Veterinary Microbiology 146 (2010) 309–313

Topical treatment of Pseudomonas aeruginosa otitis of dogs with abacteriophage mixture: A before/after clinical trial

Catherine Hawkins a, David Harper a, David Burch b, Erik Anggard a, James Soothill c,*a Biocontrol Ltd., Suites 217-218, BioCity Nottingham, Pennyfoot Street, Nottingham, NG1 1GF, UKb Octagon Services Ltd., The Round House, The Friary, Old Windsor, Berkshire, SL4 2NR, UKc Level 4 CB Laboratories, Great Ormond Street Hospital, London, WC1N 3JH, UK

A R T I C L E I N F O

Article history:

Received 29 January 2010

Received in revised form 30 April 2010

Accepted 5 May 2010

Keywords:

Dog

Otitis

Bacteriophage

Pseudomonas

A B S T R A C T

In an evaluation of a bacteriophage treatment for infection, ten dogs were included with

chronic Pseudomonas aeruginosa otitis. Each received, directly into the auditory canal of

one ear, a single dose of a topical preparation containing approximately 1� 105 plaque

forming units (PFU) of each of 6 bacteriophage strains, active against P. aeruginosa. At the

time of bacteriophage administration and 48 h later each dog’s core temperature was

taken, its ear was assigned a clinical score (higher = worse condition) and aural swabs were

taken for bacteriophage and P. aeruginosa counts. Forty eight hours after treatment the

clinical score and P. aeruginosa count of all ears had fallen (mean score fall: 30.1%, range

7.7–56.3%, p< 0.0001; mean count fall: 67%, range 29.4–96.8%, p< 0.001). The

bacteriophage counts had risen from the administered dose (mean rise: 99.1-fold, range

2.8–433.3-fold). No treatment related inflammation or other adverse events were detected

during the trial period. This is the first report of a veterinary clinical trial of a bacteriophage

treatment of infection. The results show that administration of this topical bacteriophage

mixture leads to lysis of P. aeruginosa in the ear without apparent toxicity and that it has

potential to be a convenient and effective treatment for P. aeruginosa otitis in dogs.

� 2010 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Veterinary Microbiology

journa l homepage: www.e lsev ier .com/ locate /vetmic

1. Introduction

Pseudomonas aeruginosa otitis of dogs is often difficultto clear. P. aeruginosa is usually resistant to mostcommonly used antibiotics and readily acquires newresistance (Martin-Barrasa et al., 2000; Brothers et al.,2002; Teresa-Tejedor et al., 2003; Morris, 2004), but evenwhen agents to which it is sensitive are used, infectionsoften fail to clear. Treatment involves multiple doses oftopical antibiotics and because of resistance and refractorydisease some vets resort to the use of parenteralantibiotics, which is very inconvenient as multiple visitsmay be needed. Surgery is sometimes required involving

* Corresponding author. Tel.: +44 207 405 9200x5237;

fax: +44 207 813 8268.

E-mail address: soothj@gosh.nhs.uk (J. Soothill).

0378-1135/$ – see front matter � 2010 Elsevier B.V. All rights reserved.

doi:10.1016/j.vetmic.2010.05.014

partial or full resection of the ear canal. There is a need fornew more effective and more convenient medical treat-ments.

Bacteriophages are viruses that infect and can killbacteria but not human or animal cells. The potential ofbacteriophage therapy for treatment of infection has beenestablished in animal models (Dubos et al., 1943; Smithand Huggins, 1982; Smith et al., 1983, 1987; Barrow et al.,1998; Cerveny et al., 2002; Matsuzaki et al., 2003; Huffet al., 2004; Wills et al., 2005). Bacteriophage protectedmice (Soothill, 1992) against systemic infections by P.

aeruginosa and guinea pigs (Soothill, 1994) against localinfections by P. aeruginosa. However, there is littlepublished evidence from experimental systems to supportthe notion that bacteriophages may be used to cureinfections established for longer than a few hours. Twosingle case studies of bacteriophage treatment of P.

aeruginosa infection (Sivera-Marza et al., 2006), in a dog

C. Hawkins et al. / Veterinary Microbiology 146 (2010) 309–313310

with otitis and human burn patient demonstratedbacteriophage multiplication in each. This work underliesthe investigation reported here; indeed the bacteriophagedescribed by Sivera-Marza et al. (2006) is one of the sixbacteriophages used this study. Clinical trial reports ofphage therapy are limited. A trial of bacteriophage in thetreatment of P. aeruginosa otitis of humans has givenencouraging results (Wright et al., 2009), but we know ofno previous reports of veterinary clinical trial of phagetherapy.

