The nose is not enough: Multi‐site sampling is best for MRSP detection in dogs and households

Abstract

Background

Following recovery from meticillin‐resistant Staphylococcus pseudintermedius (MRSP) infection of any type, dogs may continue to carry MRSP asymptomatically on skin and mucosae, contributing to the spread of this multidrug‐resistant, veterinary hospital‐associated pathogen with zoonotic potential to others and into the environment.

Objectives

This study determined which canine anatomic and household environmental sites are most sensitive for sampling to identify carriage and contamination.

Methods and Materials

Fifty‐one dogs and 22 households, MRSP‐positive on at least one tested site, were sampled on 132 and 40 occasions over time, respectively. Dogs were swabbed at six sites (mouth, nose, conjunctiva, skin, prepuce/vulva, perianal area); household environments were sampled using contact plates (mannitol salt agar [MSA] and MSA + 6 mg/L oxacillin [MS+]) on five sites. MRSP was isolated after enrichment, grown on MSA/MS+ and was confirmed by PCR. Generalized estimating equations were used for calculation of sensitivity (95% confidence interval) for each site/combination.

Results

Each anatomical and environmental site yielded MRSP at least once. MRSP was isolated from only a single site in 27.3% of dogs, with the buccal mucosa showing the highest sensitivity (63.8%). Multi‐site sampling of a minimum of four canine anatomical or four environmental sites, respectively, was needed to achieve >95% sensitivity.

Conclusions and clinical relevance

The canine buccal mucosa should be included in MRSP sampling protocols, ideally in addition to at least three other anatomical sites. Likewise, environment sampling should be of multiple household sites in cases where it is used as a part of clinical case management.

Short abstract

Background – Following recovery from meticillin‐resistant Staphylococcus pseudintermedius (MRSP) infection of any type, dogs may continue to carry MRSP asymptomatically on skin and mucosae, contributing to the spread of this multidrug‐resistant, veterinary hospital‐associated pathogen with zoonotic potential to others and into the environment. Objectives – This study determined which canine anatomic and household environmental sites are most sensitive for sampling to identify carriage and contamination. Conclusions and clinical relevance – The canine buccal mucosa should be included in MRSP sampling protocols, ideally in addition to at least three other anatomical sites. Likewise, environment sampling should be of multiple household sites in cases where it is used as a part of clinical case management.

Abbreviations

CI

confidence interval

GEE

generalised estimating equations

MDR

multidrug‐resistant

MRSA

Meticillin‐resistant Staphylococcus aureus

MRSP

Meticillin‐resistant Staphylococcus pseudintermedius

MS+

Mannitol salt agar supplemented with 6 mg/L oxacillin

MSA

Mannitol salt agar

MSSP

Meticillin‐susceptible Staphylococcus pseudintermedius

PCR

Polymerase chain reaction

RVC

Royal Veterinary College

TSB

Tryptone soy broth

INTRODUCTION

Over the past 15 years, meticillin‐resistant Staphylococcus pseudintermedius (MRSP) has become the major multidrug‐resistant (MDR) bacterial pathogen in canine skin and soft‐tissue infections. ¹ Like other staphylococci, MRSP adheres to squames and hair after infection has resolved. Such nonclinical MRSP carriage poses a risk to the host if infection recurs, as seen in human Staphylococcus aureus infections, ² and contributes to transmission to in‐contact humans and dogs. Implementation of infection control measures preventing nosocomial spread of MRSP requires accurate identification of carriers, of especial importance in settings with low MRSP prevalence, such as the UK.

