Investigating Coxiella burnetii infection in a breeding cattery at the centre of a Q fever outbreak

Abstract

The potential role of cats in transmitting Coxiella burnetii to humans was highlighted in a Q fever outbreak, linked to a caesarean section in a breeding queen, in an Australian small animal veterinary hospital. The objectives of this study were to evaluate the C burnetii seroreactivity of the breeding queen and other cats residing at the same breeding cattery (n = 27) and to evaluate C burnetii infection of the breeding queen by molecular and histological methods. Three assays [complement fixation test (CFT), indirect immunofluorescence assay (IFA) and enzyme-linked immunosorbent assay (ELISA)] were used for serological evaluation. Additionally, uterine and ovarian samples collected from the breeding queen 11 weeks post-parturition were assessed by routine and specialised histological methods and polymerase chain reaction. The breeding queen showed strong seropositivity using CFT (titre 1/32), IFA (titre phase I 1/8192 and phase II 1/8192) and ELISA; however, the reproductive tract showed no evidence of pathology or C burnetii infection. A number of cattery-confined cats were identified as seropositive to phase II and/or phase I C burnetii. Serological detection of C burnetii in a breeding cattery linked to a Q fever outbreak indicates likely infection by this bacterium in Australian feline populations, re-confirming the relevance of this zoonosis.

Introduction

Coxiella burnetii is the aetiological agent of Q fever, a highly significant worldwide zoonosis with a large reservoir encompassing wild and domestic mammals, birds and arthropods. ¹ It is an obligately intracellular Gram-negative bacterium with extreme environmental resilience.^1–3 Coxiella burnetii produces several potentially debilitating forms of Q fever in humans, rendering it a significant public health concern. ¹ While infection is most commonly asymptomatic, 40% of primary infections in humans are symptomatic, with serious acute or chronic illnesses possible. ¹ Death can occur in both acute and, more commonly, chronic infections. ¹

Q fever is traditionally associated with contact with cattle, sheep and goats in the livestock and meat industries.^1,3 Less commonly, cats have been suggested as sources of C burnetii in southern Africa, ⁴ Japan,^5–8 Korea, ⁵ Canada^9–15 and the USA.^16–18 In maritime Canada, exposure to parturient cats and newborn kittens has been identified as a significant risk factor for Q fever, ¹¹ with seroprevalence of C burnetii infection in cats in these regions varying from 6.2 to 24%.^12,15 In other countries, 1.9–42% of cats have been seropositive,^4,5,8,18 suggesting infection of cats may be common.

Several community-acquired Q fever outbreaks have likely resulted from direct or indirect exposure to queens and/or their kittens during or shortly after parturition,^{9–11,13,14,17} probably via the infected animal’s birth products, where up to 10⁹ bacteria per gram of placental tissue have been reported in other species.^1,19 Coxiella burnetii has been detected in the blood, urine and genital tracts of infected cats.^11,13,16,20 While it appears cats may be infected by and shed C burnetii, it is uncertain whether infection is associated with clinical disease in cats. Several reports have suggested infection in cats, as in other species, may be linked to reproductive disorders.^1,6,11

Diagnosis of C burnetii infection in animals is complex owing to the lack of known disease associations and few available sensitive, specific diagnostic techniques. In animals, diagnosis primarily relies on serology. ²¹ The World Organisation for Animal Health (OIE) reference test for serological diagnosis of C burnetii remains the complement fixation test (CFT), ²² despite various more sensitive tests, including the indirect immunofluorescence assay (IFA) and enzyme-linked immunosorbent assay (ELISA), being available for use in humans, where the gold standard is IFA.^1,21 While CFT is highly specific, IFA and ELISA have higher sensitivity, detect seroconversion earlier and are more rapidly performed.^1,21 Other techniques for demonstrating C burnetii in biological samples include polymerase chain reaction (PCR) and immunohistochemistry.^1,21 Recently, fluorescent in situ hybridisation (FISH) has also been evaluated for detecting C burnetii in the placental tissues of naturally-infected ruminants, with comparable sensitivity and specificity to immunohistochemistry. ²³

