Risk factors for urinary tract infection with multiple drug-resistant Escherichia coli in cats

Abstract

The emergence of multiple drug-resistant (MDR) bacteria is a growing public health problem. The objective of this retrospective study was to identify risk factors associated with MDR Escherichia coli infection of the urinary tract in cats. All cats presenting with an E coli urinary infection between March 2010 and December 2012 were included and divided into two groups: an MDR group and a non-MDR group. The effects of different variables on the occurrence of an MDR E coli infection were evaluated: age, sex, additional diseases, number of antibiotics and number of days of hospitalisation. Fifty-two cats were identified (10 MDR and 42 non-MDR). The number of antibiotic groups used within the last 3 months was associated with an increased risk of MDR E coli urinary infection (P = 0.007). The association of the number of days of hospitalisation within the last 3 months and the increased risk of MDR E coli urinary infection did not reach significance (P = 0.090). This study provides evidence that systematic urinary culture with antibiotic sensitivity testing should be recommended when treating urinary tract infections if antibiotics have been prescribed within the past 3 months. Moreover, the selection of MDR bacteria through antibiotic use should be considered as a potential risk associated with treatment.

Introduction

Bacterial urinary infections only represent 2–12 % of urinary tract disease in the cat.^1,2 The most commonly incriminated pathogen in the literature and in our practice is Escherichia coli.^1,3,4 A predisposition has been reported in cats that are older than 10 years.⁴ The presence of a chronic concurrent disease, such as chronic renal failure, hyperthyroidism or diabetes mellitus, predisposes the cat to urinary infection.^4,5 When evaluating cats with urinary tract disease, urinary infection should be differentiated from occult bacteriuria and bacterial colonisation. Occult bacteriuria is very rare in the cat (<0.9%).⁶ Bacterial colonisation is common and associated with long-stay urethral catheters such as those used to relieve urinary obstruction in males.⁷ Urinary infection can be considered to be clinically significant when there are cytological signs of urinary inflammation (pyuria) and a bacteriuria of ≥10³ colony-forming units (CFU)/ml of urine sampled via cystocentesis in an animal without a urinary catheter.⁸

The emergence of multiple drug-resistant (MDR) bacteria is a growing public health problem. The resistance of certain bacterial strains to several classes of antibiotics complicates the treatment of infections. The situation is particularly critical in intensive care units that treat immunodepressed patients undergoing invasive procedures over prolonged periods.⁹ In non-hospitalised humans, a certain number of factors that promote the onset of infection with MDR pathogens has been identified: the age of the patient, hospitalisation within the last year and the administration of antibiotics within the last year.^10,11 In dogs, it was demonstrated that the proportion of E coli isolated from the rectum and resistant to at least one antibiotic increased after more than 3 days of hospitalisation in a university intensive care unit.¹² In the cat, no risk factors have been associated with infection with MDR bacteria.

The objective of our study was to identify the risk factors associated with a MDR E coli infection of the urinary tract in cats. Our hypotheses were that hospitalisation and prior administration of antibiotics were associated with urinary tract infection (UTI) with MDR bacteria, but that age and sex did not have an effect on the presence of a MDR E coli.

Materials and methods

Animals

The medical records of cats diagnosed with an E coli UTI between 1 March 2010 and 31 December 2012 at the Veterinary Hospital Frégis were reviewed. To be included in the study the diagnosis had to be based on the demonstration of pyuria and significant bacteriuria ≥10³ CFU/ml of urine. Samples were systematically taken by cystocentesis. Animals that had been catheterised within 48 h prior to the sample were excluded. For each case, the history, reason for urinary culture and presence of concurrent disease were recorded, as well as the treatment and hospitalisation history. The admission and discharge dates were also recorded. One of the authors directly contacted the referring veterinarian to obtain the list of antibiotics administered to the animal throughout their life. The owners were also contacted directly by telephone to identify any other veterinarians who may have seen the animal to record any other possible hospitalisations and/or antibiotic treatments. Cats with an incomplete medical history were excluded from the study.

The antibiotics used were also recorded, and the number of antibiotic groups (penicillins, cephalosporins, tetracyclines, quinolones, aminoglycosides, sulfonamides, etc) counted.

