Foodborne illness is a major cause of illness worldwide. In the United States alone, there are an estimated 9.4 million cases occurring every year (Scallan et al., 2011). A considerable number of these are attributed to outbreaks involving contaminated fresh produce (Lynch et al., 2008; Bennett et al., 2018; Chatziprodromidou et al., 2018; Carstens et al., 2019). From 1998 to 2013, a total of 972 outbreaks involving fresh produce were reported through U.S. Centers for Disease Control and Prevention’s Foodborne Disease Outbreak Surveillance System, resulting in a total of 34,674 illnesses, 2,315 hospitalizations, and 72 deaths (Bennett et al., 2018).

Pathogens commonly associated with fresh produce outbreaks include bacteria such as Listeria monocytogenes, Escherichia coli and Salmonella enterica (Carstens et al., 2019), as well as viral pathogens such as human norovirus (hNoV) and hepatitis A virus (HAV) (Chatziprodromidou et al., 2018). Human noroviruses are the leading cause of foodborne disease outbreaks, with over 5 million food-associated cases estimated per year (Scallan et al., 2011). Leafy vegetables, fruits and nuts are the general categories of produce items associated with hNoV outbreaks (Hall et al., 2012). Although often the source of contamination in hNoV-produce outbreaks is not definitively established, infected food handlers at the preparation phase are a commonly implicated route (Hall et al., 2012). Hepatitis A virus, while a less common cause of produce-associated outbreaks, causes a more severe disease and has been associated with fresh produce including, green onions (Centers for Disease Control and Prevention, 2003) and lettuce (Rosenblum et al., 2011). As described for hNoV, the most common source of contamination is the hands of infected food handlers who have not practiced adequate personal hygiene.

Of the produce items most often implicated in virus outbreaks, berries stand out. In fact, virus-contaminated raspberries, strawberries, and blueberries have been the cause of notable hNoV and HAV outbreaks, some very large and high profile (Bozkurt et al., 2021). For instance, the largest recorded foodborne outbreak in German history, with nearly 11,000 cases, was caused by hNoV contamination of frozen strawberries (Bernard et al., 2014). Similarly, a large and prolonged foodborne outbreak in Italy was caused by HAV contamination of frozen berries, resulting in greater than 1,800 cases spanning over 12 months (Scavia et al., 2017). Virus outbreaks associated with tomatoes, while occurring less frequently, have also been reported (Donnan et al., 2012; Shieh et al., 2014), including a multistate Australian HAV outbreak associated with semi-dried tomatoes and resulting in over 500 cases (Donnan et al., 2012).

Human norovirus and HAV are environmentally stable and are quite recalcitrant to many processes and sanitizers/disinfectants used routinely in food processing and preservation, including pH and water activity manipulation, high pressure, alcohols and quaternary ammonium compounds (Li et al., 2012; Bosch et al., 2018). Chlorine and chlorine dioxide (ClO₂) have shown relatively broad efficacy against various microorganisms (Park et al., 2018), including hNoV and HAV (and their surrogates) (Li et al., 2004; Lim et al., 2010; Kingsley et al., 2014; Song et al., 2022). In fact, CDC recommends high concentrations of hypochlorite (1,000-5,000 ppm) when cleaning surfaces suspected of being contaminated with hNoV. Chlorine dioxide gas has demonstrated efficacy for decontamination and preservation of candidate food products such as fruits, vegetables and raw meats (Gómez-López et al., 2009; Singh et al., 2021). Unfortunately, ClO₂ gas presents unique challenges including difficulty of use in the supply chain, elevated cost, and effects on food product quality, all of which can limit its current commercial use (Ridenour and Ingols, 1947; Singh et al., 2021). Organoleptic issues can include the potential to promote product bleaching and/or browning (Bunzel and Ralph, 2006; Goméz-Lopez et al., 2008). Nonetheless, ClO₂ has its advantages, including its rapid dissipation and reduced corrosivity when compared with other forms of chlorine (Singh et al., 2021).

