Since January Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Ultraviolet (UV)-C light for disinfection has experienced a surge in popularity since the outbreak of COVID-19. Currently, many different UV-C systems, with varied properties that impact disinfection performance, are available on the market. Therefore this review aims to bundle the available information on UV-C disinfection to obtain an overview of its advantages, disadvantages, and performance-influencing parameters. A literature search was performed using the snowball search method in Google Scholar and PubMed with the following keywords: UV-C disinfection, UV-C dose, UV-C light source, UV-C repair mechanism, UV-C photoreactivation, and UV-C disinfection standards. The main parameters of UV-C disinfection are wavelength, dose, relative humidity, and temperature. There is no consensus about their optimal values, but, in general, light at a high dose and a spectrum of wavelengths containing 260 nm is preferred in an environment at room temperature with low relative humidity. This light can be generated by mercury-vapour, light-emitting diode (LED), pulsed-xenon, or excimer lamps. Multiple factors are detrimental to disinfection performance such as shadowing, a rough surface topography, a high level of contamination, repair mechanisms, and the lack of standardization. Also, there are health and safety risks associated with the UV-C technology when used in the proximity of people. UV-C disinfection systems have promising features and the potential to improve in the future. However, clarifications surrounding the different parameters influencing the technologies' effectiveness in hospital environment are needed. Therefore UV-C disinfection should currently be considered for low-level rather than high-level disinfection.
After UV-C irradiation, micro-organism repair can occur, resulting in a reduced final inactivation by partially reversing the induced disruption after irradiation. Repair mechanisms can be subdivided into two classes, namely photoreactivation and dark repair, of which the first type has the largest impact. It operates through the enzyme photolyase and requires visible light as an energy source. Dark repair, on the other hand, requires multiple enzymes and nutrients for energy [ 6 ]. It is important to know whether final inactivation results have taken into account the occurrence of reactivation since it may result in 60% of the achieved inactivation being reversed [ 7 ]. Moreover, mutations can arise upon UV-C exposure since this exposure can result in the origination of intra-strand cyclobutyl-pyrimidine dimers in DNA [ 6 ]. Therefore, one should be careful not to induce unwanted side-effects such as rendering micro-organisms more pathogenic or resistant by using UV-C devices.
The amount of inactivation achieved by UV-C light depends on multiple variables such as wavelength, dose, relative humidity, and temperature. The wavelengths that can be emitted are determined by the light source. Every type of light source has its own properties and thus a specific spectrum.
DNA, RNA, or proteins of a micro-organism absorb UV light, with a peak absorbance around 260 nm [ 6 ]. This results in the disruption of DNA or RNA, leading to the inactivation of the micro-organism. UV-C-induced DNA disruption often consists of the bonding of two neighbouring thymine (or cytosine) bases instead of the conventional linking of a base with its complementary base on the other strand. The result is known as thymine (or cytosine) dimers [ 6 ]. In RNA a similar process occurs with uracil bases. For DNA viruses, bond formation between thymine bases and proteins in the virus capsid can occur as well, whereas for RNA viruses, the capsid may be damaged by uracil dimers. The virus capsid, therefore, does not provide much additional protection against UV-C disinfection [ 6 ].
Over the last three years the medical community has been, and still is, fighting a difficult battle against COVID-19 while experiencing a lot of shortages, especially in personal protective equipment (PPE). This led to an increased interest in ultraviolet (UV)&#;C light due to its disinfection potential for PPE but also for air, surfaces, and other materials [ 1 ]. However, UV-C disinfection is not a novel concept as its effects were first observed in by Downes and Blunt when they discovered that sunlight can avert the growth of micro-organisms [ 2 ]. A second breakthrough occurred in when Niels Finsen won the Nobel Prize for his research on the use of UV light against tuberculosis [ 3 ]. Another application arose in when UV light was used to disinfect water. However, the technology was not very reliable at the time and it took further technological developments before UV water disinfection became popular again in the s [ 2 ]. Nowadays, UV light is used for water, air, food, surface, and medical equipment disinfection. Some companies even offer UV-C disinfection systems for semi-critical medical equipment, such as ultrasound probes and endoscopes [ 4 , 5 ].
A literature search was conducted to investigate UV-C light for disinfection and to discover relevant parameters, its advantages and drawbacks. To achieve this, the snowball search method was used in Google Scholar and PubMed, starting from information supplied by different manufacturers and following keywords: UV-C disinfection, UV-C dose, UV-C light source, UV-C repair mechanism, UV-C photoreactivation, and UV-C disinfection standards. Only English-language sources were included. The search was conducted in August and publication dates of the reviewed sources range from to .