In light of the intractable nature of many cases of canineotitis involving P. aeruginosa, we describe a pilot studyexploring the efficacy and safety of topical bacteriophagetherapy for this condition. Although bacteriophagesusually are active against a limited range of strains of agiven bacterial species, the mixture used in this studycontains 6 bacteriophages that have been selected to givebroad activity against a wide range of P. aeruginosa strainswhich have caused canine otitis.

In a previous study with a single bacteriophage strain inone dog (Sivera-Marza et al., 2006), substantial clinicalimprovement and bacteriophage multiplication were seenin an ear 48 h after treatment. In light of this, we havecompared before, and 48 h after bacteriophage treatment,the clinical condition and P. aeruginosa counts from theears of dogs. At this stage we expected a fall in P. aeruginosa

counts rather than complete P. aeruginosa clearance butthe approach enabled us to seek evidence for therapeuticbacteriophage multiplication: for this a low dose of phagewas needed. It has been suggested that endotoxin releaseresulting from bacteriophage lysis might lead to sideeffects (Dixon, 2004). Were this the case, effects might beexpected 48 h after treatment, so this is also a suitable timeto seek any signs of toxicity.

Some of these data were included in a published shortletter (Soothill et al., 2004).

2. Materials and methods

2.1. Bacterial strains

The P. aeruginosa strains used for the propagation andcounting of the bacteriophages were human isolates kindlyprovided by Health Protection Agency, London, UK andcanine isolates kindly provided by the Royal VeterinaryCollege, London; Idexx Laboratories Ltd., Wetherby, UK;Axiom Veterinary Laboratories Ltd., Devon, UK and theVeterinary Laboratories Agency, UK. Bacterial strains wereidentified as P. aeruginosa by colonial morphology oxidasetesting and by API 20 NE identification kit (Biomerieux).

2.2. Bacteriophage preparation

The bacteriophage preparation was stored frozen at�70 8C and consisted of 0.2 ml doses each containing1� 105 plaque forming units (PFU) of each of the sixtherapeutic bacteriophages (designated BC-BP-01–BC-BP-06) in 10% (v/v) glycerol/phosphate buffered saline (PBS).All phages had icosahedral heads and tails. Together thesephages have in vitro activity against 90% of P. aeruginosa

strains from dog ear infections.

2.3. Study design

This was a before/after study in companion dogs withnaturally occurring, chronic, antibiotic-resistant infections.Informed, written consent was obtained from each dog’sowner prior to enrolment using approved documentation.

2.4. Study population

Dogs were eligible for the trial if they had suffered fromotitis of at least 3 months duration, which had beenrefractory to antibiotic therapy (at least 3 courses) andrepeated flushing. There were no age, breed or sexrequirements. Subjects were eligible only if P. aeruginosa,sensitive to at least one of the bacteriophages in themixture was the predominant bacterial isolate andhaemolytic Streptococci of groups A, B, C and G had notbeen detected from ear swabs. It was a requirement thatthe dog had not received any topical or systemic antibiotictherapy within a week before bacteriophage administra-tion.

