Viruses on some foods can be inactivated by exposure to ultraviolet (UV) light. This green technology has little impact on product quality and, thus, could be used to increase food safety. While its bactericidal effect has been studied extensively, little is known about the viricidal effect of UV on foods. The mechanism of viral inactivation by UV results mainly from an alteration of the genetic material (DNA or RNA) within the viral capsid and, to a lesser extent, by modifying major and minor viral proteins of the capsid.
1. Introduction
Ultraviolet (UV) light refers to the portion of the electromagnetic spectrum that falls in the wavelength range of 100–400 nm. This band is divided into vacuum UV (100–200 nm), UV-C (200–280 nm), UV-B (280–315 nm) and UV-A (315–400 nm). The range useful for inactivating viruses and microorganisms includes UV-C and UV-B up to about 300 nm. The 100–200 nm sub-band can render viruses non-infectious, but only in a vacuum, since these wavelengths are readily absorbed by air.
Artificial UV light can be provided by lamps. The most common type is the low-pressure mercury lamp, which emits a quasi-monochromatic output at 253.7 nm. Medium-pressure mercury lamps with polychromatic emission, light-emitting diodes that generate narrow bandwidth emissions in different ranges and excimer lamps containing a rare gas halide such as KrCl* (222 nm) or XeBr* (282 nm) are also available
[1].
A special case of light-based disinfecting technology is pulsed light. These systems generate flashes of high-intensity polychromatic light ranging from UV to infrared by storing electric energy in capacitors, which are discharged to a xenon lamp
[2]. Pulsed light can inactivate viruses faster than conventional light sources and is generally more efficient than low or medium pressure lamps on a time and fluence basis
[3][4].
Viral inactivation by UV light is a function of the energy impinging on the targeted surface, called fluence (
F, J/cm
2), defined formally as “the total radiant energy traversing a small transparent imaginary spherical target containing the point under consideration, divided by the cross section of this target”
[5]. If the fluence rate (
E) is constant over time, the fluence is determined according to the following equation:
where t is the exposure time in seconds. For the specific case of pulsed light technology,
where
Fp is the fluence per pulse and
n the number of pulses
[6].
The use of UV light for water disinfection, including destruction of viruses, is a well-established industrial practice that has been in continuous use for more than 50 years
[7] to process as much as 8.5 billion gallons of potable water per day
[8]. The use of UV light to inactivate bacteria in foods has been studied extensively
[9][10], while its potential for inactivating viruses in or on foods has received less attention.
According to the latest World Health Organization (WHO) report on the subject, foodborne diseases are a significant global burden causing high rates of morbidity and mortality
[11]. At least 31 foodborne agents, mainly microbial, often viral and sometimes chemical, caused an estimated 600 million illnesses in 2010
[11]. In addition to disastrous consequences for public health, contamination of food by viruses remains a significant economic issue for stakeholders in the food industry, where recall costs and drops in sales can be financially ruinous
[12]. In the United States and Canada, more than 20 recalls due to foodborne viruses were issued from 2017 to 2021, specifically for frozen berries and shellfish
[13][14][15]. Despite the prevalence of foodborne viral illness, interest in this domain has emerged relatively recently. It was not until 2008 that the Food and Agriculture Organization (FAO) in collaboration with the WHO concluded that viruses were causing an increasing proportion of foodborne infectious intestinal diseases. The scarcity of epidemiological data, the difficulty of detection, persistence on food contact surfaces and limited post-harvest decontamination options were all identified as contributors to the emerging problem. Two viruses stood out for their number of cases and the severity of their symptoms: human noroviruses (HuNoV) and hepatitis A virus (HAV)
[16]. It is estimated that the number of foodborne illnesses caused by HuNoV annually has passed 124 million cases with 34,000 deaths worldwide
[11]. HuNoV appears to be causing 58% of foodborne illnesses in the USA
[17] and 64% in Europe and the British Isles
[18]. For HAV, the number of foodborne cases is smaller (14 million worldwide), but the mortality rate is 10 times higher (28,000 per year) and its incidence is increasing at a worrying pace in and around Europe
[11][19]. HAV is transmitted via various routes starting from feces, including contaminated irrigation water, fishing sites or infected food handlers (symptomatic or not)
[20][21][22]. Ready-to-eat and minimally processed foods such as shellfish, vegetables, and fruits are often implicated in HuNoV and HAV outbreaks
[16]. During the past decade, berries, particularly frozen, have been at the origin of several major outbreaks of both viruses around the world
[23][24], including one of the largest outbreaks ever of HuNoV (>11,000 cases), which occurred in Germany in 2012, believed to be caused by strawberries from China
[25]. In 2013, frozen berries were the source of a large and prolonged multistate European outbreak of HAV centralized in Italy with more than 1800 cases
[26][27]. Green onions were the source of an outbreak of HAV that led to 601 cases and 3 deaths in the United States in 2003
[28]. The mollusc origin of several epidemics around the world has shown the significance of contaminated water as a vector of HuNoV
[29][30][31] and HAV
[32][33]. Involvement of food handlers served in public is also reported frequently
[19][34]. In recent years, public health authorities have noted an increase in cases of hepatitis E virus (HEV), an emerging foodborne virus, particularly in countries with high standards of hygiene
[35][36]. Compared to HuNoV and HAV, outbreaks of HEV tend to be small and isolated
[19]. HEV has been found in seafood
[37][38][39][40], fresh and frozen produce
[40][41][42] and, more recently, in the milk of different animals
[43][44]. However, pork and game such as wild boar and deer are the sources most often implicated in foodborne cases
[19][45]. Although HEV is the most common hepatitis virus associated with meat products, HAV has been detected in processed and raw meats in South America
[46] and 117 cases of HuNoV were traced to an infected handler employed in the production of a ready-to-eat meat product
[47].
