Sage Journals: Discover world-class research

Abstract

This study investigated 339 cases of feline mycobacterial disease from cats with cutaneous lesions or masses found at exploratory laparotomy. Tissue samples were submitted to the Veterinary Laboratories Agency for mycobacterial culture over a 4-year period to December 2008. The study assessed which species of culturable mycobacteria were involved, where the cats lived, and their clinical presentation (physical findings, serum biochemistry, radiography, feline leukaemia virus and feline immunodeficiency virus status). Mycobacterium microti was cultured from 19%, Mycobacterium bovis 15%, Mycobacterium avium 7%, non-M avium non-tuberculous mycobacteria 6%, with no growth in 53% of samples. M microti, M bovis and M avium were found in almost mutually exclusive clusters within Great Britain (GB) (ie, M bovis in South-West England/Wales/Welsh Border, M avium in eastern England and M microti south of London and in South-West Scotland). While differences were seen in the clinical presentation and distribution of lesions caused by the different infections, these were not sufficiently different to be diagnostic. Cats commonly presented with single or multiple cutaneous lesions (74%), which were sometimes ulcerated or discharging, located most frequently on the head (54%). Lymph nodes were usually involved (47%); typically the submandibular nodes. Systemic or pulmonary signs were rarely seen (10–16%). When a cat is suspected of having mycobacteriosis, accurate identification of the species involved helps to determine appropriate action. Our findings show that knowing the cat's geographic location can be helpful, while the nature of the clinical presentation is less useful. Most cases of feline mycobacterial disease in GB are cutaneous.

Mycobacterial infections are recognised as a global health concern, both in humans and other animals.^1–3 One species that is known to be infected by a number of different mycobacteria is the domestic cat. Unfortunately, many aspects of mycobacterial infections in this species remain unknown; there have been few recently published research papers on feline mycobacteriosis in general, and even fewer on feline tuberculosis in particular.⁴

Mycobacterial disease in the domestic cat can result in several different syndromes including tuberculosis (typically caused by Mycobacterium bovis or Mycobacterium microti), feline leprosy (Mycobacterium lepraemurium, and other similar bacteria), and non-tuberculous mycobacteriosis caused by non-tuberculous mycobacteria (NTM) (Mycobacterium fortuitum, Mycobacterium aviume intracellulare complex [MAC], and others).^5–14 In the UK, the majority of recently reported cases of feline mycobacterial disease have been primarily cutaneous in nature and they presented with nodules, draining tracts, ulceration and local lymphadenopathy.⁴ Where systemic disease is seen, infection with a member of the tuberculosis group or a MAC organism is most likely^8,15 although occasional cases have been seen with other NTM.¹⁶

Cats may become infected via a number of different routes. Historically, tuberculosis in cats resulted from the ingestion of tuberculous milk and was seen as gastrointestinal disease.^17–19However, following the introduction of pasteurisation of milk and tuberculous testing of cattle, gastrointestinal disease is now a rare presentation and most cases affect the skin, at least initially, but may later spread to the lungs.⁴ The current epidemiology of feline tuberculosis is unclear but infection could occur through a number of possible routes. Direct spread from wild rodents has been suggested because hunting rodents has been shown to be a risk factor for tuberculosis in cats^13,15 and some small mammal species in the UKn are naturally infected with either M microti^20,21 or M bovis.^22,23 Direct spread of M bovis from badgers following interspecific aggression is also possible. Alternatively, M bovis-infected cattle and badgers could cause environmental contamination²⁴ and cutaneous wounds on cats (eg, from fighting) could become secondarily infected. This latter method of infection is also believed to occur with feline leprosy which arises from contaminated rodent bites or following soil or plant contamination of cutaneous wounds.^4,5,7,9,10 As NTM are typically found in soil, water and decaying vegetation, NTM infections are also believed to be secondary to wound contamination. ^6,11,12,25 (For DEFRA Guidance notes on tuberculosis in cats go to: CatsTBbriefing (VIPER23 App Y5)_March 08 update.doc.)

Given the paucity of our knowledge about the nature of current feline mycobacterial infections in Great Britain (GB), the primary aims of this study were to use the exceptional number of feline mycobacterial cases collected by the Veterinary Laboratories Agency (VLA) to determine which mycobacterial species are present in cats in GB and where they occur, and how these infections most commonly present. Knowing which bacteria are present permits determination of which cases are appropriate to treat, which are more likely to respond to treatment and how best to tailor the treatment protocols. In addition, it is particularly important to identify cats infected with M bovis and M microti as these have the most significant potential zoonotic risk. As culture can take up to three months¹³ and access to molecular diagnostics is currently still limited and is expensive, the secondary aim of the study was to determine if analysis of the data could enable prediction of which mycobacterial species is present based on a cat's geographical location within GB and its clinical presentation.

Materials and methods

Tissue samples

Between January 2005 and December 2008, 339 feline samples were submitted to the VLA Weybridge by veterinary surgeons in GB for mycobacterial culture.²⁶ The samples came from cats that had been found to have cutaneous lesions or suspicious masses at exploratory laparotomy, and when formalin-fixed samples were sent to private pathology laboratories for histopathology the tissue was found to have lesions suggestive of mycobacterial infection with typical granulomatous or pyogranulomatous inflammation, consisting of multifocal to coalescent infiltration with large numbers of foamy macrophages containing variable numbers of acid-fast bacilli. The veterinary surgeons then took a second sample and submitted it without fixation to the VLA for mycobacterial culture.

