Phage Therapy for a Refractory and Multidrug-resistant Equine Surgical Site Infection Authors: Donald Alverson, Greta Gustafson, Katja F. Duesterdiec-Zelmer, J. Lacy Kamm, Lia Danelishvili Published: 2026 ABSTRACT A 14-year-old American Quarter Horse mare developed a refractory deep surgical site infection following emergency celiotomy for acute right dorsal large colon displacement with 180° volvulus. Initial cultures revealed a high-burden polymicrobial infection with Escherichia coli and Enterococcus faecium, associated with delayed healing, and later followed by multidrug-resistant Pseudomonas aeruginosa, further limiting therapeutic options. Phage therapy was initiated using a customized lytic phage cocktail administered topically on a daily basis over a 35-day period, with cocktail composition guided by daily wound culture results. Phage therapy markedly reduced colony-forming units of targeted pathogens and promoted progressive healing. Wound microbial communities varied during treatment, demonstrating selective pathogen control while preventing residual populations from recolonizing the tissue. By the time of hospital discharge, the wound surface area had decreased by more than 85%, and the body wall had healed without evidence of acute herniation. This case highlights the feasibility, safety, and clinical potential of adaptive phage therapy for managing refractory multidrug-resistant equine wound infections. Introduction Colic, a general term for any condition resulting in abdominal pain, is one of the most common complaints seen by equine veterinarians. Up to 10% of colic cases do not respond to medical management and require surgical intervention.[1] Surgical site infection (SSI) is one of the most frequently observed postoperative complications of equine patients, occurring in 27%–40.4% of cases following a single celiotomy.[23] Horses appear to have a higher incidence of SSI when compared to cattle (12.8%)[4] or canine and feline patients (7%).[5] Repeated celiotomy has a more than two-fold increased risk of SSI, with rates as high as 76%. SSI increases the risk of abdominal herniation by 17.8 times and is associated with a poorer overall prognosis and return to athleticism.[6] To reduce the incidence of SSIs, perioperative antimicrobial prophylaxis is the standard recommendation for an exploratory celiotomy, which is considered an emergency “clean contaminated” procedure.[7] The most frequently used antimicrobials selected in surgeries are potassium penicillin and gentamicin,[8] with the latter being more frequently used by specialists in high-risk surgeries that involve enterotomy.[9] In a study of 287 horses undergoing celiotomy, 25.4% developed an SSI, and 92% of the isolated organisms inciting the infection were resistant to penicillin, while 18% were resistant to gentamicin.[10] There has been no observed benefit to continuing this antimicrobial prophylaxis for more than 72 h postsurgery when compared to horses receiving it for 120 h for the prevention of infection.[11] Bacteriophages (phages), or “bacteria eaters,” are ubiquitous viruses of prokaryotes that selectively infect and kill bacteria during their replication cycle. Phages have a narrow host range, often targeting only certain strains of a particular species of bacteria.[12] In recent years, there has been a renewed effort to characterize and use phages as therapeutic agents to combat drug resistance and refractory bacterial infections, especially in the face of antimicrobial resistance.[13] With antimicrobial-resistant infections expected to claim nearly 2 million lives annually by 2050,[14] the development and rigorous evaluation of alternative therapies, including phage treatment, are urgently needed to mitigate this global threat. Although there is renewed interest in phage therapy, it remains relatively rare in veterinary medicine. In horses, only one known double-blind, placebo-controlled clinical trial was conducted, which treated superficial pyoderma with a Staphylococcus aureus phage cocktail. This study showed no significant difference between treatments and controls, but it is important to note that only 27% of the horses demonstrated a positive S. aureus culture before treatment.[15] Several studies indicate the tremendous potential of phage treatments for animals, including treating urinary infections in dogs and cats[16] and treating mastitis in ruminants.