RETAIL SURVEY KEY FINDINGS
The original report of this study was published in the Journal of Food Protection (Ref. 5). The full report contains a complete description of the study, data collected, and results. In this section, we extract a subset of key findings from that report that address, in part, data gaps identified in the 2013 version of this risk profile.
Samples for FDA’s retail spice sampling survey were collected at a variety of retail establishments throughout the United States during 1 of 2 periods: November 2013-September 2014 or October 2014-March 2015.
The retail study surveyed 11 spice types: basil leaf (whole, ground, crushed or flakes), black pepper (whole, ground or cracked), coriander seed (ground), cumin seed (whole or ground), curry powder (ground mixture of spices), dehydrated garlic (powder, granules or flakes), oregano leaf (whole, ground, crushed or flakes), paprika (ground; note this spice is only sold ground), red pepper (hot red pepper, e.g., chili, cayenne; ground, cracked, crushed or flakes), sesame seed (whole, not roasted or toasted, not black), and white pepper (ground or cracked) and included 7,250 retail samples of packaged (dried) spices offered for sale at retail establishments in the United States. Samples of various spice types were collected and Salmonella prevalence and aerobic plate count levels were determined.
A summary of the Salmonella testing results and prevalence estimates for each spice type is provided in Table 1.For each of the 11 spice types examined, estimated prevalence (based on 125 g spice analyzed) was lessthan 1% with all upper 95% confidence intervals less than 2%. Among the spice types, no positive samples were found for cumin (whole/ground/cracked), sesame seed (whole), or white pepper (ground). These prevalence estimates are not corrected for the sampling design (not weighted for market share) because the detailed information needed was not available.
Of particular interest in this study, was whether the Salmonella prevalence estimates for each spice type at the point of entry to the United States were different from those for the same spice type at the point of retail purchase by U.S. consumers, particularly for the spices where the U.S. supply is overwhelmingly imported, as is the case for at least seven of the spices examined in in this study: basil, black pepper, coriander, cumin, curry powder, oregano, and white pepper (Ref. 7). For red pepper, paprika, and sesame seed, imports are also the major source of the U.S. supply but domestic production is significant (Ref. 7-8). The Salmonella prevalence estimates for spices offered for sale at retail establishments for all of the spice types examined except dehydrated garlic and basil (for which statistical power was limited) were significantly smaller than estimates for shipments of imported spice offered for entry. Among the spice types for which significant differences in prevalence were found, examination of a smaller sample size (mass) for samples from retail than from point of entry to the United States likely contributed to the apparent decrease in prevalence (Ref. 9-11)but cannot fully explain the observations.
The results of this study are consistent with the assumption that most (bulk) shipments of spice undergo a pathogen reduction treatment following entry to the United States and prior to releasing for retail sale, as recommended in industry guidance such as the “Clean, Safe, Spices Guidance Document” by American Spice Trade Association (Ref. 12). No Salmonella-positive samples were found for cumin, sesame seed, or white pepper, which would be expected if all shipments had undergone a highly efficient pathogen reduction treatment after entry and before being offered for sale at retail establishments. The change in observed Salmonella prevalence between point of entry to the United States and retail sale for the other spice types examined was smaller and ranged from a factor of 0.33 (dehydrated garlic; not significant) to 0.015 (oregano). Assuming the spice sampled at the point of entry to the United States and retail sale similarly represented the available supply, these results may indicate that pathogen reduction treatments for some shipments were not effective or not applied (Ref. 9-11) or that post-treatment contamination had occurred. For red pepper and dehydrated garlic, where the supply is a combination of imported and domestically produced spices, differences in Salmonella prevalence pre-treatment may also have played a role in the results. The decreases observed (except for dehydrated garlic) are consistent with more than 90% of contaminated shipments offered for entry to the U.S. having been treated with an efficient pathogen reduction treatment prior to being offered for sale in retail establishments. These data cannot be used to provide a reliable estimate of the mean log reduction in these spices because of uncertainties involved, especially the lack of data on pathogen levels in these contaminated spices at the point of entry to the United States and retail sale.