Veterinary surgeons that submitted the samples to the VLA were contacted by one of the authors (SMcF) and asked to provide information on where the cat lived (ie, the postcode of the owner's house), plus the cat's signalment (age, breed, gender) and its clinical presentation including the diagnostics that had been performed to assess possible systemic involvement (serum biochemistry, radiography, feline leukaemia virus and feline immunodeficiency virus [FeLV/FIV] status). Information on the histopathological findings, treatment (surgery, drugs given, and duration of treatment) and eventual outcome are presented in the accompanying paper.²⁷ In some cases the requested information was incomplete or not available so where data were missing the number of samples included in the analysis has been noted. Complete postcodes, where available, were converted into Ordinance Survey eastings and northings. For incomplete postcodes (ie, information only available up to postcode district level), the average easting and northing for the postcode district was taken.

Statistical analyses

Geographical Information System (GIS) analyses incorporating the SaTScan (v7.0.3 www.satscan.org) statistic were carried out to ascertain whether there were any statistically significant clusters of samples that could be cultured and whether the different cultured species clustered in particular parts of GB. Two groups of factors were considered for analyses: (i) signalment and (ii) clinical presentation. For each group standard univariate logistic regression risk factor analysis was performed to see whether particular factors were associated with whether culture and classification of the sample was possible. In addition, odds ratio (OR) and associated 95% confidence interval (CI) were calculated. The culture results were divided into four groups: (i) M bovis, (ii)M microti, (iii) NTM and (iv) no growth, and Fisher exact tests were carried out to identify any association with the signalment or clinical presentation. The NTM group was also divided further into M avium and non-M avium NTM for some of the comparisons. In all cases, statistical significance was set at P< 0.05.

Results

Mycobacterial species identified and the geographical location of the infected cats

Table 1 summarises the culture results obtained for the 339 samples. Mycobacteria could be cultured from 159 (47%: 95% CI 41–52) of the samples. Of these 159 samples, three species made up 87% of the isolates (M microti: 40%, M bovis: 33% and M avium: 15%, Table 1).

Table 1.

Mycobacterial culture results. The samples had histopathological changes indicative of mycobacterial infection and were submitted to the Veterinary Laboratories Agency for mycobacterial culture between January 2005 and December 2008

Culture results	Number	Percentage (%) of cultured

M microti	63	39.6
M bovis	52	32.7
Non-tuberculous mycobacteria (NTM)
M avium	24	15.1
M malmoense	4	2.5
M fortuitum	4	2.5
M celatum	1	0.6
M intracellulare	1	0.6
Unclassified	10	6.3
Cultured total	159
No growth	180
Grand total	339

Complete postcodes were available for 277 samples and incomplete postcodes for 49 samples, with no postcode available for 13 samples. SaTScan analysis of the 326 samples with complete or incomplete postcodes revealed that there were no apparent clusters in terms of being able to culture a sample (P = 0.544, Fig 1a). The samples with no postcode information were two M bovis and 11 no growth.

Fig 1.

Map of GB showing the location of 326 feline samples obtained between January 2005 and December 2008 for which the VLA tried to culture mycobacteria. (a) Samples have been subdivided into samples that could (red symbols) and could not be cultured (blue symbols). (b) Samples which could be cultured were subdivided into either the Mycobacteria species isolated or as unclassified mycobacteria. Also indicated is whether the position on the maps is from a complete postcode (•) or the mean easting and northing of the postcode district (▄). The coloured shaded areas correspond to predominance by one Mycobacteria species and the coloured circles the spatial clusters identified by the SaTScan analysis.

The spatial distribution for the M microti samples and some of the M bovis samples have already been published by Smith et al,¹³ who focused on M microti-positive samples from cats collected by the VLA over the last 14 years, and they found there were two areas where M microti dominated; Northern England/Southern Scotland and Southern England (Fig 1b). The current SaTScan analysis identified two smaller more well defined M microti clusters within these areas; one south of London and the other in South-West Scotland (P < 0.015). All of the 10 culture-positive samples in the Scottish cluster were M microti, and all but one of the 28 culture-positive samples in the south of London cluster were M microti, with the other cultured sample being the M bovis isolate described below. While there were four M microti culture-positive samples apparently tightly clustered together in Cumbria this cluster was not identified as statistically significant by the SaTScan analysis, nor were they included in the Southern Scotland cluster. The two other main Mycobacteria species also separated into quite distinct areas in GB (Fig 1b). This was confirmed by the SaTScan analysis that identified a large M bovis cluster in South-West England/Wales/Welsh Border (P< 0.001), with all but two of the 50M bovis samples (where location was known) being within this cluster, andM bovis samples comprising 76% (95% CI 63–86) of the cultured samples within the cluster. One of the M bovis isolates out with the main cluster coincides with a reported M bovis cluster in cattle in Sussex.²⁸ The other M bovis isolate out with the cluster is from a cat reported in Lincolnshire. This cat was a stray cat handed into the local Cats Protection centre with no previous history available. Finally, there was a large M avium cluster in Eastern England (Fig 1b), with 66% (44–84) of the Mycobacteria species culture-positive samples in the cluster being M avium (P< 0.001).

Case signalment and risk factors

Male cats constituted 61% of the samples (Table 2), with 93% of cats having been neutered (92% male, 94% female). Eighty-three percent of the cats were either domestic shorthair (DSH) or domestic longhair (DLH). The remaining breeds identified were Siamese (n = 18), Burmese (five), Bengal (three), British Shorthair (BSH) (three), Persian (three), Maine Coon (two), Abyssinian (one), Burmilla (one), Oriental (one), Russian (one) and Ragdoll (one) cats. Sixty-five percent of the cats were known to have outdoor access, and a further 33% were presumed to have outdoor access as they had a history suggestive of this, with yearly administration of vaccines, regular flea and worming treatment and/or treatment for cat bite abscesses. Only 2% of the cats were reported as living exclusively indoors.

Table 2.

