West Nile Virus in Minnesota: 2003 Update

Go to full issue: DCN, May/June 2004: Volume 32, Number 3 (PDF: 136KB/12 pages)

West Nile virus (WNV) was first documented in Minnesota in July 2002. (See “West Nile Virus: An Update for Minnesota Medical Providers” in the June 2003 issue [vol. 31, No. 4] of the Disease Control Newsletter for a review of 2002 surveillance findings and a discussion of WNV epidemiology.) Since then, the Minnesota Department of Health (MDH) and other agencies have continued WNV surveillance efforts and worked to clarify the local field ecology of the virus. During 2003, Minnesota was on the eastern edge of a large WNV outbreak in the Great Plains states*. Most of the 9,858 cases (including 262 fatalities) reported nationally in 2003 occurred in a small number of these states (Colorado, Nebraska, North Dakota, South Dakota, Texas, and Wyoming). A summary of national surveillance data for 2003 can be found at the Centers for Disease Control and Prevention (CDC) West Nile Web site. This update describes Minnesota WNV surveillance findings for 2003.

Human Case Surveillance

In 2003, 148 human cases of WNV disease were reported among residents of 56 Minnesota counties (Figure 1). Ninety-nine (67%) of 148 case-patients were diagnosed with West Nile fever (WNF), the less severe form of the disease; 25 (17%) had aseptic meningitis; and 24 (16%) had encephalitis. Four elderly WN encephalitis patients (median age, 83 years; range, 73-86 years) died from their illness. While many patients reported temporary numbness in their extremities, no cases of acute flaccid paralysis were identified. Ninety-five (64%) case-patients were male. The overall median case-patient age was 47 years (range, 2-96 years). WN encephalitis patients tended to be older (median, 74 years; range, 38-96 years), but WN meningitis patients were younger (median, 38 years; range, 2-80 years). The earliest case-patient had onset of symptoms on June 18; the latest, on October 1 (Figure 2). Similar to 2002, the peak in illness onset was from August 15, 2003 through September 15, 2003 (107 of 148 [72%] cases).

Most cases occurred among residents of western and central Minnesota (Figure 1), but not all of these patients became exposed to infected mosquitoes near their residences. Twenty-three (16%) case-patients traveled outside of Minnesota during the entire 2 weeks (the possible incubation period for WNV disease) prior to the onset of the illness. Most of these people traveled to other Great Plains states. Another 16 (11%) case-patients traveled to other states during part of the 2-week period prior to illness onset; thus, they may have been exposed either in Minnesota or outside the state. Many of these people regularly traveled to North Dakota or South Dakota. After excluding cases in which exposure to the virus may have occurred outside of Minnesota, WNV disease incidence rates were significantly higher in western and southwestern Minnesota than in eastern Minnesota, including in the Twin Cities Seven County Metropolitan Area (Figure 3).

Because 23 transfusion-associated WNV cases were identified in the United States during 2002, the blood products industry and the Food and Drug Administration created a program to screen all units of donated blood in the United States during 2003 with a newly established WNV nucleic acid test (NAT). As of January 2004, 1,027 presumptive viremic donors (PVDs) had been identified nationally in 2003 (from 6.2 million donations), and 6 probable cases of transfusion-associated WNV infection were found in blood recipients. Twenty-three PVDs were Minnesota residents. Five of them subsequently developed WNF. There has been no evidence of transfusion-associated WNV infections in Minnesota to date. Equine Surveillance During 2003, the first evidence of WNV in Minnesota was an equine WN case in Crow Wing County; its reported onset was April 10. This case was the first reported evidence in 2003 of WNV transmission to mammals in the United States. Unfortunately, equine WN surveillance data were not very useful for human risk assessment purposes in 2003, largely because WNV activity in Minnesota horses during 2003 (74 reported cases) was much reduced from 2002 (992 reported cases). Several factors likely combined to substantially reduce reported case numbers in horses, including widespread usage of the WNV equine vaccine, a greater number of horses which have built up natural antibody protection to WNV disease, and changes in diagnosis and/or reporting of equine WNV. The majority of cases reported in 2003 occurred in unvaccinated horses that lived in counties just north and/or east of the areas of highest WNV activity in humans (Figure 1). Equine WN cases peaked in late August through early September, the same time that onset of illness peaked in humans.