All cases were divided into two groups: an MDR group and a non-MDR group.

Bacteriological analysis

Urine samples were stored at 4°C and underwent bacterial culture within 6 h of collection. Specimens (10 µl) were plated on MacConkey agar. Inoculated plates were incubated at 37°C (98.6°F) and examined for bacterial growth 24 and 48 h later. Escherichia coli isolates were identified using biochemical tests on pure colonies that were derived from urine. Lactose-fermenting and non-lactose-fermenting isolates were tentatively identified as E coli if they were indole-positive and oxidase-, citrate-, urease- and hydrogen sulphide-negative, and had an acid slope and acid butt in triple sugar iron medium. If reactions to any of these tests were equivocal, an additional panel of tests was performed with a commercial biochemical system to confirm the identity of atypical E coli isolates. The number of CFU/ml of urine was calculated by counting E coli colonies on the plate and multiplying by dilution factor 10². Susceptibility testing of urinary E coli isolates was performed via the disk diffusion method according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI; http://www.clsi.org). The following antimicrobial discs were used: penicillin, amoxicillin, amoxicillin–clavulanate, cephalothin, cefoxitin, cefotaxime, ceftiofur, trimethoprim–sulfamethoxazole, chloramphenicol, gentamicin, doxycycline, enrofloxacin, marbofloxacin and nalidixic acid. Inhibition zone diameters were recorded and interpreted according to the guidelines of the CLSI. Escherichia coli isolates with intermediate resistance to an antimicrobial were classified as resistant, so outcome was binary (resistant or susceptible). Isolates that were resistant to ≥3 antimicrobial classes were considered to be MDR.

Statistical analysis

Statistical analysis was performed using Systat 11. The descriptive statistical analyses were expressed as a mean ± SD. A logistic regression was used to determine the variables associated with an increased risk of developing a UTI associated with a MDR E coli. Sex, age, presence of chronic renal failure, previous history of urethral obstruction, presence of neoplasia, presence of hyperthyroidism, days of hospitalisation (during lifetime, within the last 6 months and within the last 3 months) and number of groups of antibiotics (during lifetime, within the last 6 months and within the last 3 months) were selected as the candidate explanatory variables. The presence of UTI associated with a MDR E coli was selected as the binary-dependent variable. Sex, presence of chronic renal failure, presence of hyperthyroidism and presence of a MDR UTI were determined as categorical variables, with the rest of the variables being numerical variables. A stepwise procedure was used to determine the effect of the potential explanatory variables with a P-value of 0.15 chosen to enter the model and a P-value of 0.3 to leave the model. Final statistical significance was set at P <0.05.

Results

Urinary infection with E coli was diagnosed in 61 cats between 1 March 2010 and 31 December 2012 at the Veterinary Hospital Frégis. Only 52 records were included to ensure completeness of the information concerning the antibiotic treatment and hospitalisations of the animals throughout their lifetime. The mean age was 9.8 ± 5 years. Females represented 40.4% (21/52) and males 59.6% (31/52). Sixty percent (31/52) of the cats were presented for clinical signs related to UTI (pollakiuria, haematuria, etc) and 40% (21/52) were presented for another reason (diabetes mellitus, chronic renal failure, hyperthyroidism, ureteral stent follow-up, etc). Concurrent disease was present in 92.3% of cases (48/52). Chronic renal failure, previous history of urethral obstruction, neoplasia and hyperthyroidism were present in 50% (26/52), 26.9% (14/52), 9.6% (5/52) and 7.7% (4/52) of cases, respectively. The complete list of concurrent pathologies is provided in Table 1. The MDR group contained 10 cats and the non-MDR group contained 42 cats. There was no statistical difference in age or sex between the non-MDR group (9.7 ± 5.1 years, 15 females and 27 males) and the MDR group (10.1 ± 4.7 years, six females and four males). The reason for urinary culture was similar in both groups (7/10 for UTI signs in the MDR group and 24/42 for UTI signs in the non-MDR group). The proportion of animals with chronic renal failure was similar in both groups (22/42 in the non-MDR group and 4/10 in the MDR group), as was the proportion of animals with a previous history of urethral obstruction (10/42 in the non-MDR group and 4/10 in the MDR group), neoplasia (4/42 in the non-MDR group and 1/10 in the MDR group) and hyperthyroidism (3/42 in the non-MDR group and 1/10 in the MDR group) (Table 1). None of the animals had received anti-neoplastic immunosuppressant chemotherapy or long-term steroids. The feline immunodeficiency virus (FIV)/feline leukaemia virus status was known in 10 animals (7/42 in the non-MDR group and 3/10 in the MDR group). Only one of these tested cats (in the MDR group) was positive for FIV.