The potential role of cats in C burnetii transmission to humans was highlighted by a Q fever outbreak in a Sydney small animal veterinary hospital following a caesarean section in a breeding queen in June 2010. ²⁴ Nine veterinary personnel, in addition to the queen’s owner, showed evidence of recent C burnetii infection. Six of the nine veterinary personnel were symptomatic with mild-to-severe ‘flu-like’ signs, with two requiring extended hospitalisation. Those most severely affected had assisted the caesarean section and performed mouth-to-mouth resuscitation on newborn kittens, and therefore likely experienced the most significant exposure to the queen’s birth products. ²⁴ To date, investigations of cats involved in Q fever outbreaks have focused on CFT and IFA to demonstrate C burnetii antibodies.^9,10,13,14 This article uses three standardised serological assays (CFT, IFA and ELISA) in combination with histological and molecular techniques (FISH and PCR) to evaluate C burnetii infection in the breeding queen identified as the probable primary source in the Q fever outbreak (index cat) and the seroreactivity of cats residing at the same breeding cattery.

Materials and methods

Samples

Serum samples were collected from the index cat and 26 other cats from the same breeding cattery. A second serum sample was acquired from three of these cats at different time points. Collection dates of all serum samples (n = 30) are given in Table 1. Whole blood (0.5–3 ml) from each cat was collected in serum separator tubes and centrifuged at 12,000 g for 10 min. Serum was harvested, divided into 300 μl aliquots and stored at −20°C until tested. Samples from the ovaries, left and right uterine horns, and uterine body were collected from the index cat at ovariohysterectomy, 11 weeks post-parturition, with half fixed in formalin and the other half frozen immediately at −20°C.

Table 1.

Collection dates and sequence of serum samples taken from cats in the breeding cattery

Collection date	Months since outbreak	Cat/s sampled	First collection	Second collection
July 2010	1.1	Index cat	✓
April 2011	11	Cats 2–22	✓
September 2011	16	Cat 23	✓
February 2012	22	Cats 2, 6, 9		✓
April 2012	23	Cats 24–27	✓

Serological testing

CFT

Serum samples (n = 30 from 27 cats) were submitted to the Elizabeth Macarthur Agricultural Institute, Menangle, Australia, for testing by CFT (OIE reference test ²² ).

IFA

A modification of a commercial human C burnetii IgG/IgM/IgA IFA (Vircell) standardised previously by the authors, was used to detect IgG antibodies to phase I and phase II C burnetii (Nine Mile strain). Briefly, feline serum samples (n = 30 from 27 cats) were initially screened at 1/256 dilution using 5% skim milk powder (SMP) in phosphate buffered saline (PBS; Vircell) for C burnetii seropositivity. Two sets of phase I and phase II slide wells in each run contained positive and negative human control solutions from the kit. Diluted feline serum was added to remaining wells and slides were then incubated in a humid chamber at 37°C for 30 min. Following rinsing in PBS and then water, slides were allowed to air dry. Anti-human IgG fluorescein isothiocyanate (FITC) conjugate solution from the kit was applied to positive and negative control wells and anti-feline IgG FITC conjugate solution (VMRD) to the remaining wells. Slides were incubated, washed and dried as above then read under a fluorescent microscope (Olympus BX60F-3; Olympus) at 400× shortly thereafter by two of the authors independently. Samples read as positive on phase I and/or phase II at titre 1/256 were serially diluted to end titre using the same method.