2.5. Clinical assessment

Immediately before treatment, and 48 h after it, thecondition of the affected ear was assessed and scored forthe following clinical signs:

Occlusion of ear canal (0–3); 0 = normal, 1 = occluded(but possible to insert a 6 mm otoscope nozzle into thevertical ear canal, 2 = occluded (but possible to insert a4 mm nozzle), 3 = occluded (not possible to insert a4 mm nozzle).Erythema (0–3); 0 = normal, 1 = mild, 2 = moderate,3 = severe.Discharge quantity (0–3); 0 = absent, 1 = slight,2 = moderate, 3 = profuse. Discharge type (0–3); 0 = ab-sent, 1 = waxy, 2 = exudative, 3 = purulent. Lesions/ulceration (0–3); 0 = absent, 1 = isolated, 2 = numerous,3 = confluent. Odour (0–3); 0 = absent, 1 = mild,2 = moderate, 3 = intense.

The scores for each clinical indicator were addedtogether to give a total score (out of 18) for the ear. Theear was re-assessed 48 h post-therapy.

Core temperature recordings were taken on day 1 andday 3.

2.6. Treatment

On study day 1, a vial of the bacteriophage mixture wasremoved from �70 8C storage and allowed to reach roomtemperature. A single 0.2 ml dose was then instilleddirectly into the external auditory canal using a sterile1 ml syringe and the ear canal was massaged to promotedeep penetration.

2.7. Microbiology

Pre-weighed swabs were used to sample detritus in theears immediately before and 48 h after treatment. Sam-

C. Hawkins et al. / Veterinary Microbiology 146 (2010) 309–313 311

pling did not remove sufficient material to be consideredpart of the treatment. Swabs were re-weighed aftersampling and the (known mass of) detritus was suspendedin PBS. Counts of bacteriophage and P. aeruginosa in thedetritus suspensions were then performed, in triplicate, onthe same day as sample collection. P. aeruginosa counts(colony forming units [CFU]/g detritus) before and aftertherapy were determined by plating 0.1 ml volumes ofserial ten fold dilutions on Pseudomonas-selective agar(cetrimide agar, Oxoid) and counting colonies after 24 hincubation at 37 8C. Bacteriophages were counted by theagar overlay technique as described in Adams (1959) using0.5% agar (Oxoid) for the overlay on nutrient agar plates(Oxoid); 0.1 ml portions of the serial dilutions of detritusspecimens were added to the agar overlays that alsocontained P. aeruginosa. Each dilution was added to six setsof plates with each set (of 3) containing a different strain ofP. aeruginosa. The P. aeruginosa strains for bacteriophagecounting had been selected so that each would support thegrowth of one only of the bacteriophages in the mixture.

2.8. Statistical methods

Changes in clinical scores and in bacterial countsbetween day 1 and day 3 were compared using the pairedt-test.

3. Results

3.1. Patient characteristics

Initially, thirteen dogs were enrolled in the study. Basedon experience with the first dog the clinical scoring systemfor subsequent dogs was changed, so data from this dogwas excluded. A pre-set criterion for inclusion of a dog inthe study was the non-detection of b-haemolytic strepto-cocci in ear swab cultures. Ear swabs taken at the time oftreatment of each of two dogs grew group G haemolyticStreptococcus that had not been isolated from previousswabs, so data from these dogs were excluded from theanalysis. The remaining ten dogs included in the studywere of eight breeds including two Basset hounds, twoSpinones and two Spaniels and their ages ranged from 3 to

Table 1

Change in indicators from before to 48 h after bacteriophage treatment.

Clinical condition

aReduction in

clinical score

Reduction in clinical score (

(score on day 1

considered as 100%)

Dog 1 4 30.8

Dog 2 5 50

Dog 3 6 54.5

Dog 4 4.5 56.3

Dog 5 3 21.4

Dog 6 1.5 12.5

Dog 7 3 24

Dog 8 1 7.7

Dog 9 (L ear) 2 28.6

Dog 10 (R ear) 2.5 17.9

Mean (all dogs) 3.3 30.1a Out of a total of 18, based on 5 clinical indicators where higher scores ind

13 years. Two dogs had bilateral otitis and in the others itwas unilateral. Atopy had been diagnosed in three of thedogs, two of which (and one other dog) were known tohave ruptured tympanic membranes. All P. aeruginosa

isolates were at least partially resistant to two or more ofthe anti-pseudomonal agents tested. From the ears of fourdogs Staphylococcus spp were also isolated and from one,Enterobacteriaceae.

