American Journal of Epidemiology Copyright © 2002 by the Johns Hopkins Bloomberg School of Public Health All rights reserved 1060 Vol. 155, No. 11 Printed in U.S.A. La Crosse Encephalitis Erwin et al. La Crosse Encephalitis in Eastern Tennessee: Clinical, Environmental, and Entomological Characteristics from a Blinded Cohort Study Paul C. Erwin, 1 Timothy F. Jones, 2 Reid R. Gerhardt, 3 Sandy K. Halford, 1 A. Brent Smith, 3 Lori E. R. Patterson, 4 Kristy L. Gottfried, 5 Kristen L. Burkhalter, 5 Roger S. Nasci, 5 and William Schaffner 6 A blinded cohort study was conducted in 2000 to better understand the emergence of La Crosse virus infection in eastern Tennessee, with special emphasis on the potential mosquito vector(s). Children with suspected central nervous system infection were enrolled at the time of clinical presentation at a large pediatric referral hospital. Clinical, environmental, and entomological data were collected prior to case confirmation. Sixteen of the 40 children included in the final analysis were confirmed to have La Crosse infection by a fourfold increase in antibody titers between collection of acute- and convalescent-phase sera. Factors significantly associated with La Crosse infection included average number of hours per day spent outdoors (5.9 for La Crosse virus cases vs. 4.0 for noncases, p = 0.049); living in a residence with one or more tree holes within 100 m (relative risk = 3.96 vs. no tree holes within 100 m, p = 0.028); and total burden of Aedes albopictus (number of female and male larvae and adults collected at a site), which was more than three times greater around the residences of La Crosse virus cases versus noncases (p = 0.013). Evidence is accumulating that the newly introduced mosquito species Ae. albopictus may be involved in the emergence of La Crosse virus infection in eastern Tennessee. Am J Epidemiol 2002;155:1060–5. Aedes; arbovirus infections; arboviruses; central nervous system infections; immunologic surveillance; virus diseases Received for publication May 25, 2001, and accepted for publica- tion January 11, 2002. Abbreviation: CDC, Centers for Disease Control and Prevention. 1 East Tennessee Regional Office, Tennessee Department of Health, Knoxville, TN. 2 Tennessee Department of Health, Nashville, TN. 3 Department of Entomology and Plant Pathology, University of Tennessee, Knoxville, TN. 4 East Tennessee Children’s Hospital, Knoxville, TN. 5 Division of Vector-Borne Infectious Diseases, Centers for Disease Control and Prevention, Fort Collins, CO. 6 Department of Preventive Medicine, Vanderbilt University School of Medicine, Nashville, TN. Reprint requests to Dr. Paul Campbell Erwin, East Tennessee Regional Office, Tennessee Department of Health, 1522 Cherokee Trail, Knoxville, TN 37920 (e-mail: [email protected]). La Crosse virus infection is the most commonly reported arboviral disease, and the most frequently identified cause of pediatric encephalitis, in the United States (1, 2). La Crosse is the predominant virus of the California serogroup of Bunyaviruses and, in 2000, accounted for 109 reported cases of encephalitis (Centers for Disease Control and Prevention (CDC) provisional data; (3)). Although most infections with La Crosse virus are thought either to be asymptomatic or to result in mild flulike illnesses, the majority that come to the attention of the public health com- munity are serious and result in emergency room visits and intensive care admissions (4, 5). Since the initial isolation of La Crosse virus from a human in the 1960s, most cases of La Crosse encephalitis have been reported from the upper Midwestern states, primarily Illinois, Iowa, Indiana, Minnesota, Ohio, and Wisconsin (4, 6). Over the past decade, the majority of reported cases have occurred in West Virginia; during this same time period, La Crosse encephali- tis has also been reported from North Carolina (7, 8). Most recently, La Crosse infection has been newly identified in eastern Tennessee, with a range of six to 10 cases reported annually between 1997 and 1999 (9, 10). Numerous studies have established Ochlerotatus ( Aedes) triseriatus (the eastern tree hole mosquito) as the primary vec- tor for La Crosse virus (4, 11). This species is native to the geographic areas in which La Crosse virus infections have been reported. The virus can be transmitted within this mos- quito species both vertically (transovarially) and horizontally (venereal); in addition, La Crosse virus may overwinter in Oc. triseriatus (12–14). A variety of mammals may serve as amplification hosts, including the eastern chipmunk ( Tamias striatus), grey squirrel (Sciurus carolinensis), and red fox (Vulpes fulva) (4). Humans and white-tailed deer (Odocoileus virginianus) are considered incidental, or “dead-end,” hosts that do not support viremia sufficiently well enough to infect other mosquitoes; however, since a blood meal may result in a gonadotropic cycle whereby eggs are infected, even these incidental hosts may serve to amplify the virus (15). Recently, La Crosse virus was isolated for the first time from naturally infected Aedes albopictus (Asian tiger mos- quito) reared from eggs collected around residences in east- ern Tennessee and western North Carolina where La Crosse encephalitis cases have been documented (16). This is a at Pennsylvania State University on February 27, 2014 http://aje.oxfordjournals.org/ Downloaded from
La Crosse Encephalitis in Eastern Tennessee: Clinical, Environmental, and Entomological Characteristics from a Blinded Cohort Study
Jul 24, 2022
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
erwin.qxdAmerican Journal of Epidemiology Copyright © 2002 by the Johns Hopkins Bloomberg School of Public Health All rights reserved
1060
La Crosse Encephalitis Erwin et al.