ELISA

A modification of the commercially available Panbio Q fever IgG ELISA (Alere) (previously standardised by the authors) was used to detect IgG antibodies to phase II C burnetii (Henzerling strain) in feline serum samples (n = 30 from 27 cats). With the exception of microwells containing control and calibrator samples, microwells were incubated at room temperature for 1 h with 5% SMP in PBS. Dilution of positive and negative control and calibrator samples to 1/100 was performed using Tris-buffered saline (pH 7.2–7.6) (Alere). To dilute feline serum samples to 1/100, 5% SMP in PBS was used. Positive and negative control samples were run singly, the calibrator sample in triplicate and feline serum samples in duplicate. The plate was covered and incubated for 30 min at 37°C after applying samples to microwells. The microwells were washed six times using PBS containing 0.05% Tween 20. To microwells containing control and calibrator samples, horseradish peroxidase (HRP) conjugated anti-human IgG was applied. The same volume of HRP conjugated anti-feline IgG (Peroxidase-conjugated AffiniPure Goat Anti-Cat IgG; Jackson ImmunoResearch Laboratories) was applied to microwells containing feline samples. The plate was covered, incubated and manually washed as above. Tetramethylbenzidine was applied to each microwell, the plate incubated for 10 min and then the reaction stopped by adding 1 M phosphoric acid. Within 30 min plate microwells were read in a microtitre plate reader at wavelength 450 nm with reference filter 600 nm to determine optical density (OD).

Two methods were used to assess OD results. According to the manufacturer’s instructions, a sample absorbance/calibrator absorbance ratio (index value) was calculated, with calibrator absorbance determined by calculation of the average absorbance of the calibrator in triplicate and multiplication of this by the calibration factor supplied with each kit batch. Positive samples were defined as having an index value >1.1. Those that were negative were <0.9. Results between these values were equivocal, with these samples re-tested and, if still equivocal, considered negative. A S/P% was also calculated using the positive and negative control samples from the kit by the formula:

$$\text{S/P\%}=100\times \left(\text{OD samples}-\text{OD negative control}\right)/\left(\text{OD positive control}-\text{negative control}\right).$$

$$\text{Positive samples had S/P\% greater than 50\%}.$$

Comparisons between IFA and ELISA results

Owing to the lack of a ‘gold standard’, agreement between results from IFA and ELISA (by index value and S/P%) for anti-phase II C burnetii antibodies was assessed by calculating Cohen’s kappa coefficient and its approximate 95% confidence interval (CI). ²⁵ Minitab Version 15 (Minitab) was used for analysis.

Histological examination and FISH

Formalin-fixed, paraffin-embedded 5 μm sections of the index cat’s ovaries, left and right uterine horns, and uterine body were stained using haematoxylin and eosin, Gram Twort and Giemsa. To assess for C burnetii within the index cat’s ovaries, left and right uterine horns, and uterine body, formalin-fixed, paraffin-embedded histological sections (4 µm) from each tissue were mounted on Probe-On Plus slides (Fisher Scientific) and evaluated by FISH using a eubacterial probe (EUB-338; GCTGCCTCCCGTAGGAGT) by a previously described technique. ²⁶ Slides were examined under a fluorescent microscope (Olympus BX60F-3) at 400×. Control samples were formalin-fixed, paraffin-embedded tissues prepared by forming a ‘bacterial sandwich’ using lung tissue. ²⁷ Control bacteria in these preparations were Staphylococcus pseudintermedius and Pseudomonas aeroginosa. Coxiella species controls could not be prepared owing to the lack of physical containment 3 facilities required for handling this bacterial species.

PCR

Frozen samples of the index cat’s ovaries, left and right uterine horns, and uterine body were submitted to the Australian Rickettsial Reference Laboratory, Geelong, Australia, for real-time PCR testing for C burnetii.

Results

Summary of Q fever outbreak in a small animal veterinary hospital

Sydney South West Public Health Unit conducted an epidemiological investigation into the cluster of human Q fever cases at a small animal veterinary hospital. ²⁴ All animals that underwent caesarean section in the identified period of interest were assessed by CFT to detect the presence of C burnetii antibodies, with this group comprising seven dogs and one cat (the index cat). Five of the seven dogs were seronegative, one dog was equivocal (titre 1/8) and one was unable to be tested owing to death from neoplastic-related disease. The index cat was seropositive to C burnetii (titre 1/32).

Nine veterinary personnel belonging to the veterinary hospital were confirmed as cases of Q fever on the basis of positive C burnetii serological and/or PCR results. Of the nine veterinary personnel, eight had worked on the day the caesarean section was performed on the index cat, while the ninth person handled equipment used during the caesarean section the following morning. The degree of clinical disease correlated with known exposures to the index cat and her kittens.