3.2. Clinical outcome

Clinical scores (out of a maximum of 18) were based onfive clinical indicators. Reduction in score denotesimprovement in the ear’s condition. Scores fell in all tendogs 48 h after treatment, compared to baseline measure-ments (Table 1), with a mean fall in score of 3.3(p< 0.0001; 95% C.I. 4.4, 2.1). The mean percent reductionin clinical score was 30.1%. Re-growth of the tympanicmembrane was noted in one dog. All individual clinicalscores either fell or remained unaltered. Clinical scores foreach of erythema and odour improved in 7 dogs in eachcase. Scores for each of discharge type and ulceration/lesions fell in 5 dogs in each case. The occlusion score fell inone dog only. The discharge quantity score fell in 6 dogs. Allcore temperatures measured on the treated dogs were inthe normal range of 37.9–39.9 8C (Wolfensohn and Lloyd,2003). Those on day 1 ranged from 38.2 to 39.2 8C (mean38.6 8C), while those on day 3 ranged from 38.0 to 39.4 8C(mean 38.6 8). In the bilateral cases, only one ear (selectedat random) was treated during the study period; theresults were treated in the same manner as the unilateralcases. The second ears were also treated with bacterio-phage and as with the first ears the second treated earsimproved similarly after 48 h (data not shown).

3.3. Bacterial and bacteriophage counts

The P. aeruginosa counts in ear swabs before therapyranged from 1.5� 107 to 1� 1010 cfu/g detritus, with amean of 2.9� 109 cfu/g detritus. The counts fell in all cases48 h after therapy (range 3.5� 106 to 5.4� 109, mean1.1� 109 cfu/g detritus). The mean percentage reduction incount for the ears was 67% (p< 0.0001; 95% C.I. 52%, 82%).

P. aeruginosa Bacteriophage

%) Reduction in P. aeruginosa

count (%) (count on day 1

considered as 100%)

pfu/swab (6� 105 pfu

administered on day 1)

74 8� 107

58.6 4.7� 106

75.8 1.4� 108

46 1.7� 106

54 5.5� 106

63.6 2.6� 108

89 1� 107

29.4 4.5� 106

82.5 7.9� 107

96.8 9.4� 106

67 5.9� 107

icate worse condition.

C. Hawkins et al. / Veterinary Microbiology 146 (2010) 309–313312

No distinct plaques were detected in the bacteriophagecount plates from any ear before bacteriophage wasadministered. Counts of bacteriophage/swab 48 h aftertreatment (mean 5.9� 107 pfu/swab, range 1.7� 106 to2.6� 108 pfu/swab) were all higher than those adminis-tered (6� 105 pfu), in four cases more than 100-foldhigher. In six cases there was evidence that more than one(and in 2 cases, 5) of the individual bacteriophage strainsadministered had multiplied.

3.4. Long term follow up (18 months after treatment)

Of three dogs that were treated but excluded, one thathad had Group G Streptococcus isolated from the earcontinued to have relapses of infection but no P. aeruginosa

was isolated after the bacteriophage treatment. The otherdog with Group G Streptococcus infection developed avulvo-vaginitis from which P. aeruginosa (not a wellrecognised cause) was cultured, but also group Ghaemolytic Streptococcus, which has caused outbreaks ofsuch infection (Hare and Fry, 1938). The dog, a 15-year-old,diabetic Lhasa Apso with Cushing’s disease was eutha-nased. Its companion, also a Lhasa Apso, developedvagintis but recovered. The other excluded dog is generallywell but still has an ear infection despite surgery 1 yearafter bacteriophage therapy.

Follow up of the ten included dogs was incomplete butgenerally positive; in three, the chronic ear infectionsappear to have resolved following bacteriophage therapy.In a further three the P. aeruginosa component of the earinfection appears to have resolved. Two of the dogs havebeen euthanased (one for heart failure aged 15, the otherfor defecating in the house, a problem that pre-datedbacteriophage therapy). No side effects of the bacterio-phage therapy have been noted.

