Distribution of Cronobacter spp. in industrial batches of powdered infant formula and the impact of sampling approaches

Objectives - The three objectives of this study are: first, to investigate how Cronobacter spp. are distributed throughout a recalled and a normal batch of powdered infant formula (PIF); second, to investigate the occurrence of conglomerations of Cronobacter spp. cells, since conglomerations of cells with high concentrations may impact on risk assessment and public health; and third, to investigate the performance of typical sampling plans. Results - In the recalled batch the concentrations Cronobacter spp. versus the filling time was assessed by 415 samples of 333 g using the MPN technique. In 58 % of the samples, concentrations were below the detection limit of -2.52 log CFU/g. However, specifically within three time intervals MPN concentrations were estimated varying between -2.52 and 0.66 log CFU/g. In addition 2290 samples of 1 g were investigated by plating, and 8 were found to contain Cronobacter spp. above the detection limit of 0.52 log CFU/g. These samples were originating from packages filled at the time interval with also the highest MPN concentrations. The two largest conglomerations were 123 and 560 cells in 1 g of PIF. In 99% of the samples from the normal batch, concentrations were below the detection limit. Various sampling plans were evaluated for the contamination data from the recalled batch. Keeping the total sample weight constant at 300 g and increasing the number of random samples from 1 to 300, increased the probability of detection from 0.38 till 1.00. Conclusions - Cronobacter spp. was distributed heterogeneously within a recalled batch of PIF with parts of the batch with no detectable contamination, and parts of the batch with concentrations up to 2.75 log CFU/g. The presence of conglomerations of Cronobacter spp. cells occurred with a low frequency. Taking more and smaller samples, keeping the total sampling weight constant, considerably improved the performance of the sampling plans to detect such a type of contaminated batch.