HANTAVIRUS IN RODENTS OF SAN DIEGO COUNTY IN RELATION TO THE EL NIÑO-SOUTHERN OSCILLATION (ENSO) A Thesis. Presented to the.

Transcription

1 HANTAVIRUS IN RODENTS OF SAN DIEGO COUNTY IN RELATION TO THE EL NIÑO-SOUTHERN OSCILLATION (ENSO) A Thesis Presented to the Faculty of San Diego State University In Partial Fulfillment of the Requirements for the Degree Master of Public Health with a Concentration in Environmental Health by Sarah Elizabeth Cokely Summer 2011

2

3 iii Copyright 2011 by Sarah Elizabeth Cokely All Rights Reserved

4 iv ABSTRACT OF THE THESIS Hantavirus in Rodents of San Diego County in Relation to the El Niño-Southern Oscillation (Enso) by Sarah Elizabeth Cokely Master of Public Health with a Concentration in Environmental Health San Diego State University, 2011 Hantavirus Pulmonary Syndrome (HPS) is a viral infection carried by rodents and transmitted to humans. Increases in rodent population following El Niño precipitation in the Four Corners region of the United States and resulting increases in HPS cases have been studied; however, no such studies exist for the San Diego County region. Proportions of hantavirus positive rodents were compared for years with normal weather events and years where an El Niño event occurred between 1994 and 2009 in San Diego County. Annual precipitation was also evaluated to establish if weather effects played a significant role in proportion of hantavirus positive rodents as well. When comparing percent of rodents positive for hantavirus in the year following an El Niño event as compared to all other years sampled, rodents in the year following El Niño years were almost three times more likely to test positive for hantavirus that rodents during normal weather events (OR 2.96, 95% CI ) El Niño years did not experience a statistically significant greater rainfall than normal weather years with the mean rainfall for normal year 8.1 inches (SD 5.0 inches) and the mean rainfall for El Niño years 12.1 inches (SD 6.4 inches) and a p-value of The excess rains in a region like San Diego County likely allow for a great increase in plant growth and therefore rodent resources, allowing for an increase in the hantavirus vector. The increase in hantavirus positive rodents in any given region presents a public health concern as hantavirus has no vaccine or virus specific treatment and has a high case fatality rate. A better understanding of whether or not El Niño events correlate with increases in hantavirus positive rodents in San Diego County can allow for increased surveillance and public alerts to prevent human cases.

5 v TABLE OF CONTENTS PAGE ABSTRACT... iv LIST OF TABLES... vii LIST OF FIGURES... viii ACKNOWLEDGEMENTS... ix CHAPTER 1 INTRODUCTION...1 ENSO Events and Hantavirus...2 Objectives LITERATURE REVIEW...3 Hantavirus...3 Four Corners Sin Nombre Virus Outbreak...5 Hantavirus (HPS) Transmission...6 El Niño Southern Oscillation...9 Enso and Vector-Borne Disease...10 Hantavirus and Enso Events...11 Climate Change and Vector-Borne Disease...13 Hantavirus Vectors and Transmission Capabilities...14 Food Supply for Southern California Rodents...15 County of San Diego Surveillance and Testing METHODS...18 Abundance Data...18 Rodent Trapping...18 Rodent Sampling...19 Laboratory Testing...20 Data Maintenance...20 Climate Data...20 Statistical Analysis...21

6 vi 4 RESULTS DISCUSSION CONCLUSIONS...32 REFERENCES...33 APPENDIX A HANTAVIRUS HUMAN CASE MAP...37 B PEROMYSCUS MANICULATUS US RANGE MAP...39 C TRAPPING LOCATION PICTURE...41 D PERSONAL PROTECTIVE EQUIPTMENT...43 E CORRECT SYRINGE PLACEMENT...45 F PERCENTAGE HANTAVIRUS POSITIVES BY MONTH...47 G 2 X 2 TABLE FOR HANTAVIRUS DURING ENSO YEARS...53 H 2 X 2 TABLE FOR HANTAVIRUS FOLLOWING ENSO YEARS...55

7 vii LIST OF TABLES PAGE Table 1.Chi Squared/Odds Ratios for Hanta and El Niño Years...24 Table 2. Averaged Precipitation and Positives...24 Table 3. Hantavirus Positive Test Ratio Based on Year, ENSO Highlighted...49 Table 4. Hantavirus Positive Rodent Species...50 Table 5. Hantavirus Positive Rodent Species by Percent...51 Table 6. County Technician Staffing by Year...52 Table 7. 2X2 Table for OR Calculation of Rodent Positive during an ENSO Event...54 Table 8. 2X2 Table for OR Calculation of Rodent Positive in the Year following an ENSO Event...56

8 viii LIST OF FIGURES PAGE Figure 1. Relationship between percentage of hantavirus positive rodents and annual precipitation from the previous year Figure 2. Human HPS cases in the United States Figure 3. Range of Peromyscus maniculatus in the US and locations where hantavirus positives have been found Figure 4. County of San Diego rodent trap location Figure 5. PPE for rodent sampling Figure 6. Proper syringe placement for rodent sampling Figure 7. Distribution of hantavirus positive rodents by month....48

9 ix ACKNOWLEDGEMENTS I would like to thank the members of my committee: Dr. Quintana, Dr. Gersberg, and Dr. O Leary. All of these individuals aided me a great deal in the formulation and completion of this thesis and I am extremely grateful for all of their help and support. I would also like to thank Chris Conlan at San Diego County s Vector Control Program who took time to answer all of my questions. Additionally, the surveillance technicians Paola and Dana were kind enough to take me out during rodent trapping allowing me to truly understand the procedures followed during rodent sampling. Last, my friends and family provided an enormous amount of support during this undertaking. My husband in particular was the best cheerleader anyone could ever ask for, and I am extremely thankful to have had his constant encouragement.

10 1 CHAPTER 1 INTRODUCTION Human disease due to hantavirus has been detected in many areas of the world as either Hantavirus Renal Syndrome (HRS) or Hantavirus Pulmonary Syndrome (HPS). Hantavirus is a zoonotic disease passed from rodent to human (Simmons & Riley, 2002). In the Americas, HPS has been present for a number of years; however, it was not recognized in the North America until 1993 when an outbreak occurred in the Four Corners region of the United States. People in the Four Corners region were taken ill with a respiratory disease and 22 died (Centers for Disease Control and Prevention [CDC], 1993b). At the beginning of the outbreak, the cause of death was unknown, but following investigation, a new strain of hantavirus was revealed and named Sin Nombre Virus (SNV). The vector was later identified as the Deer Mouse or Peromyscus maniculatus (Childs et al., 1994). While the Four Corners region still has the highest incidence of hantavirus, both in the rodent population and human cases, there have been cases in many US states, including a great number in California (see Appendix A). This has prompted the state of California, and specifically the San Diego County Vector Control Program, to perform surveillance for HPS by testing the antibody levels of rodents (Conlan, personal communication, January 13, 2011). It has been claimed that the original outbreak of hantavirus in the US was the result of excess precipitation from the El Niño Southern Oscillation (ENSO) event during the winter. Prior to this year, there had been a period of drought which was followed by the excess precipitation characteristic of an ENSO event. This is thought to have prompted excess growth of the vegetation in the area which typically does not receive a great deal of water (Englethaler et al., 1999). One specific food source that flourished were the pinon nuts, the seeds from the pinon pine tree, which were a staple of the rodent diet in the region, and were noted to increase a great deal following the ENSO rains of (Stone, 1993). In turn, the rodents had a larger food supply and produced more offspring which led to greater population density. This allowed for an increase in the spread of the hantavirus present in the rodent community due to increased rodent-to-rodent contact (Root, Calisher, & Beaty, 1999).

11 2 As the percent of rodents with hantavirus increased, so did the risk for humans in the area contracting the virus as they were more likely to come into contact with an infected rodent. ENSO EVENTS AND HANTAVIRUS As the realities of global climate change begin to play out, one factor to consider is the greater number of extreme weather events expected to occur. One type of extreme weather event mentioned often are ENSOs. These events typically occur every 2-7 years; however, they may begin to occur more frequently and could become more severe (National Oceanic and Atmospheric Administration [NOAA], 2010a). While many people claim a possible link between ENSO events and rodent populations or hantavirus positives in the rodents in the Four Corners region, there is very little research on the subject, and there is no literature on the link between these two events in Southern California. Should global climate change truly suggest a greater number and intensity of these ENSO storms, the effect on vector-borne diseases will be important to understand in order to plan public health interventions. Therefore, this study will evaluate the proportion of hantavirus positive rodents tested at various trap sites throughout San Diego County following ENSO events compared to normal weather years from OBJECTIVES An anecdotal association between hantavirus and the El Nino-Southern Oscillation in the Four Corners region of the United States is often asserted. However, to date, there are only two studies regarding a link between hantavirus in humans following ENSO events in the Four Corners region (Englethaler et al., 1999; Hjelle & Glass, 2000). Another study examined seasonality and hantavirus in rodents in Montana (Luis, Douglass, Mills, & Bjornstad, 2009). However, only one study has assessed a possible link between hantavirus positive rodents and ENSO events, also in the Four Corners region (Glass et al., 2002). The purpose of this thesis project was to determine if there was a statistically significant association between ENSO events and rodent positives for hantavirus in San Diego County. The specific objectives were to: 1. Compare the percentage of positive rodents identified in El Niño episodes versus otherwise normal weather event years. 2. Determine if there is any statistically significant association between time after an El Niño episode and increase in percentage of positive rodents.