Study population

The index cat was a 4-year-old Burmese queen from an urban breeding cattery with a history prior to 2010 of having produced three litters normally, although some were of small size (one kitten only). The breeding for each of these litters occurred within the cattery. She was bred to an external Burmese male stud cat in April 2010 – the only occasion she had been outside the cattery. At this mating, she did not have contact with any animals outside that cattery. She had no known prior illnesses, and, during the pregnancy in 2010, was in good health. The index cat and cat 23 were confined indoors, while remaining cattery-confined cats had free-roaming outside access. Most cats in the cattery were fed a diet comprised predominantly of raw beef supplemented by commercial pet food. The index cat and cat 23 refused raw beef and were therefore fed commercial pet food and raw chicken necks.

The index cat commenced parturition on 31 May 2010 with a live kitten born without assistance that evening. Prolonged labour without fetal delivery and failure to respond to 2.5 IU oxytocin subcutaneously led to caesarean section the following morning at which the veterinarian assessed the uterus to be normal in view of the duration of dystocia. The placenta was unremarkable. One live and one dead kitten were extracted and received mouth-to-mouth resuscitation; both appeared small for gestational stage. Following caesarean section, the index cat was administered 75 mg amoxicillin subcutaneously and was discharged immediately. Antibacterial therapy continued with 50 mg amoxicillin clavulanate orally twice daily. After identification as the probable primary source of C burnetii in the Q fever outbreak, the index cat was administered 25 mg doxycycline orally twice daily for 14 days commencing 4 weeks after the initial caesarian. The index cat’s ovaries and uterus were grossly normal at ovariohysterectomy 11 weeks after caesarean section. The two live kittens from the gestation developed normally.

Of the 27 cats from the breeding cattery, 11 (41%) were female entire, 11 (41%) were female neutered, four were male entire (15%) and one male was neutered (3.7%). Burmese comprised five (19%) and Cornish Rex 22 (81%) of 27 cats within the cattery population. The median age at initial sampling was 6 years (range 3 months to 17 years). From initial sample collection to May 2012, three of 27 (11%) cats died. Signalment of individual cats and available further information are provided in Table 2.

Table 2.

Signalment and Coxiella burnetii serological results for all cattery-confined cats. The reciprocal antibody titre is provided for positive complement fixation test (CFT) and indirect immunofluorescence assay (IFA) results. Where a second sample was collected, results are given as first sample collection/second sample collection

Cat	Breed	Sex	Age at first sample collection (years)	Further information	CFT	IFA		ELISA
						Phase I	Phase II	(S/P%)	Index value
Index cat	Burmese	FE	4		32	8192	8192	+	+
2	Cornish Rex	FN	8		−/−	1024/512	8192/512	+ /−	+/−
3	Cornish Rex	FN	14		−	−	−	−	+
4	Cornish Rex	FE	5		−	−	−	−	−
5	Cornish Rex	MN	9		−	−	−	−	−
6	Cornish Rex	FE	6	Now deceased — feline infectious peritonitis	−/−	−/−	−/−	−/−	−/−
7	Cornish Rex	FE	6		−	512	512	+	+
8	Cornish Rex	FN	8		−	256	256	+	+
9	Cornish Rex	FE	3	Neutered after two stillborn litters	−/−	−/−	−/−	−/−	−/−
10	Cornish Rex	FE	3		−	−	−	−	+
11	Cornish Rex	FE	5		−	−	−	−	−
12	Cornish Rex	FN	15	Now deceased — congestive heart failure	−	−	256	+	+
13	Cornish Rex	FN	14		−	256	512	−	−
14	Cornish Rex	FN	4	Daughter of cat 17	−	−	−	−	−
15	Cornish Rex	ME	3		−	−	−	−	−
16	Cornish Rex	FE	9		−	−	−	−	−
17	Cornish Rex	FN	6		−	−	−	−	−
18	Cornish Rex	FE	3		−	−	−	−	−
19	Cornish Rex	FN	9		−	−	−	−	+
20	Burmese	ME	7	Source of reproductive failure — small litter size	−	512	1024	+	+
21	Burmese	FN	17		−	−	−	−	+
22	Burmese	FN	14	Now deceased — chronic renal failure	−	−	−	−	+
23	Burmese	FE	3	Daughter of cat 1	−	−	−	−	−
24	Cornish Rex	ME	0.3	Son of cat 6	8	−	−	−	−
25	Cornish Rex	ME	0.3	Son of cat 6	−	−	−	−	−
26	Cornish Rex	FE	0.3	Daughter of cat 6	−	−	−	−	−
27	Cornish Rex	FE	2	Daughter of cat 6	−	−	−	−	−

