Transcript on Human West Nile Virus Epidemics in North

The Impact of Day Length and the Dispersal of the American robin (Turdus migratorius) on
Human West Nile Virus Epidemics in North America
Abstract
The threat of West Nile Virus (WNV) has become widespread in North America,
prompting extensive research into both the prediction and prevention of the disease.
Several species are involved in the persistence and transmittance of WNV, thus
understanding the species interactions which shape the epidemiological patterns of WNV
infection is paramount. Recent research suggests that seasonal increases in human WNV
cases in North America can be attributed to the migratory exodus of the American robin
(Turdus migratorius), and a corresponding shift in feeding behavior by mosquito vectors to
humans. In this study we examine human WNV case data across North America in 2007,
compare timing of epidemics to shifts in day length (an indicator of bird migration), and
contrast WNV activity between breeding and wintering ranges of T. migratorius. Our
findings do not support the claim that increased WNV activity is due to the exodus of
robins, as we found no difference between the two ranges. Our study indicates that there is
a significant link between increased WNV activity and decreased day length and
demonstrates that epidemics occurred simultaneously in all locations in 2007. We conclude
that while both bird migration and peak WNV activity are related to seasonality, the latter is
not a result of the former.
Background
West Nile Virus: West Nile Virus (WNV) is a mosquito-borne flavivirus and human,
avian, and equine neuropathogen. The virus was first identified in the West Nile district of
Uganda in 1937 and is indigenous to Africa, Asia, Australia, and Europe. In the 1990s,
human outbreaks of West Nile occurred in Europe and the Mediterranean area. WNV was
first observed in the Western hemisphere in 1999 during a meningeoencephalitis epidemic
in New York City causing 59 hospitalizations and seven deaths. (1, 2) This outbreak was an
unpredicted and startling reminder of the impact of emerging infectious diseases, even in
the United States. Since 1999, West Nile has spread west across the continent and has been
reported in every state of the continental United States.
Sara A. Koehring ’08/’09 and Jamie Fitzgerald, Ph.D. Candidate
Clark University, Worcester, MA 01610, USA
Faculty Sponsor: Dr. Todd Livdahl, Biology Department
Purpose
Research efforts have recently focused on understanding the
patterns of WNV outbreaks observed in North America as they
relate to bird migration, and the causes contributing to these
patterns. Kilpatrick et al (9) examined a seasonal 7-fold shift in
feeding preference by C. pipiens from Turdus migratorius to humans in
Maryland, USA. PCR analysis of blood meals isolated from 148
mosquitoes indicated not only a change in host frequency, but that
T. migratorius was preferred over any other bird in the area, and that
humans were the preferred mammal. Since an increase in human
WNV cases occurs during the period of robin dispersal, the
Fig. 1: American Robin (Turdus migratorius)
authors propose that their absence leads to a shift in mosquito
(photo credit: Project Wild Bird: Field Guide
to Birds of North America)
feeding behavior from birds to humans, and is largely responsible
for human WNV epidemics.
The purpose of this study was to test these claims across a broader scale by examining WNV case
data and by relating it to bird migration across North America. We predicted that if the claims make by
Kilpatrick et al (9) are true, then peaks in human WNV cases should be timed with bird dispersal, or with
factors indicating and driving bird migration. Due to the lack of comprehensive bird migration data, we
used day length as a well-established and reliable indicator of bird dispersal (12-14). Further, we
predicted that if increased WNV activity in humans were in fact due to the exodus of T. migratorius from
an area, then the intensity of the increase should be more severe in areas where most, if not all, robins
migrate than it would be in areas where many robins remain throughout the year.
Fig. 4: Distribution of Culex pipiens
(photo: NASA/Goddard Space Flight
Center, Scientific Visualization Studio)
Fig. 5: Distribution of Aedes vexans
(photo: NASA/Goddard Space Flight
Center, Scientific Visualization Studio)
Methods
We examined data from the most recent year (2007) for which complete data sets were available
at the time of this study. We selected three geographical regions based specifically on the range of T.
