Presence of Influenza Virus on Touch Surfaces in Kindergartens and Primary Schools

Abstract

Influenza virus infections occur in all ages, but rates of infection are typically highest in children [1]. The more intense social contact among children in schools has been suggested to support rapid dispersion of infection in the community [2]. The impact of school closures (and later reopenings) on influenza epidemics is indicative of the importance of schools in influenza epidemiology [3].

Influenza can spread from one person to another through several routes, including droplets of various sizes as well as indirect contact via fomites. However, the relative importance of the various possible transmission modes in various settings including schools remains unclear [4], hampering efficient mitigation measures. Apart from droplets and direct contact, indirect contact transmission may also be important in schools, where children spend long hours in close proximity to each other and also share objects frequently, while having relatively lower personal hygiene. Children may also shed more influenza virus when infected, for a longer duration [5], possibly leaving higher dose of infectious virus on contaminated surfaces.

To understand the potential significance of indirect contact transmission among children in school, the presence of influenza virus on touch surfaces in schools serves as an important prerequisite. Our work presented here evaluated the potential of indirect contact transmission in schools by sampling common surfaces in kindergartens and primary schools for the presence of influenza viruses, and we provided insights on the contact points that are at higher risk for influenza infection for children in schools.

METHODS

We conducted a longitudinal study during the 2017/2018 winter influenza peak season in Hong Kong. Four primary schools and 3 kindergartens were recruited by convenience sampling to serve as study sites for weekly environmental sample collection. In each school, samples were collected from 1 to 2 classrooms for children 4–9 years of age. These classrooms have an average class size of 26 children, which corresponds to the normative class sizes locally. None of these participating schools were undergoing intensive hand hygiene or environmental disinfection efforts.

Surface Sampling

We referred to existing guidelines on prevention of communicable diseases in schools by health authorities and previous studies [6, 7] to select a list of commonly touched surfaces for sample collection, including doorknobs, desks and chairs of students, bookshelves, shared toys, and walls. These surfaces were categorized as “high-touch surfaces.” Spare desks and chairs in the classroom, walls at low height (15 cm from ground), and area of bookshelves deemed to be less frequently contacted by children were categorized and sampled as “low-touch surfaces.” Tops of shelves and windows were sampled as “never-touched surfaces” for comparison. In addition to surfaces inside classrooms, staircase handrails, bathroom cubicle locks and faucets (where available) in each school were also sampled as high-touch surfaces. The material of surfaces sampled were documented.

Samples were collected at the beginning of the school day before the arrival of children and 7–8 hours later at school dismissal, before daily cleaning and disinfection of classrooms and school premises by school staff. Samples were collected by swabbing 10 times vigorously from 2 directions (horizontally and vertically) in an approximately 100-cm² area using a sterile polyester-tipped applicator (Model 164KS01; COPAN) premoistened in viral transport medium. For faucets, doorknobs, and bathroom cubicle locks, the surface of the entire item was swabbed. Swabbed surfaces were wiped using tissue after sample collection to remove any nucleic acid remnants, to assess viral inoculation onto surfaces throughout the school day. Samples were stored and transported to laboratory at 2–8°C in viral transport medium and subsequently stored at −80°C until nucleic acid extraction.

In addition to surface sampling, temperature and relative humidity (RH) of the classroom environment were also logged at 15-minute intervals throughout the school day using a data logger (HOBO UX-100). Absolute humidity (AH) was calculated from logged temperature and RH using the Clausius-Clapeyron equation.

Detection of Nucleic Acid

Samples were subjected to reverse-transcription quantitative polymerase chain reaction (RT-qPCR) screening for influenza A and B virus. Samples were considered positive when the cycle threshold (Ct) value was ≤40. Copy number standard controls were included in each run for deducing the absolute copy number of the targeted gene. Positive samples were subjected to viral culture. A 100-µL aliquot of the original specimen was inoculated in Madin-Derby Canine Kidney (MDCK) cells grown in 24-well plates in minimum essential medium with trypsin (1 µg/mL). Cells were monitored under a microscope daily for cytopathic effect. Each sample was blind serial passaged in MDCK cells 3 times. At the end of the third culture, the supernatant was harvested and tested for the presence of influenza viruses by hemagglutination assay and rapid influenza antigen diagnostic test.