UV light can be subdivided into four zones: UV-A (315&#;400 nm), UV-B (280&#;315 nm), UV-C (200&#;280 nm), and vacuum UV (100&#;200 nm). The UV-C zone is used for disinfection but there is no consensus on the exact optimal wavelength. Bacterial DNA and RNA have peak absorbances of light at 260&#;265 nm and around 260 nm, respectively [6]. Therefore light at 260 nm can cause the most disruption. However, various micro-organisms are most vulnerable to slightly different wavelengths. Emitting a wavelength spectrum can thus be interesting since it can optimally target multiple micro-organisms at once [11]. It even has an additional benefit by reducing photoreactivation through a decrease of photolyase [9]. On the other hand, it has technical implications since the total energy of the light beam is then divided over all present wavelengths. Therefore, a micro-organism that is vulnerable to 254 nm light will be inactivated more by a lamp that emits solely light at 254 nm than a lamp that emits a wavelength spectrum at equal total energy.
The applied dose impacts disinfection effectiveness. In general, bacteria require a lower dose than viruses, which in turn require a lower dose than fungi or yeasts to be inactivated [12]. However, there is uncertainty about the exact required doses to achieve a certain reduction for a specified micro-organism. These variations can be explained by a lack of standardization which leads to the use of different dose measurement techniques by researchers [13]. Measurements are challenging since the dose is difficult to measure directly. In general, the intensity and exposure time are measured and then the dose is calculated thus [14,15]:
dose (mJ/cm2) = intensity (mW/cm2) × exposure time (s) (× duty rate (%)/100)
In the case of pulsed light, the duty rate, which is the ratio of the duration of the pulses to the duration of the pauses between them, influences the effective exposure time. For continuous light the duty rate is 100% [15]. Exposure times of 10&#;45 min for room disinfection and 25 s to 5 min for medical equipment were encountered in literature.
The intensity is inversely proportional to the squared distance between the light source and the surface and is therefore defined at the surface in the dose calculation equation [14]. It is clear from this definition that the UV intensity received by a surface decreases significantly the further the surface is from the UV source. Further, the output of a lamp decreases over time, so it is recommended to calculate the dose at the end of lamp life, which is representative of a worst-case scenario. The dose also influences the amount of photoreactivation. Quek et al. found that the percentage of photoreactivation decreased from 5.31% to 0% for an increase in dose from 1.6 to 19.7 mJ/cm2 [8]. Thus, a higher dose reduces photoreactivation and increases the achieved initial inactivation. There is a risk of material damage after irradiation which can be prevented by keeping the applied dose under a material-dependent threshold [16,17].
A higher relative humidity causes a lower UV efficiency due to decreasing susceptibility of micro-organisms to UV-C light and a decline in UV irradiation. The reduction in susceptibility occurs due to the formation of a water layer around the micro-organisms, protecting them against UV-C-induced DNA or RNA disruption. UV irradiation, on the other hand, decreases due to absorption, refraction, and reflection of UV light by water molecules in the air. Zhang et al. observed a change in UV irradiance of 34% when the RH increased from 50% to 90% [18]. The amount of RH influence on UV efficiency depends on the present micro-organism and is more apparent for bacteria than for viruses [16].
Lastly, the influence of temperature depends on the light source. Zhang et al. found that for mercury-vapour low-pressure (LP) lamps an optimum for UV-C efficiency is achieved around 20&#;21 °C [18], whereas for LEDs the temperature dependence is not strong but their output decreases with increasing temperature [19]. Furthermore, low temperatures lead to a reduction in cell viability, resulting in less and slower photoreactivation [10].
summarizes the optimal values of the parameters discussed.