ELISA = enzyme-linked immunosorbent assay; S/P% = sample/positive ratio; FE = female entire; FN = female neutered; MN = male neutered; ME = male entire

Serological testing

Coxiella burnetii serology results from CFT, IFA and ELISA in individual cattery-confined cats are summarised in Table 2. Of the 27 cats, two (7.4%) were positive using CFT. Using IFA, 7/27 (26%) cats were positive to anti-phase II C burnetii antibodies, of which six (22%) were also positive to anti-phase I antibodies. All cats seropositive using IFA had antibody titres to phase II C burnetii greater than or equal to their antibody titres to phase I C burnetii. Calculating ELISA results by S/P% resulted in 6/27 (22%) cats being positive to anti-phase II C burnetii antibodies; these cats were also all detected as seropositive to phase II C burnetii by IFA. Eleven of 27 (41%) cats were seropositive when ELISA results were calculated using index values. Of five additional cats detected as seropositive based on index values compared to S/P% none were detected as seropositive to phase II C burnetii by IFA.

In three cats from which a second sample was collected, there were no changes to seroreactivity using the serological assays with the exception of ELISA results for cat 2. The anti-phase I and II C burnetii antibody titres of cat 2 using IFA also decreased between the two sampling events (from 1/1024 to 1/512 and 1/8192 to 1/512, respectively).

A comparison between phase II IgG IFA and ELISA results is provided in Table 3. Results for IFA and ELISA S/P% suggest good agreement and, for IFA and ELISA index values, moderate agreement between these tests. ²⁵

Table 3.

Percent agreement and Cohen’s kappa coefficient for phase II IgG indirect immunofluorescence assay (IFA) and enzyme-linked immunosorbent assay (ELISA) results

Tests compared	Percent agreement	95% CI	Cohen’s kappa coefficient	95% CI
IFA and ELISA S/P%	93	78, 99	0.81	0.46, 1.2
IFA and ELISA index value	77	58, 90	0.47	0.12, 0.82

CI = confidence interval; S/P% = sample/positive ratio

Histological and molecular analysis

The ovarian cortical stroma contained moderate numbers of growing follicles. The endometrium was lined by simple cuboidal-columnar epithelium in longitudinal folds, with moderate hyperplasia of endometrial glands that were also tortuous and contained scant luminal secretions. There was moderate myometrial hypertrophy. Throughout the endometrium and myometrium were moderate numbers of macrophages, some of which were multinucleated, containing brown granular pigment (likely haemosiderin). There was marked congestion of vessels in the stratum vasculare of the myometrium and perimetrium. These findings were consistent with a post-partum uterus. ²⁸ Bacteria were not visualised in Giemsa or Gram Twort-stained sections. Staining using the EUB-338 probe did not reveal bacteria. Further staining using a C burnetii-specific probe was therefore not undertaken. PCR analysis of the index cat’s reproductive tissues for C burnetii was negative.

Discussion

This study was a feline-focused investigation of an important Q fever outbreak in a small animal veterinary hospital following a caesarean section on a breeding queen, the index cat. Previous investigations of feline-associated outbreaks have demonstrated serological evidence of C burnetii infection in the cat at their centre, but focused on human perspectives.^9–11,13,14 In this study, the index cat showed marked antibody responses to phase I and phase II C burnetii using all serological assays. Extending the investigation to assess seroprevalence in cats from the same cattery demonstrated past or current C burnetii infection in 26% of these cats using IFA. This evaluation of seroprevalence was enabled by using three standardised serological assays.