Summary of the recording and occurrence (number [N], prevalence and 95% exact binomial confidence intervals [CI]) of signalment factors (age, gender, neutered status, whether they were DSH/DLH cats, and whether they had outdoor access).^*

Risk factor	N	Positive	%	P value	OR (95% CI)

Age	302			0.011	0.93 (0.88–0.98)
Sex = male	289	175	60.6 (54.6–66.2)	0.802	1.06 (0.66–1.70)
Status = meutered	258	241	93.4 (89.6–96.1)	0.326	0.60 (0.21–1.67)
DSH/DLH = yes	303	250	82.5 (77.7–86.6)	0.310	1.36 (0.75–2.49)
Cat = known outdoor access	168	109	64.9 (57.1–72.1)	0.357	1.35 (0.71–2.54)
Cat = known plus presumed outdoor access^†	168	164	97.6 (94.0–99.3)	0.981	1.02 (0.14–7.45)

Signalment data were not available in all cases with the age of 37 cats, the gender of 50, the neutered status of 81, and the breed of 50 cats being unknown. Also included is the statistical significance (Wald P value) and OR (and 95% CI) for these factors associated with mycobacterium being grown from the samples. For the continuous variable (age), the OR refers to the change in odds for an increase in age of 1 year, for the remaining categorical variables the odds of the second category relative to the first category for each factor.

†

Where it could not be confirmed that a cat had outdoors access, its file was assessed for a history suggestive of outdoor access, including the administration of yearly vaccines, regular flea and worming treatment and/or treatment for cat bite abscesses.

Table 2 summarises the univariate analyses of the relationships between being able to culture mycobacteria from the sample and the signalment of the cat. The only statistically significant relationship was with age, with a decrease in the odds of having a positive culture increasing with age (OR (years) 0.93 (0.88–0.98)). If the different species of mycobacteria were considered then differences in the age distributions of cats were observed (Fig 2), with M bovis samples coming from significantly younger cats than those with M microti or where no bacteria could be grown (P < 0.001). No significant variation was observed in the percentage of (i) males (58–65%, P = 0.934), (ii) neutered cats (92–100%, P = 0.318), (iii) DSH/DLH (76–93%, P = 0.071), or (iv) known to be outdoor cats (96–100%, P = 0.720) when the samples were divided into the different Mycobacteria species identified.

Fig 2.

Boxplot of the age distribution (in years) associated with the particular mycobacteria isolated. Horizontal thick lines indicate the median age, boxes indicate the interquartile range and whiskers the range. Also included is the number of isolates of each particular Mycobacterium species for which the age of the cat was identified.

Clinical presentation

Lumps and lesions were the most common presenting signs (85%, Table 3), with 74% of cases presenting with cutaneous lesions (where the location of the lumps/lesions was reported). Lumps/lesions were the sole presenting sign in 59% of the cats. The most common localisation of the lumps/lesions was on the head (54%), but they were also seen all over the body. Multiple lesions were present in almost half of the cats. For the 145 cats for which descriptions of the lumps/lesions were available there was no single predominate description, with ulceration, discharge and bite wound/abscess being roughly equal in occurrence (26–35%).

Table 3.

As for Table 2, but for information about clinical presentation. Presenting signs were not available for 57 cases

	N	Positive	%	P value	OR (95% CI)

Presenting information available	282
Lame	270	16	5.9 (3.4–9.4)	0.181	0.48 (0.16–1.41)
Ocular	271	18	6.6 (3.9–10.3)	0.757	0.86 (0.33–2.25)
Respiratory	272	28	10.3 (6.9–14.5)	0.846	0.93 (0.42–2.03)
Weight loss	272	43	15.8 (11.6–20.7)	0.041	2.01 (1.03–3.93)
Off colour	272	44	16.2 (12.0–21.1)	0.058	1.89 (0.98–3.67)
Lumps/lesions (LL)	281	240	85.4 (80.7–89.3)	0.266	0.68 (0.35–1.33)
LL only presenting condition		167	59.4 (53.4–65.2)	0.538	0.86 (0.54–1.39)
(a) LL location
Head	228	124	54.4 (47.6–61.0)	0.308	0.77 (0.47–1.27)
Limbs	226	70	31 (25.0–37.4)	0.974	1.01 (0.59–1.72)
Body	223	60	26.9 (21.2–33.2)	0.264	1.38 (0.78–2.45)
Not cutaneous	224
Abdomen		20	8.9 (5.5–13.5)	—	—
Eye		8	3.6 (1.5–6.9)	—	—
Oropharynx/throat		2	0.9 (0.11–3.2)	—	—
Oedema		1	0.4 (0.01–2.5)	—	—
(b) Multiple lesions	227	106	46.7 (40.0–53.4)	0.866 1.05	(0.62–1.76)
(c) LL description	145
Bite wound/abscess		39	26.9 (19.8–34.9)	0.568	1.24 (0.59–2.59)
Ulcerated		38	26.2 (19.2–34.2)	0.251	0.64 (0.30–1.37)
Discharging		51	35.2 (27.4–43.5)	0.691	1.15 (0.58–2.28)
Swelling		30	20.7 (14.4–28.2)	0.551	0.78 (0.34–1.77)
Lymph node		15	10.3 (5.9–16.5)	0.219	1.98 (0.67–5.89)
Nodular		6	4.2 (1.5–8.8)	—	—
Lymph nodes involved	223	109	48.9 (42.1–55.6)	0.004	2.19 (1.28–3.75)
Submandibular lymph nodes	220	64	29.1 (23.1–35.6)	0.035	1.89 (1.04–3.44)
Other peripheral lymph nodes	218	51	23.4 (17.9–29.6)	0.031 2.04 (1.07–0.91)
Internal lymph nodes	220	26	11.8 (7.8–16.8)	0.145	1.89 (0.80–4.44)

— = No logistic regression analysis performed.

Lymph node involvement was noted in 49% of cases (where it was possible to determine whether lymph nodes were involved), with submandibular lymph nodes affected most commonly (29%, Table 3). However, in 12% of the cats multiple lymph nodes were affected — with a combination of submandibular and other peripheral lymph nodes (typically the pre-scapular nodes) the most common multiple site combination in 18/27 cases. Internal lymph node involvement was recorded in only 12% of the cases.