The West Nile Virus is an arthropod-borne
migratorius; the northern region (Ontario, Saskatchewan, Alberta and Manitoba) is considered breeding
virus, also known as an arbovirus. Arboviruses
range for T. migratorius, the middle region (Montana, North Dakota, and Minnesota) contains areas that
survive in nature through blood-feeding
are breeding and breeding/wintering ranges, and the southern region (Colorado, Kansas, and
arthropods which transmit the virus to susceptible
Missouri) is considered breeding/wintering range (Figure 6).
vertebrate hosts. When an infected arthropod takes
Human WNV data for the northern region were collected from the Public Health Agency of
a blood meal from a vertebrate the virus is passed
Canada’s West Nile Virus Monitor webpage (15), using the West Nile National Surveillance Report for
to the vertebrate (Figure 1). WNV is transmitted
2007 (24). Data for the middle and southern regions were collected from the United States Geological
via vector mosquitoes. Birds are the reservoir or
Services (USGS) Disease Maps webpage (16). Error was estimated to be 1 for Montana, 2 for
amplifying host while humans and other mammals
North Dakota, and 2 for Colorado, as data were extracted from a graph.
are typically incidental hosts. Different mosquitoes Fig. 1: West Nile Virus Transmission Cycle
(photo credit: www.fda.gov)
Day length was calculated from sunrise and sunset times acquired from the United States Naval
are considered to be the primary vectors in
Observatory (17). Data from a centralized location served as representative for each province or state.
different geographical regions. In Canada the major mosquito species involved in the
Fig. 6: Distribution of American Robin (Turdus Rise and set times were recorded for Wednesdays only, and were used as an estimate of average day
transmission cycle of WNV are Culex pipiens and Aedes vexans while in Eastern and Western migratorius). Red outline indicates Canadian
length for each week.
provinces
used,
blue
outline
indicates
northern
US the main species are Culex pipiens and Culex tarsalis, respectively. (3) Figures 2 and 3 show
Human population data for 2007 were acquired for each state from the United States Census
US states used, green outline indicates
Culex pipiens and Aedes vexans while Figures 4 and 5 illustrate their respective geographic
southern US states used.(photo adapted from: Bureau (projected by the US Census Bureau in association with the Federal State Cooperative Program
distributions.
Canadian Wildlife Service & Canadian Wildlife
for Population Estimates) (18), and for each province from Statistics Canada (projected by Statistics
Federation © 2008, www.hww.ca)
Canada based on 2001 census data) (19).
Clinical Manifestations: Approximately 80% of persons infected with WNV are
The total number of human WNV cases in each province or state was plotted against day length in order to determine and
asymptomatic and exhibit no signs of infection (4). Mild symptoms displayed in roughly 20%
compare the day length at which peak WNV activity occurred in each region. Single factor ANOVA without replication was used to
of human WNV cases include fever, vomiting, nausea, head and body aches, and in some
compare between regions the day length at peak activity, and the number of weeks post-solstice peak activity occurred. The measure of
cases, swollen lymph glands or a dermatological rash on the back, chest, and stomach. These
WNV activity in each province and state was standardized by dividing the number of reported human cases by the population. Single
symptoms typically last only a few days, but even healthy people may be ill for several weeks.
factor ANOVA without replication was used again to determine differences in this standardized measure between regions.
Approximately 1 in 150 people infected with WNV will develop severe neurological
syndromes such as meningitis, encephalitis, and flaccid paralysis (5). Symptoms resulting
from acute cases of WNV include high fever, headache, neck stiffness, disorientation, stupor,
Table 1: Summary of data and ANOVA analysis for north, middle and south regions
Peak WNV activity in all provinces and states occurred 6-9 weeks after the
North Region
Middle Region
South Region
ANOVA
coma, muscle weakness, tremors, convulsions, vision loss, numbness, and paralysis. These
summer solstice week, or the week with the longest calculated day lengths,
Ontario
14:27 Montana
13:55 Colorado
13:40
F=16.35
symptoms can persist for several weeks and neurological damage is usually permanent (4).