InvisiShield^™ technology (hereby referred to as “the product”) is a novel ClO₂ gas delivery system for active antimicrobial packaging. The product material contains a base polymer, a channeling agent, and the active ingredient, ClO_2. The release of the ClO₂ gas is triggered by the relative humidity in the package, which then causes the ClO₂ to migrate through the channeling agent into the package in a controlled manner. InvisiShield^™ is Generally Recognized as Safe (GRAS) by an expert panel and has regulatory allowance on all fresh fruits and vegetables in the U.S. and Canada. For the EU, data on the technology has been submitted to the European Food Safety Authority’s Active and Intelligence body for review. However, the efficacy of this product against microbial pathogens on foods is yet to be reported in the literature. The purpose of this study was to evaluate this novel ClO₂-based antimicrobial packaging system for its efficacy against hNoV and HAV in refrigerated fresh produce.

Chlorine dioxide gas, when delivered using InvisiShield^™ Technology (the product), showed efficacy against both hNoV and HAV on the fresh produce items included in this study (grape tomatoes and blueberries) when stored under refrigeration conditions (7°C). Using log₁₀ GEC reduction as proxy for infectivity, hNoV was inactivated such that reduction approached or reached the limit of detection (LOD; 2.7-3.7 log₁₀ hNoV GEC reduction) on both produce items after 7 days of refrigerated storage. Hepatitis A virus was more difficult to inactivate, but a 2.1 log₁₀ GEC reduction was reached on tomatoes and a 3.1 log₁₀ HAV GEC reduction was observed for blueberries within 7 days. Only the latter was reflective of the assay LOD.

While antimicrobial active packaging is not a new concept, advancements in nanotechnology and developments in packaging technologies has allowed for growth in this field. The most common gaseous compounds used in antimicrobial active packaging include ethanol, carbon dioxide and ClO₂ (Chen et al., 2019; Firouz et al., 2021). As reviewed by Firouz et al. (2021), the release of ethanol into a package headspace has been shown to inhibit the growth of yeasts, bacteria and molds on mango, sliced bread and bayberry. Active packaging containing carbon dioxide gas has also been shown to extend the shelf-life of many food products, including meat, fish and fruits (Firouz et al., 2021). The release of ClO₂ from a modified film was also been shown to reduce counts of Salmonella spp. and E.coli O157:H7 on tomatoes (Chen et al., 2019). While the efficacy of ClO₂ and ClO₂ gas against viruses has been described, as detailed below, to our knowledge, this is the first study to evaluate its efficacy against viral pathogens under active packaging conditions.

In general, chlorine and ClO₂ gas have been shown to be efficacious against viral pathogens such as HAV and hNoV on fresh produce (Li et al., 2004; Lim et al., 2010; Kingsley et al., 2014; Song et al., 2022), but oftentimes at the sacrifice of organoleptic properties. The novel ClO₂ delivery system utilized by the product evaluated in this study produced no significant differences in appearance, flavor or texture attributes of sliced tomatoes as compared to an untreated control sample after a 3-day refrigerated hold time. This demonstrates the potential for this technology as an effective antiviral treatment for refrigerated tomatoes, while maintaining product sensory qualities. Due to funding restrictions, the organoleptic properties of blueberries following treatment were not evaluated in this study, and it is unknown if these results can be extrapolated to this commodity. Treatments with ClO₂ gas have been reported to have differing effects on a variety of fruit and vegetables, and these have primarily been related to appearance. One such defect has been the development of browning. This is likely related to oxidation of mono- and di-phenols into quinones that condense to form brown polymeric pigments (Gómez-López et al., 2009). Typically, this is mediated by the presence of oxygen and the enzyme polyphenol oxidase (PPO) and is most prevalent when fruit or vegetables have been sliced or ground. However, some evidence suggests that ClO₂ can directly oxidize these compounds and cause browning even without PPO.