[17] A clinical trial using a phage cocktail treatment for canine otitis caused by Pseudomonas aeruginosa demonstrated high efficacy and was briefly marketed as a commercial product in Europe in 2010, but was removed it same year after manufacturing challenges.[181920] Here, we demonstrate the complexity of wound infections, where interactions between pathogenic and commensal organisms influence disease progression. Accurate identification of the primary pathogens is critical to guide targeted phage therapy, prevent pathogen colonization, and, in cooperation with the innate immune system, achieve successful healing outcomes. We report the use of customized phage cocktails to treat a challenging coinfection with Escherichia coli and Enterococcus faecium, a pattern associated with increased virulence, biofilm formation,[21] and delayed wound healing,[22] complicated by the emergence of multidrug-resistant P. aeruginosa in an equine SSI following emergency celiotomy. This case highlights the potential of phage therapy to manage complex, refractory coinfections and underscores the importance of adaptive, pathogen-specific approaches in promoting effective wound healing, while supporting antimicrobial stewardship. Case Report A 14-year-old American Quarter Horse mare (538 kg) was referred to the Oregon State University Veterinary Teaching Hospital, for acute colic. Rectal palpation showed a mildly distended, fluid-filled cecum, and ultrasonography suggested two populations of small intestine. Abdominocentesis produced peritoneal fluid of normal color, with an elevated abdominal lactate indicating possible intestinal compromise. Due to a lack of improvement following medical management, including intravenous fluids (lactated Ringer’s solution [LRS]: 10 L bolus followed by 1 L/h), approximately 1.5 L of mineral oil administered via nasogastric tube, and analgesia with flunixin meglumine (1.1 mg/kg IV, q12 h), an exploratory celiotomy was performed. Preoperative antibiotic prophylaxis consisted of gentamicin (6.6 mg/kg IV, q24 h) and potassium penicillin G (22,000 IU/kg IV, q6 h), which was continued for 4 days postoperatively. The mare was also managed with flunixin meglumine (1.1 mg/kg IV, q12 h), intravenous fluid therapy (Lactated Ringer’s Solution (LRS): 1.5 L/h), analgesia and antiendotoxic therapy (2% lidocaine, 3 mg/kg/h Constant Rate Infusion (CRI), cryotherapy due to laminitis history (ice boots applied to the front limbs), and pentoxifylline (7.5 mg/kg by mouth, q12 h). Surgery revealed a right dorsal displacement of the large colon with a 180° volvulus, right dorsal colon impaction, and secondary distension of the small intestine and cecum. A pelvic flexure enterotomy was performed to remove the impacted ingesta, and the colon was successfully repositioned to its normal anatomic location. Postoperatively, the mare continued to receive antimicrobial prophylaxis with the previously mentioned antibiotics and dosages, intravenous fluid therapy (LRS: 2 L/h with 0.18 mEq/kg KCl), flunixin meglumine (1.1 mg/kg IV, q12 h), analgesia and antiendotoxic therapy (2% lidocaine, 3 mg/kg/h CRI), and pentoxifylline (7.5 mg/kg PO, q12 h) as part of the standard supportive care. Four days postoperatively, the mare developed mild ventral abdominal incisional discharge, and ultrasonography revealed a localized fluid pocket, indicating an early SSI. Figure 1 presents the integrated diagnostic and treatment timeline for the SSI. The wound samples submitted to the veterinary diagnostic bacteriology laboratory (OVDL) initially revealed a polymicrobial infection with lactose-fermenting E. coli [3+, penicillin-resistant, gentamicin-susceptible; Table 1] and E. faecium [3+, penicillin-susceptible, gentamicin-resistant; Table 1]. Despite 17 days of intensive management, including drainage of the fluid pocket, daily wound lavage with diluted povidone-iodine (0.1%), topical applications of triple antibiotic ointment (neomycin/polymyxin B/bacitracin [28.4 g/tube]), Neogen®Vet AluSpray®Aluminum Powder, medical-grade honey, hypertonic saline lavage, and continued systemic antibiotics including potassium penicillin (22,000 IU/kg IV, q6 h) and gentamicin (6.6 mg/kg IV, q24 h), followed by oral doxycycline (10 mg/kg PO, q12 h), the infection persisted, with no appreciable improvement