La Crosse Encephalitis in Eastern Tennessee: Clinical, Environmental, and Entomological Characteristics from a Blinded Cohort Study
Paul C. Erwin,1 Timothy F. Jones,2 Reid R. Gerhardt,3 Sandy K. Halford,1 A. Brent Smith,3 Lori E. R. Patterson,4
Kristy L. Gottfried,5 Kristen L. Burkhalter,5 Roger S. Nasci,5 and William Schaffner6
A blinded cohort study was conducted in 2000 to better understand the emergence of La Crosse virus infection in eastern Tennessee, with special emphasis on the potential mosquito vector(s). Children with suspected central nervous system infection were enrolled at the time of clinical presentation at a large pediatric referral hospital. Clinical, environmental, and entomological data were collected prior to case confirmation. Sixteen of the 40 children included in the final analysis were confirmed to have La Crosse infection by a fourfold increase in antibody titers between collection of acute- and convalescent-phase sera. Factors significantly associated with La Crosse infection included average number of hours per day spent outdoors (5.9 for La Crosse virus cases vs. 4.0 for noncases, p = 0.049); living in a residence with one or more tree holes within 100 m (relative risk = 3.96 vs. no tree holes within 100 m, p = 0.028); and total burden of Aedes albopictus (number of female and male larvae and adults collected at a site), which was more than three times greater around the residences of La Crosse virus cases versus noncases (p = 0.013). Evidence is accumulating that the newly introduced mosquito species Ae. albopictus may be involved in the emergence of La Crosse virus infection in eastern Tennessee. Am J Epidemiol 2002;155:1060–5.
Aedes; arbovirus infections; arboviruses; central nervous system infections; immunologic surveillance; virus diseases
Received for publication May 25, 2001, and accepted for publica- tion January 11, 2002.
Abbreviation: CDC, Centers for Disease Control and Prevention. 1 East Tennessee Regional Office, Tennessee Department of
Health, Knoxville, TN. 2 Tennessee Department of Health, Nashville, TN. 3 Department of Entomology and Plant Pathology, University of
Tennessee, Knoxville, TN. 4 East Tennessee Children’s Hospital, Knoxville, TN. 5 Division of Vector-Borne Infectious Diseases, Centers for
Disease Control and Prevention, Fort Collins, CO. 6 Department of Preventive Medicine, Vanderbilt University
School of Medicine, Nashville, TN. Reprint requests to Dr. Paul Campbell Erwin, East Tennessee
Regional Office, Tennessee Department of Health, 1522 Cherokee Trail, Knoxville, TN 37920 (e-mail: [email protected]).
La Crosse virus infection is the most commonly reported arboviral disease, and the most frequently identified cause of pediatric encephalitis, in the United States (1, 2). La Crosse is the predominant virus of the California serogroup of Bunyaviruses and, in 2000, accounted for 109 reported cases of encephalitis (Centers for Disease Control and Prevention (CDC) provisional data; (3)). Although most infections with La Crosse virus are thought either to be asymptomatic or to result in mild flulike illnesses, the majority that come to the attention of the public health com- munity are serious and result in emergency room visits and intensive care admissions (4, 5). Since the initial isolation of La Crosse virus from a human in the 1960s, most cases of La Crosse encephalitis have been reported from the upper
Midwestern states, primarily Illinois, Iowa, Indiana, Minnesota, Ohio, and Wisconsin (4, 6). Over the past decade, the majority of reported cases have occurred in West Virginia; during this same time period, La Crosse encephali- tis has also been reported from North Carolina (7, 8). Most recently, La Crosse infection has been newly identified in eastern Tennessee, with a range of six to 10 cases reported annually between 1997 and 1999 (9, 10).
Numerous studies have established Ochlerotatus ( Aedes) triseriatus (the eastern tree hole mosquito) as the primary vec- tor for La Crosse virus (4, 11). This species is native to the geographic areas in which La Crosse virus infections have been reported. The virus can be transmitted within this mos- quito species both vertically (transovarially) and horizontally (venereal); in addition, La Crosse virus may overwinter in Oc. triseriatus (12–14). A variety of mammals may serve as amplification hosts, including the eastern chipmunk (Tamias striatus), grey squirrel (Sciurus carolinensis), and red fox (Vulpes fulva) (4). Humans and white-tailed deer (Odocoileus virginianus) are considered incidental, or “dead-end,” hosts that do not support viremia sufficiently well enough to infect other mosquitoes; however, since a blood meal may result in a gonadotropic cycle whereby eggs are infected, even these incidental hosts may serve to amplify the virus (15).
Recently, La Crosse virus was isolated for the first time from naturally infected Aedes albopictus (Asian tiger mos- quito) reared from eggs collected around residences in east- ern Tennessee and western North Carolina where La Crosse encephalitis cases have been documented (16). This is a
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
nonnative mosquito but one that has spread rapidly after it was introduced in Texas in 1985 (17, 18).