12 3 CHAPTER 2 LITERATURE REVIEW HANTAVIRUS Hantavirus belongs to the Bunyavirdae family of viruses which contains bunyavirus, nairovirus, plebovirus, tospovirus and hantavirus. Hantavirus is the only virus to have nonarthropod vectors in this family. Within each virus strain, there is a large range in infection capability as well as fatality, and this is very evident in the hantavirus example (Childs, Glass, Korch, & Arthur, 1988; Nichol et al., 1993). The hantavirus genus contains properties that require it to be regarded as a bio-safety level 4 virus, indicating handling of this virus requires the highest precaution as it is known to be aerosol transmitted. Those who handle viruses at this level require special training, and labs require additional features to accommodate the possible hazards (Botten, Nofchissey, Kirkendoll-Ahern, & Rodriguez- Moran, 2000). Hantavirus is a relatively small, single stranded RNA negative virus (Dearing & Dizney, 2010). There are currently 40 differently named hantavirus genotypes in the Americas alone, seventeen of which are seen in North America, and six of which seventeen strains are known to infect humans (Mills, Amman, & Glass, 2009). In the Americas, Hantavirus Pulmonary Syndrome (HPS) is the primary concern. However, hantavirus originated as, and is still observed in other world regions, as Hantavirus Renal Syndrome (HRS), a hemorrhagic fever. The virus was given its name from the Hantan River Valley in South Korea where the first cases were seen in the 1950 s where UN soldiers were infected. Approximately 3200 soldiers became very ill with an unknown hemorrhagic fever (Hart & Bennett, 1999). Most cases were observed in the spring and summer, but no cause was identified until 1978 by Dr. Ho-Wang Lee and his colleagues. HRS identification occurred more than 20 years after the first cases. The virus was identified as a member of the Bunyavirdae family and named hantavirus (Lee, Lee, Johnson, 1978). The largest differences between HPS and HRS exist in the manifestations of the diseases, the geographical occurrences, and the vectors responsible for the spread. While both HPS and HRS begin with fairly non-specific symptoms and a febrile illness, they

13 4 progress into two very different areas. As suggested in the name, HPS is a disease largely affecting the pulmonary region. This is observed on a chest X-ray by clouding of fluid. Without proper and immediate supportive care, patients can face pulmonary failure (Schmalijohn & Hjelle, 1997). HRS on the other hand, moves from being a febrile illness to creating large gastrointestinal issues for the patient with some neurological involvement. Patients with HRS may experience headaches, low blood pressure and possible kidney failure that may result in shock (Zheniqiang, Formenty, & Roth, 2008). Since the two manifestations attack different systems in the body, fatality rates differ. The range for HRS fatality is typically 1-12% and depends on the strain of hantavirus (Zheniqiang et al., 2008), HPS typically has a higher fatality rate around 35% (Lambin, Tran, Vanwambeke, Linard, & Soti, 2010; Mills, Ksiazek, Peters, & Childs, 1999). The two manifestations of HPS and HRS are the result of not only different viral strains of hantavirus, but different vector species that are capable of carrying the hantavirus strains. There is often one specific rodent species for one specific virus strain. Many times spillover is observed, and more than one rodent becomes capable of carrying the virus (Schmalijohn & Hjelle, 1997). HRS is most often transmitted by a rat or something from the vole species (Lambin et al., 2010), on the other hand, HPS is typically carried and transmitted by mice (Childs et al., 1994). Hantaviruses have been noted in many geographic locations; however, HPS and HRS are not observed everywhere. Typically, only one or two strains of hantavirus are ever seen in one geographical location. Additionally, HPS and HRS are not often seen in the same region, resulting in a regionality of hantaviruses. As a result, most HRS is observed in the Korean Peninsula, Russia, and North Western Europe (Zheniqiang et al., 2008). Conversely, HPS is commonly observed in South America, specifically Chile and Argentina, as well as the United States (with a favor for the western states). There are very few cases of HPS or HRS seen in climatic or geographic extremes (Mills et al., 1997). Ultimately, the expression of the virus by means of number of positive rodents, will be dependent on the specific location and the specific year since there are a number of factors to consider for any location (Mills et al., 1999).

14 5 FOUR CORNERS SIN NOMBRE VIRUS OUTBREAK HPS has been present in the United States for a number of years; however, most cases have been sporadic. In 1993 an outbreak of a disease, pulmonary in nature, was seen in the Four Corners region of the United States where the states of Utah, Colorado, New Mexico, and Arizona meet. The onset of HPS is nonspecific, and it was difficult to identify HPS as the reason for this particular outbreak (Centers for Disease Control and Prevention [CDC], 1993b). In addition, this outbreak was the result of an HPS strain that had previously never been seen, was later named Sin Nombre virus (SNV), and is still only detected in North America (CDC, 2004). The very first human case in the Four Corners outbreak, was a young male that had been previously very healthy. He passed away due to the pulmonary complications of HPS. The second case was the fiancé of the first man, and without this connection it would have taken much longer for proper attention to be given to the cases. This connection between the two cases prompted greater review of the case histories and involvement of state government. As more cases arose and the state agencies could not detect the agent responsible, the CDC special pathogens branch was brought in to aid in the identification of the responsible agent (CDC, 1993a). The initial cases continued to infect otherwise healthy, young individuals, and had an extremely high mortality rate of 75% due to the difficulty providing supportive care to an unknown virus (Nichol et al., 1993). Soon the CDC was involved, the virus was identified as a hantavirus but the particular strain had not been observed before. The strain was then labeled as SNV or virus without a name. The original proposed name had been Four Corners Virus but due to inaccuracies since the virus is present in other regions, and objections from residents, SNV was used instead (Nichol et al., 1993). The entire procedure occurred over a matter of weeks. Although weeks allowed for a number of human cases to occur, the original hantavirus took over 20 years to identify, in comparison, this procedure happened incredibly quickly (CDC, 1993a). This outbreak was the first time this strain had been seen on a large scale in the United States but, there have been indications that singular cases have resulted in human fatalities prior to the 1993 outbreak. Reverse transcriptase PCR has been used in order to retrospectively diagnose cases from years past (Simmons & Riley, 2002). Using this method

15 6 a human death from 1975 (Wilson, Hjelle, & Jenison, 1994) as well as one from 1978 were able to be given a cause of death of hantavirus SNV strain (Simmons & Riley, 2002). The same reverse transcriptase PCR process used on humans was to test the blood of mice that had died in Kern County, CA in From this there were 75 mice that tested positive for the SNV strain as far back as These positive mice have allowed for comparison of the virus strain from what was present in 1975 to what is currently present in the United States. While most viruses experience natural mutations over time, the rate of mutation for the SNV strain has proven to be extremely high. Coupled with the lack of data and information about the virus prior to 1993, it very possibly existed in the United States and caused human fatalities even prior to what is currently understood (Wilson et al., 1994). To this day, the Four Corners region and the four states that comprise the Four Corners have the highest human cases of SNV, or hantavirus in general across the country (see Appendix A; Englethaler et al., 1999). In addition to having the greatest number of cases, this region is also where the bulk of the SNV research is based. Much of this has to do with natural reservoirs in the animals that occupy the area, coupled with the presence of the ideal habitat for the main vectors of SNV. Both Peromyscus maniculatus and Reithrodontomys megalotis favor pinyon juniper woodland areas, open grassland or desert scrubland along with limited human contact, the Four Corners region provides for these needs of the species (Schmalijohn & Hjelle, 1997). Although there is a preference for the Four Corners region, there have been SNV human cases in most states with a favoring of the western United States. Despite the dramatic favoring of the west coast, there is a large unexplained paradox in relation to Oregon, which sees far fewer cases than any of the states surrounding it (see Appendix A). An understanding of local SNV patterns and presence is advised since host avoidance is currently still the only prevention method as no vaccine is available (CDC, 2009). HANTAVIRUS (HPS) TRANSMISSION The transmission of hantavirus from rodent to human occurs through aerosolization of virus particles, often from feces or urine. Humans inhale the aerosolized virus particles and become infected (Mills et al., 1997; Boone et al., 2002). The chance of human infection depends on the amount of interaction with the vectors responsible for hantavirus in the area

16 7 and the proportion of positive rodents. Each area normally has only one strain of hantavirus observed, and only one vector species is capable of transmitting each strain. Transmission to other species does exist so that more than one rodent species can transmit the same hantavirus strain (Shmalijohn & Hjelle, 1997). It is suggested this interaction between species and virus strain exists because the particular virus strains of hantavirus and the rodents species capable of transmitting them have co-evolved. This has helped the virus expand with the rodent species as they move to other locations (Simmons & Riley, 2002). The general trend observed, however, is that rats and voles are capable of transmitting viruses in the HRS family (Lambin et al., 2010), and mice are capable of transmitting viruses in the HPS family (Childs et al., 1994). Within HPS, SNV is the most commonly observed strain in the western United States with the typical 35% fatality of HPS cases. The main vector is the Deer Mouse (Peromyscus maniculatus) with spill over into the Western Harvest Mouse (Reithrodontomys megalotis). These mice generally avoid people and crowded areas. They favor being in the wild or at the very least, living in the urban fringe. Other mice in the Peromyscus family have been known to carry hantavirus, but these two species have been responsible for a great majority of positive antibody tests (Childs et al., 1994; Schmalijohn & Hjelle, 1997). Although the infection is potentially fatal to humans, it presents a chronic, but not life threatening infection to its rodent hosts. Rodents transmit hantavirus to other rodents through horizontal means, indicating the virus goes from rodent to rodent but is not passed from parent to child. Activities that bring rodents into contact with one another increase the chance of hantavirus transmission within rodents. Some of the factors that relate to a heightened chance of a rodent testing positive for the hantavirus antibody are: (1) being male, (2) presence of scars, (3) weight and (4) older age. If a rodent is male, he is more likely to have a greater travel distance, cover more land, and come into contact with more mice, specifically, more hantavirus positive mice. Presence of scars on a rodent indicate history of fighting with other mice which easily lends to transmission of the virus. Older age accounts for the fact that the longer a mouse lives the more likely he or she would encounter and contract the virus in time since it is a chronic infection. On the same note, the heavier a rodent is, the more likely the rodent is older and has come into contact with the virus (Mills et al., 1999). The increase in a rodent s need to forage for food also increases chance of contracting the virus.