Table 1

Identified concurrent diseases or previous medical history

Disease	MDR group (n)	Non-MDR group (n)	Percentage of cases (n)
Chronic renal failure	40% (4/10)	52% (22/42)	50% (26/52)
Previous history of urethral obstruction	40% (4/10)	24% (10/42)	27% (14/52)
Neoplasia	10% (1/10)	10% (4/42)	10% (5/52), 3 fibrosarcomas, 1 mammary carcinoma, 1 intestinal adenocarcinoma
Hyperthyroidism	10% (1/10)	7% (3/42)	8% (4/52)
Inflammatory bowel disease	10% (1/10)	7% (3/42)	8% (4/52)
Diabetes mellitus	10% (1/10)	5% (2/42)	6% (3/52)
Pyelic and urethral lithiasis	0% (0/10)	7% (3/42)	6% (3/52)
Chronic rhinitis	10% (1/10)	2% (1/42)	4% (2/52)
Idiopathic hypercalcaemia	0% (0/10)	2% (1/42)	2% (1/52)
Suppurative pericarditis	0% (0/10)	2% (1/42)	2% (1/52)
Cryptococcosis	0% (0/10)	2% (1/42)	2% (1/52)
Feline triad	10% (1/10)	0% (0/42)	2% (1/52)
Chronic pancreatitis	10% (1/10)	0% (0/42)	2% (1/52)

The number of antibiotic groups used within the last 3 months was associated with an increased risk of MDR E coli urinary infection (mean antibiotic groups per cat of 0.17 in the non-MDR group and 1.30 in the MDR group; P = 0.007). The sensitivity of MDR strains of E coli, as well as the list of groups of antibiotics to which the animals had been exposed prior to the isolation, are shown in Table 2. It appears that all of the MDR strains of E coli are resistant to penicillins and tetracyclines. Resistance to cephalosporins, quinolones, sulfonamides and phenicols is also very common (between seven and eight isolates out of 10). Only sensitivity to aminoglycosides remained high, with nine out of 10 strains sensitive. The antibiotics to which the cats were the most exposed were penicillins, cephalosporins and quinolones (Table 3). The antibiotics to which the cats were the least exposed were the aminoglycosides. No link could be demonstrated between exposure to a group of antibiotics and the subsequent isolation of a resistant strain to this same group. Information regarding the dose, duration and treatment compliance were sparse and too incomplete to be used. Forty percent (20/52) of the cats were hospitalised when E coli was isolated. The proportion of animals hospitalised was similar between the two groups (16/42 in the non-MDR group and 4/10 in the MDR group). The association between the number of days of hospitalisation within the last 3 months and the increased risk of MDR E coli urinary infection did not reach significance (4.7 days in the MDR group versus 1.19 days in the non-MDR group; P = 0.090). Of the four hospitalised cats infected with a MDR strain, two had undergone urethral catheterisation to relieve obstructions (cats 1 and 3 in Table 4). The bacteriological examination on admission was negative and the MDR pathogen was isolated on the days 8 and 11 of hospitalisation, respectively (Table 4). These two infections with MDR E coli were therefore definitely contracted at the hospital. The same observation was made in the non-MDR group, with four cats whose infection was probably acquired at the hospital (bacteriological examination on admission negative, catheterisation and infection with E coli between 3 and 6 days thereafter). The proportion of infections acquired at the hospital was similar in the MDR and no-MDR group.

