Validation of an automated colony counting system for group A Streptococcus

BACKGROUND: The practice of counting bacterial colony forming units on agar plates has long been used as a method to estimate the concentration of live bacteria in culture. However, due to the laborious and potentially error prone nature of this measurement technique, an alternative method is desirable. Recent technologic advancements have facilitated the development of automated colony counting systems, which reduce errors introduced during the manual counting process and recording of information. An additional benefit is the significant reduction in time taken to analyse colony counting data. Whilst automated counting procedures have been validated for a number of microorganisms, the process has not been successful for all bacteria due to the requirement for a relatively high contrast between bacterial colonies and growth medium. The purpose of this study was to validate an automated counting system for use with group A Streptococcus (GAS). METHODS: Twenty-one different GAS strains, representative of major emm-types, were selected for assessment. In order to introduce the required contrast for automated counting, 2,3,5-triphenyl-2H-tetrazolium chloride (TTC) dye was added to Todd-Hewitt broth with yeast extract (THY) agar. Growth on THY agar with TTC was compared with growth on blood agar and THY agar to ensure the dye was not detrimental to bacterial growth. Automated colony counts using a ProtoCOL 3 instrument were compared with manual counting to confirm accuracy over the stages of the growth cycle (latent, mid-log and stationary phases) and in a number of different assays. The average percentage differences between plating and counting methods were analysed using the Bland-Altman method. RESULTS: A percentage difference of ±10 % was determined as the cut-off for a critical difference between plating and counting methods. All strains measured had an average difference of less than 10 % when plated on THY agar with TTC. This consistency was also observed over all phases of the growth cycle and when plated in blood following bactericidal assays. Agreement between these methods suggest the use of an automated colony counting technique for GAS will significantly reduce time spent counting bacteria to enable a more efficient and accurate measurement of bacteria concentration in culture.