Having lumps/lesions as a presenting sign was not associated with being able to culture mycobacteria from the sample, either if considering lumps/lesions as the sole presenting condition or where there was more than one presenting condition (P > 0.266, Table 3). Furthermore, there was no association between being able to culture mycobacteria and the description of the lesion, its location, or whether multiple or single lesions were observed (P > 0.219). The only presenting sign that was associated with being able to culture mycobacteria was weight loss, but this was not a strong association (P = 0.041, OR 2.01 (1.03–3.93)), with 63% of the samples from cats that presented with weight loss being able to be cultured, compared to 46% of the samples from cats that did not present with weight loss.

A stronger association was observed with whether lymph nodes were involved, (P = 0.004, OR 2.19 (1.28–3.75)); with 61% of samples from cats where lymph nodes were involved being able to be cultured, compared to 42% of samples from cats where lymph nodes were not involved. This association was also observed if just submandibular or other peripheral lymph nodes were considered, though the associations were much weaker (P< 0.036).

If the culture results were divided into which mycobacterial species could be grown (M bovis, M microti, NTM, or no growth) there was little overall difference in the percentage of samples with any of the presenting conditions (P > 0.097). That said, there was variation between some of the groups for some of the findings, for example, the percentage of samples where lumps/lesions were observed on the head (P = 0.036), with NTM having a lower percentage (31%) compared to M bovis (52%) or M microti (62%) and no growth (55%) (Fig 3). The same was not true for lumps/lesions observed either on limbs or the body (P > 0.112). The percentage of samples with multiple lesions also differed between the four groups (P = 0.002), with M microti having a much higher percentage (70%), compared to M bovis (34%), NTM (30%) and no growth (46%). There was little variation in the percentage of samples in all but one of the descriptions of the lumps/lesions (P > 0.123). The exception was a much lower percentage of M microti samples (16%) having a discharging lesion (34–52%, P = 0.026) (Fig 3). There was also marked variation in the percentage of samples where the lymph nodes were thought to be involved (P = 0.004), ranging from 73% of M bovis samples to 39% with no growth (Fig 3). If just whether the submandibular lymph nodes were involved was analysed, there was little difference between M bovis (41%) and M microti (46%), but the other two groups had much lower percentages (17–22%, P = 0.005). Finally, internal lymph node involvement was more likely with M bovis and M avium (21% and 27%, respectively). This was also reported, but to a lower degree, with M microti (4/45 cases; 9%) and cases with no growth (9/106 cases; 8.5%), but it was not reported in any of the cats with non-M avium NTM.

Fig 3.

Pictorial representations of the (a) prevalence of lesions due to the different infections on different parts of the cats' body, (b) the overall prevalence of cutaneous lesions, and whether they were (c) ulcerated or (d) discharging lesions, and (e) whether there was lymph node involvement.

Diagnostics tests

Further diagnostic tests were performed in some of the cases to determine if the cats were concurrently infected with FeLV and/or FIV and to see if there was evidence of systemic dissemination of mycobacterial infection. Seventy-two cats were tested for FIV/FELV, of which two were positive for FIV. Blood calcium concentration was only assessed in 39 cats, and was elevated in nine (calcium concentration up to 3.95 mmol/l; reference interval 2.00–2.95 mmol/l). Of these nine cats, four had systemic disease and three had respiratory disease (all of which had M microti); while M fortuitum was cultured from one cat with extensive cutaneous disease overlying its thorax (and concurrent sternal lymph node involvement); and the ninth case involved the inguinal fat pads but the bacteria failed to grow. Thirty-six of 72 cats that had chest radiographs taken were found to have pulmonary pathology, but this was not associated with increased odds of successful culture (P = 0.633, OR 1.26 (0.49–3.21)). Nor was there any association between which mycobacterium was isolated and the percentage of chest radiographs with pathology (P = 0.319) M bovis 50%, M microti 67%, 3/8 NTM (two of which were M avium), and 14/30 with no growth.

Discussion

This paper (and its sister paper focussing on histopathology, treatment and outcome²⁷) reports on the largest study of feline mycobacterial disease reported to date. The largest previous studies looked at 179 cases of feline leprosy from New Zealand;²⁹ 52 cases of M bovis from the US;³⁰ 49 cases of NTM from Australia; ⁶ 101 cases of M microti from GB;¹³ and 10 cases of M avium from Australia and the US.⁸ From GB, the previous largest studies of M bovis infection included 13¹⁷ and 12 cases,³¹ and 43 cases where the focus was purely on the regional distribution.¹³ Analysis of the unique data set available in the current study has generated some fascinating results. These relate to the frequency with which different mycobacterial species cause disease in cats in GB, their geographical distribution, and their clinical presentation.

The study revealed which culturable mycobacterial species are currently causing disease in cats in GB. Mycobacterium microti and M bovis are responsible for similar levels of infection, and between them they account for a third of the submitted cases. While there have been few recently reported cases of feline tuberculosis, those that have been published were also caused by either M microti or M bovis, or they quote or present VLA data.^{4,13,15,31–33} No cases of Mycobacterium tuberculosis were identified, probably because cats are thought to be naturally resistant to this infection,^18,34 in addition to which this infection is now much less prevalent in the human population of GB.³⁵ Importantly, our findings confirm that M microti is a significant pathogen of cats in GB.^13,15,36 Confusingly, M microti infection in cats was previously termed M microti-like as it was unclear at the time that it was actually the same organism. ^15,37,38 In addition, some reports have discussed cases where the infection was reported to be M tuberculosis³⁹ or M tuberculosis var bovis,^15,40 which on further investigation appear to have been M microti. The current study isolated M avium less frequently than M microti or M bovis and this is in agreement with current literature where few cases of this infection have been reported in cats.^8,41–43

The data in the current study was also divided into four groups so that the two most important infections (M microti and M bovis) could be clearly defined, and then compared to the more heterogeneous NTM and the no-growth group. Since successful culture of M bovis can take up to two months and M microti can take up to three months¹³ and access to molecular diagnostics is currently still limited and expensive, one aim of the study was to determine if it was possible to predict which mycobacterial species was present based on the cat's geographical location within GB and its clinical presentation.