17 8 While males in general are more likely to forage, decreased food availability and need to travel further for food proves to increase chances of hantavirus transmission (Root et al., 1999). Pregnancy provides a protective factor to mice in utero. Mothers produce specific antibodies to hantavirus to keep their babies free from the virus through weaning (Buceta, Escudero, de la Rubia, & Lindenberg, 2004). This also allows for no mother-to-child transmission of the virus to occur should the mother be hantavirus positive, therefore, all rodents are born hantavirus free (Simmons & Riley, 2002). In addition to individual factors playing a role in increased rates of HPS presence, seasonality influences when HPS will peak in a rodent population. A large part of this has to do with the nature of the breeding cycle of rodents and increased population density that results from then end of the breeding season. As a result, the highest host population for HPS is noted at the end of the breeding season in the summer. As previously mentioned, these individuals will be born HPS free due to the presence of the HPS antibody passed along by the mother so the population at this point in time, will be largely young and uninfected individuals. Therefore, the highest potential for host population infection will be noticed in spring, before the young enter into the population, and the age structure is shifted more towards elder rodents who are more likely to be infected (Mills, 2005). There will be a varied time lag between introduction of new rodents into the community after the breeding season and the peak in HPS presence in the rodents. This period allows for a three-step process used to describe how HPS is acquired in a population. Following the summer breeding, recruitment of young uninfected rodents is observed, followed by winter. The second phase occurs once winter mortality ceases and the population stabilizes out once again. The final phase occurs when the rodents who have survived the winter begin to spread out, foraging for food, allowing for horizontal spread of the virus (Madhav, Wagoner, Douglass & Mills, 2007). As the seasons change and food becomes more or less available, the rodents will have an increased need to forage for their food. Generally, mice eat insects in the spring and seeds in the fall, largely due to seasonal availability (Klien & Calisherz, 2007). However, in the majority of San Diego County, there is not as significant seasonality to always dictate seasonal eating patterns as in other areas of the country (Bennett et al., 1999). The harsher

18 9 the season acts on food sources, the more the rodents will need to travel great distances to acquire food. The further the rodents travel the more likely they will come into contact with other infected rodents and become HPS positive themselves (Root et al., 1999). In addition to considering the effects of seasonality and climate on the hosts themselves, there is a significant role that predators play on the rodent population and thus on the amount of HPS in a population (Buceta et al., 2004). A loss or decline of a predator species will result in a rodent population larger than expected for the region. This increased population size will become increasingly difficult to control. The increased density of the rodents in the population will in turn result in a higher prevalence of HPS in the community (Dearing & Dizney, 2010). These changes in HPS presence in rodents from the change in the predator population or to extreme weather, such as El Niño events, are all ultimately considered irregular disturbances. The typical disturbances observed year to year are considered to be from normal weather patterns (Zell, Krumbholz, & Wutzler, 2008). The importance of understanding seasonal patterns of infection is translated into human risk of HPS acquisition should they come into contact with rodents. The greater the presence of HPS rodents in an area, or during a season, the more likely human cases will occur (Calisher et al., 2007). Although the species largely responsible for HPS tends to stay away from heavily populated areas, there are instances where the rodents will enter human structures especially in rural areas. After large rodent population growth, rodents are more likely to enter sheds, barns, cabins and homes when seeking shelter. Similarly, during human habitat expansion, rodents occasionally will not leave an area as quickly as humans enter into it and end up in human structures. Lastly, during cases of flooding, rodents are often driven from their burrows and occasionally seek shelter in the higher ground of man-made structures (Gubler et al., 2001). It is during these times that monitoring for rodent activity in buildings is particularly important in an area known to have HPS. EL NIÑO SOUTHERN OSCILLATION The El Niño Southern Oscillation (ENSO) is a cyclical warming and cooling of sea surface temperatures in the Pacific Ocean. The warming of the El Niño events rotates with the cooling of the La Niña events; these cycles typically take two-to-seven years to complete.

19 10 When the average temperature of three months exceeds the typical average surface temperature, an El Niño event is said to occur (Dearing & Dizney, 2010). Warming of ocean surface temperatures in the Pacific Ocean translates to large-scale weather changes in the areas around the Pacific. The winters of an ENSO season see excess precipitation in dry climates, specifically in the desert regions of the southwest United States. At the same time, areas that often experience high levels of precipitation, like the tropics, see drier winters during ENSO seasons. The effect of a La Niña season is the reverse due to the cooling of the sea surface temperatures. As a result, the dry areas remain dry and the wetter areas keep high precipitation (NOAA, 2010a). The basis of ENSO events are the sea surface temperatures. To date, monitoring of sea surface temperature is still the best way to predict ENSO events. Currently there are several high tech models that use a variety of inputs to predict when an ENSO will occur. However, it has been proven that low tech monitoring of sea surface temperatures averaged over time from various locations still provide the best prediction of ENSO events. These models have displayed an ability to predict ENSO events up to twelve months in advance. Accurate and advanced prediction can allow for appropriate preparation to be made for precautions relating to ENSO events (Halide & Ridd, 2008). ENSO events are rated by departure from an average sea surface temperature. Once the three-month sea surface temperature average increases by 0.5 C or more, an ENSO event is said to be occurring. ENSOs are rated as mild, moderate or severe based on how much of a departure from the average occurs. A mild ENSO occurs if the departure is C from the average, a moderate ENSO is C from the average and a severe ENSO is 1.4 C or more from the average (NOAA, 2010a). ENSO AND VECTOR-BORNE DISEASE A similar thesis project was performed by Cheng (2009) regarding ENSO events and mosquito vectors for West Nile Virus in San Diego County. Cheng (2009) demonstrated an increase in mosquito populations following the heavy ENSO precipitation in San Diego County. ENSO events are often associated with an increase in vector-borne diseases due to the resulting increase in flooding observed in many regions. Of all vectors, the mosquito is the most temperature and weather sensitive, responding quickly to changes in precipitation

20 11 and temperature highs and lows with ticks being a close second in most weather sensitive species (Klempa, 2009). While rodents are not as weather sensitive as mosquitoes and ticks, flooding can often drive them from their burrows creating a greater contact with humans (Epstein, 2002). Additionally, ENSO events create a plant surge that allows for an increase in rodent population from the excess food resources (Gubler et al., 2001). The excess precipitation from ENSO events provides locations for female mosquitoes to lay eggs, and the warmer temperatures allow the life cycle to progress faster producing more mosquitoes in the area (Epstein, 2002). Due to the increased breeding, vector-borne diseases like malaria and dengue fever often see large spikes after ENSO flooding events (Hu, Clements, Williams, & Tong, 2010; Lafferty, 2009). HANTAVIRUS AND ENSO EVENTS Although rodents do not have the same sensitivity to temperature and precipitation mosquitoes experience, they do demonstrate that they are affected by extreme weather events. Most vectors possess a seasonal breeding pattern that is affected by temperature, rodents are no exception to this as seen in their seasonal breeding patterns (Gubler et al., 2001). Additionally, rodents are more likely to have a period of hibernation in areas where temperatures get cold enough to warrant it. When events occur that prolong the breeding season of rodents, such as warmer temperatures, the rodents presence in an area will increase, and their breeding success will likely increase as well (Glass et al., 2002). In areas like desert grassland, similar to much of San Diego s ecosystem, rodent populations tend to peak during winter months and decrease in the summer during typical weather events (Mills et al., 2009). Rodent populations have been found to be highest in the year following the heavy precipitation associated with an ENSO event (Glass et al., 2002; Luis et al., 2009). Since HPS is horizontally transmitted, the increase in HPS in mice would not occur immediately but could be observed in the following calendar year when the rodent population increases. Human infection begin to occur when infected rodents begin to travel into human structures seeking shelter coincides with increases in rodent population density. Interviews with hantavirus-infected individuals suggest that most infections occur inside structures but, not in open areas (Englethaler et al., 1999).