Currently, four types of light sources are used to emit UV-C light, as indicated in . Mercury-vapour lamps are the conventional choice but they are being replaced by newer generations of UV-C instruments. These newer lamp types do not contain mercury, which makes them more environmentally friendly [20]. Apart from that, LEDs and pulsed-xenon lamps, unlike mercury-vapour lamps, have no warm-up time [20,21]. Lastly, the output of mercury-vapour lamps varies noticeably with temperature whereas this effect is much smaller for the other light sources [19,22]. Mercury-vapour lamps can be divided into three classes, namely low-pressure (LP), medium-pressure (MP), and high-pressure (HP) lamps, of which the LP lamps have the highest UV-C efficiency and are thus most used [23]. These classes have different emission spectra, as can be appreciated in [9,23,24]. The values for LP lamps indicate that almost all the light is emitted at 185 nm or 254 nm, with a peak emission of 254 nm. MP and HP lamps, on the other hand, emit a discontinuous spectrum. For LEDs, the light emission peak can be modified by the manufacturer to a value between 255 and 275 nm [21]. By using LEDs at different wavelengths in one system, a wavelength spectrum can be emitted. Pulsed-xenon lamps always generate a spectrum, ranging between 200 and nm, which often undergoes filtering to emit mainly UV-C light [21,25]. Lastly, there are different types of excimer lamps, all with their characteristic spectrum. For UV-C disinfection, krypton chloride (KrCl) lamps that emit light at 222 nm are the most common [21]. This is known as far UV-C technology and is a relatively new disinfection method with limited knowledge about its effectiveness. This makes it riskier to completely rely on this technology for disinfection in the hospital and it is therefore not widely used [21]. However, it also has benefits such as a reduced risk for use near people due to a lower penetration depth into the skin and eyes [26]. Although the FDA also mention this potential, they commented that currently there are not enough long-term data available to be sure of its long-term effects on people [25]. Furthermore, ozone will be produced due to the emittance of wavelengths <240 nm, resulting in other health risks [21,26].
Another difference between the light sources is the emittance of continuous or pulsed light. According to some investigators, pulsed light is preferred since it has a higher penetration depth and can produce higher light energy, increasing intensity [27]. In research, the results on pulsed versus continuous UV-C disinfection efficiency vary. When comparing pulsed and continuous light it is important to keep other variables such as wavelength and dose constant. Nyangaresi et al. and Sholtes et al. both found that pulsed or continuous light emitted by LEDs led to comparable log10 reductions [15,28]. Luo et al., on the other hand, found pulsed light, at certain frequencies, to be more bactericidal [29]. Further, Nyangaresi et al. found that pulsed and continuous irradiation led to comparable photoreactivation [15].
UV-C light and ozone may be associated with human health risks. UV-C light poses a risk to the skin and eyes, whereas ozone is mainly harmful to the respiratory tract. Both can be solved by using a UV-C disinfection system that does not require human intervention. Otherwise, protective clothing can be worn against UV-C light [16]. In case ozone is not required for disinfection, a modified lamp can be used. For mercury-vapour lamps, doped quartz glass or specialized soft glass can filter out short-wave UV-C light. For pulsed-xenon, doped quartz can be used as well [30].
UV-C has promising features for disinfection such as automatic disinfection, being less time-consuming than widely used manual or chemical disinfections, leaving no harmful residuals, and being environmentally friendly (if no mercury-vapour lamps are used) [31,32]. However, it also has limitations which can be found in together with their proposed (partial) solutions.
The first difficulty is that UV-C light needs a direct path to an object to be able to disinfect it. Thus, additional dirt on the surface is inconvenient but it can often be solved by a water-based pre-cleaning. However, it is also possible that the light becomes obstructed by other objects or that it only irradiates one side of an object. This is known as shadowing and indicates the increased risk of active micro-organisms remaining in non-illuminated areas. A first solution is using a reflective chamber in which the object disinfection takes place or placing reflective surfaces in a room that needs disinfection. In this way, UV-C rays are reflected, resulting in additional pathways and thus reduced shadowing [32]. However, one should be aware of the negative effect on UV intensity of this extra distance caused by this reflection step. Another option is combining UV-C disinfection with ozone disinfection. As mentioned, ozone is produced when the emitted light spectrum contains wavelengths under 240 nm, and may be helpful since it can freely move into the shadows to ensure disinfection [33]. The surface topography can also influence disinfection performance. First, high surface roughness or irregularities can cause shadowing. Second, these irregularities lead to an increase in effective surface area and thus higher required energy to maintain the intensity. After all, intensity is defined as energy per effective area [13]. The dose, and thus the achieved inactivation, would decrease if the intensity were not preserved under unchanged exposure time.
Compared to other disinfection techniques, UV-C disinfection has a low penetration depth. It is speculated that this depth increases when using pulsed light [27]. However, there is still no consensus in the literature regarding the preference for pulsed or continuous light.
Other difficulties surrounding UV-C disinfection technology are the lack of a uniform standard applied for commercially available devices, the uncertainty about delivered doses, and the rather low quality of available research. This last drawback became clear while reviewing the literature but is also reported by others [[34], [35], [36]]. The quality is often low due to low sample sizes, bias, and conflict of interest. Further, it is also challenging to draw definite conclusions from literature data due to the heterogeneity across studies, rendering the merge of data difficult [[34], [35], [36]].