Browning has been observed in some commodities treated with ClO₂ gas, but not others. For instance, shredded lettuce exposed to ClO₂ showed evidence of browning whereas shredded cabbage and carrots did not (Gómez-López et al., 2008). Apples and peaches also showed signs of browning when treated with ClO₂ gas (Sy et al., 2005). Whether or not a food item would brown depends on the type of commodity, levels of PPO enzymes, whether the product has been cut (thus releasing enzymes), and the level and total treatment time with ClO₂ gas. As observed in this study, the absence of organoleptic changes to treated tomatoes, even when cut, may be suggestive that tomatoes are less susceptible to enzymatic browning, at least as compared to commodities such as apples or lettuce. In addition, it may be that the controlled-release of ClO₂ gas by the InvisiShield^™ system better protects the fruit.

Gaseous chlorine dioxide treatment has also been known to cause bleaching. In fact, it is used in the treatment of paper pulp for just that purpose. One theory is that it causes oxidation of oligosaccharides such as cellulose or hemicellulose, but it is unclear how this promotes lightening or translucency (Gómez-López et al., 2008). In addition, for leafy plants the oxidation of chlorophyll can cause the loss of green color. Thus, bleaching has been seen in iceberg lettuce (Mahmoud and Linton, 2008) as well as julienne carrots and strawberries (Sy et al., 2005). However, blueberries and raspberries were not affected by ClO₂ treatment nor were tomato and onions (Sy et al., 2005). In contrast, strawberries did not undergo color change when treated with 5 mg/L ClO₂ for 10 min (Mahmoud and Linton, 2008). Again, this highlights the importance of optimizing concentration and length of treatment time. Future studies should include evaluation of bleaching as a potential sensory detriment occurring as a consequence of ClO₂ treatment, especially since this phenomenon is likely to be product-specific. Collectively, further studies investigating the effects of gaseous ClO₂ on organoleptic properties of other fruits and vegetables are warranted.

This work is particularly significant given that enteric viruses are recalcitrant to most non-thermal food technology control measures when applied under conditions allowable by law (Seymour and Appleton, 2001), with the added benefit of maintaining product organoleptic properties. Chlorine solutions are one of the more commonly used antimicrobial agents in food processing, however these solutions pose issues for equipment, as they are highly corrosive to stainless-steel surfaces and can cause nose, throat, and eye irritation (Singh et al., 2021). In addition, when chlorine dioxide reacts with dissolved organic compounds in water, it forms disinfection by-products, such as chlorite and chlorate ions. Chlorine dioxide therefore breaks down into chlorate, chlorite, and ultimately chloride. This reaction is rapid and results in the development of chloride salts, very similar to table salt (Gómez-López et al., 2009). InvisiShield^™ has been evaluated by third parties using an IC column to ensure the intermediate products, chlorate, and chlorite, are at safe levels for all humans, including children. These studies showed no detectable chlorate ions were present at an LOD of 0.799 mg/kg (data not shown). Further evaluation of the technology using methods such as LC–MS/MS which have lower limits of detection are ongoing in order to ensure the chlorates and perchlorate levels are lower than 0.05 mg/kg as permitted by the European Commission (2020).

Other chemical treatments, such as organic acid washes, have also been utilized in the processing of fruits and vegetables grown organically (Seymour and Appleton, 2001). However, the ability of viruses that initiate infection through the gastrointestinal tract, including hNoV and HAV, to withstand conditions of extreme pH is well documented and such washes are unlikely to have significant effect on these viruses (Cromeans et al., 2014). Ozone inactivation of enteric viruses on produce has also been studied, with results showing food-matrix associated variability. Specifically, Hieisen et al. (2016) demonstrated that ozone treatment of 6.25 ppm for 10 min resulted in a 3 log₁₀ reduction in feline calicivirus (FCV) and murine norovirus (MNV), two cultivable hNoV surrogates, on lettuce, and close to 4 log₁₀ reduction of MNV on green onions. However, only an approximate 2 log₁₀ reduction of FCV was observed in green onions. In short, the efficacy of chemical treatments to inactivate viruses on foods appears to be quite variable between viruses and also between different food matrices.