in wound healing. After failing to respond to these interventions, the owner elected experimental phage therapy to treat the SSI. Integrated treatment timeline and culture dynamics during phage therapy of an equine surgical site infection. The black line (days 0–52) represents the overall treatment timeline, beginning with surgery at time 0. The dark purple segment (days 0–4) indicates the initial systemic treatment phase, including intravenous potassium penicillin (22,000 IU/kg IV, q6 h) and gentamicin (6.6 mg/kg IV, q24 h), anti-inflammatory therapy, IV fluid therapy, analgesia, antiendotoxic therapy, cryotherapy, and pentoxifylline. The dark blue arrow (days 5–9) indicates the period of standard wound care without antibiotics while awaiting bacterial susceptibility test results. The red arrow (days 10–52) represents daily doxycycline (10 mg/kg orally, q12 h) therapy. The yellow arrow (days 17–52 postsurgery; days 0–35 postphage therapy) denotes daily topical phage applications; initiation of phage therapy is designated as day 0 for the phage-treatment timeline, and alignment of the yellow and black arrows allows interpretation of events relative to both time postsurgery and time postphage therapy. Wound swabs were collected daily and cultured on LB, MacConkey and mannitol salt agar to monitor bacterial dynamics. Cultures revealed lactose-fermenting and nonlactose-fermenting Escherichia coli variants, as well as two Pseudomonas aeruginosa clones, including a pyocyanin-producing strain and a nonpyocyanin-producing variant that emerged during treatment. The timeline also shows the transient appearance of commensal and environmental organisms detected by veterinary diagnostic bacteriology laboratory. Emergence of multidrug-resistant P. aeruginosa and variant E. coli strains prompted iterative optimization of phage cocktails (Cocktail 1–5), including adjustments to address newly appearing bacterial clones and potential antagonistic interactions among phages. The figure includes 28 days of in-clinic phage therapy followed by outpatient care after discharge. The horse continued daily topical application of the final phage Cocktail 5 during bandage changes at home. Complete wound closure was reported by the horse’s trainer by day 39 and confirmed by photographic documentation Table 1Clinical isolate antibiotic susceptibilityAntibioticsAntibiotic ClassE. coli 47-1 E. faecium E. coli 47-2P. aeruginosa 47-1P. aeruginosa 47-2 Days postsurgery at isolation Day 5Day 5Day 22Day 22Day 26MinocyclineTetracyclineSusceptibleSusceptibleSusceptibleResistantResistantDoxycyclineTetracyclineSusceptibleSusceptibleSusceptibleResistantResistantEnrofloxacinFluoroquinoloneSusceptibleResistantSusceptibleResistantResistantChloramphenicolPhenicolSusceptibleSusceptibleSusceptibleResistantResistantCefazolinΒ-LactamSusceptibleResistantSusceptibleResistantResistantPenicillinΒ-LactamResistantSusceptibleResistantResistantResistantAmpicillinΒ-LactamSusceptibleSusceptibleSusceptibleResistantResistantTetracyclineTetracyclineSusceptibleSusceptibleSusceptibleResistantResistantAmikacinAminoglycosideSusceptibleResistantSusceptibleSusceptibleSusceptibleCeftiofurΒ-LactamNIResistantNIResistantResistantGentamicinAminoglycosideSusceptibleResistantSusceptibleSusceptibleSusceptibleRifampinRifamycin-SusceptibleResistantResistantResistantTrimethoprim/SulfamethoxazoleSulfonamide-ResistantSusceptibleResistantResistantThe table summarizes the antimicrobial susceptibility profiles of clinical bacterial isolates obtained from the patient described in this report. Susceptibility testing was performed using broth microdilution methods and interpreted according to 2024 veterinary CLSI standards[23] for veterinary pathogens. E. coli: Escherichia coli, P. aeruginosa: Pseudomonas aeruginosa, E. faecium: Enterococcus faecium, NI: No interpretation and not performed, -: Antibiotics are grouped by class Clinical isolate antibiotic susceptibility The table summarizes the antimicrobial susceptibility profiles of clinical bacterial isolates obtained from the patient described in this report. Susceptibility testing was performed using broth microdilution methods and interpreted according to 2024 veterinary CLSI standards[23] for veterinary pathogens. E. coli: Escherichia coli, P. aeruginosa: Pseudomonas aeruginosa, E. faecium: Enterococcus faecium, NI: No interpretation and not performed, -: Antibiotics are grouped by class Phage therapy Elimination of E. coli was prioritized to disrupt Enterococcus-driven support that promotes its persistence, leading to attenuation of the biofilm phenotype, reduction in total bacterial burden, and improved immune-mediated clearance.