2.1. Non-Food Liquids
Water is relatively easy to disinfect with UV since it does not absorb light and can be mixed to ensure equal exposure of suspended particles including viruses. Transparent liquids such as phosphate-buffered saline, of which the absorbance in the 200–1100 range is close to zero, are therefore preferred for in vitro tests
[48]. The amount of information available on the use of UV light to disinfect water is considerable. Indeed, phage MS2 is used for UV reactor validation in North America
[49]. A classic review article on this subject was published in 2006 by Hijnen et al.
[50] and comprehensive kinetic data were published in 2009 by Kowalski
[51].
Pulsed light is a more recent technology and, therefore, less is known about its effectiveness for water disinfection. The first published account of its use for inactivation of viruses, including some foodborne, appears to be that of Roberts and Hope (2003), who obtained at least 4.8 log reductions in HSV-1, HAV, poliovirus 1, canine parvovirus, bovine parvovirus and non-foodborne viruses (Sindbis, vaccinia, encephalomyocarditis and Simian virus 40) with only two pulses of 1 J/cm
2 each for less than 1 s of treatment time
[52]. Lamont et al. (2007) later obtained 4 log reductions in poliovirus 1a and Group D adenovirus infectiousness in PBS, Jean et al. (2011) at least 4 log reductions in MNV-1 and HAV in PBS and Huang et al. (2017) at least 5 log reductions in MNV-1 and TV in PBS
[48][53][54]. Inactivation was significantly less in PBS containing dissolved protein
[52][48], which absorbs in the UV range.
Table 1 summarizes the different viral inactivation results obtained using pulsed light (
Table 1).
Table 1. Inactivation of viruses by pulsed light.
2.2. Liquid Foods
The inactivation of viruses in liquid foods has been studied using coconut water
[59][60] and skim milk
[61][62] challenged with phages MS2 and T1UV in two types of experimental set-up: collimated beam in a stirred batch reactor and a Dean vortex continuous reactor. Reductions in viral titer reached 4 log cycles for MS2 and more than 5 log for T1UV in coconut water
[59][60] and at least 5 log for both MS2 and T1UV in skim milk
[61][62]. It is widely known that the efficacy of UV treatments of liquid foods decreases with turbidity and suspended solids, which absorb or scatter the light before reaching the viral particles. However, this limitation can be overcome by reactor designs that promote turbulent flow to maximize the exposure of any virus present. In the above cases, UV transmittances as low as 9.7%/cm in coconut water and 0.57–0.89%/cm in skim milk were sufficient. The formation of potentially toxic compounds due to the action of UV light on food components was also examined. No cytotoxicity was observed when skim milk was treated with up to 0.17 J/cm
2 [62] or coconut water with up to 0.4 J/cm
2 [60], fluences that were high enough to reduce viral titers substantially.
2.3. Meat Products
The use of UV light as a single method for viral inactivation in meat products is possible, although results are modest. This is likely because the irregular surface shields viral particles from exposure and because meat proteins can absorb part of the illumination
[63]. For example, only 1.23 and 1.17 log reductions in MNV-1 and HAV, respectively, were obtained on fresh chicken breast. The inactivation curve in both cases progressed without tailing, suggesting that higher fluences (up to 3.6 J/cm
2) could increase inactivation. However, the sensory quality of the chicken meat was decreased at fluences above 1.2 J/cm
2, at which only 0.5 log reductions were observed for both viruses. UV-C treatment combined with another decontamination technique would likely give better results
[64]. A still modest but higher inactivation of FCV and
Escherichia coli coliphages MS2 and φX174 (0.97–2.8 log reduction) was observed on pork liver, ham and sausage treated by pulsed light at fluences of 45–60 J/cm
2 [56].
2.4. Fruits and Vegetables
Berries are delicate products very susceptible to damage and are usually just washed in tap water or chlorinated water during processing. These processes are insufficient to eliminate human pathogens that may be present. HAV and NoV have been implicated in numerous outbreaks related to fresh and frozen berries
[65]. Decontamination of fresh and frozen strawberries, blueberries, and raspberries spiked with HAV and MNV-1 has been attempted (
Table 2).