This article describes the first known blinded cohort study examining the clinical, environmental, and entomological characteristics of the disease in an area in which it is now emerging. Data were collected immediately after a case of the disease was suspected but had not yet been confirmed.
MATERIALS AND METHODS
Active surveillance for La Crosse infection was carried out between June 1 and September 30, 2000, at a large pedi- atric referral hospital located in Knoxville, Tennessee. Children were enrolled in the study if they were aged 6 months or older, had a physician-diagnosed febrile central nervous system infection, had a lumbar puncture performed, and had no evidence of a bacterial central nervous system infection or other documented disease to explain their ill- ness. For all children meeting these criteria, an acute-phase serum specimen was obtained for testing for antibodies to La Crosse virus as part of the initial laboratory evaluation. Parents were provided information about La Crosse virus infection, and consent to participate further in the study was obtained. The involvement of human subjects (e.g., obtain- ing acute-phase and convalescent-phase serum samples) fol- lowed established guidelines for investigating reportable diseases in Tennessee.
Parents who gave their consent were interviewed by a public health nurse within 1 week of the child’s hospital dis- charge. These interviews focused on personal or household characteristics such as use of insect repellent, length of time children spent outdoors, and presence of window screens in the residence. Clinical data were obtained from hospital records. For La Crosse virus cases only, information on symptoms at discharge was obtained from medical records.
Convalescent-phase serum specimens were obtained 2–4 weeks following onset of illness. Laboratory testing of paired acute-phase and convalescent-phase serum samples was per- formed at the Knoxville Regional Laboratory of the Tennessee Department of Health. All samples were tested for immunoglobulin G antibodies to La Crosse virus, St. Louis encephalitis, eastern equine encephalitis, and western equine encephalitis viruses by use of a commercially available indi- rect immunofluorescent antibody test for arboviruses (MRL Diagnostics, Cypress, California). All samples testing posi- tive were subsequently tested for La Crosse virus at the CDC, Division of Vector-Borne Infectious Diseases, Fort Collins, Colorado, by using both enzyme-linked immunosor- bent assay and plaque reduction neutralization tests.
In addition to the tests for La Crosse virus infection, some combination of serologic and cerebrospinal fluid studies was conducted for most children enrolled in the study, including those for antibodies to Coxsackieviruses, cytomegalovirus, echoviruses, herpes, influenza, mumps, measles, and varicella-zoster. For a number of these chil- dren, their cerebrospinal fluid was also tested for enteroviruses and herpes polymerase chain reaction.
Environmental assessments were conducted around the residences of study enrollees within approximately 1 week
following hospital discharge. These assessments focused on characteristics such as the numbers and types of potential mosquito breeding containers, the maintenance condition of the house and grounds, and the predominant surrounding habitat (e.g., suburban, pasture, successional forest). Mosquitos were gathered by using standard oviposition traps to collect eggs (to rear to adults) and to collect larvae from water-filled containers (both artificial and natural); carbon-dioxide-baited traps from the CDC were used to gather adult mosquitos.
Detailed methodologies for the use of oviposition traps and for mosquito rearing from eggs have been described elsewhere (16). Briefly, 10 oviposition traps were left in place at each residence for 1 week, and one carbon-dioxide- baited CDC trap was operated for a single 24-hour period at each site. Artificial and natural containers (but not oviposi- tion traps) were sampled for mosquito larvae, and collected larvae were removed to the laboratory for rearing to adults. Eggs were gathered from the oviposition traps only. Adult mosquitoes, reared from eggs and larvae, and adults col- lected in the CDC traps were separated by species and sex and then into pools of not more than 50 mosquitoes, which were frozen at –80C for virus isolation at the CDC Division of Vector-Borne Infectious Diseases in Fort Collins. Pools of mosquitoes were processed by using the methods of Beaty et al. (19) and Dogget et al. (20) and were tested for the presence of live virus with vero cell culture plaque assay, as described by Gerhardt et al. (16).
All data were entered and analyzed by using EpiInfo soft- ware, version 6.0 (CDC, Atlanta, Georgia) (21). Categorical data were compared for confirmed La Crosse cases versus confirmed noncases by using chi-square analysis. Continuous data were analyzed by use of analysis of variance. Data shown to be not normally distributed by using Bartlett’s test for homogeneity of variance were further analyzed with the Kruskal-Wallis test for two groups.
RESULTS
Sixty children met the criteria for initial enrollment in the study. Fourteen (23 percent) were subsequently removed from the study: three children were not in eastern Tennessee during the known incubation period for La Crosse virus, other diagnoses were subsequently confirmed for four chil- dren, and the parents of seven children refused consent for further participation. Forty-six children (77 percent) were included for subsequent interviews, environmental assess- ments, and mosquito collection. Of the 42 children from whom convalescent-phase serum samples were obtained, 16 (38 percent) had laboratory-confirmed La Crosse virus infection based on a fourfold increase in their antibody titers between collection of the acute- and convalescent-phase samples. When plaque reduction neutralization tests were conducted, all 16 children were confirmed positive for La Crosse infection. Mosquitos were not collected at the resi- dences of two children because cold weather arrived as mos- quito collections were scheduled to begin.