Determining C burnetii infection in cats has been complicated by the absence of sensitive, specific diagnostics, particularly as serological tests cannot be translated directly from one species to another without standardisation. In humans, IFA and ELISA are more sensitive than CFT^1,21 (the OIE reference technique) ²² and therefore valuable for serodiagnosis in felines. The IFA and ELISA methodologies used in this study were recently standardised in cats by the authors. Coxiella burnetii seropositivity in the index cat was detected by all three serological assays. A challenge in assessing the remaining cattery-confined cats seropositive by IFA and/or ELISA, but negative by CFT, was the lack of a ‘gold standard’ against which to evaluate their validity. While inter-test agreement suggested IFA and ELISA to be sensitive and specific in the study population,^25,29 definitively establishing these as alternatives to CFT requires evaluation in large, representative feline populations. Further evaluation is also required to determine the preferred method of ELISA analysis as the small sample size in this study was limiting.

Cross reactions between C burnetii and other organisms, including Legionella species and Bartonella species have been described in human sera using certain serological techniques.^1,29–31 Cross reactions involving IgA and IgM have been described between C burnetii and Bartonella henselae and Bartonella quintana in human sera using IFA, likely owing to similarity between protein antigens.^1,30 The extent of serological cross-reactions involving C burnetii is unknown in feline sera; however, as cats are recognised as the primary reservoir hosts for several Bartonella species, including B henselae, ³² potential cross reactions may warrant consideration. In our study, the absence of cross-reactions using IFA between C burnetii and B henselae was able to be confirmed in testing by the manufacturer (Vircell).

In C burnetii-infected humans, elevated phase II IgM antibodies are detected within 14 days of signs appearing, followed later by IgM antibodies to phase I.^1,33,34 From 2–3 weeks, IgG antibodies to phase II reach high levels, with phase I IgG responses developing more slowly and at lower titres.^1,33,34 The index cat demonstrated, approximately 4 weeks post-outbreak, strong positivity to anti-C burnetii IgG antibodies, in accordance with her probable role as the source of C burnetii in the Q fever outbreak. Previously reported antibody titres using IFA in cats linked to community-acquired Q fever outbreaks have varied between 1/32 and 1/8192 to phase I and between 1/8 and 1/8192 to phase II antigens;^10,11,13,14 the index cat’s high antibody titres therefore appear consistent with those from earlier investigations. Variable anti-phase I and phase II antibody titres were detected in the remaining cattery-confined cats up to 22 months post-outbreak. This suggests such antibodies may remain raised for some time after infection in cats, though it is uncertain whether exposure to C burnetii continued in the population. In humans, anti-phase II IgG titres have been identifiable 12 years after a Q fever outbreak, though CFT antibodies generally decrease earlier.^35,36 Cell-mediated immunity, important in controlling C burnetii infection in humans, ¹ was not assessed in the study, likely limiting detection of all infected cats

Previously, seroprevalence studies have indicated stray cats or dogs may have higher rates of C burnetii seropositivity than those that are client-owned, with rodents a possible reservoir of infection.^5,18 In view of the high seroprevalence in the breeding cattery studied, cattery-confined breeding cats may also represent a population at increased risk of C burnetii infection, though this requires further investigation. Likely important risk factors within breeding catteries include more frequent exposure to parturient animals (C burnetii is abundant in the birth products of infected individuals ¹ ), as well as high population densities. The primary source of C burnetii for cats in the breeding cattery at the centre of this outbreak was unable to be established. The exposure of the index cat to the male stud at mating could not be excluded as we were not able to acquire serum samples to assess this cat’s seroreactivity. Alternatively, reactivation of previously latent C burnetii in the next cat may have occurred. The active predatory behaviours of some cats have been proposed to increase their likelihood of infection by C burnetii by close contact with reservoirs such as rodents.^37,38 Contact with rodents or other potential animal sources of C burnetii was unlikely in the index cat, but possible for cattery-confined cats with outside access. Other forms of dietary access, such as ingestion of raw meat, are possible, with ingestion of high doses of C burnetii a recognised, albeit rare, source of infection in humans. ¹

