Lyme Disease Emergence After Invasion of the Blacklegged Tick

Ixodes scapularis, Ontario, Canada, 2010-2016

Manisha A. Kulkarni; Isha Narula; Andreea M. Slatculescu; Curtis Russell

Disclosures

Emerging Infectious Diseases. 2019;25(2):328-332. 

In This Article

The Study

Our study included 3 public health units in eastern Ontario, Canada: Kingston, Frontenac, and Lennox and Addington (KFL); Leeds, Grenville, and Lanark (LGL); and Ottawa. This region spans from the St. Lawrence River in the south to the Ottawa River in the north, and has several major population centers, including Kingston (2016 population 123,798) and Ottawa (2016 population 934,243).[5] The region is largely characterized by mixed deciduous forest and agricultural land use.

We used data from the Integrated Public Health Information System database to identify human LD cases on the basis of provincial case definitions.[6] We geocoded cases to their forward sortation area (FSA) (i.e., first 3 digits of the postal code) of residence and extracted data on patient sex, age, episode date (onset of symptoms), and reported history of travel (defined as travel outside the municipality of residence within the previous 2 weeks). Data on ticks collected during 2010–2016 through passive tick surveillance activities in Ontario were obtained from Public Health Ontario (PHO).[7] We aggregated I. scapularis tick records according to the FSA of the submitter (i.e., location of residence of the person who acquired the tick) and excluded records with missing collection date, submitter FSA, or PCR test result and records with reported history of travel. We similarly excluded human LD records with missing patient FSA or with reported travel history. We obtained FSA-level population data for 2011 and FSA boundary files from Statistics Canada.[5]

To examine the association between the invasion of I. scapularis ticks and B. burgdorferi and the spread of human LD, we examined associations between FSA-level data on time to first case (in years) and several variables: time to first reported I. scapularis tick, time to first reported B. burgdorferi–infected tick, distance to FSA with highest LD incidence in 2010, and population. We constructed bivariable and multivariable linear regression models with time to first case (in years) as the outcome.

To visualize LD spread during 2010–2016, we plotted the annual FSA-level incidence of human LD and B. burgdorferi prevalence in ticks by using ArcGIS 10.4 (ESRI, https://www.esri.com). We also assessed the annual weighted mean center and distribution of human LD incidence by using ArcGIS 10.4, after spatial projection of the data to preserve distance.[8] We applied Kulldorff's spatial scan statistics[9] by using SaTScan 9.6 (https://www.satscan.org) to assess and compare spatiotemporal patterns in human LD incidence and B. burgdorferi prevalence in ticks at the FSA level (FSA centroids). (For additional methods, see the Appendix, https://wwwnc.cdc.gov/EID/article/25/2/18-0771-App1.pdf).

A higher proportion of LD cases occurred in men and in adults 50–69 years of age (Table 1), similar to patterns observed in other regions of North America.[10] LD incidence increased over time; 55% of cases occurred during 2015 and 2016 (Table 2). Roughly 70% of cases occurred during June–August, whereas ≈20% occurred during September–December. The number of collected ticks increased annually from 2010 and reached a peak in 2013, with a subsequent decrease because of reductions in passive surveillance activities in KFL and LGL;[11] Ottawa received an increasing amount of ticks over time (Table 2). The percentage of ticks testing positive for B. burgdorferi increased annually, from 12% in 2010 to 23% in 2016 (p<0.001). Infection rates were higher among regions of KFL and LGL, although FSAs with high B. burgdorferi prevalence among submitted ticks were observed in parts of Ottawa in more recent years (Figure 1).

Figure 1.

Annual prevalence of Borrelia burgdorferi in Ixodes scapularis ticks from passive tick surveillance, based on forward sortation area of tick submitter, 3 public health units, eastern Ontario, Canada, 2010–2016.

Within our study area, the first human LD case was reported an average of 2.2 years after the first reported I. scapularis tick and 1.1 years after the first reported B. burgdorferi–infected tick. Time to first case was significantly associated with time to first reported I. scapularis tick (adjusted r[2] = 0.56; p<0.001) and time to first B. burgdorferi–infected tick (adjusted r[2] = 0.67; p<0.001) after adjusting for distance to the FSA with highest LD incidence in 2010. The associated lag between each phase of ≈1 year supports the hypothesis that invasion and establishment of tick populations is followed by colonization of B. burgdorferi,[12] or it might reflect the arrival of infected ticks with subsequent increase in B. burgdorferi prevalence. However, drawing conclusions on the exact timing of tick and pathogen invasion is difficult because of the nature of passive surveillance data.

LD incidence was concentrated in southern FSAs in 2010 and 2011 but had spread in a northeasterly direction by 2013 (Figure 2). Overall, a northeast shift of 54 km occurred between mean centers during 2010–2016, with the greatest spread observed in 2011–2013 (Appendix). We detected a spatiotemporal cluster of high rates of B. burgdorferi–infected ticks in the Kingston-Gananoque region bordering the St. Lawrence River, which overlapped with 2 clusters of human LD cases (Appendix Figure 4). The overlapping clusters support the conclusion that increased tick encounter is a determinant of human LD risk. Residence in endemic areas (i.e., where infected ticks have been found) has been consistently recognized as a risk factor for LD infection.[13,14]

Figure 2.

Spatiotemporal spread of human Lyme disease incidence, 3 public health units, eastern Ontario, Canada, 2010–2016. Annual Lyme disease incidence estimated from notifiable disease surveillance and population data based on forward sortation area of patient residence.

processing....