Screening for MRSA carriage in humans before hospital admission or elective procedures is routine worldwide, within the boundaries of funding and practicability. Systematic review indicates that the most sensitive single human swabbing site is the nose, detecting 68% of carriers (34%–91%), ³ and this can be increased to almost 95% sensitivity by combining throat, groin and nose. ³

Conversely, nasal swabbing in dogs is reported to be less sensitive than in humans to identify S. pseudintermedius carriers (16%–64%), with the perineum and mouth cited as predominant carriage sites. ⁴ , ⁵ In a single study, analysis of MRSP recovery from 73 sample sets from 27 dogs covering four sites (nose, mouth, perineum, pharynx) identified similar predilection sites for MRSP to meticillin‐susceptible S. pseudintermedius (MSSP), with perineal (63%) and corner of the mouth (58%) sampling being most sensitive. ⁶

The household environment has been highlighted for potential contamination by MRSP‐carrier dogs, providing a source for future reinfection. Previous studies identified MRSP most commonly in dog‐accessible areas, such as feeding and sleeping places. ⁷ Further investigation is needed to inform on the utility of household sampling when managing recurrent MRSP infections that do not relate to MRSP carriage on mucosal sites.

Previous sampling studies for MRSP carriage in dogs have focussed on the nose and perineum, extrapolated from nasal swabbing for human MRSA carriage, and ease of sampling the perineum. This study aimed to determine the sensitivity of swab sampling six anatomical sites for detecting canine MRSP carriage, and sensitivity of contact plate sampling of five household environment sites for environmental contamination.

MATERIALS AND METHODS

Ethics

This study was approved by the Royal Veterinary College's (RVC) Clinical Research Ethical Review Board (URN 2012 1166); owners gave written consent at enrolment.

Study population and design

Dogs had been recruited as part of a study investigating the efficacy of topical antimicrobial therapy for eradicating MRSP carriage after a previous episode of MRSP infection had resolved. They had been enrolled either at the Queen Mother Hospital for Animals (RVC) or via referring veterinary surgeons between May 2016 and December 2019.

Sampling for dog carriage of and household contamination with MRSP

Six anatomical sites were sampled from each dog: buccal mucosa inside the upper lip, outside the teeth; nose including nasal planum and one or both nostrils; conjunctival mucosa uni‐ or bilaterally; axilla or groin skin; preputial or vulval mucosa; and perianal area where nonhaired meets haired skin. Dry sterile cotton swabs (charcoal transport swabs, SLS) were rolled over each site for 3–5 s.

For household contamination, five different sites in the room most frequently occupied by the dog were sampled: floor; frequently cleaned hard hand‐touch surface (e.g. kitchen work surface); infrequently cleaned inaccessible surface (e.g. top of cupboard); dog bed; and dog bowl. Owners were instructed to place paired mannitol salt agar (MSA; CM0085, ThermoScientific) and MSA supplemented with 6 mg/L oxacillin (MS+) 55 mm contact plates (ThermoScientific) on each site for 5 s.

‘Sampling event’ was used to describe each time at which a full set of samples (six anatomic or five environmental) was taken.

Isolation and confirmation of MRSP

Swabs and plates were posted to the RVC for microbiological analyses and processed immediately on receipt. Swabs were incubated individually in tryptone soy broth (TSB; CM0129, ThermoScientific) supplemented with 10% sodium chloride (Sigma‐Aldrich Ltd) at 37°C for 48 h. Broth aliquots were subcultured onto MSA and MS+ and incubated at 37°C for 24–48 h. Contact plates were incubated at the laboratory at 37°C for 48 h. If no growth was observed after 48 h, the sample was discarded.

Each distinct, presumed staphylococcal, colony from MSA and MS+ was subcultured onto blood agar base (CM0271; Thermo Scientific) containing 5% sheep blood (TCS BioScience) after morphological assessment based on size (small–medium), shape (round with regular edges) and colour (white to cream) of colonies. Growth was further characterized phenotypically for clumping factor ability using dog plasma, DNase production and by Voges‐Proskauer testing. ⁸ Suspected MRSP were confirmed by PCR, demonstrating the presence of species‐specific thermonuclease, nuc, and mecA. ⁸

Statistical methods

Results of sampling events were analysed for dogs/environments testing positive for MRSP at a minimum of one site. Generalized estimating equations (GEE), accounting for repeated measures of some dogs or environments, were used to determine sensitivity (95% confidence interval [CI]) of each site, and combinations of sites, in detecting MRSP (Spss Statistics v26, IBM). GEE used dog or environment as the subject variable, with an exchangeable working correlation matrix in a binary logistic regression model. Comparison of sensitivity of single sites used the same GEE with site included as a linear predictor, significance p ≤ 0.05 (Spss Statistics v26).