Studies describing well-defined disease associations with feline C burnetii infection are lacking. Experimental infection of cats has caused fever, inappetence and lethargy. ²⁰ Reproductive abnormalities in periparturient cats associated with Q fever outbreaks have also been described, yet such relationships are likely complex. The reported history of some queens and their kittens was unremarkable,^9,10,17 while others have exhibited signs, including bleeding per vaginum for 3 weeks prior to parturition, ¹⁴ stillbirths and death of kittens shortly post-parturiently. ¹³ In a study of Q fever cases connected to exposure to felines, 7/10 queens had one or more stillborn kittens. ¹¹ In other species, manifestations of C burnetii infection during pregnancy encompass abortion, prematurity, stillbirth, low birth weight and neonatal weakness, with these connected to placentitis and/or direct fetal injury.^1,3,39 Signs in the index cat, including stillbirth and small kittens relative to gestational stage, may have been due to C burnetii infection, though other contributing factors cannot be excluded. Of other C burnetii seropositive cattery-confined cats with known disease, it was difficult to determine the contributions of their serological status. For several years, a breeding male (cat 20) was considered a source of reproductive failure that resolved when affected females were bred to a different male. The reproductive failures associated with cat 20 have now resolved. It is questionable whether C burnetii was the underlying cause, particularly given that this cat was C burnetii seropositive. However, interestingly, in humans, C burnetii infection has been associated with orchitis, epididymitis and priapism.^1,36

Duration of C burnetii shedding from the reproductive tract of naturally infected cats is unknown, though shedding in animals is primarily recognised periparturiently. ¹ Coxiella burnetii has been isolated from feline uteruses 3 and 8 weeks post-parturition, but could not be isolated from another uterus 10 weeks post-parturition.^11,13 In the present investigation, inability to detect C burnetii histologically may represent true- or false-negative results. Several factors, particularly the interval between caesarean section and ovariohysterectomy, during which the index cat was administered doxycycline—the antibacterial of choice in empirical treatment of acute Q fever in humans ³⁶ —were likely to limit findings in this case. Although false-negative results can occur using FISH with metabolically inactive bacteria due to low ribosomal RNA content, ⁴⁰ negative PCR results on these tissues increased our suspicion that C burnetii was, indeed, no longer present.

As cats are recognised as C burnetii sources, development of infection control procedures minimising human exposure to this bacterium should be considered within veterinary hospitals. Veterinary infection control guidelines are increasingly addressing this with recommendations to wear appropriate personal protective equipment (PPE) where potential for exposure to infective materials, particularly birth products, exists and avoid direct mouth-to-mouth or mouth-to-nose resuscitation of neonates.^41–43 However, as well as needing to improve and expand these guidelines, adopting such procedures within hospitals will be dependent on adherence to and awareness of written infection control guidelines by veterinary staff, in particular the requirements for PPE.^44,45 In conjunction with infection control procedures, vaccination of all small animal veterinary personnel against Q fever may gain importance as the epidemiology of C burnetii infection in companion animals unfolds.

Future seroprevalence studies will be valuable in determining the importance of C burnetii in feline populations and enabling enhanced understanding of the risks cats pose as sources of C burnetii in Australia. Together with epidemiological data, elucidation of disease associations with C burnetii is required to advance our understanding of the pathogenesis of feline infections.

Conclusions

Serological detection of C burnetii in a breeding cattery linked to a Q fever outbreak amongst veterinary hospital personnel indicates there is likely infection by this bacterium in Australian feline populations. However, many questions remain regarding the epidemiology and disease outcomes following infection in this species, both in Australia and globally. This outbreak’s association with a caesarean section on a parturient queen, together with the elevated C burnetii seroprevalence in the cattery from which this queen arose, re-confirm the relevance of this zoonosis.