Table 2

Multiple drug-resistant (MDR) Escherichia coli isolate sensitivities and antibiotic groups that each animal had been exposed to prior to E coli isolation

Cat	MDR E coli isolate sensitivities							Antibiotic groups used prior to MDR E coli isolation
	Penicillins	Cephalosporins	Tetracyclines	Quinolones	Aminoglycosides	Sulfonamides-TM	Amphenicols	Antibiotic groups used prior to MDR E coli isolation
1^a	R	S	R	S	S	R	R	Penicillins
2	R	R	R	R	R	R	R	Penicillins/cephalosporins/quinolones/tetracyclines
3^a	R	R	R	R	S	R	R	Penicillins/tetracyclines
4	R	S	R	R	S	R	S	Cephalosporins/tetracyclines/quinolones/metronidazole/macrolides
5^a	R	R	R	R	S	S	S	Cephalosporins/quinolones/sulfonamides-TM/metronidazole/macrolides
6^a	R	R	R	R	S	R	R	Penicillins/quinolones
7^a	R	S	R	S	S	R	R	Penicillins/sulfonamides/metronidazole/macrolides
8^a	R	R	R	R	S	R	R	No antibiotic received prior to E coli isolation
9	R	R	R	S	S	S	R	Penicillins/clindamycin
10^a	R	R	R	R	S	R	R	Aminoglycosides/sulfonamides/metronidazole/macrolides
Total	10/10R	7/10 R and 3/10 S	10/10 R	7/10 R and 3/10 S	1/10 R and 9/10 S	8/10 R and 2/10 S	8/10 R and 2/10 S

TM =trimethoprim; R = resistant; S = sensitive

Cat with symptomatic UTI

Table 3

Total number of cats to receive each group of antibiotic prior to Escherichia coli isolation

	MDR (n = 10)	Non-MDR (n = 42)	Total (n = 52)
Penicillins	5	11	16
Cephalosporins	3	12	15
Quinolones	4	6	10
Metronidazole	4	5	9
Tetracyclines	3	3	6
Macrolides	4	2	6
Sulfonamides	2	2	4
Clindamycin	1	2	3
Aminoglycosides	0	1	1

MDR = multiple drug-resistant

Table 4

Duration of hospitalisation when multiple drug-resistant (MDR) Escherichia coli was isolated, or delay between discharge and MDR E coli isolation for outpatients

	Cat	Number of days of hospitalisation when MDR E coli was isolated	Reason for hospitalisation
Hospitalised patients	1	11	Urethral obstruction
	2	1	Enteritis/pancreatitis complex
	3	8	Urethral obstruction
	4	3	Anorexia
	Cat	Duration of hospitalisation (days)/delay between discharge and MDR E coli isolation (days)	Reason for hospitalisation
Outpatients	5	7/297	Cardiac failure
	6	16/9	Chronic renal failure
	7	6/346	Urethral obstruction
	8	Never hospitalised prior to isolation
	9	6/23	Urethral obstruction
	10	3/575	Complicated diabetes

Discussion

In this study, previous antibiotic therapy (last 3 months) was found to be significantly and independently associated with urinary infection by MDR E coli. This concurs with data from literature on humans.^11,13 This information justifies the use of urinary culture with antibiotic sensitivity testing whenever antibiotics have been prescribed within the last 3 months, thus preventing the use of inappropriate empirical treatment. Resistance to penicillins, cephalosporins, quinolones and tetracyclines is high (Table 2), and these are the families of antibiotics that were used the most commonly in all of the cats in our study (Table 3). One possible explanation for this finding can be seen simplistically as an equation with two main components: the antibiotic that inhibits susceptible organisms and selects the resistant ones. From a Darwinian point of view, antibiotic treatment can be considered as a form of natural selection enabling resistant bacteria to survive and prosper, as their sensitive companions are eliminated. The spread of resistance genes in plasmids bacteriophages, naked DNA or transposons, and the maintenance of these in the gut are also implicated in antibacterial resistance. An exhaustive review of antimicrobial resistance pathogenesis can be found elsewhere.¹⁴ More recently, the notion of a window of selection for mutants has been proposed.¹⁵ The emergence of resistance thus occurs in a specific ‘mutant selection window’ (MSW). The lower margin corresponds to the minimum inhibitory concentration for susceptible bacteria, whereas the upper limit, or mutant prevention concentration, restricts the growth of the entire population, including that of the resistant mutants. Early treatment to avoid an increase in the size of the inoculum, plus a protocol that restricts the time within the MSW (increased dose and/or frequency of administration depending on pharmacokinetic/pharmacodynamic considerations) will reduce the probability of the emergence of resistant mutants.¹⁶ In this study, the antibiotics administered were clearly identified, but data on the dose, frequency and duration of treatment were too incomplete to be analysed. No link could be demonstrated between exposure to a group of antibiotics and the subsequent isolation of a resistant strain to this same family.