Non-chemical approaches to controlling pathogens on fresh produce include ionizing radiation, pulsed light, and high-pressure processing (HPP). DiCaprio et al. (2016) reported that application of ionizing radiation at a dose greater than the FDA regulated limit, to fresh strawberries inoculated with hNoV and its cultivable surrogate, Tulane virus, did not effectively inactivate the viruses, resulting in less than one log₁₀ reduction. Similarly, in part due to varying surface characteristics for berries, pulsed light was found to not be an effective means by which to reduce the concentrations of HAV and MNV on blackberries, strawberries and raspberries, with only 1–2 log₁₀ reductions (Jubinville et al., 2022). Pulsed light also caused a reduction in surface luminosity (browning) for certain produce items, including blueberries and lettuce (Jubinville et al., 2022). High pressure processing has been shown to effectively inactivate the MNV cultivable surrogate, up to 5 log₁₀ on lettuce berries and in fruit purees, however the impact on product quality was variable, leading to the opinion that HPP may be better suited for products such as purees and sauces (Lou et al., 2011). As with chemical treatments, the performance of non-chemical approaches to the control of viruses in foods is quite variable.

A major barrier in studying the efficacy of antimicrobials against human enteric viruses is the absence of methods by which wild-type strains can be cultivated in the laboratory. While a cell culture-adapted strain of HAV is commercially available, its growth is slow and its cultivation expensive. As a proof-of-concept study, sufficient funds were not available to commit to the use of HAV infectivity assays in our study. Recent hNoV cultivation methods have met with some success (Ettayebi et al., 2016; Dycke et al., 2019) but they are in the early phases of adoption, and importantly, cannot provide quantitative data on log₁₀ inactivation. As described in many of the studies cited above, cultivable surrogates such as MNV, FCV and Tulane virus are often used as proxies, however they have been shown to respond differently, usually being more sensitive to inactivation strategies than are hNoVs (Cromeans et al., 2014). For these reasons, we chose to evaluate the efficacy of the product against both HAV and hNoV using the RNase-RT-qPCR method. While technically not an infectivity assay, this approach remains widely used and its efficacy in direct comparison to hNoV surrogates has been demonstrated in several studies (Knight et al., 2012). In a companion piece, Knight et al. (2016) concluded that the persistence of RT-qPCR signals for hNoV strains exposed to sodium hypochlorite was equal to or greater than that which was observed for the cultivable surrogates. This was the case regardless of inclusion of an RNase pre-treatment and may suggest that using RT-qPCR to evaluate chemical treatments such as ClO₂ may actually produce data that underestimates, rather than overestimates, the efficacy of the technology. Clearly, there is precedence in using RNase-RT-qPCR in enteric virus inactivation studies, and the data produced may be conservative.

Further study of InvisiShield^™ Technology is warranted. Future research could focus on investigating this technology relative to inactivation of other foodborne pathogens, particularly those that are non-cultivable; exploring the efficacy of the product against viral pathogens via infectivity assay or on other produce items; alternative storage conditions (such as frozen or room temperature environments); and the sensory analysis of other commodities following treatment with InvisiShield^™. Nonetheless, this study is a good first step toward proof-of-concept that active ClO₂-based packaging has antiviral efficacy against hNoV and HAV on blueberries and tomatoes, while showing little effect on sensory characteristics of tomatoes. This makes the technology a good candidate for future study.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

RG, JF, WK, AR, MJ, and L-AJ contributed to the experimental design. JF performed the microbial bench top work. RG and L-AJ performed data analysis for microbial work. WK performed the sensory analysis experiments, data analysis, and wrote sections of the manuscript related to sensory work. RG drafted the first full version of the manuscript. All authors contributed to the article and approved the submitted version.

This work was made possible by funding from Aptar CSP Technologies.

AR and MJ are employees of Aptar CSP Technologies, who produce the product tested in this study.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.