Analysis of the postcode data showed that when a cat's location was mapped within GB there was marked clustering of the three most significant infections. The two M microti clusters have previously been reported by Smith et al¹³ with one cluster south of London and the other in Northern England/Southern Scotland (Fig 1b). In addition, most of the isolates found in South-West England/Wales/Welsh Border were M bovis (Fig 1b). This confirms the reported spatial distribution from Smith et al¹³ which was based on a smaller sub-set of the current M bovis data. However, there was also a M avium cluster — where isolates were predominately found in Eastern England. Interestingly, where clusters were identified other species of mycobacteria tended to be absent.¹³ This suggests that knowing a cat's location can be useful in suggesting which infections are more likely, for example, if a cat is seen in Scotland, it is unlikely to have M bovis (unless it has moved from a high risk area).

It is unclear exactly how the cats are becoming infected with M bovis and M microti. While the current study could not determine if the cats came from urban and non-urban regions, the majority (98%) were presumed to be outdoor cats from the information obtained from their veterinary surgeons. The previous study on a larger set of VLA M microti isolates, suggested that cats withM microti infection were more likely to come from extra-urban areas.¹³ This supports the hypothesis that these infections are probably gained when the cats are outside their homes, most probably in their garden territories. Cats could become infected by a number of different routes. These include hunting small rodents (infected with either M microti or M bovis), interspecific aggression with badgers (infected with M bovis), and/or following environmental contamination, for example, M bovis being shed by badgers that have domestic gardens as part of their territory.^4,44 The current study finding that most (74%) of the cats had cutaneous lesions at ‘fight and bite sites’ correlates well with this theory. What role cats may have in the onward transmission of these mycobacteria is unknown.

The study supports previous findings that in GB cases of reported mycobacterial disease typically occurs in adult cats (mean age 7.2 years; 95% CI 6.7–7.7), that are neutered (93%), male (61%), and non-pedigree (83%), and have outdoor access (98%).⁴ While all but two of the cats were over a year of age, the mean age of the cats with the different infections did vary, as cats with M bovis infection had a mean of three years of age, compared to eight years of age for M microti (Fig 2). The percentage of cats with known outdoor access was similar to that for most domestic cats in GB (98%).⁴⁵ This was also the case for the percentage of non-pedigree cats as approximately 90% of pet cats in the GB are non-pedigree (Pet Food Manufacturers' Association [PFMA] Annual Report 2004). However, cats with M microti were slightly more likely to be pedigree (24% pedigree cats) than the cats with the other infections. The Siamese breed may be over-represented in the current study as 18/39 pedigree cats were Siamese, while this breed is only the second most popular cat breed in GB (General Council of the Cat Fancy [GCCF], data from 2008; www.gccf.org). Interestingly, while previous papers have suggested that Siamese cats may be predisposed to MAC infections,^46–48 none of the Siamese cats in the current study were found to have this infection.

Where the information was reported, the cats in this study most commonly presented with single or multiple cutaneous lumps or lesions (74%), which were sometimes ulcerated or discharging, and most frequently located on the head (54%), and multiple lesions were found in 47% of cases. When looking at the different infections, NTM were the least likely to cause lesions on the head (31% compared to 52–60% for the other groups) (Fig 3). The presence of multiple lesions differed between the groups, being more common with M microti (70%), compared to the other groups (30–46%). Having a discharging lesion also varied between groups, as it was less likely for M microti (16%) compared to the other groups (34–52%) (Fig 3). Lymph nodes were frequently involved (49%); most typically the submandibular and/or other peripheral lymph nodes; with lymph node involvement being very common with M bovis (73%) and less common with the no-growth cases (39%). When submandibular lymph node involvement was analysed M bovis (41%) and M microti (46%) were similar, and in both groups this was more likely than for the other two groups (17–22%). These findings support the theory that most cases of mycobacterial infections in cats in GB are seen as cutaneous lesions affecting ‘fight and bite sites’ and/or presenting as submandibular lymphadenopathy.

It has previously been reported that systemic involvement is more likely with infection by a member of the tuberculosis group or a MAC organism;^8,15 with only occasional cases been caused by other NTM.¹⁶ Therefore, it was hoped that the presence of systemic disease would be a useful indicator of M microti, M bovis, or M avium infections and it would make a diagnosis of non-M avium NTM less likely. In support of this, internal lymph node involvement was found more frequently with M bovis and M avium (21% and 27%, respectively) and, to a lesser extent, M microti (4/45 cases; 9%) and cases with no growth (9/106 cases; 8.5%). Importantly, it was not reported in any of the cats with non-M avium NTM. This was also supported when looking at the cases where chest radiographs were taken; the percentage of cases which showed pathology consistent with mycobacterial infection was 67% with M microti, 50% with M bovis, 47% with no growth, and 3/8 (38%) with NTM (two of these were caused by M avium and the third was unclassified). Overall, the current study supports the assumption that non-M avium NTM are less likely to spread systemically than other mycobacterial infections; however, it also shows that systemic infections can still be caused by these bacteria.