Avian Surveillance

During 2003, MDH continued an effort to determine areas of WNV activity by identifying dead WNV-infected wild birds. Because several bird species (e.g., American Crow, Blue Jay) are very susceptible to the virus and are rapidly killed by it, MDH has encouraged the public to report suspected bird cases to its epidemiology staff, either via the phone or through the MDH West Nile Web site. In 2003, as in 2002, the public reported thousands of dead birds (9,713, including 5,850 American Crows and Blue Jays). Of these birds, 754 were obtained for virus testing. The first WNV-infected bird of the year was identified on June 12. In July, the epizootic in birds increased substantially. MDH ultimately identified 433 WNV-infected birds. This total included 325 (79%) of 414 American Crows tested and 54 (47%) of 114 Blue Jays (Figure 1). WNV was found in birds from 63 counties across the state. On average, WNV-infected birds were found 30 days prior to illness onset of the first human case in a given county (2 to 3 weeks prior to human exposure). However, WNV-infected birds were found in 28 counties that did not have human cases. Also, 18 counties had human cases but no WNV-positive birds (or infected birds were not found until after the human cases had occurred). Thus, the resulting predictive values of WNV-positive birds predicting subsequent human illness in a county were low (positive predictive value, 48%; negative predictive value, 55%). This may be due, at least in part, to the transmission of the virus among birds by bird-feeding mosquitoes that rarely (if ever) feed on humans and logistical problems in acquiring birds from non-metro counties for testing. (See discussion below.)

Mosquito Surveillance

During 2003, MDH tested over 99,000 mosquitoes (divided into 6,436 species-, site-, and date-specific “pools”) for WNV. Fifty-nine pools of mosquitoes from 9 counties tested positive for the virus. Forty-two of 59 WNV-positive pools were Culex tarsalis, the suspected primary vector of WNV in Minnesota and many western states. Sixteen of the 17 remaining positive pools were other Culex genus mosquitoes that likely serve as vectors of the virus among birds, but seldom feed on humans. The final WNV-positive mosquito pool was Aedes vexans, a common pest of humans in Minnesota. The statewide minimum field infection rate of Cx. tarsalis was 3.6 infected mosquitoes per 1,000 mosquitoes (95% confidence interval, 1.8 to 7.2). However, infection rates were higher in mosquitoes from west central and southwestern Minnesota trapping sites, and peaked during the last week of August (Figure 4). In several parts of west central and southwestern Minnesota, Cx. tarsalis numbers and infection rates peaked just prior to the peak of human cases, in mid to late August (Figure 4).

Predictions for 2004 and Beyond

The field ecology of WNV is extremely complex. It is difficult to predict how many people will become infected with WNV in Minnesota or in the rest of North America in 2004. The virus appears to be established throughout Minnesota; it will probably be present in the state to some extent every year. The disease risk to humans, however, will likely continue to be higher in central and western Minnesota, where Cx. tarsalis is most abundant. The risk is also likely to be highest from August through mid-September. Ultimately, the risk in any given year will be largely dependent on: 1) the weather (warm weather, with timely wet and dry conditions, enhances the production of vector mosquitoes), 2) the presence of the virus in mosquitoes and susceptible birds that can act as reservoirs for the virus, and 3) human behavior patterns (the level of outdoor activity at dusk and dawn, when Cx. tarsalis is active, and the relative amount of personal measures taken by people to protect themselves from the vector mosquitoes).

For more information on West Nile virus, see the MDH's West Nile Web page.


*The Great Plains include parts of 13 states: Colorado, Iowa, Kansas, Minnesota, Missouri, Montana, Nebraska, New Mexico, North Dakota, Oklahoma, South Dakota, Oklahoma, and Texas.

Updated Tuesday, 02-Jul-2013 09:23:04 CDT