Successful medical management of Pythium insidiosum keratitis using a combination of minocycline, linezolid, and chloramphenicol.

PURPOSE: To report successful medical management of Pythium insidiosum keratitis using an antibiotic combination of minocycline, linezolid, and chloramphenicol. OBSERVATIONS: A 20-year-old Japanese man was referred for visual disturbance, hyperemia, and discharge from his right eye. Slit-lamp examination revealed a paracentral corneal hyphate ulcer. His visual acuity was 20/28. Smear examination of corneal scrapings revealed a filamentous fungus. Pimaricin ointment four times a day and voriconazole eye drops hourly were initially prescribed. Although intravenous liposomal amphotericin B 100 mg was added, the corneal infiltrates and ulcer worsened. The possibility of P. insidiosum keratitis was considered, and in vitro antifungal susceptibility testing were performed based on the disc diffusion method. The inhibition zones around each antibiotic disc revealed that the pathogen was susceptible to minocycline, linezolid, and chloramphenicol. Therefore, minocycline ointment four times a day, chloramphenicol eye drops hourly, and linezolid 1200 mg orally per day were also administered. Eventually, sequencing of ribosomal DNA confirmed the pathogen to be P. insidiosum. The triple regimen dramatically improved the patient's keratitis. Therapeutic penetrating keratoplasty for corneal perforation was successfully performed, and his visual acuity recovered from 20/2000 to 20/25. CONCLUSIONS AND IMPORTANCE: We have encountered a case of P. insidiosum keratitis that responded to a combination of minocycline, linezolid, and chloramphenicol. This triple combination should be considered in patients with P. insidiosum keratitis.

Introduction

Pythium insidiosum belongs to the kingdom Stramenopila, a fungus-like aquatic oomycota often found in tropical, subtropical, and temperate climates. P. insidiosum keratitis has been reported in several countries, including Thailand, India, China, Australia, Haiti, and Israel. Although P. insidiosum keratitis is often misdiagnosed as fungal keratitis clinically from corneal scrapings, it is recalcitrant to antifungal therapy.

Case report

A 20-year-old Japanese man who presented with pain, hyperemia, and discharge in the right eye was referred to our hospital after one week of administration of moxifloxacin and fluorometholone eye drops without remission. He belonged to the water polo club of his university, and played with daily disposable contact lenses. His corrected distance visual acuity (CDVA) was 20/28, and slit-lamp examination revealed a paracentral corneal hyphate ulcer and inflammatory infiltrates with feathered margins involving inflammation in the anterior chamber (Fig. 1A, B). The contralateral eye was essentially normal. A set of corneal scrapings for smear including Fungiflora Y and microbial culture was performed immediately. Branching filamentous hyphae were observed in Fungiflora Y staining but were poorly stained on gram staining. Culture of the corneal scrapings revealed growth of a filamentous organism after a few days (Fig. 2A, B). Treatment for fungal keratitis was initiated, including voriconazole eye drops and natamycin ointment. However, the symptoms and ulceration worsened, so intravenous liposomal amphotericin B was introduced. The infiltrates extended progressively and a further corneal scraping with culture yielded the same results. After 2 weeks of outpatient treatment, the patient was hospitalized (Fig. 1C).

Fig. 1

(A–F) Digital images of the right eye taken during the clinical course as observed on slit-lamp examination in a chronological sequence. (A, B) A paracentral corneal hyphate ulcer with staining surrounded by infiltrates with feathery margins associated with tentacle-like lesions observed at presentation. (C) Worsening of the paracentral lesion after one month of intensive treatment; at this time, the possibility of P. insidiosum was considered and minocycline, linezolid, and chloramphenicol were started. (D) Progressive regression of the corneal infiltration was observed. (E) The patient presented with corneal perforation 1 month after changes to the systemic and topical triple regimen. (F) Eleven months after therapeutic penetrating keratoplasty, the patient's corneal perforation had resolved completely.

