Seroprevalence of Powassan Virus Infection in an Area Experiencing a Cluster of Disease Cases: Sussex County, New Jersey, 2019

Abstract

In 2019, a geographically focal cluster of 3 Powassan virus neuroinvasive disease cases occurred in New Jersey. We conducted a serosurvey of 273 adult area residents and estimated that immunoglobulin M seroprevalence was 0.31% (95% confidence interval [CI], .04%–1.00%) and 23% (95% CI, 7%–100%) of infections result in neuroinvasive disease.

A serosurvey of 273 residents of a New Jersey community with several recent Powassan virus neuroinvasive disease cases estimated IgM seroprevalence at <1% and that roughly one-quarter of infections result in neuroinvasive disease.

Powassan virus (POWV) is a ribonucleic acid virus in the family Flaviviridae. Powassan virus is maintained in a transmission cycle between Ixodes species ticks (eg, Ixodes scapularis) and small mammal amplifying hosts, with humans being dead-end hosts [1]. Many human POWV infections are asymptomatic, but infection can also result in febrile illness or neuroinvasive disease, including meningitis and encephalitis [2]. The proportion of POWV infections that result in neuroinvasive disease is unknown.

During 2009–2018, an average of 15 US cases were reported annually [3]. Most US cases of POWV disease occur in the Northeast and Great Lakes regions, and cases usually occur from late spring through mid-fall when ticks are most active [2]. There is no vaccine to prevent POWV infection, so prevention relies on preventing tick bites.

During May–September 2019, 4 neuroinvasive POWV disease cases detected through serologic testing were reported in Sussex County, New Jersey. Two of the cases resulted in death. Before 2019, a total of 8 POWV disease cases had been reported in New Jersey, with the first reported in 2013 and 5 (63%) occurring in residents of Sussex County [3]. The first 3 2019 Sussex County POWV disease patients lived in 2 adjacent census tracts, and the fourth patient lived just outside these census tracts. Because of this geographic clustering and the unknown risk of POWV infection in endemic areas, we conducted a serosurvey to evaluate POWV seroprevalence and estimate the proportion of infections resulting in neuroinvasive disease in adults in this geographic area.

METHODS

Sample

The survey was implemented in 2 census tracts in Sussex County, New Jersey where the first 3 POWV disease patients from 2019 resided. The census tracts have a combined adult population of 4133 adults, with approximately 2 adults per household [4]. We used tax records to randomly select 423 households from our sampling frame of 2513 households. From October 11 to 22, 2019, we attempted to enroll all adults in the selected households.

Patient Consent Statement

Written consent was obtained from all participants. The Centers for Disease Control and Prevention deemed the project to be nonresearch public health surveillance, and New Jersey’s Institutional Review Board approved the project as minimal risk.

Data Collection

English-speaking adults (aged ≥18 years) who lived in the area since at least January 1, 2019 were eligible. Consenting adults were administered a survey that included demographics, potential tick exposures in 2019, and select medical information (eg, encephalitis or meningitis diagnosis in the previous 3 years). Participants also provided a blood specimen to test for POWV-specific neutralizing antibodies (evidence of past infection) and, if positive, for POWV-specific immunoglobulin (Ig)M antibodies (evidence of acute infection), as previously described [2, 5]. To rule out false-positive results because of cross-reactive antibodies to related flaviviruses, specimens with POWV-specific antibodies were also tested for antibodies to West Nile (WN) and St. Louis encephalitis (SLE) viruses.

Analysis

Categorical variables are reported as counts and proportions with 95% confidence intervals (CIs). Continuous variables are reported as medians, interquartile ranges (IQRs), range, and 95% CI. Seroprevalence estimates were calculated using exact methods. We adjusted for clustering within households and age-adjusted with poststratification using the known age distribution of the census tracts [4]. To estimate the number of expected infections in these 2 census tracts in 2019, we applied the final adjusted IgM seroprevalence estimate to the total adult population. To estimate the proportion of recent infections resulting in neuroinvasive disease, we divided the number of reported neuroinvasive disease cases in 2019 by the estimated number of total expected infections for these census tracts.

RESULTS

Of the 423 randomly selected households, 179 (42%) participated, 150 (36%) declined to participate, 76 (18%) were unreachable, and 18 (4%) were ineligible. From the 179 participating households, we enrolled 274 participants and collected blood from 273. Of the 274 participants, 140 (51%) were female, and the median age was 63 years (IQR, 50–70 years) (Table 1). Our sample skewed older than the overall population of the 2 census tracts (Figure 1). The median length of residence at participants’ current addresses in Sussex County was 20 years (IQR, 9–33 years) and 264 (96%) lived in the area year-round. Of the 274 participants, 197 (72%) reported finding a tick crawling on them and 115 (42%) reported finding a tick at least once since January 2019.

Table 1.

Characteristics of POWV Serosurvey Participants: Sussex County, New Jersey, 2019

Characteristic	All Participants (N = 274) n (%)	IgM Positive Participants (N = 2) n (%)
Age (Years)
20–29	8 (3)	—
30–39	22 (8)	—
40–49	37 (14)	—
50–59	43 (16)	—
60–69	95 (35)	1 (50)
70–79	51 (19)	1 (50)
80–89	17 (6)	—
≥90	1 (<1)	—
Sex
Male	135 (49)	2 (100)
Female	139 (51)	—
Reported crawling tick	197 (72)	2 (100)
Reported attached tick	115 (42)	2 (100)
Year-round resident	264 (96)	2 (100)

Abbreviations: Ig, immunoglobulin; POWV, Powassan virus.