Cats with systemic involvement are believed to have a poorer prognosis and may be more resistant to treatment.^{5,7–9,12,15,46} Because of this it is important to assess each cat for possible systemic spread. At presentation, systemic signs (such as weight loss, involvement of internal lymph nodes and/or respiratory disease) were seen in only 10–16% of cases. No cats were FeLV positive and only 2/72 cats tested were FIV positive. As discussed above, where chest radiographs were taken (possibly because of a suspicion of pulmonary involvement) pulmonary pathology was seen most commonly with M microti cases, to a lesser extent with M bovis and no-growth cases, and least frequently with NTM. Where blood calcium concentration was assessed (again, possibly because of suspicion of systemic involvement) hypercalcaemia was found in 9/39 (23%) cases; seven of which were infected with M microti and had systemic or pulmonary disease, and two had extensive cutaneous disease (one caused by M fortuitum and one from which mycobacteria could not be grown). Hypercalcaemia has previously been reported in only four cats with mycobacterial disease, all of which had systemic M avium infection.⁸ This study adds to our understanding of hypercalcaemia associated with mycobacterial infection as this has not previously been documented in so many cats, in cats with M microti or M fortuitum infections, or in cats with only cutaneous (albeit extensive) disease. These data show that evidence of systemic spread of mycobacterial infection was not uncommon. Unfortunately, when reviewing the presentation data, no group of cats was found to be significantly more likely to have either pulmonary and/or systemic involvement, although, they tended to occur less frequently when infection was caused by the non-M avium NTM.

Overall, by reviewing the data, the different infections were found to be associated with slightly differing patient and disease profiles (Fig 3):

M microti: Cats with M microti infections were most typically from South-West Scotland and Northern England or south of London,¹³ with a median age of eight years. They were slightly more likely to be pedigree cats (24% pedigree cats) than those with M bovis 7% or NTM 12%, and they usually had multiple lesions (70%), particularly on the head (60%), which were typically non-discharging (only 16% had discharging lesions), and submandibular lymph nodes were involved in 46% of cases. While only 39 cats were assessed for serum total calcium concentration nine were found to have elevated levels, and seven of these had M microti infections, which in each case involved either pulmonary and/or systemic infection.

M bovis: Cats with M bovis infections were typically from the South-West England/Wales/Welsh Border (in agreement with reference¹³) with a median age of three years. They were unlikely to be pedigree cats (93% DSH/DLH), and they more typically had single lesions (only 34% had multiple lesions), which were on the head (52%) more frequently than anywhere else. Lymph nodes were frequently affected (73%), with the submandibular lymph nodes being involved in 41% of cases.

NTM and no-growth groups: This data is more difficult to interpret because the culture system used by the VLA has been optimised to detect M bovis, and is, therefore, not ideal for detecting many species of NTM. Therefore, in addition to the NTM that did grow, others may have been contained in the no-growth group. The spatial distribution of the no-growth samples was no different from that of all the samples that could be cultured. This could suggest that many of the no-growth samples were NTM. However, false negative results could also have occurred, with the failure to culture M bovis or M microti resulting from the paucity of organisms and/or, reduced viability of organisms due to recent antimicrobial treatment or decomposition of the sample.

NTM: When looking at the non-M avium NTM, four cats with Mycobacterium malmoense came from the same part of England, as did 3/4 cats with M fortuitum. While this is an interesting observation, care should be taken not to over-interpret this as the numbers involved are small. The cats in the NTM group were generally similar to the cats in the three other groups. However, they tended to have multiple lesions less frequently than the other groups (30% compared to 34–70%), which affected the head less frequently than other groups (30% compared to 52–60%), and there was slightly less lymph node involvement than with M bovis and M microti (46% compared to 57% with M microti and 73% with M bovis). When considering only the cats with M avium infection, they tended to be located in eastern England, had a median age of six years, and had an increased risk of internal lymph node involvement (27% of M avium cases had internal lymph node involvement, compared to 21% of M bovis cases, 9% of M microti cases, 8% of no-growth cases). This is in contrast to the rest of the NTM (after removal of the M avium cases), where none of the cases had systemic lymph node involvement. One footnote to the current study is that Mycobacterium celatum is reported in this study for the first time in a cat. The cat was a 3-year-old neutered female cat from Kent with a 2-year history of pyrexia and recurrent full-thickness punctuate lesions over her dorsum and right thigh that oozed seropurulent material.

No-growth group: The cats in this group had similar clinical presentations to the other groups, although peripheral or systemic lymph node involvement was less typical. Peripheral lymphadenopathy was only seen in 39% of cases (compared to 73% with M bovis) and internal lymphadenopathy was only seen in 8% of cases (compared to 21% with M bovis and 27% with M avium).

Interestingly, the cats in the no-growth group were statistically older than the cats with M bovis; although there was no difference between them and the cats with M microti. The increased likelihood of gaining a positive culture from samples from younger cats could perhaps result from the younger cats having a less effective immune response and so being more permissive for mycobacterial growth, as has previously been hypothesised. ^10,49 This could, in turn, result in higher bacterial numbers and a greater chance of a positive culture.

Clearly, there are a number of caveats to this study. These include: (i) being unable to accurately define the population from which the cases came (see below), (ii) having to rely on clinical files for case data some of which was sometimes absent, (iii) using a culture system that is optimised for M bovis and (iv) not having the resources to undertake molecular diagnostics to identify which bacteria were involved in the cases where the mycobacteria did not grow in culture. The prevalence of mycobacterial infections in cats in GB remains unclear because we were unable to accurately define the population to which these cases belong. This is because a large number of steps have to be taken before a sample is submitted to the VLA. These include: the lesions have to be large enough for the owner to notice and/or the cat needs to be sufficiently ill, the owner has to be able to afford to take their cat to a veterinary surgeon, the veterinary surgeon has to send the sample off to a diagnostic laboratory for histopathology, the laboratory has to recommend that the primary clinician collect a second sample and send it to the VLA for culture, the cat has to have another lesion suitable for biopsy, and the owner has to be willing to pay for the repeat procedure. Unfortunately, few veterinary surgeons have mycobacterial infections on their differential diagnosis list when they collect the primary sample so it is only when the diagnostic laboratory reports their findings that the veterinary surgeon knows they need a fresh sample for culture. Ideally, when a veterinary surgeon takes a biopsy from a suspicious lesion or an enlarged lymph node, they would section it and send one sample for histopathology and store the second sample in the freezer pending the histopathology result. Despite these caveats, this study is the largest study of feline mycobacteriosis ever published and it has produced some fascinating results.