[2425] To generate a five-phage lytic cocktail, the initial clinical E. coli isolate was obtained from the OVDL (designated as E. coli 47-1) and screened against our in-house phage library, which comprises 31 candidates previously isolated using an E. coli ATCC (NCTC 9001) reference strain as the host. This initial screening against E. coli 47-1 did not identify potent phages from the original phage library, prompting a real-time phage discovery effort using a repository of more than 1,000 wastewater and environmental filtrates,[26] curated for the discovery of phages targeting multidrug-resistant ESKAPE pathogens. Screening E. coli 47-1 yielded five lytic phages [ECO22–ECO26; Figure 2a], which were combined to formulate Cocktail 1 [Table 2], assembled over 4 days. Phages were propagated using a modified Phage on Tap protocol,[24] concentrated with 30 kDa Amicon® filters, and washed five times with phage buffer. Endotoxin levels were confirmed below 0.5 EU/mL using Qubit Endotoxin Detection Assay kit, and purified stocks were stored at 1011 PFU/mL before resuspension in 5 mL phage buffer for wound application. Efficiency of plating (EOP) of therapeutic phages against clinical Escherichia coli and Pseudomonas aeruginosa isolates, (a) Serial dilution spot assays showing EOP of six E. coli phages - ECO22, ECO23, ECO24, ECO25, ECO26, and ECO30 (respectively, 1–6) on the clinical E. coli 47-1 isolate recovered from the wound. Phage lysates were serially diluted and 4 µL of each dilution was spotted onto bacterial lawns to assess lytic activity and relative plating efficiency. Clear zones (plaques) across dilution series indicate efficient bacterial lysis, (b) Serial dilution spot assays showing EOP of five P. aeruginosa phages - PAD4, PA-711, TC-1, FG-2, and PAD11 (respectively, 1–5) against the clinical P. aeruginosa 47-3 isolate. Plaque formation across serial dilutions demonstrates the susceptibility of the isolates to the selected phages and supports their inclusion in the therapeutic phage cocktails Table 2Evolution of the phage cocktail during treatmentPhage CocktailsTreatment daysE. coli PhagesP. aeruginosa PhagesNumber of phagesCocktail 10–8ECO22, ECO23, ECO24, ECO25, ECO26None5Cocktail 29–15ECO22, ECO23, ECO24, ECO25, ECO26PAD4, PA-711[26]7Cocktail 316–22ECO22, ECO23, ECO24, ECO25, ECO26PAD4, PA-711, TC-1, FG-29Cocktail 423–27NonePAD4, PA-711, TC-1, FG-24Cocktail 528–35ECO30PAD4, PA-711, TC-1, FG-2, PAD116Modifications to the cocktail were guided by wound culture findings and the observed efficacy of individual phages over time. E. coli: Escherichia coli, P. aeruginosa: Pseudomonas aeruginosa Evolution of the phage cocktail during treatment Modifications to the cocktail were guided by wound culture findings and the observed efficacy of individual phages over time. E. coli: Escherichia coli, P. aeruginosa: Pseudomonas aeruginosa Despite daily doxycycline, which both organisms demonstrated susceptibility to in culture, wound swabs taken on day 0 (17 days postsurgery) showed heavy bacterial growth: E. coli 4+, E. faecium 1+, and Staphylococcus sciuri 1+, indicating a marked increase in E. coli and the emergence of commensal S. sciuri [Figure 1]. The mucoid colony morphology of E. coli suggested high virulence and a strong capacity for biofilm formation [Figure 1; day 17]. Treatment with E. coli Cocktail 1 was started on day 17 postsurgery. A protocol was implemented for daily wound swab collection and culturing in the research laboratory to assess bacterial loads (1–4+ scale) and weekly or as needed cultures by the OVDL for quality control, identification, and antibiotic susceptibility testing. For daily assessments, wound swabs were suspended in 500 μL Hank’s Balanced Salt Solution (HBSS, Corning), vortexed, and streaked onto Luria-Bertani (LB, Hardy Diagnostics), mannitol salt, and MacConkey agar plates, which were incubated at 37°C for 24 h. These cultures were used to monitor bacterial dynamics and guide phage therapy efficacy. The phage cocktail was prepared daily (200 μL of each lysate