To determine the presence of infectious virus in vomit, we used the human intestinal enteroid (HIE) culture system to culture vomit samples positive for norovirus. The system was previously used to replicate human noroviruses from fecal samples (9). HIE cultures were established using biopsy specimens from patients who underwent gastric bypass (ethics permission no. 2019–00600, Linköping Ethical Board, Linköping, Sweden). Written informed consent was obtained from all participants. We obtained 28 PCR-positive norovirus vomit samples collected for routine diagnosis from persons with acute gastroenteritis from Karolinska University Hospital (Stockholm, Sweden) and University Hospital of Linköping. The vomit samples were anonymized when received, and only information regarding the initial cycle threshold (C_t) value was provided. Decoded clinical samples without person-related data and traceability that have not been taken for research purposes do not require ethics or legal clearance according to The Swedish Ethics Review Authority.

The norovirus C_t values in the diagnostic PCRs ranged from 13.4 to 31.7. A previous study using fecal samples observed that the replication rate dropped substantially when 1.9 × 10³ GEq/mL were used as inoculum for infection (10), whereas another study reported loss of infectivity at higher C_t values (11). Of 28 vomit samples, 20 that had C_t values of <26 had 8.9 × 10⁶ to 1.6 × 10¹⁰ GEq/mL (Table); the remaining 8 vomit samples had <1 × 10⁶ GEq/mL (in undiluted vomit) and were excluded from further evaluation.

Figure

Figure. Replication in human intestinal enteroids of norovirus from vomit samples. Differentiated human intestinal enteroid monolayers were inoculated with norovirus-positive vomit samples. The number of norovirus GEq was quantified by reverse...

Infectivity was tested on 5-day-old differentiated HIEs established from the jejunum of persons who had undergone gastric bypass surgery. Initial screening to determine infectivity of vomit samples was done with 2 different HIEs (HIE 003 and HIE 004) isolated from secretor-positive persons (i.e., having a functional fucosyltransferase 2 gene). Both HIEs showed similar replication for the same 5 vomit samples. Next, we used HIE 003 for infection in triplicates with 2 technical repeats during quantitative reverse transcription PCR (qRT-PCR) (Figure). Norovirus genotypes in the vomit samples were determined by nucleotide sequencing. We defined infection as a >10-fold increase in GEq 72 hours postinfection (hpi) compared with 2 hpi, determined by qRT-PCR. We compiled details regarding the qRT-PCR method and the isolation, culturing, genotyping or phenotyping, and infection of HIEs (Appendix).

Partial nucleotide sequencing of the norovirus capsid region showed that 16 of the 20 vomit samples contained GII.4 norovirus genotype (belonging to GII.4 genotype Sydney 2012 variant), 3 contained GII.2, and 1 contained GII.17 norovirus genotype (Table). In the HIE infectivity assay, 5 of the vomit samples resulted in an increase in GEq, ranging from 13-fold to 714-fold at 72 hpi compared with 2 hpi; all these samples contained GII.4 noroviruses (Figure). The percentage of vomit samples (31.2%) containing GII.4 norovirus that successfully replicated is similar to that reported by Constantini et al. (10) using fecal samples positive for norovirus by PCR (25.6%). Of the 4 vomit samples containing GII.2 (n = 3) and GII.17 (n = 1), none demonstrated any replication in HIE, despite 2 GII.2 and 1 GII.17 vomit samples having similar or higher GEq in the inoculum compared to the fecal samples that could be successfully replicated in Constantini et al. (10). Of note, this finding might be because of the small number of GII.2-containing vomit samples and GII.17-containing vomit samples tested; not all fecal samples with high viral loads can be successfully replicated (10).

A previous study reported that fecal suspensions that showed successful norovirus replication in HIE cultures contained 1.9 × 10³ to 1.7 × 10⁷ GEq in the inoculum, regardless of genogroup or genotype (10). In our study, the GII.4 norovirus that could be successfully replicated contained a similar viral load (9.55 × 10⁴ to 1.61 × 10⁷ GEq) in the inoculum used for infection. Vomit samples that failed to show norovirus replication had of 8.91 × 10³ to 1.66 × 10⁶ GEq in the inoculum used for infection (Table), which suggests that viral load is not the sole criterion for successful infection in HIEs, as has been reported for norovirus cultured from feces (10). Because the vomit samples in this study were anonymized, no information beside the initial norovirus C_t value was available. Factors such as long-term storage (12) and the time of collection postinfection (13) might affect infectivity and cannot be ruled out. Repeated freeze-thaw cycles could also influence the infectivity of viruses, possibly because of the disruption of the capsid proteins, which could degrade the viral genome. However, Richards et al. (12) reported that norovirus capsid integrity is not compromised after repeated freeze-thaw cycles. Therefore, despite not knowing the exact long-term storage conditions in the 2 hospitals that provided the vomit samples (although most were stored at −70°C for <3 years), variation in infectivity should not have been caused by multiple freeze-thaw cycles. The time of sample collection also might influence infectivity. Samples should be collected within the first 24 hours after symptom onset. Norovirus can be shed in feces for >7 days, but no studies report infectivity after the initial 48–72 hours after symptom onset (13). Although qRT-PCR is standard for detecting norovirus RNA, it does not distinguish infectious virus particles from noninfectious virus particles (10).

Although an estimate of the 50% human infectious dose (HID₅₀) in vomit containing virus is unknown, it has been calculated to be ≈2,800 GEq for secretor-positive persons challenged with the Norwalk virus (7). Comparing the RNA levels in vomit and feces (on the basis of human challenge studies with the Norwalk virus), it was estimated that 1 mL of vomitus contained up to 9,000 HID₅₀ of virus (7). The combination of a low infectious dose and a large quantity of virus in vomit led to the suggestion that each vomiting event has the potential to infect >150,000 persons (4). In our study, we found that >95,500 GEq per inoculum was sufficient for infection of HIEs. Considering the different models studied (human vs. in vitro), the use of strains from different genogroups, and fecal versus vomit inoculum, the similarity in infectious dose is noteworthy.

In conclusion, this study demonstrates that norovirus contained in vomit is infectious. Aerosols and droplets from vomiting could be a source of norovirus transmission.