Table 1 summarizes the demographic and clinical charac- teristics of the 40 children for whom full data could be
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
obtained. The mean age of children with La Crosse virus infection was 7.5 years, and 60 percent were male. Fever, headache, vomiting, and behavioral changes were the pre- dominant symptoms, while loss of consciousness and seizures were observed less frequently. No clinical or cere- brospinal fluid findings were significantly different between confirmed La Crosse infection cases and noncases.
As shown in table 2, La Crosse infection cases differed from noncases in the reported average number of hours per day they spent outdoors (5.9 for La Crosse infection cases vs. 4.0 for noncases, p 0.049). No statistically signifi- cant differences were noted regarding type of clothing worn, use of insect repellent, or how often windows were left open.
La Crosse virus cases were almost four times more likely than noncases to live in a residence with one or more tree holes nearby (within 100 m) (table 3) (relative risk 3.96, p 0.028). No statistically significant differences were noted for the number or type of potential artificial mosquito- breeding containers, other household characteristics (such as use of window screens), maintenance of house and grounds, predominant surrounding habitat, or distance from the forest line or road.
Results of mosquito collection activities are summarized in table 4. For three children, mosquito collection data were analyzed for a site other than the primary residence; in these situations, parents noted another more likely exposure site. The total burden of Ae. albopictus (defined as the number of
TABLE 1. Demographic and clinical characteristics of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
7.5 60
7.2 76
0.64 0.29, 1.40
Symptoms Fever Headache Vomiting Stiff neck Photophobia Behavioral changes Confusion Loss of consciousness Seizures
15 15 14 9
100 100 93 60 73 87 60 13 27
21 23 20 15† 12 15 9‡ 5† 5
84 92 80 63 48 60 39 21 20
Undefined; chi square = 0.28 Undefined; chi square = 0.51
2.47 0.94 2.03 2.79 1.67 0.70 1.25
0.39, 15.46 0.42, 2.10 0.78, 5.29 0.74, 10.49 0.74, 3.76 0.20, 2.44 0.52, 2.99
Mean value (range) Mean difference 95% CI
176 (15–683) 49 (18–86) 64 (49–84)
Cerebrospinal fluid White-cell count/mm3
Protein (mg/dl) Glucose (mg/dl)
43 7.32 2.81
128, 215 9.90, 24.54 8.10, 13.72
* CI, confidence interval. † Data missing for one child. ‡ Data missing for two children.
TABLE 2. Family and household characteristics of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean difference
Length of time at current residence (years) Average no. of hours outdoors Average no. of daylight hours outdoors
7.2 5.9 5.1
5.1 4.0 3.6
2.03 1.93 1.47
Use of insect repellent (never vs. always or sometimes)
Window screens absent in the residence 10 13
67 86
10 19‡
40 79
0.50 0.62
* p = 0.049. † CI, confidence interval. ‡ Data missing for one child.
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
female and male larvae and adults collected at the site) was more than three times greater around the residences of La Crosse virus cases versus noncases (p 0.013). No statisti- cally significant differences were noted regarding the total mosquito burden of Oc. triseriatus. Other mosquitoes col- lected at the study sites, and submitted for viral isolation, included Culex pipiens, Culex pipiens/restuans, Culex species, and Orthopodomyia signifera. During this study, no viruses were isolated from any of the mosquitoes collected at any of the sites.
DISCUSSION
This is the first known blinded cohort study of La Crosse virus that includes three major areas of investigation: clini- cal, environmental, and entomological. To our knowledge, Tennessee and North Carolina are the first and only loca- tions in which La Crosse virus has been isolated from wild populations of Ae. albopictus. The entomological data from this study suggest that a high level of exposure to Ae. albopictus may be a risk factor for La Crosse infection,
TABLE 3. Environmental assessment, comparing site of likely exposure, for children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean site elevation (feet‡) Permanent containers (mean no.) Disposable containers (mean no.)
1,273 12.47 41.85
0.92, 4.63 0.63, 3.10
0.33, 1.56
Tree holes nearby (≥1) House condition: not well maintained Yard condition: untidy Shade condition: forest line <50 feet
from house Distance from road: >100 feet Classification of surrounding habitat:
suburban
* p = 0.028. † CI, confidence interval. ‡ 1 foot = 30.5 cm. § Data missing on one site.
TABLE 4. Mean numbers of Aedes albopictus and Ochlerotatus triseriatus mosquitoes collected as adults, larvae, or eggs at primary exposure sites of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean difference
Ae. albopictus (mean no.) Female adults Female larvae‡ Male adults Male larvae‡
Overall mosquito burden§
Eggs from oviposition traps
Oc. triseriatus (mean no.) Female adults Female larvae‡ Male adults Male larvae‡
Overall mosquito burden§
0.59, 5.92
3.95, 19.91**
130, 147
95% CI†Variables
* Data shown to be not normally distributed by using Bartlett’s test for homogeneity of variance; Kruskal-Wallis test for two groups showed p > 0.05; ** p = 0.013.
† CI, confidence interval. ‡ Larvae collected from preexisting larval habitats. § Defined as the number of total adults trapped plus larvae.
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
which may explain in part the recent emergence of La Crosse infection in Tennessee and its appearance in subur- ban in addition to rural areas.
