among many such findings that take place. The concern that this study raises and also accepts is that there might be high recall bias involved. This is because control than case interviews have taken place after a longer delay. It would have been difficult for controls to recall exposures in the incubation period that is inferred. Cases could not have suffered from recall issues because they were the ones who acquired the disease. Even as the study has not implicated presence of wallabies and kangaroos around households as risk factors, sensitivity analysis could be suggestive of the same. One milestone that the study has achieved is with reference to the camping being as one of the major risk factors. It is because results for this year have been consistent with results the next year. That the study was first to examine household and individual risk factors certainly goes in its advantage. And the emergence of the conclusion that low cost protection is the safest as well is of pivotal importance. Mosquito coils, citronella candles and repellents, which all fall under the category of personal protective measures, reduce the risk by two times. And there is an astounding 8-time reduction in the risk by earning clothes which are light coloured. The best part, as revealed by the study, is that even measures as these when employed during camping drastically reduce the risk of exposure to the virus. In Australia, where camping forms major part of leisurely activities this comes as an important suggestion. Similarly the observation about bromeliad plant presence is worth introspection. These plants in coastal areas have a role in becoming breeding sites for Aedes aegypti, which is a potential vector and also for Oc. notoscriptus, another potential vector. The paper has hinted at an important point, which is RRV transmission's vector succession. The paper has highlighted this with specific reference to the waning effect of either repellents or protective clothing in the latter part of the transmission season. This means that the virus can shift its transmission to later transmission either by nocturnal or creuscular biters like Cx. annulirostris. This could be an area of concern in future and possibly worth exploring in the researchers of this kind that might take place now. It has been seen mosquitoes that bite in the days are attrracted to dark colours and mostly include Ochlerotatus vigilax, Verrallina carmenti and Oc. notoscriptus (Lee et al, 1982; Lee et al, 1984; Harley et al, 2000). Most of this research has focused on RRV's vector association but camping-related results have come as a real eye-opener in as far as RRV disease and its spread is concerned. The negative side of the research can be traced to the observer bias. As can be seen the interviewer has not be blinded to the status of the disease of the person he was interviewing. Another negative aspect of the study has been very small size of cases and controls. This would have resulted in limited analysis and limited control of confounding. Other things that resulted in confounding were use of protective measures but that could be, as was done by the researchers, adjusted. Even though cases were matched with controls rather intelligently, but the same thing seems to have prevented confounding. The matching of controls and cases for sex, age, geographical area, and exposure period prevented confounding by ecological determinants, matched factors and by such factors as rainfall and temperature. Also the study seems to have in mind only a few mosquito species while assessing risks associated with RRV disease spread. This is irrespective of the fact that the virus has been isolated from a number of other mosquito species. The study could have thrown some light on the same even if at present evidence shows only a few species exhibiting virus-vector association. The establishment of this association and studying the risk thereby is of importance because mosquitoes are the most proximal determinants which can result in the infection being transmitted. Conclusion Given the results that the study has provided and the extent to which they can be used for future research, it can be deduced that selection of case-control methodology for studying the risk is efficient and feasible and ideal for further prevention-oriented research. The study is of significant value since this method has not so far been used to study arbovirus risks. References Blumer C, Roche P, Spencer J et al. (2003). Australia’s notifiable diseases status, 2001. Commun Dis Intell; 27:1–78. Edman JD, Spielman A. (1998). Blood-feeding by vectors: physiology, ecology, behaviour, and vertebrate defense. In: Monath TP (ed.).The Arboviruses: Epidemiology and Ecology. 1st edn. Boca Raton, Florida: CRC Press, pp.153–89. Harley D, Sleigh A, Ritchie S. (2001). Ross River virus transmission, infection and disease: a cross-disciplinary review. Clin Microbiol Rev;14:909–32. Harley D, Bossinghmam D, Purdie DM, Pandeya N, Sleigh AC. (2002). Ross River virus disease in tropical Queensland: evolution of rheumatic manifestations in an inception cohort followed for six months. Med J Aust;177:352–55. Han LL, Popovici F, Alexander JP Jr et al. (1999). Risk factors for West Nile virus infection and meningoencephalitis, Romania, 1996. J Infect Dis;179:230–33. Harley D, Ritchie S, Phillips D, van den Hurk A. (2000). Mosquito isolates of Ross River virus from Cairns, Queensland, Australia. Am J Trop Med Hyg; 62:561–65. Jouan A, Coulibaly I, Adam F et al. Analytical study of a Rift Valley fever epidemic. (1989). Res Virol;140:175–86. Lee DJ, Hicks MM, Griffiths M, Russell RC, Marks EN. (1982). The Culicidae of the Australasian Region, Volume II. Canberra: Australian Government Publishing Service. Lee DJ, Hicks MM, Griffiths M, Marks EN. (1984). The Culicidae of the Australasian Region, Volume III. Canberra: Australian Government Publishing Service. Morrison TE, Fraser RJ, Smith PN, Mahalingam S, Heise MT (2007). "Complement contributes to inflammatory tissue destruction in a mouse model of Ross River virus-induced disease". J. Virol. 81 (10): 5132–43 Murray-Smith S, Weinstein P, Skelly C. (1996). Field epidemiology of an outbreak of dengue fever in Charters Towers, Queensland: are insect screens protective? Aust NZ J Public Health;20:545–47. National Notifiable Diseases Surveillance. (2008). Australia: Department of Health and Ageing. Australia. Russell RC & Doggett SL. (2008). Ross River & Barmah Forest". Department of Medical Entomology, University of Sydney. Rothman KJ, Greenland S. Modern Epidemiology 2nd edn. (1998). Philadelphia: Lippincott-Raven. SPSS. Release 10.0.5 & 11.5.0. Chicago, Ill: SPSS Inc., 1989–1999 & 2002, respectively. Thomson J, Lin M, Halliday L et al. (1999). Australia’s notifiable diseases status, 1998: Annual report of the National Notifiable Diseases Surveillance System. Commun Dis Intell;23:277–305. Read More