Figure 1.

Age of Powassan virus (POWV) serosurvey participants and census tract residents: 2 census tracts in Sussex County, New Jersey, 2019.

Among the 273 participants with a blood specimen, 3 (1.1%; 95% CI, .00%–2.30%) had POWV-specific neutralizing antibodies, including 2 (0.73%; 95% CI, .09%–2.62%) with POWV-specific IgM antibodies. One of these 2 participants was a known 2019 POWV disease patient whose residence was in our random sample. The 2019 POWV disease patient was the only participant to report any encephalitis or meningitis in the past 3 years. All seropositive participants were aged ≥60 years, the participant with only neutralizing antibodies was female, and both IgM-positive participants were male. None of the participants with POWV-specific antibodies had antibodies to WN or SLE virus.

After adjusting for clustering within households and applying poststratification, we estimate an adjusted IgM seroprevalence of 0.31% (95% CI, .04%–1.00%), resulting in 13 (95% CI, 2–41) expected recent POWV infections in these 2 census tracts, and an estimated 23% (3 of 13; 95% CI, 7%–100%) of POWV infections resulting in neuroinvasive disease.

DISCUSSION

Despite several recent POWV disease cases and high prevalence of tick exposure reported by participants, we estimated a low POWV IgM seroprevalence (<1%) and identified very few individuals with evidence of past exposure to POWV. There have been at least 2 previous estimates of POWV infection prevalence in specific populations. A serologic review of 2963 patients with central nervous system infections in New York state during 1966–1977 found evidence of previous POWV infection in 8 (0.27%) patients [6]. In contrast, a 2015 assessment found higher prevalence of previous POWV infection in Wisconsin patients being tested for Lyme disease than those undergoing routine blood chemistry (10% [9 of 95] versus 4% [2 of 50], respectively) [7].

Although our estimate of 23% of POWV infections resulting in neuroinvasive disease has a high degree of uncertainty due to the small number of infections detected, it is comparable to estimates for a related tick-borne flavivirus, tick-borne encephalitis virus (~25%) [8]. For mosquito-borne viruses, the estimates are lower; ~1% of WN virus infections [9] and ~4% of Eastern equine encephalitis virus infections are estimated to result in neuroinvasive disease [10].

In our survey, both IgM-positive participants, one of whom had neuroinvasive POWV disease, were male and aged ≥60 years. Although risk factors likely differ for infection compared with disease, this is consistent with previous descriptions of POWV disease cases, 73% of which were male and 59% were aged ≥60 years [2]. This is also consistent with other flavivirus diseases, including WN and SLE viruses [2]. Occupational or behavioral factors might contribute to the increased risk of exposure in males, and age-related factors might increase the risk of developing neuroinvasive disease once infected.

The number of reported POWV disease cases has increased over time, with an average of 8 US cases reported annually during 2006–2015, compared with 30 annually during 2016–2019 [2, 3]. It is unclear whether the increase in reported cases is because of increased awareness, testing, and surveillance, a true increase in disease incidence, or a combination of these factors. Our IgM prevalence was low despite high prevalence of tick bites, which suggests that there is likely a low POWV infection rate among ticks in this region as has been documented in other studies [11, 12]. The low POWV antibody prevalence coupled with the finding that our participants generally have lived in the area for a long time also suggests that POWV might be emerging in this area. In addition to expanding geographic distribution of I scapularis ticks, reported incidence of multiple other I scapularis-vectored diseases has increased over time, including Lyme disease, anaplasmosis, and babesiosis [13]. There is also evidence of increasing POWV infection in nonhuman hosts, including a study of white-tailed deer in Connecticut that found seroprevalence increased over time, from <10% in 1979 to >80% in 2010 [14].

This investigation is subject to several limitations. First, residents of participating households may have differed from those that declined; 1 notable difference is that our sample was biased towards older adults. We adjusted for this with poststratification, but if potential risk factors for infection (eg, outdoor activities or recreational areas visited) differ by age, then we might have missed infections in younger residents. Second, although IgM detection is routinely used as an indication of recent infection, POWV IgM persistence is not well defined, which could result in under- or overestimation of infections occurring in 2019. Third, the number of recent infections in our sample was lower than assumed during study planning, which resulted in wide CIs. Finally, our results are not generalizable to other areas.

CONCLUSIONS

Based on the increasing presence of tick-borne diseases, we recommend strengthening tick prevention messaging (eg, using Environmental Protection Agency-registered insect repellents, and treating clothing and gear with permethrin). Because POWV transmission can occur within 15 minutes of tick attachment in contrast to Borrelia burgdorferi, which usually takes ≥48 hours, prevention messaging should emphasize the importance of tick bite avoidance in addition to performing tick checks and showering soon after being outdoors [15]. Clinicians in areas where POWV disease occurs should consider POWV in the differential diagnosis for patients with history of tick bites and febrile illness, meningitis, or encephalitis.