Epidemiology and Spatial Emergence of Anaplasmosis, New York, USA, 2010–2018

Alexis Russell; Melissa Prusinski; Jamie Sommer; Collin O'Connor; Jennifer White; Richard Falco; John Kokas; Vanessa Vinci; Wayne Gall; Keith Tober; Jamie Haight; JoAnne Oliver; Lisa Meehan; Lee Ann Sporn; Dustin Brisson; P. Bryon Backenson


Emerging Infectious Diseases. 2021;27(8):2154-2162. 

In This Article


Anaplasmosis Cases

Human anaplasmosis cases reported to the NYSDOH were analyzed retrospectively for 2010–2018 for all NYS counties, excluding the 5 boroughs of New York City (NYC). Provider-diagnosed anaplasmosis cases and positive laboratory test results for anaplasmosis were reported to the NYSDOH as mandated by NYS public health law.[14,15] Both provider-reported cases and those with positive laboratory test results were investigated by NYS local health departments; clinical and demographic information for each case was entered into the NYSDOH Communicable Disease Electronic Surveillance System. Reports were assigned a case status on the basis of the 2008 Centers for Disease Control and Prevention case definition for anaplasmosis.[16] Reports with case status of either confirmed or probable were included as cases in this study. Cases with the diagnosis of ehrlichiosis/anaplasmosis undetermined were excluded.

Tick Collection and Testing

Host-seeking ticks were collected from publicly accessible lands across NYS during 2010–2018 by using standardized drag surveys as described.[17] Collection sites were selected on the basis of tick habitat suitability and risk for human exposure (e.g., presence of leaf litter and hiking trails). I. scapularis nymphs were collected during April–September by dragging a 1-m2 piece of white flannel through leaf litter and low-lying vegetation. I. scapularis adults were collected during September–December by flagging a 1-m2 piece of white canvas over edge ecotone and understory vegetation up to 1 m high. Ticks were stored in 100% ethanol at 4°C until they were sorted by developmental stage and identified to species by using dichotomous keys, placed into sterile microcentrifuge tubes containing 100% ethanol, and stored at −20°C until DNA extraction.[18,19] Individual I. scapularis ticks underwent total genomic DNA extraction as previously detailed and were tested for (target gene) A. phagocytophilum (major surface protein 2), Babesia microti (18S rDNA), Borrelia burgdorferi (16S rDNA), and Borrelia miyamotoi (16S rDNA) by using a quadplex real-time PCR.[17,20]

Data Analysis

We analyzed case reports meeting criteria for inclusion by using SAS version 9.2 (https://www.sas.com). Incidence rates were aggregated by NYS regions (Capital, Central, Metro, and Western), and ZIP code tabulation area (ZCTA) by using patient address and 2010 US Census population data and shapefile.[21] We used ArcGIS version 10.7[22] to map incidence at the ZCTA level. Spatial autocorrelation at the ZCTA level was determined by using Moran I analysis for each year. We determined spatial clusters by using Getis-Ord Gi* hot spot analysis (https://pro.arcgis.com) at the ZCTA level for each year. Getis-Ord Gi* analysis generated statistically significant hot spots and cold spots on the basis of the local sum of the incidence rates for each ZCTA and its neighbors within a fixed distance band at peak z-score spatial increments. We assessed temporal changes in hot spot coverage by using a 2-tailed z-test for proportions (α = 0.05).

We analyzed tick collection and pathogen testing data by using in R Studio version 1.2[23] and mapped data by using ArcGIS. Tick population density was calculated for each collection site visit as the total number of target ticks (adult or nymphal I. scapularis) collected per 1,000 m2 sampled. We calculated pathogen prevalence as the proportion of ticks positive for A. phagocytophilum among those tested by PCR for each collection site visit and region. Temporal changes in pathogen prevalence were assessed by using a 2-tailed z-test for proportions (α = 0.05).

We used the entomologic risk index (ERI), a measure of population density of pathogen-carrying ticks, to estimate human risk for an infected tick bite.[24] ERI was calculated as the product of tick population density (ticks per 1,000 m2 sampled) and A. phagocytophilum prevalence at each collection site for each life stage (nymph and adult) and each year. We calculated ZCTA-level ERI as the average ERI of all sites within the ZCTA for each life stage and year. Correlation of anaplasmosis incidence and ERI at the ZCTA level was analyzed for each year by using the Spearman rank correlation. We mapped collection sites with circles sized according to ERI magnitude and then overlaid them onto the anaplasmosis incidence Getis-Ord Gi* hot spot analysis map of the corresponding year to identify common patterns in ERI and human incidence clusters.