21 12 Only one study (Glass et al., 2002) examined the increase in hantavirus positive rodents in relation to ENSO events, and that study noted this hantavirus rodent positive increase to occur the year following the ENSO event. This time delay allows for the plant increase from the excess precipitation to support a larger rodent community, which in turn fosters the increase of hantavirus since increased rodent density is associated with increased hantavirus (Glass et al., 2002). In other regions, precipitation and temperature play a more crucial role in the survival of the rodents. Studies performed in regions that experience freezing temperatures display a temperature effect on rodent survival (Luis et al., 2009), something not often observed in temperate climates such as San Diego County. Additionally, these regions experience winter snow. Therefore, precipitation effects can remain important for a longer window of time, since it can take months for the snow to fully melt. However, the central Montana region used in the study by Luis et al., 2009 still experienced the increase in rodent food sources following precipitation events and an increase in rodent populations. In the one study that utilized hantavirus positive rodents and ENSO events (Glass et al., 2002) rodents in the Four Corners region were trapped using a live capture method as was done in San Diego County and tested for hantavirus antibodies. That study used satellite imagery to establish low, moderate and high risk regions based on the levels of herbaceous vegetation and put traps out overnight at a mix of the risk sites. A total of 15,042 trap nights occurred at 38 trap sites during 1998 and 1999 following an ENSO event that occurred in The majority of the rodents were of the genus Peromyscus (77.8%). During 1998, the percentage of rodents positive for hantavirus was 6.6% for all risk sites, but the year following the ENSO event (1999) the presence increased to 8.3% in moderate risk areas and 30.8% in high risk areas. Human cases experienced a similar time delay between when an ENSO event occurred and when infection cases increased. Two studies evaluated the connection between increase in human hantavirus cases and ENSO events (Englethaler et al., 1999; Hjelle & Glass, 2000). Similar to rodent positives, human cases have been found to increase in the year following the ENSO event, particularly during late spring through summer. Englethaler et al. (1999) examined the initial 1993 outbreak of hantavirus in relation to the ENSO event prior to it. The researchers plotted the human cases of the outbreak by date of

22 13 occurrence and noted an initial bump a year following the ENSO event, then a negative correlation between human cases and time after the ENSO event. The ENSO event allowed for this relationship between an ENSO event and increase in HPS cases, that were evaluated. It also helped that this event was the strongest ENSO event that has been observed in the last 50 years. As a result, the typical six cases seen in the area per year increased more than five-fold to 33 cases (Hjelle & Glass, 2000). In this study the researchers assessed the precipitation at locations where SNV infection was presumed to occur, and also noted the deviation of the precipitation during the ENSO event from average precipitation at that location was plotted against the distribution of SNV cases. Human cases were plotted by month to estimate the lag time between ENSO event and beginning of human outbreak. However, a 2-sample t-test calculated compared number of SNV cases against precipitation departure from the average, the total cases expected to be attributed to the outbreak were used. CLIMATE CHANGE AND VECTOR-BORNE DISEASE Global climate change indicates a change in average global temperatures. On the most basic level this will lead to a change in average ocean temperatures and weather patterns (Epstein, 2005). Temperature and precipitation changes may be drastically different from area to area with some locations far more vulnerable than others. The effects could ultimately impact land use and encourage the habitats of various vectors in new regions (Beniston, 2002). As previously mentioned, the mosquito vector is extremely temperature and precipitation sensitive. Small increases in average temperature can prolong mosquito breeding seasons resulting in increased mosquito presence in an area (Epstein, 2005). As a result of the warming trends already observed, the species responsible for malaria and dengue fever have been able to increase their worldwide presence, living in areas once too cold for them to survive in (Beniston, 2002; Sutherst, 2004). It has been proposed that global climate change is largely responsible for the increase in deaths from infectious disease since the 1970 s. This is said to occur because of the link between climate and the host defense, immunity, pathogens, vectors, and habitats (Epstein, 2002).

23 14 As non-vector species have difficulty surviving in an area, the diversity of an area begins to decline, and the vector species takes over. With this, the vector species is able to increase its density due to lack of predators and competition for resources. The greater the population density of a vector species in an area, specifically with horizontally transmitted viruses, the greater the presence of the virus in the vector population and the more dangerous this vector population becomes to the humans in surrounding areas (Dizney & Ruedas, 2009). Another outcome of global climate change to consider are the increased extreme weather events projected to occur. It has been proposed that the numerous heat waves, floods, and severe hurricanes during the past decade, have been early effects of global climate change (Epstein, 2002). One of the severe weather events of concern regarding global climate change are ENSO events. As global average temperature increases sea surface temperatures will also increase and the possibility of meeting the temperature requirements for an ENSO event will occur more often. Additionally, temperatures will likely stray from the normal range a greater amount, thus producing more severe ENSO events than previously seen. Lastly, there is a possibility these ENSO events may not produce heavy precipitation in the same locations they currently do but may alter their patterns (Haines & Patz, 2004; Sutherst, 2004). The increased likelihood and severity of ENSO events are an important consideration of global climate change in relation to vector-borne diseases, particularly hantavirus. The connection between ENSO events and hantavirus has been proposed since the 1993 outbreak. The initial outbreak was said to have occurred because of an ENSO event providing increased food supply for rodents in the Four Corners region as a result of the excess precipitation (Klempa, 2009). HANTAVIRUS VECTORS AND TRANSMISSION CAPABILITIES In the United States there are 40 hantavirus genotypes known with six known to cause HPS, but SNV is still the largest strain of concern in the country (Mills, 2009). Given the SNV strain, Peromyscus maniculatus and Reithrodontomys megalotis are the vectors largely responsible for carrying and transmitting the virus with some spillover to other species (Childs et al., 1994; Schmalijohn & Hjelle, 1997). Both Peromyscus maniculatus and Reithrodontomys megalotis have been observed across the US with P. maniculatus

24 15 displaying a very wide range (see Appendix B) and often in San Diego County. Through regular trapping of rodents in the County by the County Vector Control Program, a consistent presence of the Peromyscus maniculatus and Reithrodontomys megalotis has been noted (County of San Diego Vector Control Program, ). Although these species have a general dislike for highly populated areas, they are still often found close to human developments as long as there is some space for the rodents to burrow and a food supply. In Figure 4 (see Appendix C) a trapping location where Peromyscus maniculatus were trapped is displayed. The region in which the rodents were trapped is relatively close to a housing development, and thus positive rodents would have provided a human health risk. This example of urban fringe is where the majority of rodents are trapped by the San Diego County Vector Control Program. Rodent age, gender, size, foraging, and scaring all tend to be correlated with whether or not the rodent is positive for hantavirus since it is a horizontally transmitted virus (Mills et al., 1999). As such, the older a rodent is, the more rodents it has come into contact with (indicated by need to forage, scars and male sex) the more likely the rodent will test positive for hantavirus. This is not to say males and females do not have the same inherent possibility for infection. However, the males higher amount of interaction with other rodents and aggression that puts them at greater risk for hantavirus infection (Clay, Lehmer, Previtali, St. Jeor & Dearing, 2009). It should also be noted that not all infected rodents are equally responsible for infecting other rodents. There is an 80/20 rule of HPS infection in mice indicating that roughly 20% of the infected mice are responsible for 80% of the hantavirus infections (Clay et al., 2009). While the reasoning for this is not completely understood, it has been proposed that the 20% have greater rodent contacts, not that they possess any more of an infectious strain of HPS than the other rodents (Clay et al., 2009). FOOD SUPPLY FOR SOUTHERN CALIFORNIA RODENTS In the Four Corners region, where SNV was first identified, the rodent diet is largely comprised of the piñon nut (Stone, 1993). San Diego County rodents feed on a large variety of plants and grasses throughout the area with their diet typically insect heavy in the spring and focusing on nuts and seeds in the fall (Ramirez-Hernandez, & Herrera, 2010).

25 16 San Diego County has a highly variable rainfall based on location. The coastal region sees an average of ten inches (254mm) of rain per year (NOAA, 2010b). This low level of precipitation requires plants that are largely drought tolerant. However, when larger storms do hit the area, such as ENSO events, there is a large pulsed productivity of plant growth and large increase in plant cover from the excess precipitation. The pulsed productivity also allows for an increase in the insect population as well, another food source for the rodents (Polis, Hurd, Jackson & Pinero, 1997). During an ENSO year the flowering cycle of plants will often speed up, and in turn, increase the overall population of these plants during the ENSO event. As these are part of the rodent diet, the increase in this food supply can aid in increasing rodent populations. The cycle of plant growth following ENSO rains will be altered as a die off of the excess growth will occur in the year following without continued excess rains to sustain the increased growth (Franke et al., 2006). COUNTY OF SAN DIEGO SURVEILLANCE AND TESTING The County of San Diego Department of Environmental Health works to provide education, regulation and enforcement of various environmental health issues ranging from hazardous materials to vector-borne diseases. The Vector Control Program (VCP) under the Department of Environmental Health educates the public on vector-borne diseases, responds to public need for vector inspections, controls vector populations and performs routine surveillance for vector-borne diseases known to exist in the County. Following the outbreak of HPS in the Four Corners region of the United States the San Diego County VCP, under the division of Environmental Health, began performing surveillance for HPS in rodents. Surveillance had been performed for other vector-borne diseases by the division prior to the HPS outbreak. Other vector-borne diseases consistently monitored have been: plague in ground squirrels, tularemia in the Pacific Coast Tick, and Lyme disease in the Western Black-legged tick. Following the onset of HPS surveillance, West Nile virus (WNV) was discovered in San Diego County, and the seasonal testing began for WNV in mosquitoes and sentinel chickens in 2003 (Conlan, personal communication, January 13, 2011). Although the most recent human case of HPS in San Diego County was in 2004, surveillance is still regularly done to be mindful of where the virus may be present.