In conclusion, this large study looking at 339 cases of feline mycobacterial infection was able to show which species of mycobacteria most commonly cause disease in cats in GB, geographically where the infected cats are most likely to live, and how M microti and M bovis infections occur in almost mutually exclusive spoligotype clusters (and see Smith et al¹³). The study found M microti in 19% of reported cases, with M bovis in 15%, M avium in 7%, non-M avium NTM in 6%, and no growth in 53%. While there were some differences between the patterns of disease caused by the different mycobacterial species, these were not diagnostic, and regardless of the mycobacteria involved, affected cats most commonly presented with single or multiple cutaneous lumps or lesions, which were sometimes ulcerated or discharging, and most frequently found on the head. Lymph nodes were usually involved; most typically the submandibular nodes. Therefore, clinical signs and lesion distribution are useful at suggesting that a diagnosis of mycobacterial infection should be considered. However, accurate identification of the species involved is needed to determine the appropriate course of action. Our findings show that while knowing the geographic location of the cat can be very helpful in this respect, the nature of the clinical signs helps less. Since culture takes up to three months and is frequently unsuccessful, we recommend concurrent use of molecular diagnostics (such as polymerase chain reaction and sequencing) and further development of cat specific assays (such as the interferon (IFN)-gamma test), as these can result in more precise and faster identification of the mycobacteria.

Footnotes

Acknowledgements

Thank you to all of the veterinary surgeons, nurses and reception staff who helped in this study. The VLA cultures and histopathology were funded by DEFRA under the project SB4510.

References

Glaziou

Floyd

Raviglione

. Global burden and epidemiology of tuberculosis. Clin Chest Med 2009; 30: 621–36.

Lobue

Enarson

Thoen

. Tuberculosis in humans and animals: an overview. Int J Tuberc Lung Dis 2010; 14: 1075–8.

Shiloh

DiGiuseppe

Champion PA

. To catch a killer. What can mycobacterial models teach us about Mycobacterium tuberculosis pathogenesis? Curr Opin Microbiol 2010; 13: 86–92.

Gunn-Moore

. Mycobacterial infections in cats and dogs. In: Ettinger

Feldman

, eds. Textbook of veterinary internal medicine. 7th edn. Philadelphia: WB Saunders, 2010: 875–81.

McIntosh

. Feline leprosy: a review of forty-four cases from Western Canada. Can Vet J 1982; 23: 291–5.

Malik

Wigney

Dawson

Martin

Hunt

Love

. Infection of the subcutis and skin of cats with rapidly growing mycobacteria: a review of microbiological and clinical findings. J Feline Med Surg 2000; 2: 35–48.

Malik

Hughes

James

. Feline leprosy: two different clinical syndromes. J Feline Med Surg 2002; 4: 43–59.

Baral

Metcalfe

Krockenberger

. Disseminated Mycobacterium avium infection in young cats: over-representation of Abyssinian cats. J Feline Med Surg 2006; 8: 23–44.

Greene

Gunn-Moore

. Infections with slow-growing mycobacteria. In: Greene

, ed. Infectious diseases of the dog and cat. 3rd edn. St Louis: Saunders, Elsevier, 2006: 462–77.

10.

Malik

Hughes

Martin

Wigney

. Feline leprosy syndromes. In: Greene

, ed. Infectious diseases of the dog and cat. 3rd edn. St Louis, Missouri: Saunders, Elsevier, 2006: 477–80.

11.

Malik

Martin

Wigney

Foster

. Infections caused by rapidly growing mycobacteria. In: Greene

, ed. Infectious diseases of the dog and cat. 3rd edn. St Louis, Missouri: Saunders, Elsevier, 2006: 482–8.

12.

Horne

Kunkle

. Clinical outcome of cutaneous rapidly growing mycobacterial infections in cats in the south-eastern United States: a review of 10 cases (1996–2006). J Feline Med Surg 2009; 11: 627–32.

13.

Smith

Crawshaw

Parry

Birtles

. Mycobacterium microti: more diverse than previously thought. J Clin Microbiol 2009; 47: 2551–9.

14.

Gunn-Moore

Dean

Shaw

. Mycobacterial infections in cats and dogs. In Pract 2010; 32: 444–52.

15.

Gunn-Moore

Jenkins

Lucke

. Feline tuberculosis: a literature review and discussion of 19 cases caused by an unusual mycobacterial variant. Vet Rec 1996; 138: 53–8.

16.

Couto

de Bolla

Artacho

. Mycobacterium fortuitum pneumonia in a cat and the role of lipid in the pathogenesis of atypical mycobacterial infections. Vet Pathol 2007; 44: 543–6.

17.

Jennings

. The distribution of tuberculous lesions in the dog and cat, with reference to the pathogenesis. Vet Rec 1949; 27: 380–4.

18.

Smith

. Symposiumon diseases of cats e III. Some pathogenic bacteria of cats with special reference to their public health significance. J Small Anim Pract 1965; 5: 517–24.

19.

Parodi

Fontaine

Brion

Tisseur

Goret

. Mycobacterioses in the domestic carnivora e present day epidemiology of tuberculosis in the cat and dog. J Small Anim Pract 1966; 6: 309–17.

20.

Cavanagh

Begon

Bennett

. Mycobacterium microti infection (vole tuberculosis) in wild rodent populations. J Clin Microbiol 2002; 40: 3281–5.

21.