in 5 mL HBSS) and sterility-tested on LB plates. Wounds were treated every 24 h: dressing removed, discharge assessed, wounds cleaned with povidone-iodine (0.1%) and sterile gauze, lavaged via syringe with saline, photographed, and swabbed for culturing. The phage cocktail was then sprayed onto the wound and surrounding skin, with any remaining solution applied to a fresh dressing. After 24 h of Cocktail 1 treatment, lactose-fermenting E. coli 47-1 dropped from 4+ to 2+, with mucoid morphology disappearing. By day 5, new blue-green and nonlactose-fermenting colonies appeared [Figure 1; day 22 postsurgery], and OVDL identified multidrug-resistant, pyocyanin-producing P. aeruginosa 47-1, and nonlactose-fermenting E. coli 47-2 [Table 1]. P. aeruginosa 47-1 clone was screened against our in-house library of 61 P. aeruginosa phages,[26] resulting in the identification of two potent lytic phages [PA-711 and PAD4; Table 2]. E. coli 47-2 was screened against the five phages in Cocktail 1 and, while still susceptible, exhibited increased resistance, resulting in reduced phage activity. Incorporating these phages into Cocktail 2 on day 9 of postphage therapy reduced P. aeruginosa 47-1 from 4 + to 1 + and E. coli 47-2 from 2+ to 1+ by day 11, demonstrating the effectiveness of the expanded cocktail. A newly emerged, nonpyocyanin-producing P. aeruginosa, designated as the 47-2 clone, was detected in cultures on day 11 of phage therapy [Figure 1]. This prompted us to enrich our P. aeruginosa phage collection by isolating two additional phages in real-time from wastewater [TC-1 and FG-2; Figure 2b], which demonstrated strong lytic activity against both the 47-2 and 47-1 clones. These phages were incorporated into the cocktail on day 16 of phage therapy, resulting in the formulation of Cocktail 3. Interestingly, it was observed that while individual phages and phage combinations in Cocktail 3 were effective against their respective hosts, the combined presence of E. coli and P. aeruginosa phages in Cocktail 3 reduced efficacy against P. aeruginosa 47-2. To address potential antagonism among phages in Cocktail 3, on day 23, the E. coli phages were removed, and only P. aeruginosa phages PA-711, PAD4, TC-1, and FG-2 [termed Cocktail 4; Table 2] were applied to the wound from day 23 to 27 of phage therapy. A final wound swab sample submitted to OVDL on day 23 of phage treatment demonstrated 1+ levels of both E. coli and P. aeruginosa, along with several skin commensals, including Staphylococcus intermedius and Streptococcus equi subspecies zooepidemicus [Figure 1]. The environmental organism Providencia stuartii was also observed. Both E. coli and P. aeruginosa clones were designated as 47-3 clones, obtained from OVDL and used to retest phages from Cocktail 3. These results prompted real-time phage isolation from wastewater, leading to the discovery of PAD11 [targeting P. aeruginosa 47-3; Table 2] and ECO30 [targeting E. coli 47-3; Table 2], both exhibiting strong lytic activity. They were incorporated into Cocktail 5 alongside PA-711, PAD4, TC-1, and FG-2 phages. The patient was discharged after 28 days of phage therapy, and the 7-day supply of the final phage Cocktail 5 was sent home in a spray bottle for daily application during bandage changes, while oral doxycycline therapy continued. At this point, daily photography and wound culturing were discontinued. By day 35 of treatment, the owner reported complete closure of wounds. On day 39, a photo from the horse’s trainer confirmed full resolution of the incision [Figure 3a]. Progress of equine surgical wound healing, (a) Representative images of the infected wound at 4–12 days postsurgery (before phage therapy), at discharge on day 45 (corresponding to day 28 postphage therapy), and at day 52 (corresponding to day 35 postphage therapy completion), (b) Daily wound evaluations over a 4-week postphage therapy period, performed after cleaning and immediately before phage application. Images include a ruler for accurate assessment of both wound sizes, (c) Quantitative analysis of wound closure. The cranial wound area decreased by 44% by day 5 and was fully covered with granulation tissue by day 14. The caudal wound area decreased by 22% by day 7 and by 85% by day 28, at the time of discharge Wound healing During treatment, the wound surface area progressively decreased. Two