regardless of location or day length specific to location (see Figures 7, 8, and
15:22 N.Dakota
14:16 Kansas
13:40
p=0.002
Day length at peak Manitoba
WNV activity
Saskatchewan
15:47 Minnesota
14:16 Missouri
13:25
WNV Activity in 2007: During 2007, 3,630 human West Nile
9). While a significant difference was found between regions in regards to the
Alberta
15:22
cases with 117 fatalities were reported across 775 counties in
day length at which peak WNV activity occurred (F=16.35, p=0.002), no
Ontario
9 Montana
9 Colorado
8
F=1.62
Weeks past
Manitoba
7 N.Dakota
8 Kansas
8
p=0.264
44 states and the District of Columbia, indicating that the
significant difference was found between regions in regards to the timing of
solstice at peak
Saskatchewan
6 Minnesota
8 Missouri
9
WNV activity
virus has become well established in the United States. (6)
peak activity (F=1.62, p=0.264), measured by weeks past solstice, indicating
Alberta
7
Ontario
0.001 Montana
0.211 Colorado
0.119
F=0.86
that the difference in day length is likely a product of latitude alone. No
Human WNV
Manitoba
0.493 N.Dakota
0.577 Kansas
0.014
p=0.462
WNV and Migratory Birds: It is likely that migratory birds
cases/population
significant difference in intensity of epidemics, measured as human WNV per
Saskatchewan
1.451 Minnesota
0.019 Missouri
0.013
(1000)
Fig. 2: Culex pipiens
were responsible for the initial spread of West Nile given that
Alberta
0.092
population, occurred between regions in 2007. (see Table 1)
(photo credit: Joseph Hlasek,
the increasing number of human WNV cases along the
hlasek.com)
Atlantic seaboard from 1999 to 2000 correlate with common
migratory pathways for avian populations that summer in the
Northeastern US (7, 8). The migration patterns of birds that
are WNV reservoir and amplification hosts are important for
two reasons: (i) their role in carrying the disease across
geographical ranges and (ii) because the timing of bird
presence and absence in a given area plays a role in which
Fig. 3: Aedes vexans
(photo credit: Troy Bartlett
hosts mosquito species may have access to (9-11).
Results
© 2004, BugGuide.net)
Fig. 7: Number of Human WNV Cases Per Day Length
(in Hours) in Northern Region
Black vertical line indicates summer solstice
Fig. 8: Number of Human WNV Cases Per Day Length
(in Hours) in Middle Region
Black vertical line indicates summer solstice
Fig. 9: Number of Human WNV Cases Per Day Length
(Hours) in Southern Region
Black vertical line indicates summer solstice
Discussion
Peak WNV activity does not appear to be driven by bird dispersal in the regions
examined in this study. Human WNV cases peaked simultaneously in all three regions,
regardless of geographic location or day length, despite the expectation that the timing of
bird dispersal varies geographically, and that since birds alter their migratory behaviors in
response to local events, some irregularity in bird dispersal would be expected. We propose
that while it is likely that both bird dispersal and a shift in mosquito feeding behavior
occur in response to similar cues, the latter is not a direct result of the former. Further, if
T. migratorius dispersal is responsible for a shift in target hosts of Culex spp., and
corresponding increase in human WNV cases, then more extensive increases would be
expected in regions of heavier migration, and less extensive increases expected in regions
where birds remain. Our findings do not support bird dispersal as a cause of increased
WNV incidence, as no difference in human WNV cases per population occurred between
breeding and wintering ranges of T. migratorius.
It is, however, noteworthy that in all cases, peak WNV activity occurred 6-9 weeks
post-solstice, and that timing in relation to solstice, or the number of days post-solstice,
was more important than actual hours of light per day in determining WNV activity. The
regularity of this observation is indicative of a link between WNV activity and decreasing
day length, and warrants further investigation.
A change in host preference observed by Kilpatrick et al (16) can not be
disregarded, and further exploration may lead to an alternate explanation of the
phenomenon, or its relationship with WNV activity. Increasing mosquito numbers and
geographic range would undoubtedly prove useful, and new techniques would allow more
accurate and definitive repetition of earlier analysis (e.g., 17, 18) and its relation to modern
WNV epidemics. Studies should be performed in areas of low to no bird migration, as a
decline in avian blood meals corresponding to increases in human blood meals in such
area indicates host preferences change regardless of bird migration.
Future research efforts may also be directed at quantifying mosquito activity, and
the number of WNV positive mosquito pools, as increased WNV incidence in humans
may simply be due to a larger population or proportion of infected mosquitoes. If no
change is observed these parameters, then a change in WNV incidence may implicate
other causes such as host availability or mosquito feeding preferences. Further, seasonal
abundance of mosquitoes varies geographically and in relation to climate, thus disease is
more likely to persist in areas where mosquito activity is constant.
This study examined data spanning only one year (2007). The expansion of these
efforts to a broader temporal scale may illuminate additional factors driving the seasonality
of peaks in WNV activity, such as the influence of temperature or precipitation, or may
alter our perception of the patterns examined here.
Further investigation into mosquito responses, decisions, and actions are
undoubtedly necessary to understand the factors driving WNV epidemiology. It is possible
that the mosquitoes are responding directly and actively to decreases in day length, rather
than indirectly or passively to changes in host availability. Other behavioral changes could
account for the shift in host preference, such as slower and lower flight or activity related
to changes in temperature, humidity, day light changes, or other unknown factors.