The number of antibiotic groups used within the last 6 months was similar in MDR and non-MDR groups. The dilution of resistant populations by ingestion of susceptible organisms over time may be an explanation.

The association between hospitalisation within the last 3 months and urinary infection with MDR E coli did not reach significance (P = 0.09). This variable may have reached the threshold for significance if the sample size had been larger. Further testing is required to enable conclusions. This factor is also well documented in human medicine.¹¹ It results from the association between hospitalisation and exposure of sick patients to bacteria that have been selected from a milieu that is intrinsically subjected to a high selection pressure (intensive use of antibiotics). Furthermore, the hospital is where invasive procedures are performed (intravenous catheter, urinary catheterisation) potentially leading to bacterial contamination. In medical centres, direct and indirect contacts are the main routes for the transmission and dissemination of MDR bacteria.¹⁷ In our study, six cats (2/10 in the MDR group and 4/42 in the non-MDR group) had clearly acquired their infection during hospitalisation, as the bacteriological examination on admission had been negative. These six cats had also undergone procedures to relieve urethral obstruction, increasing the chances of bacterial contamination.

More than 60% of the animals underwent urinary culture for clinical signs related to UTI (pollakiuria, haematuria, dysuria, etc). The remaining 40% did not present signs of UTI, and it was the identification of systemic disease (chronic renal failure, diabetes mellitus, hyperthyroidism, etc), potentially associated with infection, that justified urinary culture. More than 90% of the animals in our study presented with concurrent disease or a previous history of pathology. None of these diseases was identified as being a risk factor for the development of a urinary infection with MDR E coli. Only one cat in the MDR group was seropositive for FIV out of the 10 cats with a known serological status (3/10 in the MDR group and 7/42 in the non-MDR group). The frequency of urinary infections in the context of chronic renal failure has already been reported and is probably related to the isosthenuria and the reduction of local and systemic immune defences.¹⁸ Diabetes mellitus is also a known risk factor for infection.³

According to some publications, urinary infections are more common in females.⁴ In our study, females only represented 40.4% of the animals. This percentage is undoubtedly owing to the significant number of cases of previous urethral obstruction in the males (14/52). If the 14 cases of previous urethral obstruction in the males were excluded, the percentage of females would become 78.9% (30/38), which concurs with published data. This difference is undoubtedly due to anatomical considerations (shorter, wider urethra). Gender was not identified as a risk factor for infection with MDR E coli.

In this study, the distinction between colonisation and infection was well respected. For cases of urethral obstruction, in particular, it was definitely urinary infections and not colonisation, as only animals that had not been catheterised for more than 48 h and presenting pyuria, and a bacteriuria ≥10³ CFU/ml sampled by cystocentesis were included.⁴

The limitations of this study result from its retrospective nature, which could, theoretically, adversely affect the exhaustiveness of the information collected. However, the telephone survey was undertaken studiously by one of the authors, and only cases with reliable and complete information were included. Information regarding the dose, duration of treatment and compliance were not exhaustive enough to be included. The size of the MDR group (n = 10) also limits the conclusions that can be drawn from this study. With more cases, hospitalisation may have been identified as a risk factor for the occurrence of MDR E coli urinary infection.

Conclusions

We demonstrated that antibiotic therapy in the past 3 months is a risk factor for MDR E coli urinary infection. This information leads us to recommend systematic urine culture and antibiotic sensitivity testing in this specific context. Moreover, the selection of MDR bacteria through the use of antibiotics should be considered as a potential risk associated with treatment. It is therefore legitimate to limit antibiotic treatment to situations in which it is absolutely essential.