26 17 In order to perform surveillance for HPS surveillance technicians set live-capture traps regularly at various spots throughout the county. The following day the technicians return to collect the traps and any rodents that may have been captured. Only Peromyscus maniculatus and Reithrodontomys megalotis species are known to carry HPS in the San Diego region. But due to spillover, any species under the Peromyscus genus, along with the Reithrodontomys megalotis and the Myodes califoricus are tested and are used for surveillance. Any other trapped rodents are set free. Captured rodents are transported to a secluded region where they can be euthanized and have a blood sample collected to be submitted to the County Veterinarian. The County Veterinarian will perform an enzyme linked immunosorbent assays (ELISA) test for the IgG antibody in order to assess whether or not the rodent was positive for hantavirus. All data regarding species, sex, reproductive status, location of capture and antibody test result are recorded in a County database for record keeping. In the event of a positive rodent a press release is sent out and the case is reported to the California State Health Department.

27 18 CHAPTER 3 METHODS ABUNDANCE DATA The Vector Control Program of San Diego County s Department of Environmental Health conducts surveillance for a variety of vector-borne diseases. Hantavirus surveillance is performed year-round through the trapping and sampling of species capable of carrying hantavirus. The goal of surveillance is to understand where hantavirus is present in the County and when the number of positive rodent cases may increase to ultimately reduce the number of human cases of hantavirus. RODENT TRAPPING Trap sites were located at a variety of locations throughout the county. Surveillance personnel set traps two days of the week at two different trap locations. Sherman live traps were used, baited with grain containing weevils and left overnight in the field in order to take advantage of the rodents nocturnal nature. Traps are placed in runs of eight with each trap roughly five feet away from other traps. The traps were best placed when hidden from human view, low in the grass, in shrubs, or with branches covering them. Placement was also suggested to be close to possible water sites, rodent tracks, or burrows which would indicate the area had rodents in the vicinity. All traps had proper warning labels on them indicating they were for a disease study in an attempt to keep people away and discourage trap theft. Traps were collected approximately 24 hours after they were set out. Proper personal protective equiptment (PPE) for those removing traps were long sleeves and pants, gloves and a mask to cover the nose and mouth. Only certain species of rodents are capable of carrying and transmitting hantavirus, therefore species known to not carry hantavirus, such as the Neotoma cinerea, or wood rat, were released if found in a trap. The only species sampled were Myodes califoricus, Reithrodontomys megalotis, and any Peromyscus species. Any traps found with a live animal meeting the species criteria for hantavirus testing were

28 19 separated from empty traps and taken to a remote location where the bleeding procedure could be performed. RODENT SAMPLING Locations for sampling, or bleeding the rodents, were often separate from the trapping sites as it was vital to keep bleeding procedures out of public view and as private as possible. Once a remote location was established and secured, tools for bleeding were prepared, specifically a biohazard collection container and a container for used sharps. Those bleeding rodents were required to wear personal protective equiptment (ppe) consisting of: Tyvex coveralls, Tyvex boot covers, a Tyvex hood with under layers tucked into the coveralls, two pairs of latex gloves (one pair with the coveralls tapped around and the other over these gloves), and a powered air purifying respirator (PAPR) attached to the Tyvex hood with a fully charged battery (see Appendix D). Transferring a rodent for the bleeding procedure was accomplished by placing a thick plastic bag over the end of the trap and opening the door slightly, shacking to first remove any excess food in the trap. The food was then put in the biohazard waste receptacle. Following food removal the rodent was taken out by putting the plastic bag over the trap again, but opening the door completely while shaking out the rodent into the bag. In order to euthanize the rodent, two q-tips were soaked in halothane and dropped into the bag with the rodent until movement ceased. The rodent was then removed from the bag for bleeding which was performed by grabbing the skin on the back of the neck tightly holding the rodent ventral side up. The syringe was inserted at a shallow angle under the sternum until it reached the heart where blood was then drawn (Appendix E). In the event lung fluid was drawn the syringe was discarded and a new syringe was used. Blood samples were taken until 200ul of blood was collected. After bleeding, the rodent was combed for fleas and ticks. If ticks or fleas were found, they were collected and submitted for testing as well, fleas would be tested for plague, ticks would be tested for Lyme disease or tularemia depending on the species. Following bleeding, information on the rodent s species, sex and reproductive status were recorded. Disposal of the syringe was done in the sharps container and the rodent s body was placed in a plastic bag with the rodent s identification number

29 20 written on the front. The rodent s body was submitted to the County Veterinarian along with the blood sample collected. Once all rodents were bled the blood samples and rodents were secured in a cooler that was turned into the County Veterinarian. Cleanup procedures were able to begin once all rodents and blood were secured. All containers, tools, trays and the Tyvex hoods were sprayed with a disinfectant in order to kill and possible virus before the PPE was removed. Once all tools used were scrubbed and put away, the PAPR could be turned off and disconnected and the hood removed which was be placed in the sun during the remaining part of clean up. The gloves, Tyvex coveralls and Tyvex booties all went into a biohazard sealed bag while the hood was dried of disinfectant and used again. LABORATORY TESTING The mice that had been bled along with the corresponding blood samples were stored in the County of San Diego Veterinary refrigerator until proper antibody testing on the blood could be performed. The County of San Diego ran ELISA tests which assessed presence of IgG antibodies in the blood samples from the tested rodents. An antibody kit produced by the University of New Mexico for the purpose of hantavirus surveillance was used by the County (Conlan, personal communication, January 13, 2011). A rodent testing positive with IgG could have indicated a very recent exposure or from years ago. Currently the IgG method is the only method to assess presence of hantavirus in rodents (Bego et al., 2008). DATA MAINTENANCE All data on rodent testing was entered into the County s surveillance database. Information regarding species, sex, reproductive status, location of trap and whether the antibody test was positive or negative for hantavirus was collected. In the event of a positive hantavirus test, the titer level was also recorded. CLIMATE DATA Climate data regarding precipitation were obtained though the archived data from the National Oceanic and Atmospheric Administration (NOAA) Western Regional Climate Center. The data used came from San Diego s Lindbergh Field weather station, located

30 N, W at 15ft (4 m) elevation, office # This location was fairly central to many of the testing sites and had the most consistent readings of all the stations in San Diego County with only four data days missing from the period of data used. Annual precipitation was calculated based on the precipitation-year which is counted from October 1 st through September 30 th (NOAA, 2010b). ENSO events were defined by the NOAA with the Oceanic Niño Index (ONI) as an increase in the mean sea surface temperature across the central Pacific. The El Niño is the warming phase of the ENSO where La Niña is the cool phases where sea surface temperatures are lower than usual. An El Niño event is said to be occurring when the three month average sea surface temperature is higher by more than 0.5 o C from the average. Episodes are said to be weak ( departure), moderate ( departure) or strong (>1.4) (NOAA, 2010a). Since La Niña events were not considered in this project, years were classified either as El Niño, when the ONI was greater than 0.5, or a normal weather year, anything less than 0.5 on the ONI. Variation in the severity of the El Niño event was not considered for evaluation. A year was classified and coded as an ENSO year if it was the latter of the two years contained in an ENSO season. ENSO events occur during the winter that spans two calendar years, the second of the two years was coded an ENSO event. Given the data from NOAA, (2010a), these years would be: 1995, 1998, 2003 and The year following the second year was the year where rodent populations would display increased hantavirus positives due to the lag time required in their breeding cycles to respond to the extra food produced by the extra precipitation (Buceta et al., 2004). STATISTICAL ANALYSIS All statistical analyses were done using SPSS Version 17 and data were not considered to be statistically significant unless p<0.05. In order to evaluate the possible effect of an El Niño season on hantavirus positives, a calculation of odds ratio was made. Data were categorized as El Niño years or normal weather events allowing for construction of a 2X2 table of the total positive and negative rodents tested from these years. The odds ratio was calculated with a 95% confidence interval, i.e., as long as all values in this interval were above 1.0 the results could be said to

31 22 be significant given the p<0.05. In order to determine if there was a time lag between El Niño event and increase in hantavirus, an odds ratio was also calculated for the year following the ENSO event. Comparison of the precipitation was calculated through two sample t-tests. For precipitation contrasts, the respective means of years, both normal and during El Niño events, were used lending this data to be used in a parametric test. The two sample t-test examined the differences between the means of the two groups to see if a significant difference in precipitation or temperature could be found. All calculations were performed using the data from all years To determine the average percent of hantavirus present during ENSO, non-enso and the years following an ENSO, an average of those years were taken. During the period, four years occurred during this time frame where there were no hantavirus positive rodents trapped. Since hantavirus never left San Diego County, evident in the fact positives were displayed in years following, it can be assumed the percentage of positive rodents was simply too small to be accounted for with the sample sizes collected during these years. Given this, it was assumed that the percentage of hantavirus positive rodents in the population was less than whatever one positive would have been for the number of rodents tested in those years. This turned out to be: 7.7%, 5.9%, 1.0% and 1.2% for the years 1997, 1998, 2003 and 2005, respectively. In order to perform statistical testing a concrete percentage for these years needed to be assigned, not to simply assume less than a certain value. The average annual percentage (2.7%) was assigned to years 1997 and For Years 2003 and 2005, the percentage of hantavirus rodents was estimated at 0.5% as it was lower than the tested proportion but still allowed for hantavirus to exist in the County.