UV-C disinfection standardization is a new field, resulting in a variability of standards being used by providers to prove efficacy. However, not all standards have the same evidential value. In Europe, EN is often used, although it has been developed for chemical disinfectants. It discusses the test standards and required log10 reductions to claim bactericidal, virucidal, fungicidal, sporicidal, and mycobactericidal activity [37]. If this standard is followed entirely and properly it can be used to claim high-level disinfection [38]. However, it is unclear how this standard and its corresponding tests are adapted to UV-C disinfection by companies that use it. Therefore it is uncertain whether the evidential value of the standard remains as powerful for UV-C disinfection as for chemical disinfectants. For example, it does not take the risk of repair after UV-C disinfection into account, which can impact the final inactivation efficacy and thus the classification as a high-level, intermediate-level, or low-level disinfectant. However, there are superior test standards available that apply specifically to UV-C disinfection, namely the ASTM E-18 for general use and the ASTM E-18 for the influenza virus. They define how to test UV-C light against micro-organisms on carriers but they do not describe how to handle shadowing [39,40]. The results of these tests can be compared to the requirements set by the US Environmental Protection Agency and the US Food and Drug Administration (FDA) to evaluate disinfection performance. For low-level disinfection a 3-log10 reduction for viruses and a 5-log10 reduction for bacteria are dictated [41]. For high-level disinfection a 6-log10 reduction of mycobacteria is required [42]. One should be aware that manufacturers can select for themselves the micro-organisms to be tested from a list available in the E-18 standard. It is important to ask the manufacturer against the micro-organisms its UV-C device has passed compliance with the standard. This can make a difference for a specific healthcare institution aiming to remediate a micro-organism specific problem.
A recent standard EN :, derived from the French standard NFT 72&#;281, considers tests to be used by suppliers of automated airborne disinfection systems to claim defined antimicrobial activity. The standard focuses on devices distributing chemicals by air diffusion and describes requirements such as uniformity of biocide distribution, total airborne disinfection contact time, the specific micro-organisms used and the test volumes of the enclosure. The objective of the described processes of distribution of chemicals by air diffusion is to disinfect the surfaces of the overall area including the external surfaces of the equipment contained in such rooms (distribution test) [43].
The most recent standard, BS :, is specifically developed to cover the requirements and methodology for testing the efficacy of UV devices. The standard is based on EN : with some minor differences specific to UV devices. By contrast with EN :, the method consists of only the efficacy test. The distribution test is not performed since UV devices are not intended to decontaminate the whole room but rather &#;close by&#; surfaces [44].
The reported efficacy in literature varies, partially due to the heterogeneity that is present in UV-C research. After all, the large diversity in UV-C disinfection devices leads to different applications and parameter values. This makes it difficult to compare effectiveness across studies [34]. However, Alvarenga et al. state, in their systematic review, that most studies report a 1- to 2-log10 reduction for UV-C disinfection combined with manual precleaning for surfaces. Other review articles also recommend or describe UV-C disinfection rather as a supplementary method to standard cleaning than as a stand-alone disinfection procedure [35,45]. Further, Alvarenga et al. mention that in laboratory experiments disinfection between 2- and 6-log10 was obtained. However, in real life, the achieved reductions might be lower due to suboptimal conditions including the presence of shadows and larger distances between the surface and the lamp [34]. It is therefore recommended to verify whether the UV-C device performs as well for the intended application as under test conditions, since RH and temperature can vary [46].
Commercial qualitative colorimetric indicator cards are available to assess delivery of UV-C light on surfaces. Although they provide rapid and easy-to-interpret information in a rather cheap manner, they provide only rough estimates of UV-C delivery. They confirm that in-use devices are working correctly and are helpful during initial training and verification for determination of the ideal cycle time and positions of the devices in a specific setting. When looking for a reliable colorimetric indicator when using a pulsed xenon UV-C device, it is important to check their susceptibility to change colour by UV-A and UV-B light exposure [14,47].
Some disinfection devices make use of a built-in dose-monitoring system that continuously measures the UV-C dose given during the disinfection cycle [4]. This system is used to replace colorimetric indicator cards. An example of how this can be achieved is by using two photodiodes for monitoring and a third independent photodiode for validation of the disinfection cycle. However, no irrefutable proof of the proper functioning of these types of systems was found during the literature search.
For semi-critical devices, high-level disinfection is recommended [42]. Although some UV-C disinfection companies claim to offer this, no definite proof of high-level disinfection with UV-C systems was found in the literature. There are studies reporting up to 7-log10 reductions for artificially contaminated flexible endoscopes [48]. However, they often do not take the risk of genomic repair into account or have a conflict of interest. Furthermore, UV-C light is not FDA-cleared for high-level disinfection and the US Centers for Disease Control and Prevention only mentions it in non-critical applications [49,50]. It should be borne in mind that these latter sources have not been updated since and , respectively.