In this study, no clinical or laboratory characteristics dis- tinguished La Crosse encephalitis from the undiagnosed central nervous system infections in the enrolled children. The recent report from McJunkin et al. (22) on a case series of 127 children with La Crosse encephalitis also did not reveal any uniquely distinguishing clinical feature of the acute illness. In the present study, seizures occurred in 25 percent of La Crosse infection cases, a rate somewhat lower than the 40 percent noted in a study by Gundersen and Brown of 178 patients (23), the 46 percent noted by McJunkin et al., and the 42–60 percent reported by Rust et al. (24). Although seven of the 10 La Crosse infection cases for whom we had medical records data were noted to have some symptoms at discharge, only one (with tremor, decreased manual dexterity and balance, and difficulty in thinking) could be characterized as having a true neurologic deficit. McJunkin et al. reported primarily on children who had true neurologic deficits at discharge, and they noted that 12 percent of their patients had such sequelae. Rust et al. described a broader array of outcomes, including seizures, migrainous headaches, and behavioral disturbances (such as emotional lability).
Although the appearance of encephalitis may truly be similar for a variety of etiologies, there are at least two pos- sible explanations for the absence of any uniquely distin- guishing clinical characteristics of La Crosse encephalitis. First, the vast majority of non–La Crosse encephalitis cases might indeed be La Crosse encephalitis cases, not diagnosed either because they failed to develop antibodies or because the tests currently available might not be sufficiently sensi- tive. A second possibility is that the non–La Crosse encephalitis cases of disease represent one or more illnesses identical to La Crosse encephalitis that are also arthropod- borne. Both of these areas are important for further research.
Children with La Crosse infection spent a greater total number of hours outdoors—and a greater number of day- light hours outdoors—than children who were not infected, important since both Ae. albopictus and Oc. triseriatus are daytime feeders. Grimstad et al. (25) have shown that a lab- oratory population of Ae. albopictus was capable of trans- mitting La Crosse virus at rates equal to or greater than the rate observed for Oc. triseriatus. Anecdotal evidence exists that Ae. albopictus may be a more aggressive (human) biter than Oc. triseriatus (Reid Gerhardt, University of Tennessee, personal communication, 2000), further height- ening the concern about this mosquito species.
La Crosse-infected children were much more likely to live in a residence with one or more tree holes nearby. This find- ing is consistent with that of a case-control study in West Virginia conducted by Woodruff et al. (26), which found an increase in risk for La Crosse infection in children who lived in households with one or more tree holes within 300 feet (90 m) of the house. The degree to which Ae. albopictus may competitively displace Oc. triseriatus for similar breeding habitat such as tree holes is not clear. Given the rapid spread of Ae. albopictus, this is an important area for future research.
The total Ae. albopictus burden was significantly higher around residences…
1060
La Crosse Encephalitis Erwin et al.
La Crosse Encephalitis in Eastern Tennessee: Clinical, Environmental, and Entomological Characteristics from a Blinded Cohort Study
Paul C. Erwin,1 Timothy F. Jones,2 Reid R. Gerhardt,3 Sandy K. Halford,1 A. Brent Smith,3 Lori E. R. Patterson,4
Kristy L. Gottfried,5 Kristen L. Burkhalter,5 Roger S. Nasci,5 and William Schaffner6
A blinded cohort study was conducted in 2000 to better understand the emergence of La Crosse virus infection in eastern Tennessee, with special emphasis on the potential mosquito vector(s). Children with suspected central nervous system infection were enrolled at the time of clinical presentation at a large pediatric referral hospital. Clinical, environmental, and entomological data were collected prior to case confirmation. Sixteen of the 40 children included in the final analysis were confirmed to have La Crosse infection by a fourfold increase in antibody titers between collection of acute- and convalescent-phase sera. Factors significantly associated with La Crosse infection included average number of hours per day spent outdoors (5.9 for La Crosse virus cases vs. 4.0 for noncases, p = 0.049); living in a residence with one or more tree holes within 100 m (relative risk = 3.96 vs. no tree holes within 100 m, p = 0.028); and total burden of Aedes albopictus (number of female and male larvae and adults collected at a site), which was more than three times greater around the residences of La Crosse virus cases versus noncases (p = 0.013). Evidence is accumulating that the newly introduced mosquito species Ae. albopictus may be involved in the emergence of La Crosse virus infection in eastern Tennessee. Am J Epidemiol 2002;155:1060–5.
Aedes; arbovirus infections; arboviruses; central nervous system infections; immunologic surveillance; virus diseases
Received for publication May 25, 2001, and accepted for publica- tion January 11, 2002.
Abbreviation: CDC, Centers for Disease Control and Prevention. 1 East Tennessee Regional Office, Tennessee Department of
Health, Knoxville, TN. 2 Tennessee Department of Health, Nashville, TN. 3 Department of Entomology and Plant Pathology, University of
Tennessee, Knoxville, TN. 4 East Tennessee Children’s Hospital, Knoxville, TN. 5 Division of Vector-Borne Infectious Diseases, Centers for
Disease Control and Prevention, Fort Collins, CO. 6 Department of Preventive Medicine, Vanderbilt University
School of Medicine, Nashville, TN. Reprint requests to Dr. Paul Campbell Erwin, East Tennessee
Regional Office, Tennessee Department of Health, 1522 Cherokee Trail, Knoxville, TN 37920 (e-mail: [email protected]).