primary wounds were noted along the ventral midline incision: a smaller caudal wound (~1 cm2) and a larger cranial wound (>5 cm2), as shown in Figure 3b. The cranial wound demonstrated the most rapid reduction in area during the initial phase of phage treatment, with a 44% decrease in size between day 0 and day 5. Healing progressed steadily thereafter, and by day 14 of phage treatment, the cranial wound was considered fully covered by granulation tissue, indicating the successful proliferative phase of wound healing [Figure 3b and c]. The caudal wound, being larger, exhibited a slightly slower initial healing rate, with the greatest reduction in area occurring during the first 7 days of phage therapy, corresponding to a 22% decrease from baseline. Despite the slower start, healing continued consistently over the following weeks. By the time of discharge from the hospital on day 28 postphage treatment, the caudal wound had decreased in surface area by more than 85% [Figure 3b and c]. The horse’s trainer reported that the caudal wound was completely closed and covered with fresh epithelium by day 35, marking the end of visible wound healing [Figure 3a]. This timeline also coincided with the final day of phage cocktail application. Discussion and Conclusions This case represents the first demonstration of successful phage therapy for treating a complex polymicrobial SSI in an equine patient, highlighting its potential as a targeted and effective alternative to conventional antibiotics. The infection initially was dominated by E. coli and E. faecium, with subsequent emergence of multidrug-resistant P. aeruginosa. Despite systemic doxycycline, antiseptic wound management, and topical antibiotics, the infection persisted, highlighting the limitations of conventional therapy for biofilm-associated infections. Targeting E. coli was prioritized to indirectly suppress Enterococcus, which depends on it for metabolic support and biofilm scaffolding,[2425] thereby reducing overall microbial burden and enabling host immunity to clear the remaining bacteria. The application of phage therapy allowed targeted suppression of pathogenic organisms while preserving beneficial commensal flora. Initial phage cocktails derived from our in-house library enabled rapid deployment within 2 days for most clinical isolates, illustrating the importance of maintaining a diverse and ecologically rich phage repository. In cases where preexisting library phages exhibited low activity, real-time isolation of phages directly from environmental sources was necessary, extending assembly time to 4 days but ensuring high lytic efficiency. This dual approach emphasizes that while comprehensive libraries accelerate therapy, real-time isolation remains a critical tool for difficult or resistant strains. Over the course of treatment, wound microbial populations displayed dynamic shifts. The initial E. coli and E. faecium strains were treated with doxycycline before phage therapy; however, despite in vitro susceptibility, E. coli persisted and bacterial loads in the wound increased. This is likely due to its mucoid biofilm phenotype, which reduces antibiotic efficacy in vivo and reflects the well-recognized challenge of biofilm-mediated tolerance in E. coli.[27] Application of Cocktail 1 effectively suppressed the initial E. coli strain; however, complete infection clearance was slower compared to that observed for E. faecium, due to higher bacterial burden and biofilm-associated protection. While no phages were applied against E. faecium, its clearance was likely driven by doxycycline treatment and may have been further facilitated by disruption of interspecies interactions caused by E. coli-targeting phages among the coinfecting pathogens. In addition, although in vitro testing did not reveal a synergistic effect between phage cocktails and doxycycline, phage-mediated enhancement of antibiotic efficacy in vivo cannot be ruled out.[28] Over the course of phage therapy, emerging clones from the dominant population required iterative phage modifications to maintain efficacy. The later emergence of P. aeruginosa further restricted antibiotic options, as doxycycline has no established breakpoints and is ineffective against this pathogen, while aminoglycosides such as gentamicin or amikacin can cause nephrotoxicity and have unfavorable pharmacokinetics.