RESULTS

In total, 1352 samples were collected from 29 sampling sessions from week 3–6 (January 16–February 7) and week 9 (February 26–March 2) of 2018. No samples were collected from week 7 and 8 because schools were closed for the Lunar New Year holiday. The sampling time frame corresponded to the peak of influenza B activity in Hong Kong according to local surveillance data (Figure 1) [8]. The mean temperature, RH, and AH in the sampling site classrooms were 19.0°C (range, 12.1–24.7°C), 62.0% (range, 28.9%–89.5%), and 10.5 g/m³ (range, 3.9–15.8 g/m³), respectively. The temperature and AH remained relatively constant throughout the study (Supplementary Figure 1).

Figure 1.

Sampling time frame during the 2017/2018 winter influenza season and weekly detection of influenza virus. Lines indicate local influenza activity based on surveillance data from the Centre for Health Protection.

Open in new tabDownload slide

An average of 31 samples were collected from each classroom during each sampling session. Seven of the 11 classrooms sampled had ≥1 touch surface that tested positive for influenza virus ribonucleic acid (RNA) from at least 1 sampling session. Overall, 12 of 1352 samples were positive (<1%). The Ct values ranged from 25.7 to 39.9 (median, 39.0), corresponding to viral load of 5 to 2 × 10⁶ RNA copies/mL (median, 151 RNA copies/mL). None of the 12 PCR-positive samples were positive by viral culture. Influenza B virus RNA was detected in 8 samples, and influenza A virus RNA was detected in 4 samples. None of the samples had RNA of both influenza A and B virus. Eight of twelve (66.7%) positive samples were collected at the beginning of school day, and separately only 2 of 12 (16.7%) samples were collected from kindergartens.

Influenza virus RNA was detected in samples from 6 of 916 high-touch (0.7%) and 6 of 266 low-touch (2.3%) surfaces, respectively. All of these surfaces were either made of plastic, metal, or wood with lacquer finishing. Influenza virus RNA was more frequently detected in samples from bookshelves and doorknobs/door surfaces inside classrooms. Table 1 shows the frequency of positives among all samples collected from high-touch and low-touch surfaces before and after school hours. None of the never-touched surfaces sampled, ie, windows (n = 85) and top of bookshelves (n = 85), were positive.

Discussion

Surface sampling studies have been previously conducted in locations where individuals mix, such as transportation hubs, religious pilgrimage sites, and homes, in addition to schools and kindergartens [6, 7, 9–12]. These studies have detected influenza and other respiratory viruses on environmental surfaces; however, the level of contamination varies between studies. Our work in this study supplemented these findings, especially so for school settings. Overall, influenza virus RNA was detected on less than 1% of touch surfaces in kindergarten and primary schools during the peak of a local influenza B epidemic, suggesting infrequent contamination.

The low presence of influenza virus in our study was similar to other surface sampling studies conducted in airports and households, which detected viral nucleic acid on 0%–10% of surfaces sampled [9–12]. However, our findings differ considerably from 2 studies conducted in day care centers and elementary schools in the United States, in which one quarter to one half of the surface samples collected in fall and spring were positive [6, 7]. Influenza virus RNA was detected on at least 20% of each type of items sampled in these 2 studies except for soap dispenser and water fountain toggle in elementary school, whereas in our study the detection of influenza virus was low (≤2.9%) across different items sampled. Such differences could be due to different sample collection systems and laboratory analyses, time of sample collection, differences in temperature and humidity, and the proportion of present children who were ill. The sociocultural differences in hygiene practice of children in the 2 geographical locations, as well as school cleaning and disinfection practices, may also play a role in the observed difference.