RESULTS

One‐hundred thirty‐two sampling events from 51 dogs (≤12 repeated samples from the same dog) and 40 sampling events from 22 household environments (up to four repeated samples from a single environment) were available for analyses (for sensivities of all sites and combinations, see Table S1).

Each of the six anatomical sites and five environmental sites yielded MRSP at least once, although the pattern of MRSP recovery varied between individuals and within repeated samples from the same individuals.

Meticillin‐resistant Staphylococcus pseudintermedius was isolated from only a single canine carriage site in 36 of 132 (27.3%) sampling events (24 of 51 dogs). In 16 of 132 (12.1%) sampling events, all six swabs yielded MRSP (13 of 51 dogs). The buccal mucosa most frequently yielded MRSP (Table 1) and was a significantly more sensitive sampling site than either axilla/groin skin (p < 0.0005) or prepuce/vulva (p = 0.008). Nose, conjunctiva and prepuce/vulva were significantly more likely to yield MRSP than axilla/groin skin (p = 0.002, p = 0.016 and p = 0.011, respectively). No other significant differences were seen. To achieve sensitivity ≥95%, at least four sites needed to be swabbed, always including both buccal mucosa and nose (Table 2).

TABLE 1.

Sensitivity of sampling individual canine anatomical and household environmental sites for detecting MRSP carriage or contamination in 51 dogs (132 sampling events) and 22 households (40 sampling events), respectively

Site	Number of positive sampling events in this site/Total number of MRSP‐positive dogs or households a	Sensitivity (%) (95% CI)
Dog
Buccal	83/132	64 (54–72)
Nasal	77/132	60 (37–58)
Conjunctival	67/132	48 (37–58)
Axilla/ groin skin	48/132	36 (28–45)
Prepuce/ vulva	63/132	48 (39–56)
Perianal	62/132	44 (33–56)
Environment
Dog's bed	24/40	55 (37–72)
Dog's bowl	12/40	30 (18–46)
Floor	15/40	36 (22–53)
Frequently cleaned	2/40	5 (1–18)
Infrequently cleaned	8/40	21 (10–37)

Abbreviation: CI, confidence interval; MRSP, meticillin‐resistant Staphylococcus pseudintermedius.

^a Every sampling event recovered MRSP from at least one site, and thus, this column represents traditional sensitivity of (number true positive)/(total number positive).

TABLE 2.

Combinations of sampling sites (canine anatomical and household environmental) to achieve ≥95% sensitivity for detecting MRSP carriage or contamination in 51 dogs (132 sampling events) and 22 households (40 sampling events), respectively

Combination of sites	Number of positive sampling events in this combination of sites/Total number of MRSP‐positive dogs or households a	Sensitivity (%) (95% CI)
Dog
4 sites sampled
B N C S	128/132	97 (93–99)
B N C PrV	126/132	96 (87–99)
B N S PrV	125/132	95 (91–98)
B N P PrV	124/132	95 (86–98)
5 sites sampled
B N C S PrV	131/132	99 (95–100)
B N C S P	129/132	98 (94–99)
B N S PrV P	129/132	97 (92–99)
B N C PrV P	127/132	97 (87–99)
Environment
4 sites sampled
Dog's bed, dog's bowl, floor, infrequently cleaned	39/40	98 (85–100)

Abbreviations: B, buccal; C, conjunctival; CI, confidence interval; MRSP, meticillin‐resistant Staphylococcus pseudintermedius; N, nasal; S, axilla/groin skin; P, perianal; PrV, prepuce/vulva.

^a Every sampling event recovered MRSP from at least one site, and thus, this column represents traditional sensitivity of (number true positive)/(total number positive).