32 23 CHAPTER 4 RESULTS The number of animals trapped per year ranged from 13 to 978, and percentages of hantavirus positive rodents ranged from less than 1% to 7.2% (Table 3; see Appendix F). Twelve different species of rodents tested positive for hantavirus. However, P. maniculatus and R. megalotis were the most common with 34 and 35 positive rodents, respectively, out of the 107 total positives from the time period. There were no other notable patterns for positive species (Table 4; see Appendix F). During the years 1994 to 2009 four ENSO events occurred, and as precipitation-years are calculated from October 1 st to September 30 th these events were assigned to years 1995, 1998, 2003 and 2007, which is the year following the start of the ENSO event. An odds ratio for hantavirus positives was calculated using a 2 X 2 table first for the number of positive rodents during an ENSO event by calendar year compared to all other weather years (see Appendix G) (OR 1.35, CI , not significant). Following this, the number of positive rodents for the years following ENSO events (1996, 1999, 2004, and 2008) were compared to all other weather years (see Appendix H) to assess if there was a greater likelihood for rodents being hantavirus positive in the year following an ENSO event. When comparing the number of rodents positive for hantavirus in the year following an El Niño event as compared to all other years sampled, rodents in the year following El Niño years were almost three times more likely to test positive for hantavirus than rodents during normal weather events (OR 2.96, 95% CI ). (Table 1). ENSO events are typically associated with increased precipitation in areas that do not normally see a great deal of precipitation. San Diego County is an example of a region that routinely does not receive a great deal of precipitation, but during an ENSO event, is believed to receive a higher than normal amount (NOAA, 2010a). In order to determine if this was the case for the ENSO years in this study, precipitation from ENSO years was compared using the October 1 st to September 30 th period against precipitation from all other

33 24 Table 1.Chi Squared/Odds Ratios for Hanta and El Niño Years Years Compared Odds Ratio 95% CI Lower Boundary 95% CI Upper Boundary ENSO Event Year vs. Normal Weather Year After ENSO Event vs. All Other Years years. The mean precipitation during normal weather years was 8.1 inches (206mm) +/- SD 5.0 inches (127mm) 4.0 inches (102mm) less than the mean precipitation during ENSO event years of 12.1 inches (308mm) +/- 6.4 inches (163mm), but this difference was not significant (p = 0.35) (Table 2). Table 2. Averaged Precipitation and Positives Normal Weather Events El Niño Southern Oscillations Year After ENSO Events Average Precipitation Average Average Average Average Percent Precipitation SD Positives # Tested # Tested Percent Positive Inches SD Positive SD 8.05 (206mm) 5.0 (127mm) (308mm) (163mm) (153mm) (25mm) Precipitation in inches was evaluated for the correlation with hantavirus positive rodents in the following year independent of the status of the year as an ENSO or non-enso event. From this plot a regression analysis was performed indicating that the greater the precipitation, the higher the percentage of hantavirus positive rodents in the following year

34 25 (R 2 = , p < 0.05). It should be noted that the precipitation data for 2005 where there was 22.6 inches of rain but only 3.3% of rodents positive for hantavirus the following year is very much an outlier (see Figure 1). It could be due to the fact that the County only had two surveillance personnel working in 2006 to test rodents (Table 5, Appendix F), and that blood was not drawn if it was raining. If this point was removed the new R 2 value becomes October to October Rainfall and Percent Hantavirus Positive Rodents of Following Year 8 Percentage Positive Rodents Rainfall in Inches Figure 1. Relationship between percentage of hantavirus positive rodents and annual precipitation from the previous year.

35 26 CHAPTER 5 DISCUSSION With San Diego County data, there was not a significant increase in the percentage of mice positive for hantavirus during El Niño years (OR 1.35, 95% CI , Table 1). However, the year following an ENSO event did display a significant increase in hantavirus positive rodents trapped (OR 2.96, 95% CI ). These results are similar to those in the one other study found that examined the relationship between hantavirus positive rodents and ENSO events (Glass et al., 2002). Glass et al. (2002) studied hantavirus positive rodents in the Four Corners in relation to the ENSO event of the years and found an increase in infected rodents. Glass et al. (2002) also used live trap sampling and antibody testing; however, satellite imagery was also used in order to assess vegetation of various locations to determine the risk level of the different regions where trapping occurred. Additionally, random sampling from the various trap sites was done in order to randomize the testing. The other two studies regarding human hantavirus cases and El Niño events have been performed in the Four Corners region, and the effects of heavy precipitation may have a more drastic effect on the vegetation and hantavirus vectors in this region than in San Diego County (Englethaler et al., 1999; Hjelle & Glass, 2000). The increase of hantavirus cases observed following El Niño years, has been attributed to pulsed vegetation production. This allows for an increased food supply for the rodents which provides them the resources to increase their populations. With larger populations, population density increases and hantavirus is more likely to be spread within the rodent community by horizontal means. Therefore, increased rodent density often results in increased hantavirus prevalence amongst rodents. Additionally, it has been suggested that heavy precipitation can draw rodents from their burrows (Epstein, 2002). In this instance that could have increased the numbers by allowing rodents to come into contact with other rodents and increase transmission of hantavirus. There are very few studies that have studied the association between hantavirus and ENSO events, and none that have looked at this through long-term rodent testing in a California region. The current research either uses records of human cases to compare with

36 27 ENSO events (Englethaler et al., 1999; Hjelle & Glass, 2001), monitors the overall rodent population (Luis et al., 2009), or has done very short periods of rodent testing for hantavirus (Glass et al., 2002). In Hjelle and Glass (2000), the number of human HPS cases in the Four Corners region were observed in comparison to excess precipitation during the ENSO event. In this region, the University of New Mexico Medical School is regarded as an excellent HPS diagnostic and treatment facility. Due to this, the majority of HPS cases are taken to this hospital if possible. Data for their study were taken from UNM s Medical School s records as well as from the state health departments for New Mexico, Arizona, Utah, and Colorado. Clearly there is a large difference between the study performed using data from rodents trapped in San Diego County and human case data. However, in the Four Corners region, HPS is far more common, with typically four to six human cases occurring per year. At the same time, San Diego County has not recorded a human case since 2004, so this would not have been a very useful measurement tool for a study based out of San Diego. While markers for HPS presence differ a great deal between these two studies, hantavirus positive rodents for the current study as opposed to human cases of hantavirus, weather data were collected in a very similar fashion. Hjelle and Glass (2000) utilized NOAA s website and database as well in order to determine ENSO events and used 310 different weather stations to determine if precipitation during the ENSO event was greater than the 20 year average. From this, the number of cases was compared against the deviation from the average precipitation. If a correlation existed between HPS and ENSO events, then the greater the deviation the more HPS cases there should be. As expected, during the ENSO event HPS cases rose dramatically, and there were 33 compared to the expected six. These 33 cases included a one year lag time from the end of the ENSO event, only slightly longer than the lag time observed in the present study. Another study utilizing human HPS cases was undertaken (Englethaler et al., 1999) focused on the HPS cases in the Four Corners region during the initial 1993 outbreak that also followed an ENSO event. Climate data were used from 52 locations that were thought to be probable exposure sites, and at these locations precipitation and temperature from the previous ten years was assessed. A Wilcoxon matched pairs analysis was performed for the difference in precipitation between the ENSO event and years after ( ) and monthly

37 28 precipitation averages for A Spearman s correlation between time after the ENSO event and number of human HPS cases indicated a negative correlation, revealing that the more time passed after the ENSO event, the fewer HPS cases were seen. It should be noted this decline with time did not occur immediately after the ENSO event occurred, but following the initial surge in HPS cases. A similar rise and decline was observed in the data from the present study where the year following an ENSO event saw increases in rodent positives, but no other time period shared this phenomena. In addition a definite spring/summer peak was noted for human cases, which has often been associated with humans entering into summer cabins and doing spring cleaning. In the current study the peak in rodent hantavirus positives was observed earlier in the year during the months of January, February and March (see Appendix F). It is possible the high rodent positives occur as a result of increase in density which prompts the rodents to expand their habitat into human structures where humans then become exposed. However, since the present study and Englethaler et al., (1999) evaluated two different regions, it is also possible the seasonal patterns of rodent and human infection differ between the two regions. In one study that attempted to asses the effects of seasonality on deer mouse population in Montana, seasonality played a very large role, although this did not particularly look at ENSO events and their effects in the population (Luis et al., 2009). In this study, live traps were set out for three nights every month between 1994 and 2004 in Montana grassland. Climate data were gathered from a nearby weather station to determine the temperature and amount of precipitation. The area s mouse population was estimated based on the number of captures and their ages. Luis et al. s, (2009) use of age class structure provided a picture of the entire mouse population. From this study it was discovered that any precipitation in the previous five months appeared to affect rodent populations. However, temperatures varied a great deal at these sites, and much of the precipitation came in the form of snow which could take months to fully thaw. Therefore, the time for precipitation effects to play out in Montana would greatly differ from those in San Diego County because the timeline for precipitation presence differs greatly. The one study that looked at hantavirus in populations of mice in the Four Corners region, happened to do so during and after the ENSO event (Glass et al., 2002). This study relied on satellite imagery to indicate whether specific regions would be at low,