Footnotes

Acknowledgements

We would particularly like to thank Marion Farbos for all her hard work and research.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflicts of interest

The authors do not have any potential conflicts of interest to declare.

Accepted: 7 August 2013

References

Eggertsdottir

Lund

Krontveit

Sorum

. Bacteriuria in cats with feline lower urinary tract disease: a clinical study of 134 cases in Norway. J Feline Med Surg 2007; 9: 458–465.

Saevik

Trangerud

Ottesen

Sorum

Eggertsdottir

. Causes of lower urinary tract disease in Norwegian cats. J Feline Med Surg 2011; 13: 410–417.

Bailiff

Nelson

Feldman

Westropp

Ling

Jang

, et al. Frequency and risk factors for urinary tract infection in cats with diabetes mellitus. J Vet Intern Med 2006; 20: 850–855.

Labato

. Uncomplicated urinary tract infection. In: Bonagura

Twedt

(ed). Kirk’s current veterinay therapy XIV. St Louis, MO: Saunders, Elsevier, 2009, pp 918–921.

Freitag

Squires

Schmid

Elliott

Rycroft

. Antibiotic sensitivity profiles do not reliably distinguish relapsing or persisting infections from reinfections in cats with chronic renal failure and multiple diagnoses of Escherichia coli urinary tract infection. J Vet Intern Med 2006; 20: 245–249.

Eggertsdottir

Saevik

Halvorsen

Sorum

. Occurrence of occult bacteriuria in healthy cats. J Feline Med Surg 2011; 13: 800–803.

Hugonnard

Chalvet-Monfray

Dernis

Pouzot-Nevoret

Barthelemy

Vialard

, et al. Occurrence of bacteriuria in 18 catheterised cats with obstructive lower urinary tract disease: a pilot study. J Feline Med Surg 2013; 15: 843–848.

Weese

Blondeau

Boothe

Breitschwerdt

Guardabassi

Hillier

, et al. Antimicrobial use guidelines for treatment of urinary tract disease in dogs and cats: antimicrobial guidelines working group of the international society for companion animal infectious diseases. Vet Med Int 2011; 2011: 263768.

Nseir

Di Pompeo

Diarra

Brisson

Tissier

Boulo

, et al. Relationship between immunosuppression and intensive care unit-acquired multidrug-resistant bacteria: a case-control study. Crit Care Med 2007; 35: 1318–1323.

10.

Sahm

Thornsberry

Mayfield

Jones

Karlowsky

. Multidrug-resistant urinary tract isolates of Escherichia coli: prevalence and patient demographics in the United States in 2000. Antimicrob Agents Chemother 2001; 45: 1402–1406.

11.

Cardoso

Ribeiro

Aragao

Costa-Pereira

Sarmento

. Additional risk factors for infection by multidrug-resistant pathogens in healthcare-associated infection: a large cohort study. BMC Infect Dis 2012; 12: 375.

12.

Ogeer-Gyles

Mathews

Sears

Prescott

Weese

Boerlin

. Development of antimicrobial drug resistance in rectal Escherichia coli isolates from dogs hospitalized in an intensive care unit. J Am Vet Med Assoc 2006; 229: 694–699.

13.

Wright

Wrenn

Haynes

Haas

. Prevalence and risk factors for multidrug resistant uropathogens in ED patients. Am J Emerg Med 2000; 18: 143–146.

14.

Levy

Marshall

. Antibacterial resistance worldwide: causes, challenges and responses. Nat Med 2004; 10 (12 Suppl): S122–129.

15.

Blondeau

. New concepts in antimicrobial susceptibility testing: the mutant prevention concentration and mutant selection window approach. Vet Dermatol 2009; 20: 383–396.

16.

Canton

Morosini

. Emergence and spread of antibiotic resistance following exposure to antibiotics. FEMS Microbiol Rev 2011; 35: 977–991.

17.

Strausbaugh

Siegel

Weinstein

. Preventing transmission of multidrug-resistant bacteria in health care settings: a tale of 2 guidelines. Clin Infect Dis 2006; 42: 828–835.

18.

Lees

. Epidemiology of naturally occurring feline bacterial urinary tract infections. Vet Clin North Am Small Anim Pract 1984; 14: 471–479.