Mosquitoes may sense or derive different elements from mammals than birds, and may
prepare for winter by altering feeding behaviors. While speculative and hypothetical
examples, these possibilities illustrate the limited knowledge that exists regarding mosquito
behavior. It is imperative that future research focus on the drivers of mosquito behaviors
if we are to truly understand mosquito-borne disease.
Conclusion
While a definite seasonal pattern to WNV incidence has been established, our data do not
support that it is the direct result of bird dispersal. It is likely that additional and, as yet,
unknown factors are responsible for the observed pattern of epidemics, and indeed that
these factors are multiple and dynamic. Given the prevalence and epidemiological
implications of WNV, the understanding of these phenomena is an important priority for
science and medicine. As always, more research is both warranted and needed, not only to
advance our understanding of West Nile Virus, but of all vector-borne diseases, and the
organisms and systems involved in their persistence.
References
1. Campbell GL, Marfin AA, Lanciotti RS, Gubler DJ. West Nile Virus. The Lancet. 2002; 2(9):519-29.
2. Huhn GD, Sejvar JJ, Montgomery SP, Dworkin MS. West Nile virus in the United States: an update on an emerging infectious disease. American Family Physician.
2003; 68(4):653-60.
3. Hayes EB KN, Nasci RS, Montgomery SP, O'Leary DR, Campbell GL,. Epidemiology and transmission dynamics of West Nile virus disease. Emerging Infectious
Diseases. 2005; 11(8):1167-73.
4. Center for Disease Control and Prevention. West Nile Virus Fact Sheet. 2005.
5. Watson JT, Pertel PE, Jones RC, Siston AM, Paul WS, Austin CC, et al. Clinical Characteristics and Functional Outcomes of West Nile Fever. Annals of Internal
Medicine. 2004; 141(5):360.
6. Center for Disease Control and Prevention. Statistics, Surveillance, and Control: Maps and Data. 2007 [cited 2007; Available:
http://www.cdc.gov/ncidod/dvbid/westnile/surv&control.htm#surveillance]
7. Reed KD, Meece JK, Henkel JS, Shukla SK. Birds, Migration and Emerging Zoonoses: West Nile Virus, Lyme Disease, Influenza A and Enteropathogens. Clinical
Medicine and Research. 2003; 1(1):5-12.
8. Rappole JH, Derrickson SR, Hubalek Z. Migratory Birds and Spread of West Nile Virus in the Western Hemisphere. Emerging Infectious Diseases. 2000; 6(4):319-28.
9. Kilpatrick AM, Kramer LD, Jones MJ, Marra PP, Daszak P. West Nile Virus Epidemics in North America Are Driven by Shifts in Mosquito Feeding Behavior. PLoS
Biology. 2006; 4(4):606-10.
10. Tempelis CH, Francy DB, Hayes RO, Lofy MF. Variations in Feeding Patterns of Seven Culicine Mosquitoes on Vertebrate Hosts in Weld and Larimer Counties,
Colorado. American Journal of Tropical Medical Hygiene. 1967; 16(1):111-9.
11. Tempelis CH, Reeves WC, Bellamy RE, Lofy MF. A Three-Year Study of the Feeding Habits of Culex Tarsalis in Kern County, California. American Journal of
Tropical Medical Hygiene. 1965; 14(1):170-7.
12. New World Encyclopedia. Bird Migration. 2008 [cited Dec 2008]; Available: http://www.newworldencyclopedia.org/entry/Bird_migration?oldid=678335
13. Pough FH, Janis CM, Heiser JB. Vertebrate Life. 6th ed. Upper Saddle River, NJ: Prentice-Hall, Inc. 2002.
14. Sparks TH. Phenology and the changing pattern of bird migration in Britain. International Journal of Biometeorology. 1999; 42(3):134-8.
15. Public Health Agency of Canada. West Nile Virus MONITOR. 2008 [cited Dec 2008; Available: http://www.phac-aspc.gc.ca/wnv-vwn/index-eng.php]
16. United States Geological Services. USGS Disease Maps. 2007 [cited 2008; Available from: http://diseasemaps.usgs.gov/index.html]
17. United States Naval Observatory: Astronomical Applications Department. Information Center: Sun or moon rise/set table for one year. 2008 [cited Dec 2008];
Available: http://aa.usno.navy.mil/data/docs/RS_OneYear.php]
18. United States Census Bureau. Population Finder. 2008 [cited Dec 2008]; Available: http://www.census.gov/index.html]
19. Statistics Canada. CANSIM Table: Population by year, province and territory. 2007 [cited 2008]; Available: http://www40.statcan.gc.ca/lo1/cst01/demo02a-eng.htm]