[29] Notably, phage therapy selectively controlled pathogenic populations, allowing wound healing to progress, as evidenced by healthy granulation tissue, reduced discharge, and eventual complete closure. Mixed commensal skin flora was observed in later cultures, consistent with a normal healing wound and suggesting that phage therapy selectively targets pathogenic organisms without sterilizing the wound environment. Five sequential cocktails were ultimately applied, achieving successful infection control and wound closure. This case highlights key principles for phage therapy in complex infections: Rapid assembly of diverse phage cocktails is critical; a well-curated, rich library enables fast deployment and can often eliminate the need for real-time isolationIterative monitoring and cocktail adaptation are essential, as suppression of primary pathogenic strains can allow secondary clones to emergeTargeting dominant pathogens can indirectly suppress secondary species by disrupting ecological dependenciesPhage-phage antagonism may occur, emphasizing the need to test and refine cocktailsPhage therapy can control biofilm-associated infections that are refractory to antibioticsBy selectively targeting pathogens while preserving commensal microbiota, phages reduce bacterial burden and work synergistically with the host immune system, preventing tissue colonization, promoting full clearance of the infection and supporting tissue healing Rapid assembly of diverse phage cocktails is critical; a well-curated, rich library enables fast deployment and can often eliminate the need for real-time isolation Rapid assembly of diverse phage cocktails is critical; a well-curated, rich library enables fast deployment and can often eliminate the need for real-time isolation Iterative monitoring and cocktail adaptation are essential, as suppression of primary pathogenic strains can allow secondary clones to emerge Iterative monitoring and cocktail adaptation are essential, as suppression of primary pathogenic strains can allow secondary clones to emerge Targeting dominant pathogens can indirectly suppress secondary species by disrupting ecological dependencies Targeting dominant pathogens can indirectly suppress secondary species by disrupting ecological dependencies Phage-phage antagonism may occur, emphasizing the need to test and refine cocktails Phage-phage antagonism may occur, emphasizing the need to test and refine cocktails Phage therapy can control biofilm-associated infections that are refractory to antibiotics Phage therapy can control biofilm-associated infections that are refractory to antibiotics By selectively targeting pathogens while preserving commensal microbiota, phages reduce bacterial burden and work synergistically with the host immune system, preventing tissue colonization, promoting full clearance of the infection and supporting tissue healing By selectively targeting pathogens while preserving commensal microbiota, phages reduce bacterial burden and work synergistically with the host immune system, preventing tissue colonization, promoting full clearance of the infection and supporting tissue healing In conclusion, phage therapy provided a safe, targeted, and effective approach for managing a complex coinfection involving E. coli, E. faecium, and P. aeruginosa in a postsurgical equine wound. Stepwise suppression of dominant and emerging pathogens enabled complete wound healing where conventional antibiotics failed due to biofilm formation and drug resistance. This case demonstrates that, when carefully designed and dynamically adapted, phage therapy can successfully control challenging, multidrug-resistant, and biofilm-associated infections while reducing reliance on systemic antibiotics, supporting antibiotic stewardship. It underscores the value of maintaining rich phage libraries, monitoring microbial evolution, and optimizing cocktails in real time, providing a practical model for precision phage therapy in veterinary medicine and beyond. Ethical approval All treatment procedures were conducted under an approved protocol by the Oregon State University Institutional Animal Care and Use Committee (IACUC; protocol no. IACUC-2025-0605), in accordance with institutional guidelines and applicable animal welfare regulations. Conflicts of interest There are no conflicts of interest.