La Crosse virus infection is the most commonly reported arboviral disease, and the most frequently identified cause of pediatric encephalitis, in the United States (1, 2). La Crosse is the predominant virus of the California serogroup of Bunyaviruses and, in 2000, accounted for 109 reported cases of encephalitis (Centers for Disease Control and Prevention (CDC) provisional data; (3)). Although most infections with La Crosse virus are thought either to be asymptomatic or to result in mild flulike illnesses, the majority that come to the attention of the public health com- munity are serious and result in emergency room visits and intensive care admissions (4, 5). Since the initial isolation of La Crosse virus from a human in the 1960s, most cases of La Crosse encephalitis have been reported from the upper
Midwestern states, primarily Illinois, Iowa, Indiana, Minnesota, Ohio, and Wisconsin (4, 6). Over the past decade, the majority of reported cases have occurred in West Virginia; during this same time period, La Crosse encephali- tis has also been reported from North Carolina (7, 8). Most recently, La Crosse infection has been newly identified in eastern Tennessee, with a range of six to 10 cases reported annually between 1997 and 1999 (9, 10).
Numerous studies have established Ochlerotatus ( Aedes) triseriatus (the eastern tree hole mosquito) as the primary vec- tor for La Crosse virus (4, 11). This species is native to the geographic areas in which La Crosse virus infections have been reported. The virus can be transmitted within this mos- quito species both vertically (transovarially) and horizontally (venereal); in addition, La Crosse virus may overwinter in Oc. triseriatus (12–14). A variety of mammals may serve as amplification hosts, including the eastern chipmunk (Tamias striatus), grey squirrel (Sciurus carolinensis), and red fox (Vulpes fulva) (4). Humans and white-tailed deer (Odocoileus virginianus) are considered incidental, or “dead-end,” hosts that do not support viremia sufficiently well enough to infect other mosquitoes; however, since a blood meal may result in a gonadotropic cycle whereby eggs are infected, even these incidental hosts may serve to amplify the virus (15).
Recently, La Crosse virus was isolated for the first time from naturally infected Aedes albopictus (Asian tiger mos- quito) reared from eggs collected around residences in east- ern Tennessee and western North Carolina where La Crosse encephalitis cases have been documented (16). This is a
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
nonnative mosquito but one that has spread rapidly after it was introduced in Texas in 1985 (17, 18).
This article describes the first known blinded cohort study examining the clinical, environmental, and entomological characteristics of the disease in an area in which it is now emerging. Data were collected immediately after a case of the disease was suspected but had not yet been confirmed.
MATERIALS AND METHODS
Active surveillance for La Crosse infection was carried out between June 1 and September 30, 2000, at a large pedi- atric referral hospital located in Knoxville, Tennessee. Children were enrolled in the study if they were aged 6 months or older, had a physician-diagnosed febrile central nervous system infection, had a lumbar puncture performed, and had no evidence of a bacterial central nervous system infection or other documented disease to explain their ill- ness. For all children meeting these criteria, an acute-phase serum specimen was obtained for testing for antibodies to La Crosse virus as part of the initial laboratory evaluation. Parents were provided information about La Crosse virus infection, and consent to participate further in the study was obtained. The involvement of human subjects (e.g., obtain- ing acute-phase and convalescent-phase serum samples) fol- lowed established guidelines for investigating reportable diseases in Tennessee.
Parents who gave their consent were interviewed by a public health nurse within 1 week of the child’s hospital dis- charge. These interviews focused on personal or household characteristics such as use of insect repellent, length of time children spent outdoors, and presence of window screens in the residence. Clinical data were obtained from hospital records. For La Crosse virus cases only, information on symptoms at discharge was obtained from medical records.
Convalescent-phase serum specimens were obtained 2–4 weeks following onset of illness. Laboratory testing of paired acute-phase and convalescent-phase serum samples was per- formed at the Knoxville Regional Laboratory of the Tennessee Department of Health. All samples were tested for immunoglobulin G antibodies to La Crosse virus, St. Louis encephalitis, eastern equine encephalitis, and western equine encephalitis viruses by use of a commercially available indi- rect immunofluorescent antibody test for arboviruses (MRL Diagnostics, Cypress, California). All samples testing posi- tive were subsequently tested for La Crosse virus at the CDC, Division of Vector-Borne Infectious Diseases, Fort Collins, Colorado, by using both enzyme-linked immunosor- bent assay and plaque reduction neutralization tests.
In addition to the tests for La Crosse virus infection, some combination of serologic and cerebrospinal fluid studies was conducted for most children enrolled in the study, including those for antibodies to Coxsackieviruses, cytomegalovirus, echoviruses, herpes, influenza, mumps, measles, and varicella-zoster. For a number of these chil- dren, their cerebrospinal fluid was also tested for enteroviruses and herpes polymerase chain reaction.
Environmental assessments were conducted around the residences of study enrollees within approximately 1 week
following hospital discharge. These assessments focused on characteristics such as the numbers and types of potential mosquito breeding containers, the maintenance condition of the house and grounds, and the predominant surrounding habitat (e.g., suburban, pasture, successional forest). Mosquitos were gathered by using standard oviposition traps to collect eggs (to rear to adults) and to collect larvae from water-filled containers (both artificial and natural); carbon-dioxide-baited traps from the CDC were used to gather adult mosquitos.
