Estimation of Bacillus stearothermophilus Kinetic Parameters of Inactivation Using a Sub-lethal Adaptation Model in Conduction-heated Foods

Introduction: Estimation of inactivation kinetic parameters for spores is typically done using isothermal experiments with rapid come-up times in small samples of uniform laboratory media.  These parameters are then used to predict log reductions of spores in commercial conditions that may be very different from those in the isothermal simulations. Some isothermal studies have shown that spores may become more resistant to inactivation as they spend more time in the sub-lethal temperature range.  

Purpose: The purpose of this study was to determine if the inactivation kinetics of a spore under commercial thermal processing conditions was better represented by a sub-lethal adaptation model than by the first-order or Weibull models.  

Methods: Previous dynamic data were used.  Cans (6 cm dia., 3.48 cm height) were filled with pea puree inoculated with Bacillus stearothermophilus ATCC 7953 spores at ~9.5 x 10⁵ CFU/ml.  The cans were heated in duplicate at three different steam temperatures (104.4, 112.8, 120.6°C) for heating times from 21 to 305 min, for a total of 24 samples.  Each can’s mass-average spore concentration was both enumerated using a non-selective media and incubation at 55°C, and computed using three different models.  Parameters were estimated using an inverse method.  

Results: Log-reductions ranged up to 2.5.  The first-order model performance was poor, as shown by a 0.24 log CFU/ml RMSE and AICc = -62.8.  The Weibull model was significantly better, with RMSE = 0.12  log CFU/ml, AICc = -95.6.  The best model was the Weibull with sub-lethal term:  RMSE = 0.098 log CFU/ml, AICc = -100.6.  The parameter errors for d105°C = 3.2%, for z = 2.8%, and for n = 9.4%.  

Significance: The results of this work indicate that for thermally-processed foods with different heating rates, sub-lethal history may affect prediction of survivors. Process designers and researchers may consider using experiments more representative of commercial processes.