38 29 moderate or high risk for hantavirus presence in the mice due to the vegetation presence. Pixels were used to categorize the risk of different areas. People who were unaware of the risk level of the various locations were used to pick rodent sampling sights. Forty sites with different risk levels were chosen and standardized trapping protocols used to live trap rodents at those sites. All traps were placed ten m apart, and longitude and latitude were recorded. Any rodents trapped were tested using antibody testing as is performed in the County of San Diego. Since this sampling extended over an ENSO event, patterns of infection were able to be assessed during and after the excess precipitation. During the sampling time period 15,042 trap nights were conducted at 38 of the original 40 different trap sites. From these trap sites 13 species of rodents were found with 77.8% being Peromyscus with P. maniculatus the most common. The capture rates of the traps was greater in the year following the ENSO event (1999) with a trap success of 6.4% as opposed to 4.4% trap success during Along with trap success, infection presence was also greater during the year following the ENSO event. In 1998 the infection presence in rodents was 6.6% in the high and low risk areas but in 1999 the moderate risk areas moved to 8.3%, and the high risk areas rose to an infection presence of 30.8%. This can easily be contrasted with the current study where infection presence never exceeded a high of 7.2% and averaged at 2.7%. One reason for this difference arose from the difference between the ecosystems in the Four Corners region and that of San Diego County. When surveillance was done of the vegetation in the various risk areas, the high risk was associated with Ponderosa pine and Pinon pine. On the other hand, the low and moderate risk areas had a great deal of sagebrush, and saltbrush which are common plants at trap sites for the San Diego County Vector Control Program. The differences in the ecosystems between the Four Corners and San Diego County are possible reasons why the Four Corners is at higher risk for hantavirus and why hantavirus is not only not as common in San Diego, but why there may not be as strong of a increase after an ENSO event. A number of the limitations of this study arose from the fact that secondary data were utilized. Due to this, a number of factors were not as standardized as they would normally be for a study specifically designed to assess hantavirus infection in regards to ENSO events. For example, the County of San Diego does not have set policies and procedures regarding how rodent traps are supposed to be set up in order to maximize the number of rodents

39 30 trapped at any given time. Training of new technicians is done by the senior technicians, and much of what is conveyed is simply what people perceive to work best. This leads to each individual utilizing slightly different methodologies for trap placement and different returns on trap successes for every technician. Had every technician followed the same guidelines for trap locations placement the resulting number of rodents trapped may have been slightly different and the types of rodents trapped (rodents with certain behaviors) may have been more or less likely under more standardized conditions. Although San Diego County currently employs three surveillance technicians, this has certainly not always been the case. Until 2000 there was only one individual responsible for all surveillance in the County. Although WNV surveillance did not begin until it was noted in 2004, this still meant one person was responsible for hantavirus, plague, Lyme disease, and tularemia surveillance. It is hard to believe that one individual was able to adequately perform surveillance for all of these vector-borne diseases, partially evident in the years where fewer than 100 rodents were tested for hantavirus. These low numbers are cause for concern when using any statistical testing, creating worry over the accuracy of the results. In addition, the years following 2000 resulted in the hiring and loss of surveillance technicians almost every year (table 6; Appendix F). As previously mentioned, new staff do not have any policies and standardized procedures to follow regarding rodent trap placement. Since much of the training is through learning on the job, the first few months often result in few captures and many missed surveillance opportunities. Another limitation arose from the County policy regarding what to do with rodents in traps when it is raining. Since all rodent bleeding is performed outside, bleeding cannot be done if it is raining. Should traps be set and it is raining the following day, a surveillance technician will retrieve the traps and release any possible rodents in the traps since they cannot be bled. Due to the increased precipitation during ENSO events, there is a possibility that a greater number of rodents were released during those years than other years, artificially bringing down the proportion of rodents positive for hantavirus since they could not be tested. One of the biggest factors regarding the County s trapping procedures that limited the research findings, were the methods used to decide on locations for setting traps. Currently, it is up to the discretion of the technicians to decide where traps will be set and collected, there

40 31 is no set list of sites that must be visited or method to assess the entire county is being represented by this sampling. As a result, many of the rodents tested are trapped in areas near the County Operations Center in Kearny Mesa, or near the North County Office in San Marcos. Setting traps near the offices ensures the technicians are able to do more than simply set traps on a given day. However, many of the areas of concern for hantavirus, are in the East County region where there are fewer people and a geography that favorable to the vector species of hantavirus. It is anecdotally believed that the proportion of hantavirus positive rodents is very high in this region and had the County been including this region in its surveillance testing the overall positive proportion for hantavirus could have been far higher. In addition, since this region experiences even more precipitation than the coastal areas of San Diego, the effects of an ENSO event, would likely produce an even greater response from the rodent community. Lastly, weather data were taken from only one weather location for statistical sampling. The County of San Diego is a fairly mild but does experience some variability, and the use of one very coastal weather station may not have been the most representative of the entire County. Considering the probable association between increased hantavirus positives and El Niño events future research is necessary. In future research it would be helpful to develop set policy and procedure for rodent trap laying as well as determined trapping locations and schedule for trapping. Some studies ensured proper surveillance of a region by sectioning it into a grid and creating a schedule to make sure each grid had rodents trapped and sampled (Glass et al., 2002; Luis et al., 2009). This would eliminate the bias towards only sampling the western San Diego County and make sure the entire county could be represented and accounted for. It would also be prudent to mandate at least one trapping day per week in order to eliminate the chance of sampling a low number of rodents. Lastly, creating a method to bleed rodents in the event of rain would reduce the number of captured rodent that would have to be released and increase the number that could be sampled in any given year.

41 32 CHAPTER 6 CONCLUSIONS The main goal of this thesis was to examine if there was a significant difference between the number of hantavirus positive rodents during ENSO years and normal weather years. This study found that hantavirus positive rodents during El Niño periods were not significantly greater than positives during normal weather events based on an odds ratio of 1.35 and 95% confidence interval extending from 0.82 to Additionally, the percentage of positive rodents was associated with time after an El Niño event. The year following the El Niño events experienced a significant increase in hantavirus positive rodents noted by the odds ratio f 2.96 and 95% confidence interval The fact that this association was found indicates precautions regarding hantavirus exposure should be taken the year following El Niño events. Additionally, the proposed increase in ENSO events with global climate change and the association of hantavirus positives following ENSO events indicate global climate change could have an effect on rodent based vector-borne diseases and have large scale public health impacts.

42 33 REFERENCES Bego, M. G., Bawiec, D., Dandge, D., Martino, B., Dearing, D., Wilson, E., & St. Jeor, S. (2008). Development of an ELISA to detect Sin Nombre virus-specific IgM from deer mice (Peromyscus Maniculatus). Journal of Virological Methods, 151, Beniston, M. (2002). Climatic change: Possible impacts on human health. Swiss MedicineWeekly, 132, Bennett, S. G., Webb, J. P., Madon, M. B., Childs, J. E., Ksiazek, T. G., Torrez-Martinez, N., & Hjelle, B. (1999). Hantavirus (Bunyaviridae) infections in rodents from Orange and San Diego counties, California. American Journal of Tropical Medicine and Hygiene, 60(1), Boone, J. D., McGwire, K. C., Otteson, E. W., Debaca, R. S., Kuhn, E. A., & St. Jeor, S.C. (2002). Infection dynamics of sin nombre virus after a widespread decline in host populations. American Journal of Tropical Medicine and Hygiene, 67(3), Botten, J., Nofchissey, R., Kirkendoll-Ahern, H., & Rodriguez-Moran, P. (2000). Outdoor facility for quarantine of wild rodents infected with hantavirus. Journal of Mammology, 81, Buceta, J., Escudero, C., de la Rubia, F. J., & Lindenberg, K. (2004). Outbreaks of hantavirus induced by seasonality. American Physical Society, 69, 1-8. Calisher, C. H., Wagoner, K. D, Amman, B. R., Root, J. J., Douglass, R. J., Kuenzi, A. J., Mills, J. N. (2007). Demographic factors associated with prevalence of antibody to Sin Nombre virus in deer mice in the western United States. Journal of Wildlife Diseases, 43(1), Centers for Disease Control and Prevention. (1993a). Emerging infectious diseases update: Hantavirus disease United States, MMWR Weekly, 42(31), Centers for Disease Control and Prevention. (1993b). Outbreak of acute illness southwestern United States, MMWR Weekly, 42(22), Centers for Disease Control and Prevention. (2004). Tracking a mystery disease. Retrieved from Centers for Disease Control and Prevention. (2006). Distribution of Peromyscus maniculatus and location of HPS cases as of May 9, Retrieved from Centers for Disease Control and Prevention. (2009). Hantavirus pulmonary syndrome in five pediatric patients four states, MMWR Weekly, 58(50), Centers for Disease Control and Prevention (2010). Hantavirus pulmonary syndrome (HPS) cases by state. Retrieved from noframes/epislides/episls.htm.