Most of the current supporting evidence for indirect contact transmission came from the ability of influenza virus to survive and persist on surfaces and hands of volunteers in experimental settings [13]. The temperature and humidity recorded on most sampling days in our study sites were within the range that is feasible for influenza virus survival [14]. Yet, no viable influenza virus was recovered, and most of the PCR-positive samples had low viral load. Killingley et al [12] proposed a conceptual estimation of the infectious dose threshold for direct nasal inoculation to be 4 × 10⁴ to 4 × 10⁵ copies/mL. Taking into account some loss of virus during the transfer from surfaces to hands [13], to have an infectious dose for nasal inoculation at the end of the transmission chain, viral load on surfaces should be higher than the specified range. However, 11 of the 12 positive samples collected in our study had viral load below the lower bound of the range. In addition, among the few positive samples that we have collected, we did not identify substantial differences in contamination between surfaces of different touch frequencies, nor contamination by influenza A or B virus even though the season was dominated by influenza B virus. As such, the presence of virus on surfaces seem unlikely to contribute significantly to the total transmission via fomites in local school settings.

Some uncertainties remain in the findings. First, we were unable to perform viral subtyping because the viral load for many positive samples were low. Second, the detection of viral RNA and recovery of viable virus can be affected by various factors, such as the duration between virus deposition and sample collection, the initial titer and viability at deposition, and contact dynamics (in particular for the deposition, dilution, and reinoculation of virus onto surfaces). This information would help in estimating the potential of indirect contact transmission more accurately compared with the presence of viral RNA on surfaces alone; however, it is difficult to obtain such information in surface sampling studies in natural settings. Pairing surface sampling work with contact observation or concurrent swab sampling of children present in the study sites may provide more insights. Third, detection of viable virus may have been limited by the efficiency of recovering live virus from swabs. Currently documented efficiency varies, from no significant loss for influenza virus to approximately 50% for PR8 virus during immediate recovery from surfaces using swabs in laboratory settings [13, 15], but the recovery efficiency would be lower in field settings. It would be ideal to have a sampling system identified for the optimal recovery of viable influenza virus and nucleic acid for surface sampling, including the swab applicator (size, material, and alternative methods of sample collection), transport medium, and eluent.

Conclusions

In conclusion, the presence of viral RNA on touch surfaces reaffirms the potential of indirect contact transmission of influenza. Our work identified that communal items inside classrooms such as bookshelves and surfaces on doors are more frequently contaminated. These items may pose greater potential risks in viral transmission during an influenza epidemic. We also detected contamination more frequently on low-touch surfaces than high-touch surfaces. Therefore, cleaning and disinfection of low-touch and communal items should be given similar importance as high-touch and personal items such as desks and chairs of students. In addition, the surfaces that tested positive for viral RNA in our study were all inside classrooms, hence interventions such as placing alcohol sanitizers inside classroom for hand hygiene usage of children may be able to reduce the potential risk of indirect contact transmission in schools. The presence of influenza virus on touch surfaces before the beginning of school day was unexpected, and it may indicate insufficient disinfection of surfaces on the preceding day. Another less likely explanation is that these originated from the settling of airborne virus particles overnight. More evidence is needed to elucidate the relative significance of different transmission modes of influenza in different school settings.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Supplementary Figure 1. Daily mean and range of temperature and absolute and relative humidity in classrooms on sampling days. These environmental conditions were logged at 15-minute intervals from the beginning until the end of school hours. Schools were closed from February 12 to February 25 for the Lunar New Year holiday.

Notes

Acknowledgments. We thank all of the school principals and administrators of the 7 kindergartens and primary schools for allowing this study to be conducted in their schools. We also thank Leo Ho, Nicole Tsang, and Tramie Wu for research support.

Financial support. This work was funded by the Collaborative Research Fund from the University Grants Committee of Hong Kong (Project No. C7025-16G).

Potential conflicts of interest. B. J. C. reports honoraria from Sanofi Pasteur and Roche. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References