Purpose: The purpose of this study was to evaluate whether product aw significantly affects the validity of five different process validation methods for Salmonellainactivation on almonds.
Methods: Almonds were inoculated with Salmonella Enteritidis PT30 or Enterococcus faecium (NRRL B-2354) as a Salmonella surrogate, at ~108 CFU/g and equilibrated to 0.45 or 0.60 aw, and then heated in a pilot-scale moist-air impingement oven (dry bulb 121, 149, or 177 °C; dew point < 33, 69.4, 81.6, or 90.6°C; vair = 2.7 m/s) to a target lethality of 4 log. Surviving Enterococcus and Salmonella were enumerated (3 reps per treatment) by plating on deMan, Rogosa and Sharpe agar or trypticase soy agar modified with sodium thiosulfate and ammonium iron citrate (35°C, 48 h), respectively. Almond surface temperatures were measured (9 reps per treatment) using surface thermocouples (Tsurf) on almonds and aluminum almonds (TAl), with these temperatures then used to calculate Salmonellainactivation using a traditional (D, z) model and a previously published modified model accounting for process humidity.
Results: Among all process validation methods, E. faecium yielded the lowest root mean squared error in predicting Salmonella inactivation (RMSE = 1.03 log (CFU/g), n = 12, aw = 0.45). For the same data, the modified model yielded an RMSE of 2.08, and the traditional model exhibited unacceptably high error (RMSE > 16). Even with the modified model accounting for process humidity, aw significantly affected model accuracy, with the RMSE doubling between 0.45 and 0.65 aw for the model developed with low awdata.
Significance: Overall, product aw is a critical factor that affects the accuracy of process validation methods. Therefore, thermal inactivation models for moist-heat conditions ideally should account for both process humidity and product aw.