Navigation

P2-140 Evaluation of Analyst Proficiency and Laboratory Performance in Testing Pathogens in Infant Formula Based on a Proficiency Study

Monday, July 27, 2015
Exhibit Hall (Oregon Convention Center)
Ravinder M. Reddy , U.S. Food and Drug Administration , Bedford Park , IL
Hossein Daryaei , Illinois Institute of Technology/IFSH , Bedford Park , IL
Robert Newkirk , U.S. Food and Drug Administration , Bedford Park , IL
Samantha Lindemann , U.S. Food and Drug Administration , Bedford Park , IL
Christopher Powers , Illinois Institute of Technology/IFSH , Bedford Park , IL
Introduction: Understanding analyst/laboratory performance in the detection of pathogens in infant formula can affect enforcement of regulations. Sources of error in detection can be traced to analytical techniques, analysts, and random effects. Quantifying these sources can be used to improve quality assurance and safety.

Purpose: Evaluate proficiency test (PT) data to assess analytical performance in detecting pathogens.

Methods: Commercial liquid and powdered infant formula samples (LIF, PIF) were inoculated at two levels in duplicate with cocktails of Gram positive and negative organisms and with one or more of the following pathogens: Staphylococcus aureus, Bacillus cereus, Salmonella Typhimurium, Listeria monocytogenes, Shigella flexneri, Cronobacter sakazakii, and Escherichia coli O157:H7.  The inoculation procedures, mixing methods, and homogeneity/stability were tested according to ISO 17043 and ISO 13528.  A total of 135 analysts from 59 laboratories (LIF) and 29 analysts from 19 laboratories (PIF) tested for one or more pathogens based on preferred methods.

Results: Interlaboratory comparisons from proficiency testing data were performed based on ISO 22117 guidelines. Qualitative results are summarized as percentages of correct results with total observations in parentheses for LIF/PIF: Staphylococcus 100%/88% (292/60), Bacillus 88%/88% (200/40), Salmonella 98%/94% (380/72), Listeria 100%/95% (380/64), Shigella 90%/83% (220/60), Cronobacter 99%/96% (104/24), and O157 96%/73% (368/48). Specificity (rSP) and sensitivity rates (rSE) were calculated according to ISO 22117. For low inoculation levels, rSP ranged from 96.3 - 100% and rSE ranged from 75.6 - 100%.  For high inoculum levels rSE was 92 - 100%.  The lowest rSE values were found in low inoculum samples of Bacillus, 84%, and Shigella, 75.6%; indicating these samples were the most challenging.

Significance: This study shows how PT data can be used to ensure laboratories have the capability to detect key food pathogens. It also demonstrates the value of expansive, matrix-relevant PTs with realistic microbiological composition to generate method performance data in rapid and economical ways.

Back to: Poster Session 2 – Laboratory and Detection Methods; Meat, Poultry and Eggs; Dairy and Beverages; Food Law and Regulation; General Microbiology; Modeling and Risk Assessment
<< Previous Abstract | Next Abstract >>