Frontiers in Microbiology (Apr 2020)

“These Aren’t the Strains You’re Looking for”: Recovery Bias of Common Campylobacter jejuni Subtypes in Mixed Cultures

  • Benjamin M. Hetman,
  • Benjamin M. Hetman,
  • Steven K. Mutschall,
  • Catherine D. Carrillo,
  • James E. Thomas,
  • Victor P. J. Gannon,
  • G. Douglas Inglis,
  • Eduardo N. Taboada

DOI
https://doi.org/10.3389/fmicb.2020.00541
Journal volume & issue
Vol. 11

Abstract

Read online

Microbiological surveillance of the food chain plays a critical role in improving our understanding of the distribution and circulation of food-borne pathogens along the farm to fork continuum toward the development of interventions to reduce the burden of illness. The application of molecular subtyping to bacterial isolates collected through surveillance has led to the identification of strains posing the greatest risk to public health. Past evidence suggests that enrichment methods for Campylobacter jejuni, a leading bacterial foodborne pathogen worldwide, may lead to the differential recovery of subtypes, obscuring our ability to infer the composition of a mixed-strain sample and potentially biasing prevalence estimates in surveillance data. To assess the extent of potential selection bias resulting from enrichment-based isolation methods, we compared enrichment and non-enrichment isolation of mixed subtype cultures of C. jejuni, followed by subtype-specific enumeration using both colony plate-counts and digital droplet PCR. Results differed from the null hypothesis that similar proportions of C. jejuni subtypes are recovered from both methods. Our results also indicated a significant effect of subtype prevalence on isolation frequency post-recovery, with the recovery of more common subtypes being consistently favored. This bias was exacerbated when an enrichment step was included in the isolation procedure. Taken together, our results emphasize the importance of selecting multiple colonies per sample, and where possible, the use of both enrichment and non-enrichment isolation procedures to maximize the likelihood of recovering multiple subtypes present in a sample. Moreover, the effects of subtype-specific recovery bias should be considered in the interpretation of strain prevalence data toward improved risk assessment from microbiological surveillance data.

Keywords