In the household environment, MRSP was identified from only a single site in 23 of 40 (57.5%) of sampling events (17 of 22 households), and none yielded MRSP from all sites at the same time. MRSP was most frequently recovered from the dog's bed (Table 1). The bed was more sensitive than the bowl (p = 0.037) and infrequently cleaned area (p = 0.003), and the bed, bowl and floor were more sensitive than the frequently cleaned site (p < 0.0005, p = 0.003 and p = 0.001, respectively). A sensitivity ≥95% was achieved only by combining at least four sites: the dog's bed, bowl, floor and infrequently cleaned site (Table 2).

DISCUSSION

These results confirm that the buccal mucosa is comparable to the human nose ³ as the most sensitive sampling site for investigating canine MRSP carriage. This is encouraging as using a swab from the inside of the lip will be better tolerated by most dogs and be safer for the sampling person than inserting a swab into nostrils. However, desirable sensitivities of ≥95% were achieved only by combining results from at least four anatomical sites. This increased sensitivity may be a consequence of either different niches being accessed or larger total surface areas being sampled. This corroborates earlier MRSP screening recommendations ⁷ and mirrors findings from human medicine regarding MRSA. ³

In general, recovery of MRSP was comparable to that reported previously for MSSP (60% vs. 16%–64% nasal; 44% vs. 28%–72% perineum). ⁴ Combining buccal and perineal swabbing resulted in a much lower sensitivity than that reported previously (76% vs. 90%) ⁹ ; further screening of MRSP versus MSSP would be needed to confirm whether this is a true difference in carriage site preference. Variability in the data across study groups indicates the importance of multi‐site swabbing to detect all carriers. If the nose cannot be sampled, our data indicate that 9% of canine MRSP carriers would be missed despite sampling all other sites, which is comparable to the reported figure of 5%–7% of humans who had MRSA recovered only from nasal swabs. ¹⁰

Although not investigated in this study, pooling samples from different sites for processing where only a binary outcome report of MRSP carriage is needed, may be considered to reduce cost. Comparable results (93%–97% agreement) to individual culture have been reported for pooled MRSA swab processing, ¹¹ and further confirmation of this approach for MRSP is warranted.

For environmental sampling, best sites remain uncertain. Although all sites yielded MRSP at least once, the dog bed showed a moderate sensitivity of 55%, while yield from other sites was comparatively low. ‘Hand‐touch areas’ in human medicine have been identified as preferred sampling sites owing to their importance in MRSA‐transmission. ¹² However, extrapolation to ‘nose‐touch sites’ in a dog‐MRSP setting cannot be supported by these findings. Household sampling may be desirable for research into the role of the environment in pathogen dissemination, or for control of recurrent MRSP infections that do not appear to relate to carriage of the isolate and may be related to environmental contamination. Thus, use of environmental screening may be a rare consideration in management of clinical cases (e.g. a human at high risk of MRSP infection within the household). Sensitivities identified for environmental sites in this study indicate that multi‐site sampling is needed.

In conclusion, buccal mucosa, nose and at least two additional sites should be swabbed to maximize detection of MRSP carrier dogs. Overall, these findings should be incorporated into veterinary infection control protocols to minimize the impact of canine MRSP carriers within the veterinary practice through accurate detection.

AUTHOR CONTRIBUTIONS

Sian‐Marie Frosini: Conceptualization; formal analysis; investigation; methodology; writing – original draft; writing – review and editing. Ross Bond: Conceptualization; funding acquisition; investigation; methodology; writing – review and editing. Ruth H King: Investigation; writing – review and editing. Anette Loeffler: Conceptualization; funding acquisition; investigation; methodology; writing – original draft; writing – review and editing.

FUNDING INFORMATION

SMF and this study were funded by a Biotechnology and Biological Sciences Research Council (Swindon, UK; BBSRC) industrial CASE scholarship in partnership with Dechra Veterinary Products Limited (Shropshire, UK; grant no. BB/K011952/1).

CONFLICT OF INTEREST

Design of the study; collection, analysis, and interpretation of data; and writing the manuscript, were undertaken independently of the funding body.

Supporting information

Table S1

Frosini S‐M, Bond R, King RH & Loeffler A. The nose is not enough: Multi‐site sampling is best for MRSP detection in dogs and households. Vet Dermatol. 2022;33:576–580. 10.1111/vde.13118