Burthe

Bennett

Kipar

. Tuberculosis (Mycobacterium microti) in wild field vole populations. Parasitology 2008; 135: 309–17.

22.

Delahay

De Leeuw

Barlow

Clifton-Hadley

Cheeseman

. The status of Mycobacterium bovis infection in UK wild mammals: a review. Vet J 2002; 164: 90–105.

23.

Delahay

Smith

Barlow

. Bovine tuberculosis infection in wild mammals in the South-West region of England: a survey of prevalence and a semiquantitative assessment of the relative risks to cattle. Vet J 2007; 173: 287–301.

24.

Biet

Boschiroli

Thorel

Guilloteau

. Zoonotic aspects of Mycobacterium bovis and Mycobacterium aviume intracellulare complex (MAC). Vet Res 2005; 36: 411–36.

25.

Jang

Hirsh

. Rapidly growing members of the genus Mycobacterium affecting dogs and cats. J Am Anim Hosp Assoc 2002; 38: 217–20.

26.

Daniel

Evans

Rolfe

. Outbreak of tuberculosis caused by Mycobacterium bovis in golden Guernsey goats in Great Britain. Vet Rec 2009; 165: 335–42.

27.

Gunn-Moore

McFarland

Brewer

. Mycobacterial disease in cats in Great Britain: II. Histopathology, treatment and outcome of 339 cases. J Feline Med Surg 2011; 13: 945–52.

28.

Green

Cornell

. Investigations of cattle herd breakdowns with bovine tuberculosis in four counties of England and Wales using VETNET data. Prev Vet Med 2005; 70: 293–311.

29.

Thompson

Little

Cordes

. Observations of cat leprosy. N Z Vet J 1979; 27: 233–5.

30.

Snider

Cohen

Reif

Stein

Prier

. Tuberculosis in canine and feline populations. Study of high risk populations in Pennsylvania, 1966–1968. Am Rev Respir Dis 1971; 104: 866–76.

31.

Monies

Jahans

de la Rua

. Bovine tuberculosis in cats. Vet Rec 2006; 158: 245–6.

32.

Gunn-Moore

Shaw

. Mycobacterial disease in the cat. In Pract 1997; 19: 493–501.

33.

Anon. Animal Health 2007: The report of the Chief Veterinary Officer. London: DEFRA Publications 2008: 34–40, http://www.defra.gov.uk/animals/cvo/report.htm (accessed Nov 11, 2008).

34.

Aranaz

Liébana

Pickering

Novoa

Mateos

Domínguez

. Use of polymerase chain reaction in the diagnosis of tuberculosis in cats and dogs. Vet Rec 1996; 138: 276–80.

35.

Grange

Gandy

Farmer

Zumla

. Historical declines in tuberculosis: nature, nurture and the biosocial model. Int J Tuberc Lung Dis 2001; 5: 208–12.

36.

Rüfenacht

Bögli-Stuber

Bodmer

Bornand

Jaunin VF

Gonin

Jmaa DC

Gunn-Moore

. Mycobacterium microti infection in the cat: a case report, literature review and recent clinical experience. J Feline Med Surg 2011; 13: 195–204.

37.

de Bolla

. Tuberculosis in a cat. Vet Rec 1994; 134: 336.

38.

Blunden

Smith

. A pathological study of a mycobacterial infection in a cat caused by a variant with cultural characteristics between Mycobacterium tuberculosis and M bovis. Vet Rec 1996; 138: 87–8.

39.

Kipar

Schiller

Baumgärtner

. Immunopathological studies on feline cutaneous and (muco)cutaneous mycobacteriosis. Vet Immunol Immunopathol 2003; 91: 169–82.

40.

Orr

Kelly

Lucke

. Tuberculosis in cats. A report of two cases. J Small Anim Pract 1980; 21: 247–53.

41.

Gow

. What is your diagnosis? Mycobacterial infection. J Small Anim Pract 2006; 47: 484–5.

42.

de Groot

van Ingen

de Zwaan

Mulder

Boeree

van Soolingen

. Disseminated Mycobacterium avium subsp avium infection in a cat, the Netherlands. Vet Microbiol 2010; 144: 527–9.

43.

Rivière

Pingret

Etievant

. Disseminated Mycobacterium avium subspecies infection in a cat. J Feline Med Surg 2011; 13: 125–8.

44.

Monies

Cranwell

Palmer

Inwald

Hewinson

Rule

. Bovine tuberculosis in domestic cats. Vet Rec 2000; 146: 407–8.

45.

Murray

Roberts

Whitmarsh

Gruffydd-Jones

. Survey of the characteristics of cats owned by households in the UK and factors affecting their neutered status. Vet Rec 2009; 164: 137–41.

46.

Hix

Jones

Karlson

. Avian tubercle bacillus infection in a cat. J Am Vet Med Assoc 1961; 138: 641–7.

47.

Drolet

. Disseminated tuberculosis caused by Mycobacterium avium in a cat. J Am Vet Med Assoc 1986; 189: 1336–7.

48.

Jordan

Cohn

Armstrong

. Disseminated Mycobacterium avium complex infection in three Siamese cats. J Am Vet Med Assoc 1994; 204: 90–3.

49.

Gross

Ihrke

Walder

Affolter

. Infectious nodular and diffuse granulomatous and pyogranulomatous diseases of the dermis. Skin disease of dog and cat: clinical and histopathological diagnosis. 2nd edn. St Louis: Mosby Year Book, 2005: 272–340.

Mycobacterial Disease in Cats in Great Britain: I. Culture Results,Geographical Distribution and Clinical Presentation of 339 Cases

Abstract

Materials and methods

Tissue samples

Statistical analyses

Results

Mycobacterial species identified and the geographical location of the infected cats

Case signalment and risk factors

Clinical presentation

Diagnostics tests

Discussion

Footnotes

Acknowledgements

References