43 Cheng, T. J. (2009). Effect of the El Niño-Southern Oscillation (ENSO) on mosquito vector of the West Nile Virus (WNV) in San Diego, California. (Unpublished master s thesis). San Diego State University, San Diego, California. Childs, J. E., Glass, G. E., Korch, G. W., & Arthur, R. R. (1988). Evidence of a human infection with a rat-associated hantavirus in Baltimore, Maryland. American Journal of Epidemiology, 127, Childs, J. E., Ksiazek, T. G., Spiropoulou, C. F., Krebs, J. W., Morzunov, S., Maupin, G. O., Enscore, R. E. (1994). Serologic and genetic identification of Peromyscus Maniculatus as the primary rodent reservoir for a new hantavirus in the southwestern United States. Journal of Infectious Disease, 169(6), Clay, C. A., Lehmer, E. M., Previtali, A., St. Jeor, S., & Dearing, M. D. (2009). Contact heterogeneity in deer mice: Implications for Sin Nombre virus transmission. The Royal Society, 276, County of San Diego Vector Control Program. ( ). [Animal testing for vector-borne diseases]. Unpublished raw data. Dearing, M. D., & Dizney, L. J. (2010). Ecology of hantavirus in a changing world. Annals of the New York Academy of Sciences, 1195, Dizney, L. J., & Ruedas, L. A. (2009). Increased host species diversity and decreased prevalence of Sin Nombre virus. Emerging Infectious Diseases, 15(7), Englethaler, D. M., Mosley, D. G., Cheek, J. E., Levy, C. E., Komatsu, K. K., Ettestad, P., Bryan, R. T. (1999). Climatic and environmental patterns associated with hantavirus pulmonary syndrome, four corners region, United States. Emerging Infectious Disease, 5(1), Epstein, P. R. (2002). Climate change and infectious disease: Stormy weather ahead? Epidemiology, 13(4), Epstein, P. R. (2005). Climate change and human health. New England Journal of Medicine, 353(14), Franke, D. M., Ellis, A. G., Dharjwa, M., Freshwater, M., Fujikawa, M., Padron, A., & Weis, A. E. (2006). A steep cline in flowering time for Brassica Rapa in southern California: Population-level variation in the field and the greenhouse. International Journal of Plant Science, 167(1), Glass, G. E., Yates, T. L., Fine, J. B., Shields, T. M., Kendall, J. B., Hope, A. G., Mills, J. N. (2002). Satellite imagery characterizes local animal reservoir populations of Sin Nombre virus in the southwestern United States. Ecology, 99(26), Gubler, D. J., Reiter, P., Kristie, L. E., Yap, W., Nasci, R., & Patz, J. A. (2001). Climate variability and change in the United States: Potential impacts on vector-and rodentborne diseases. Environmental Health Perspectives, 109(2), Haines, A., & Patz, J. A. (2004). Health effects of climate change. Journal of the American Medical Association, 291(1),

44 Halide, H., & Ridd, P. (2008). Complicated ENSO models do not significantly outperform very simple ENSO models. International Journal of Climatology, 28, Hart, C. A., & Bennett, M. (1999). Hantavirus infections; Epidemiology and pathogenesis. Microbes and Infections, 1, Hjelle, B., & Glass, G. (2000). Outbreak of hantavirus infection in the four corners region of the United States in the wake of the El Niño-Southern Oscillation. Journal of Infectious Diseases, 181, Hu, W. B., Clements, A., Williams, G., & Tong, S., (2010). Dengue fever and El Niño/southern oscillation in Queensland, Australia: A time series predictive model. Occupational and Environmental Medicine, 67(5), Klempa, B. (2009). Hantaviruses and climate change. European Society of Clinical Microbiology and Infectious Diseases, 15, Klien, S. L. & Calisherz, C. H. (2007). Emergence and persistence of hantaviruses. Current Topics in Microbiology and Immunology, 315, Lafferty, K. D. (2009). The ecology of climate change and infectious diseases. Ecology, 90(4), Lambin, E. F., Tran, A., Vanwambeke, S. O., Linard, C., & Soti, V. (2010). Pathogenic landscapes: Interactions between land, people, disease vectors, and their animal hosts. International Journal of Health Geographics, 9(54), Lee, H., Lee, P., & Johnson, K. (1978). Isolation of the etiologic agent of Korean hemorrhagic fever. Journal of Infectious Disease, 137, Luis, A. D., Douglass, R. J., Mills, J. N., & Bjornstad, O. N. (2009). The effect of seasonality, density and climate on the population dynamics of Montana deer mice, important reservoir hosts for Sin Nombre hantavirus. Journal of Animal Ecology, 79(2), Madhav, N. K., Wagoner, K. D., Douglass, R. J., & Mills, J. N. (2007). Delayed densitydependent prevalence of Sin Nombre Virus antibody in Montana Deer Mice (Peromyscus Maniculatus) and implications for human diseases. Vector-Borne and Zoonotic Diseases, 7(3), Mills, J. N. (2005). Regulation of rodent-borne viruses in the natural host: Implications for human disease. Achieves of Virology, 19, Mills, J. N., Amman, B. R., & Glass, G. E. (2009). Ecology of hantaviruses and their hosts in north America. Vector-Borne and Zoonotic Diseases, 10(6), Mills, J. N., Ksiazek, T. G., Ellis, B. A., Rollin, P. E., Nichol, S. T., Yates, T. L., Childs, J. E. (1997). Patterns of association with host and habitat; antibody reactive with Sin Nombre virus in small mammals in the major biotic communities of the southwestern United States. American Journal of Tropical Medicine and Hygiene, 56(3), Mills, J. N., Ksiazek, T. G., Peters, C. J., & Childs, J. E. (1999). Long-term studies of hantavirus reservoir populations in the southwestern United States: A synthesis. Emerging Infectious Disease, 5(1),

45 Nichol, S. T., Spiropoulou, C. F., Morzunov, S., Rollin, P. E., Ksiazek, T. G., Feldmann, H., Peters, C. J. (1993). Genetic identification of a hantavirus associated with an outbreak of acute respiratory illness. Science, 262(5), National Oceanic and Atmospheric Administration [NOAA]. (2010a). ENSO impacts: Cold and warm episodes by season. Retrieved from products/analysis_monitoring/ensostuff/ensoyears.shtml. National Oceanic and Atmospheric Administration [NOAA]. (2010b). San Diego, CA cooperative weather stations. Retrieved from cpm/station.php?wfo=sgx. Polis, G. A., Hurd, S. D., Jackson, C. T., & Pinero, F. S. (1997). El Niño effects on the dynamics and control of an island ecosystem in the gulf of California. Ecological Society of America, 78(6), Ramirez-Hernandez, G., & Herrera, M. L. G. (2010). Nutritional importance of seeds and arthropods to painted spiny pocket mice (Lyomis pictus): The effects of season and forest degradation. Canadian Journal of Zoology, 88(12), Root, J. J., Calisher, C. H., & Beaty, B. J. (1999). Relationships of deer mouse movement, vegetative structure, and prevalence of infection with sin nombre virus. Journal of Wildlife Diseases, 35(2), Schmalijohn, C., & Hjelle, B. (1997). Hantaviruses: A global disease problem. Emerging Infectious Disease, 3(2), Simmons, J. H., & Riley, L. K. (2002). Hantaviruses: An overview. Comparative Medicine, 52(2), Stone, R. (1993). The mouse-pinon nut connection. Science, 262(5135), 833. Sutherst, R. W. (2004). Gobal climate change and human vulnerability. Clinical Microbiology Review, 17(1), Wilson, C., Hjelle, B., & Jenison, S. (1994). Probable hantavirus pulmonary syndrome that occurred in New Mexico in Annals of Internal Medicine, 120(9), 813. Zell, R., Krumbholz, A.,& Wutzler, P. (2008). Impact of global warming on viral diseases: What is the evidence? Current Opinion in Biotechnology, 19, Zhenqiang, B., Formenty, P. B. H., & Roth, C. E. (2008). Hantavirus infection: A review and global update. Journal of Infection in Developing Countries, 2(1),

46 37 APPENDIX A HANTAVIRUS HUMAN CASE MAP

47 Figure 2. Human HPS cases in the United States. Source: Centers for Disease Control and Prevention (2010). Hantavirus pulmonary syndrome (HPS) cases by state. Retrieved from episls.htm. 38

48 39 APPENDIX B PEROMYSCUS MANICULATUS US RANGE MAP

49 Figure 3. Range of Peromyscus maniculatus in the US and locations where hantavirus positives have been found. Source: Centers for Disease Control and Prevention (2006). Distribution of Peromyscus maniculatus and location of HPS cases as of May 9, Retrieved from noframes/epislides/episl9b.htm. 40

50 41 APPENDIX C TRAPPING LOCATION PICTURE

51 Figure 4. County of San Diego rodent trap location. 42

52 43 APPENDIX D PERSONAL PROTECTIVE EQUIPTMENT

53 Figure 5. PPE for rodent sampling. 44

54 45 APPENDIX E CORRECT SYRINGE PLACEMENT

55 Figure 6. Proper syringe placement for rodent sampling. 46