Cross-Sectional Serosurvey of Companion Animals Housed with SARS-CoV-2–Infected Owners, Italy

We conducted a serologic survey among dogs and cats in Italy to detect antibodies against severe acute respiratory syndrome virus 2 (SARS-CoV-2). We found that SARS-CoV-2 seroprevalence was higher among cats (16.2%) than dogs (2.3%). In addition, seroprevalence was higher among animals living in close contact with SARS-CoV-2–positive owners.

COVID-19 incidence rates and the country has >32 million companion animals, health authorities were interested in examining virus transmission between humans and animals. We conducted a cross-sectional serologic survey among domestic dogs and cats in Italy to identify a possible association between SARS-CoV-2 infection in humans and animals. We used serologic tests to detect specifi c antibodies from animals living in close contact with SARS-CoV-2-positive human patients.

The Study
Blood was collected from pets during routine activities performed by veterinary practitioners, who shared serum samples with us. Owners provided written consent for research purposes. We used 198 samples, 130 from dogs and 68 from cats, collected during the March-June 2020 COVID-19 epidemic in Italy and 100 serum samples, 65 from dogs and from 35 cats, collected in different regions of Italy before 2019 as prepandemic controls.
A recombinant antigen corresponding to the nucleocapsid (N) protein of SARS-CoV-2 has been expressed in human embryonic kidney 293T cells, which have been used to develop Eradikit COVID19-IgG (IN3diagnostic, https://www.in3diagnostic. com) a sensitive and specifi c ELISA to detect SARS-CoV-2 antibodies in human serum samples. However, our initial attempts to validate the specifi city of this ELISA on pet serum samples were unsuccessful. We switched the reaction from solid-phase to solution-phase kinetics using the same antigen and the specifi city improved. Thus, we used a novel immunoassay, xMAP (Luminex Corp., https://www. luminexcorp.com), which is based on paramagnetic beads. We developed a fl ow cytometry-based system and applied it to serum samples from cats and dogs.
To defi ne the test's specifi city, we analyzed preepidemic samples and expressed results as the mean We conducted a serologic survey among dogs and cats in Italy to detect antibodies against severe acute respiratory syndrome virus 2 (SARS-CoV-2). We found that SARS-CoV-2 seroprevalence was higher among cats (16.2%) than dogs (2.3%). In addition, seroprevalence was higher among animals living in close contact with SARS-CoV-2-positive owners.

Cross-Sectional Serosurvey of Companion Animals Housed with SARS-CoV-2-Infected Owners, Italy
fluorescence intensity (MFI) ratio of a sample-to-positive control. On the basis of reactivity distribution, we set the discriminative cutoff to 40% MFI of the positive control. Using these specifications, we recorded diagnostic specificities of 96.5% (95% CI 87.9%-99.6%) for dog serum and 100.0% (95% CI 90.0%-100.0%) for cat serum. Our choice of the viral N protein might raise concern because dogs and cats are susceptible to speciesspecific coronaviruses. The amino acid similarity between SARS-CoV-2 and the canine betacoronavirus, canine respiratory coronavirus, is slightly higher than canine and feline alphacoronaviruses (10), which could explain the suboptimal specificity obtained in pre-epidemic dog samples. In fact, 2 serum samples gave reactivity slightly over the cutoff value. However, when potential cross-reactivity of the N protein between SARS-CoV-2 and endemic human coronaviruses was evaluated, no reactivity was shown against human coronaviruses 229E, OC43, HKU1, or NL63 by western blot or ELISA (11), suggesting that similar results might be expected from phylogenetically related feline and canine coronaviruses (12).
Exact logistic regression analysis indicated a positive association between owners' infections and seropositivity in individual animals, after adjusting for animal species (Figure 2). The odds of finding >1 seropositive animal in a household were positively associated with owners' infection and with an increasing number of tested cats (Table 2; Appendix, https:// wwwnc.cdc.gov/EID/article/27/7/20-3314-App1. pdf). The association with owner infection was only statistically significant based on a 1-tailed hypothesis, whether the outcome was measured at the animal or household level.
Using exact logistic regression, we noted the percent of positive results was greater for animals living indoors only than for animals with access to the outside (odds ratio 3.4, 95% CI 0.71-35.9), but the association was not statistically significant (p = 0.15). Because information on living conditions was missing  for 60 animals, we did not include this factor in the exact logistic regression analysis (Tables 1, 2). We found the proportion of serologic positivity increased with increasing length of exposure. We recorded the first SARS-CoV-2-positive animals 10 days after owners' diagnoses and all 14 seropositive cases were classified as positive after >54 days of exposure (Appendix).
Among 5/14 positive animals, owners reported that their pets experienced clinical signs concurrent with the owner's COVID-19 illness. In particular, a 10-year-old male dog showed respiratory signs (cough, sneezing) after which he had vomiting and diarrhea in concomitance with the onset of the owners' symptoms; a 1-year-old dog showed mild respiratory signs characterized by cough and sneezing; a 12-year-old female cat showed respiratory signs characterized by rhinitis with abundant nasal discharge. Furthermore, a 13-year-old female cat was hospitalized for a brachial cephalic thrombosis and a 3-yearold male cat was hospitalized for interstitial pneumonia. Of note, 3 asymptomatic SARS-CoV-2-positive cats belonged to a single-family cluster in which both owners tested positive and hospitalized.

Conclusions
We detected antibodies against the SARS-CoV-2 N protein in pets living with SARS-CoV-2-infected owners. A higher percentage of feline samples tested positive, confirming a higher susceptibility and prevalence in cats than in dogs reported in previous experiments (10,13). The susceptibility of cats to SARSrelated human coronaviruses also was reported in 2003 when a study confirmed that cats were susceptible to infection and could transmit the virus to other in-contact animals (14). The association between seropositivity in animals and the confirmed SARS-CoV-2 infection in >1 of the animal's owners was statistically significant (p<0.05) based on a 1-sided test assuming the owner's infection could not reasonably exert a protective effect on pets' infection.
We could not draw conclusions concerning the direction of viral transmission in this cross-sectional study. Nevertheless, our results, coupled with the direction of the association between seropositivity and length of exposure to an infected owner and living indoors, suggest that the development of antibodies in pets might be a consequence of viral transmission from their owners. Additional studies with more statistical power could confirm these relationships.
Based on our results, future studies should focus on overcoming test limitations by improving specificity in dog serum samples through detailed epitope mapping of the N protein. Additional studies also should examine routes and risk factors for transmission of SARS-CoV-2 from infected persons to susceptible pets and the potential role of pets in the COVID-19 pandemic. Clinical and pathological consequences of SARS-CoV-2 infection in cats and dogs also warrant further research.
In conclusion, our study on companion animals housed with SARS-CoV-2-infected humans confirms the susceptibility of domestic cats under natural exposure. Our data statistically support other findings that cats are more susceptible than dogs Figure 2. Distribution of sample-to-positive severe acute respiratory syndrome coronavirus 2 serology results among dogs and cats exposed and unexposed to positive owners, Italy, March-June 2020. Horizontal dashed line represents the positive-negative discriminatory cutoff. and that living in contact with >1 SARS-CoV-2-infected person increases the risk for infection in pets. These results justify the need to adopt control measures in SARS-CoV-2-infected pet owners to reduce viral transmission to their companion animals.