Spatial and molecular characterization of methicillin-resistant Staphylococcus aureus infection in relation to high-density livestock production in Pennsylvania
Embargo until
Date
2014-03-18
Authors
Casey, Joan A.
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Background: Epidemiologic studies of methicillin-resistant Staphylococcus aureus (MRSA) infection in the U.S. have been limited by a focus on the healthcare setting or by the evaluation of specific populations such as military, inmate, athlete, or urban dwellers, thus providing little information about the general U.S. population. Several lines of evidence, including epidemiologic, environmental, and microbiologic, suggest that high-density livestock production may be contributing to regional MRSA epidemics. However, most studies of high-density livestock production have evaluated colonization, not infection, and few studies have been conducted in the U.S.
Objectives: To describe the epidemiology of MRSA infection and skin and soft tissue infections (SSTIs) in patients residing in a large, diverse (i.e. urban and rural) geographic area; to develop exposure models for swine and dairy/veal high-density livestock production (considering both operations where animals are raised and the crop fields to which manure is subsequently applied); and to evaluate associations of exposure models with MRSA and molecular subgroups.
Methods: We estimated the incidence of MRSA and skin and SSTIs over a 9-year period using electronic health records from the Geisinger Health System in Pennsylvania. We identified incident MRSA cases and classified them as community-associated or healthcare-associated. We frequency- matched them to both randomly selected SSTI patients without a history of MRSA and controls without a history of MRSA or SSTI, and conducted a population-based, nested case-control study from 2005-2010. We used nutrient management plans produced by operators to create exposure variables based on: patient home address; location of livestock production and number of animals; and location, area, and manure volume of crop field manure application. The primary exposure variables of interest were gravity models of seasonal crop field manure application and number of livestock animals at the operation. To evaluate associations of interest with MRSA or SSTI, we used multinomial multilevel and multilevel logistic regression, respectively. In a sub-study, we prospectively collected 196 MRSA isolates from patients stratified by diagnosis setting and residential location, then used logistic regression, accounting for sampling design, to examine associations of exposure metrics with MRSA molecular types.
Results: From 446,480 patients, 1539 community-associated MRSA (CA-MRSA), 1335 healthcare- associated MRSA (HA-MRSA) and 78,216 SSTI cases were identified. Additionally, we randomly selected 2914 controls for the case-control study. From 2005-2009, CA-MRSA, HA-MRSA, and SSTI incidence increased annually by 34% (range, 6% to 94%), 6% (-15% to 51%), and 3.6% (1% to 8%), respectively. There was a trend of increasing odds of CA-MRSA across quartiles of swine crop field manure and swine livestock exposure (both P<0.05); adjusted odds ratios for the fourth quartile were 1.38 (95% CI, 1.13-1.69) and 1.25 (0.99-1.58), respectively. Adjusted odds ratios for the randomly selected SSTIs (n = 2895) were similar. Using the population attributable fraction, approximately 11% of both CA-MRSA and SSTI cases in the study population could be attributed to crop field application of swine manure. In the molecular types sub-study, we did not identify any strain type 398 (ST398) MRSA, but found that three of four exposure variables (swine livestock, dairy/veal livestock, and dairy/veal crop field exposure) were each associated with increased odds of community-onset Panton-Valentine leukocidin-negative MRSA infection.
Conclusions: In a region with significant increases in CA-MRSA incidence from 2005 to 2009, we found consistent associations of two primary aspects of high-density livestock production – where livestock are raised and where manure is applied – with CA-MRSA, HA-MRSA, and SSTI risk. We note the association with SSTI because between 2005 and 2009 there were over 50,000 SSTI infections compared to 1539 CA-MRSA infections. The molecular type sub-study requires replication, but has implications for future research, suggesting that colonization and infection studies could be studying two very different aspects of MRSA epidemiology. Strains suited for colonization in animals (ST398) may not regularly cause infection in humans. Overall, our findings contribute evidence for the growing concern about the impact of high-density livestock production on public health.
Description
Keywords
Agriculture, concentrated animal feeding operations (CAFOs), environmental epidemiology, geographic information systems, methicillin-resistant S. aureus (MRSA), skin infections