Detailed methodologies for the use of oviposition traps and for mosquito rearing from eggs have been described elsewhere (16). Briefly, 10 oviposition traps were left in place at each residence for 1 week, and one carbon-dioxide- baited CDC trap was operated for a single 24-hour period at each site. Artificial and natural containers (but not oviposi- tion traps) were sampled for mosquito larvae, and collected larvae were removed to the laboratory for rearing to adults. Eggs were gathered from the oviposition traps only. Adult mosquitoes, reared from eggs and larvae, and adults col- lected in the CDC traps were separated by species and sex and then into pools of not more than 50 mosquitoes, which were frozen at –80C for virus isolation at the CDC Division of Vector-Borne Infectious Diseases in Fort Collins. Pools of mosquitoes were processed by using the methods of Beaty et al. (19) and Dogget et al. (20) and were tested for the presence of live virus with vero cell culture plaque assay, as described by Gerhardt et al. (16).
All data were entered and analyzed by using EpiInfo soft- ware, version 6.0 (CDC, Atlanta, Georgia) (21). Categorical data were compared for confirmed La Crosse cases versus confirmed noncases by using chi-square analysis. Continuous data were analyzed by use of analysis of variance. Data shown to be not normally distributed by using Bartlett’s test for homogeneity of variance were further analyzed with the Kruskal-Wallis test for two groups.
RESULTS
Sixty children met the criteria for initial enrollment in the study. Fourteen (23 percent) were subsequently removed from the study: three children were not in eastern Tennessee during the known incubation period for La Crosse virus, other diagnoses were subsequently confirmed for four chil- dren, and the parents of seven children refused consent for further participation. Forty-six children (77 percent) were included for subsequent interviews, environmental assess- ments, and mosquito collection. Of the 42 children from whom convalescent-phase serum samples were obtained, 16 (38 percent) had laboratory-confirmed La Crosse virus infection based on a fourfold increase in their antibody titers between collection of the acute- and convalescent-phase samples. When plaque reduction neutralization tests were conducted, all 16 children were confirmed positive for La Crosse infection. Mosquitos were not collected at the resi- dences of two children because cold weather arrived as mos- quito collections were scheduled to begin.
Table 1 summarizes the demographic and clinical charac- teristics of the 40 children for whom full data could be
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
obtained. The mean age of children with La Crosse virus infection was 7.5 years, and 60 percent were male. Fever, headache, vomiting, and behavioral changes were the pre- dominant symptoms, while loss of consciousness and seizures were observed less frequently. No clinical or cere- brospinal fluid findings were significantly different between confirmed La Crosse infection cases and noncases.
As shown in table 2, La Crosse infection cases differed from noncases in the reported average number of hours per day they spent outdoors (5.9 for La Crosse infection cases vs. 4.0 for noncases, p 0.049). No statistically signifi- cant differences were noted regarding type of clothing worn, use of insect repellent, or how often windows were left open.
La Crosse virus cases were almost four times more likely than noncases to live in a residence with one or more tree holes nearby (within 100 m) (table 3) (relative risk 3.96, p 0.028). No statistically significant differences were noted for the number or type of potential artificial mosquito- breeding containers, other household characteristics (such as use of window screens), maintenance of house and grounds, predominant surrounding habitat, or distance from the forest line or road.
Results of mosquito collection activities are summarized in table 4. For three children, mosquito collection data were analyzed for a site other than the primary residence; in these situations, parents noted another more likely exposure site. The total burden of Ae. albopictus (defined as the number of
TABLE 1. Demographic and clinical characteristics of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
7.5 60
7.2 76
0.64 0.29, 1.40
Symptoms Fever Headache Vomiting Stiff neck Photophobia Behavioral changes Confusion Loss of consciousness Seizures
15 15 14 9
100 100 93 60 73 87 60 13 27
21 23 20 15† 12 15 9‡ 5† 5
84 92 80 63 48 60 39 21 20
Undefined; chi square = 0.28 Undefined; chi square = 0.51
2.47 0.94 2.03 2.79 1.67 0.70 1.25
0.39, 15.46 0.42, 2.10 0.78, 5.29 0.74, 10.49 0.74, 3.76 0.20, 2.44 0.52, 2.99
Mean value (range) Mean difference 95% CI
176 (15–683) 49 (18–86) 64 (49–84)
Cerebrospinal fluid White-cell count/mm3
Protein (mg/dl) Glucose (mg/dl)
43 7.32 2.81
128, 215 9.90, 24.54 8.10, 13.72
* CI, confidence interval. † Data missing for one child. ‡ Data missing for two children.
TABLE 2. Family and household characteristics of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean difference
Length of time at current residence (years) Average no. of hours outdoors Average no. of daylight hours outdoors
7.2 5.9 5.1
5.1 4.0 3.6
2.03 1.93 1.47
Use of insect repellent (never vs. always or sometimes)
Window screens absent in the residence 10 13
67 86
10 19‡
40 79
0.50 0.62
* p = 0.049. † CI, confidence interval. ‡ Data missing for one child.