Table 2. Inactivation of viruses on fruits or vegetables by UV-C.
Even though UV-C was able to reduce viral titers on berry surfaces by up to 3 log cycles in some cases, its efficacy was limited and prolonging the treatment beyond 1.3 J/cm
2 did not improve inactivation. Shielding by the irregular surface of berries is the likely cause of this, despite the UV reactor being designed with mirror-finish reflectors on all sides to maximize exposure. UV-C might nevertheless be effective on berries as part of a hurdle approach, since it has only a mild effect on fruit sensory quality and does not produce the toxic by-product furan
[65]. The irregular surface of strawberries was a major obstacle for viral inactivation
[67], which ranged from 1.9 to 2.6 log TCID5
0/mL for FCV, HAV and Aichi virus, several log cycles lower than achieved on foods with smoother surfaces such as lettuce and green onions.
Pulsed light has been tested in a process by which berries are suspended in agitated water to avoid heating and facilitate berry movement and rotation and thereby allow more uniform exposure. Log reductions of 1.8 and 3.6 in MNV-1 titer were obtained for strawberry and raspberry, respectively
[57], which were significantly higher than the inactivation obtained after immersion for an equivalent time (1 min) in water containing 10 ppm chlorine. As in the case of UV-C light
[65], inactivation was greater on smoother-surfaced fruit (raspberry) than on strawberry. Furthermore, the higher efficacy of this water-assisted treatment versus chlorinated water suggests using this method as an alternative to chlorine, which can generate harmful by-products such as trihalomethanes.
Different research groups have tried to implement UV-C treatments to increase the safety of another product that is consumed fresh, namely lettuce. The efficacy of UV-C in this case is high, at least 4 log TCID
50/mL for HAV, Aichi virus and FCV on Romaine lettuce
[67] but not for phage MS2 on Iceberg lettuce
[72]. In all cases, inactivation was initially rapid then tailed off. A low level of inactivation was also observed for MNV-1
[71] but with a log-linear decline.
Green onions may or may not be disinfected depending on the type of virus. High levels of inactivation (>5 log TCID
50/mL) have been reported for HAV
[66][67], but not for FCV, Aichi virus
[67], human adenovirus type 41 and MNV-1
[66], in decreasing order of inactivation. Inactivation was negligible when viruses were internalized, which can occur during hydroponic growing by uptake through roots. These results are to be expected since UV-C is a superficial treatment.
2.5. Other Food Types
Viruses are very stable in low-moisture foods. UV-C light applied in combination with ozone and hydrogen peroxide vapor can inactivate practically 100% (4 log) of FCV and MNV in chocolate, but only 1 log in pistachios and <1 log in cornflakes
[73].
Table 3 lists the inactivation of different viruses on foods other than vegetables by conventional UV-C light (
Table 3). The efficacy of pulsed light has also been tested using MS2 virus on black pepper, garlic and chopped mint, but the titer was reduced by only 1.28 log in the best case
[55].
Table 3. Inactivation of viruses by UV-C on foods other than vegetables.
2.6. Food Contact Surfaces
The virucidal efficacy of UV light on food contact surfaces has been tested using stainless steel and different types of plastics. Inactivation is theoretically favored on stainless steel because of the low degree of shielding, and was significant using UV-C, 4.4 log for MNV-1 and 2.6 log for HAV
[69]. Both inactivation curves exhibited tailing, indicating that extending the exposure time would not increase inactivation proportionately. On the other hand, inactivation was completed when using pulsed light (5 log reductions) on stainless steel and PVC
[48]. Using MNV-1 also, Vimont et al. (2015) observed reductions of 3.8 to 4.3 log on high-density polyethylene and stainless steel and total inactivation on PVC
[58]. Both studies found that viral inactivation is decreased by several orders of magnitude when the surfaces are fouled.
3. Conclusions and Future Trends
Published results on viral inactivation by UV light and pulsed or otherwise are consistent with what would normally be expected. UV light is a highly efficient virucide (directly altering the genetic material of viruses and modifying the composition of viral proteins) for transparent liquids and even opaque liquids if turbulent flow is maintained. On solid foods, surface irregularity limits inactivation. The limited amount of information available and the high variability in the response of different types of viruses to UV light preclude reaching any definitive conclusions about the commercial applicability of this technology. UV-C is somewhat less effective on meat, berries, and dry herbs. However, it may be combined with other technologies as part of a hurdle approach to achieve better results. The agitated-water-assisted UV treatment appears to be more effective than just rinsing with chlorinated water for disinfecting berries and deserves further exploration. Inactivation kinetics are, in some cases, log-linear and in others logarithmic followed by tailing. In the former case, higher fluences might yield better results for some virus–food pairs. UV can also be useful for food–contact surface decontamination. Acquiring additional knowledge on the impact of UV light on foodborne viruses and their laboratory surrogates remains a key process in our advancement towards better understanding of food safety and improving the future security of food industries and their customers.