4. Discussion

This is the first report of a veterinary clinical trial of abacteriophage treatment, and was carried out followingthe EMEA’s ‘Guideline on Good Clinical Practices’ andtherefore to EU regulatory standards (CVMP/VICH/595/98-FINAL, London, 4th July 2000). Following bacterio-phage treatment, in all dogs a rise in bacteriophage countscoincided with clinical improvement and falling bacterialcounts. Whilst the bacterial counts were made in thepresence of phage, which might be expected to lowerthem, studies have shown (Soothill, 1993) that becausedilution limits phage action, counts are not influencedunless the phage count is much higher than the bacterialcount, which was not the case in this study. Whilst in abefore/after study of this sort it is possible that at leastsome of the changes seen were due to observer bias, theresults strongly support the view that bacteriophage wasmultiplying, killing bacteria and improving the conditionsof the ears. There was no evidence that dogs infected withP. aeruginosa strains that were susceptible to manybacteriophage strains fared better than those with P.

aeruginosa susceptible to few of the bacteriophage strains.Since the ears improved following bacteriophage admin-istration, and erythema scores either fell or remained the

same, there was no evidence of local toxicity or allergy as aresult of the bacteriophage itself or from lysis of bacteria;all dogs remained afebrile. The results support the viewthat this bacteriophage mixture may well have a role inthe treatment of P. aeruginosa otitis, and indicate the needfor controlled studies. Compared to antibiotics, bacter-iophages have narrow host ranges, most being activeagainst a limited range of strains within a bacterialspecies. This means they are less likely to perturb thenormal flora, but mixtures are needed to ensure that asufficiently wide range of bacterial strains are covered.Use of mixed bacteriophages with distinct but over-lapping specificity profiles has a further advantage as itlessens the likelihood of the development of bacterio-phage resistant strains which could evolve duringveterinary or medical use and conceivably pass fromtreated animals to man. Mixtures of bacteriophages couldalso be developed for other pathogenic species, but P.

aeruginosa is a key target because of its increasingantibiotic resistance. Antibiotic resistance will hastenthe development of anti-pseudomonal phage treatments,which are very convenient compared with standardtopical agents and compliance is more certain; since thedose multiplies, only one is needed.

Conflicts of interest

E.A. is a director of Biocontrol Ltd., C.H. is a consultant toand D.H. is an employee and director of Biocontrol Ltd. C.H.,J.S., E.A. and D.H. hold shares in this company. J.S. willreceive royalty payments from Biocontrol Ltd.

Funding source

Employees of Biocontrol played key roles in the studydesign, collection and analysis of data, and in the writing ofthe manuscript and also the decision to submit it forpublication.

Acknowledgements

We thank the veterinarians who helped with this trial,the dog owners for allowing their dogs to participate in it,and Samantha Jayasekera for her help with References.

Biocontrol Ltd and SMART grant LOT/031/254, awardedto BioPhage Ltd., supported this work.

References

Adams, M.H., 1959. Bacteriophages. Interscience Publishers Inc., NewYork, pp. 447–448.

Barrow, P., Lovell, M., Berchieri Jr., A., 1998. Use of lytic bacteriophage forcontrol of experimental Escherichia coli septicaemia and meningitis inchickens and calves. Clinical and Diagnostic Laboratory Immunology5, 294–298.

Brothers, A.M., Gibbs, P.S., Wooley, R.E., 2002. Development of resistantbacteria isolated from dogs with otitis externa or urinary tractinfections after exposure to enrofloxacin in vitro. Veterinary Thera-peutics: Research in Applied Veterinary Medicine 3, 493–500.

Cerveny, K.E., De Paola, A., Duckworth, D.H., Gulig, P.A., 2002. Phagetherapy of local and systemic disease caused by Vibrio vulnificus iniron-dextran-treated mice. Infection and Immunity 70, 6251–6262.

Dixon, B., 2004. New dawn for phage therapy. The Lancet InfectiousDiseases 4, 186.