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
female and male larvae and adults collected at the site) was more than three times greater around the residences of La Crosse virus cases versus noncases (p 0.013). No statisti- cally significant differences were noted regarding the total mosquito burden of Oc. triseriatus. Other mosquitoes col- lected at the study sites, and submitted for viral isolation, included Culex pipiens, Culex pipiens/restuans, Culex species, and Orthopodomyia signifera. During this study, no viruses were isolated from any of the mosquitoes collected at any of the sites.
DISCUSSION
This is the first known blinded cohort study of La Crosse virus that includes three major areas of investigation: clini- cal, environmental, and entomological. To our knowledge, Tennessee and North Carolina are the first and only loca- tions in which La Crosse virus has been isolated from wild populations of Ae. albopictus. The entomological data from this study suggest that a high level of exposure to Ae. albopictus may be a risk factor for La Crosse infection,
TABLE 3. Environmental assessment, comparing site of likely exposure, for children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean site elevation (feet‡) Permanent containers (mean no.) Disposable containers (mean no.)
1,273 12.47 41.85
0.92, 4.63 0.63, 3.10
0.33, 1.56
Tree holes nearby (≥1) House condition: not well maintained Yard condition: untidy Shade condition: forest line <50 feet
from house Distance from road: >100 feet Classification of surrounding habitat:
suburban
* p = 0.028. † CI, confidence interval. ‡ 1 foot = 30.5 cm. § Data missing on one site.
TABLE 4. Mean numbers of Aedes albopictus and Ochlerotatus triseriatus mosquitoes collected as adults, larvae, or eggs at primary exposure sites of children confirmed to have La Crosse encephalitis or not to have La Crosse encephalitis, eastern Tennessee, 2000
Infection (n = 15)
Mean difference
Ae. albopictus (mean no.) Female adults Female larvae‡ Male adults Male larvae‡
Overall mosquito burden§
Eggs from oviposition traps
Oc. triseriatus (mean no.) Female adults Female larvae‡ Male adults Male larvae‡
Overall mosquito burden§
0.59, 5.92
3.95, 19.91**
130, 147
95% CI†Variables
* Data shown to be not normally distributed by using Bartlett’s test for homogeneity of variance; Kruskal-Wallis test for two groups showed p > 0.05; ** p = 0.013.
† CI, confidence interval. ‡ Larvae collected from preexisting larval habitats. § Defined as the number of total adults trapped plus larvae.
at Pennsylvania State U niversity on February 27, 2014
http://aje.oxfordjournals.org/ D
Am J Epidemiol Vol. 155, No. 11, 2002
which may explain in part the recent emergence of La Crosse infection in Tennessee and its appearance in subur- ban in addition to rural areas.
In this study, no clinical or laboratory characteristics dis- tinguished La Crosse encephalitis from the undiagnosed central nervous system infections in the enrolled children. The recent report from McJunkin et al. (22) on a case series of 127 children with La Crosse encephalitis also did not reveal any uniquely distinguishing clinical feature of the acute illness. In the present study, seizures occurred in 25 percent of La Crosse infection cases, a rate somewhat lower than the 40 percent noted in a study by Gundersen and Brown of 178 patients (23), the 46 percent noted by McJunkin et al., and the 42–60 percent reported by Rust et al. (24). Although seven of the 10 La Crosse infection cases for whom we had medical records data were noted to have some symptoms at discharge, only one (with tremor, decreased manual dexterity and balance, and difficulty in thinking) could be characterized as having a true neurologic deficit. McJunkin et al. reported primarily on children who had true neurologic deficits at discharge, and they noted that 12 percent of their patients had such sequelae. Rust et al. described a broader array of outcomes, including seizures, migrainous headaches, and behavioral disturbances (such as emotional lability).
Although the appearance of encephalitis may truly be similar for a variety of etiologies, there are at least two pos- sible explanations for the absence of any uniquely distin- guishing clinical characteristics of La Crosse encephalitis. First, the vast majority of non–La Crosse encephalitis cases might indeed be La Crosse encephalitis cases, not diagnosed either because they failed to develop antibodies or because the tests currently available might not be sufficiently sensi- tive. A second possibility is that the non–La Crosse encephalitis cases of disease represent one or more illnesses identical to La Crosse encephalitis that are also arthropod- borne. Both of these areas are important for further research.
Children with La Crosse infection spent a greater total number of hours outdoors—and a greater number of day- light hours outdoors—than children who were not infected, important since both Ae. albopictus and Oc. triseriatus are daytime feeders. Grimstad et al. (25) have shown that a lab- oratory population of Ae. albopictus was capable of trans- mitting La Crosse virus at rates equal to or greater than the rate observed for Oc. triseriatus. Anecdotal evidence exists that Ae. albopictus may be a more aggressive (human) biter than Oc. triseriatus (Reid Gerhardt, University of Tennessee, personal communication, 2000), further height- ening the concern about this mosquito species.
La Crosse-infected children were much more likely to live in a residence with one or more tree holes nearby. This find- ing is consistent with that of a case-control study in West Virginia conducted by Woodruff et al. (26), which found an increase in risk for La Crosse infection in children who lived in households with one or more tree holes within 300 feet (90 m) of the house. The degree to which Ae. albopictus may competitively displace Oc. triseriatus for similar breeding habitat such as tree holes is not clear. Given the rapid spread of Ae. albopictus, this is an important area for future research.
The total Ae. albopictus burden was significantly higher around residences…
Related Documents