C. Hawkins et al. / Veterinary Microbiology 146 (2010) 309–313 313

Dubos, R.J., Straus, J.H., Pierce, C., 1943. The multiplication of bacterio-phage in vivo and its protective effect against an experimental infec-tion with Shigella dysenteriae. Journal of Experimental Medicine 78,161–168.

Hare, R., Fry, R.M., 1938. Preliminary observations of an infection of dogsby beta haemolytic Streptococci. Veterinary Record 50, 213–218.

Huff, W.E., Huff, G.R., Rath, N.C., Balog, J.M., Donoghue, A.M., 2004.Therapeutic efficacy of bacteriophage and Baytril (enrofloxacin) in-dividually and in combination to treat colibacillosis in broilers.Poultry Science 83, 1944–1947.

Martin-Barrasa, J.L., Lupiola-Gomez, P., Gonzalez-Lama, Z., Tejedor-Junco,M.T., 2000. Antibacterial susceptibility patterns of Pseudomonasstrains isolated from chronic canine otitis externa. Journal of Veteri-nary Medicine B: Infectious Diseases and Veterinary Public Health 47,191–196.

Matsuzaki, S., Yasuda, M., Nishikawa, H., Kuroda, M., Ujihara, T., Shuin, T.,Shen, Y., Jin, Z., Fujimoto, S., Nasimuzzaman, M.D., Wakiguchi, H.,Sugihara, S., Sugiura, T., Koda, S., Muraoka, A., Imai, S., 2003. Experi-mental protection of mice against lethal Staphylococcus aureus infec-tion by novel bacteriophage phi MR11. Journal of Infectious Diseases187, 613–624.

Morris, D.O., 2004. Medical therapy of otitis externa and otitis media.Veterinary Clinics of North America: Small Animal Practice 34, 541–555.

Sivera-Marza, J.A., Soothill, J.S., Boydell, P., Collyns, T.A., 2006.Multiplication of therapeutically administered bacteriophagesin Pseudomonas aeruginosa infected patients. Burns 32, 644–646.

Smith, H.W., Huggins, R.B., 1982. Successful treatment of experimental E.coli infections in mice using phage: its general superiority overantibiotics. Journal of General Microbiology 128, 307–318.

Smith, H.W., Huggins, R.B., Shaw, K.M., 1983. Effectiveness of phages intreating experimental E.coli diarrhoea in calves, piglets and lambs.Journal of General Microbiology 129, 2659–2675.

Smith, H.W., Huggins, R.B., Shaw, K.M., 1987. The control of experimentalE. coli diarrhoea in calves by means of bacteriophage. Journal ofGeneral Microbiology 133, 1111–1126.

Soothill, J.S., Hawkins, C., Anggard, E., Harper, D., 2004. Therapeutic use ofbacteriophages. Lancet Infectious Diseases 4, 544–545.

Soothill, J.S., 1992. Treatment of experimental infections of mice withbacteriophages. Journal of Medical Microbiology 37, 258–261.

Soothill, J.S., 1993. Use of bacteriophage in the treatment of infections. MDthesis. University of London.

Soothill, J.S., 1994. Bacteriophage prevents destruction of skin grafts byPseudomonas aeruginosa. Burns 20, 209–211.

Teresa-Tejedor, M., Martin, J.L., Navia, M., Freixes, J., Vila, J., 2003. Mechan-isms of fluoroquinolone resistance in Pseudomonas aeruginosa isolatesfrom canine infections. Veterinary Microbiology 30, 295–301.

Wills, Q.F., Kerrigan, C., Soothill, J.S., 2005. Experimental bacteriophageprotection against Staphylococcus aureus abscesses in a rabbit model.Antimicrobial Agents and Chemotherapy 49, 1220–1221.

Wolfensohn, S., Lloyd, M., 2003. Handbook of Laboratory Animals, 3rdedition. Backwell Publishing, Oxford, UK.

Wright, A., Hawkins, C.H., Anggard, E.E., Harper, D.R., 2009. A controlledclinical trial of a therapeutic bacteriophage preparation in chronicotitis due to antibiotic-resistant Pseudomonas aeruginosa. ClinicalOtolaryngology 34, 349–357.