Fig. 2

(A, B) Direct microscopy of specimens collected by corneal scraping. (A) Branching filamentous fungus-like hyphae were poorly stained with discoloration (Gram stain, original magnification × 1000). (B) Fragments of branching filamentous hyphae were clearly stained by Fungiflora Y (original magnification × 400). (C) Confocal imaging, hyper-reflective fungus-like filamentous lesions. (D, E) In vitro antifungal susceptibility testing: antibiotic susceptibility of the isolates was performed by the disc diffusion method. The antibiotic discs used were chloramphenicol, azithromycin, tobramycin, minocycline, erythromycin, levofloxacin, moxifloxacin (Nippon Becton Dickinson Co, Ltd., Tokyo, Japan), and linezolid (Eiken Chemical Co., Ltd., Tokyo, Japan). The antibiotic susceptibility of the isolates was determined by the presence or absence of inhibition zones around each antibiotic disc at 48 hours from the start point. The strains were most sensitive towards minocycline and linezolid, sensitive to chloramphenicol, erythromycin, and azithromycin, and resistant to tobramycin and quinolone.

After 2 further weeks of intensive inpatient treatment, the possibility of P. insidiosum keratitis was considered, therefore a request was made for identification of the pathogen by polymerase chain reaction (PCR). There were no guidelines available from the Clinical and Laboratory Standards Institute for interpretation of tests for P. insidiosum, thus the antibiotic susceptibility of the isolates was interpreted by determining the inhibition zone around each antibiotic disc (Fig. 2D, E). The strains were characterized as sensitive if there was an inhibition zone around the disc or resistant if there was no inhibition zone. Minocycline and linezolid showed the largest inhibition zones, and chloramphenicol, erythromycin, and azithromycin showed inhibition zones of intermediate size. Tobramycin and quinolones did not show inhibition zones. While the PCR investigation was being performed, oral linezolid 1200 mg twice daily, topical chloramphenicol eye drops hourly, and minocycline ointment 4 times daily were initiated, with progressive regression of the corneal infiltration (Fig. 1D). The pathogen was eventually confirmed to be P. insidiosum by ribosomal DNA sequencing.

The patient was discharged from hospital when his corneal infection was noted to be settling after the change in regimen. Three weeks after discharge from hospital, he presented with a corneal perforation in his right eye accompanied by a very shallow anterior chamber and a decrease in his CDVA from 20/200 to 20/2000 (Fig. 1E). It was apparent at this time that although the patient's infectious keratitis had been successfully managed by medical treatment, the perforation had not resolved. Penetrating keratoplasty (PK) was then performed, and the CDVA recovered to 20/25. There has been no recurrence of infection in the 11 months since his surgery (Fig. 1F).

Discussion

We have encountered a patient who was initially diagnosed to have fungal keratitis and treated accordingly but was eventually found to have P. insidiosum keratitis that was managed successfully by minocycline, linezolid, and chloramphenicol. P. insidiosum keratitis is often misdiagnosed as fungal keratitis based on slit-lamp examination findings and corneal scrapings. Dense grayish-white infiltrates with feathery margins, associated with tentacle-like lesions on slit-lamp examination and filaments resembling fungi on histochemical staining and confocal microscopy, are reported to be characteristics of P. insidiosum keratitis. When such observations are made in a patient who does not respond to treatment with antifungal agents, P. insidiosum should be considered as a causative pathogen. P. insidiosum does not contain ergosterol in the cell wall, and is recalcitrant to conventional antifungal therapy. Furthermore, accurate microbiological identification is based on formation of zoospores; in the absence of sporulation, DNA sequencing is needed for identification.

A large number of fungal species are known to cause keratitis. However, 10%–23% of fungal isolates from patients with fungal keratitis is reported as unidentified from lack of sporulation in culture,. The prevalence of P. insidiosum among unidentified fungal isolates from keratitis is reported to be 5.5% in DNA sequencing of morphologically unidentified nonconsecutive fungal isolates from corneal scraping with keratitis, and 3.9% in zoospore formation of fungal colonies resembling those of P. insidiosum followed by DNA sequencing. Therefore, incidence of P. insidiosum keratitis is thought to be 1% or lower among fungal keratitis.

Currently, there is no standard protocol for P. insidiosum keratitis, and recent in vitro and in vivo susceptibility literature supports the use of antibacterial medications for this infection. Successful resolution of human P. insidiosum keratitis using a combination of linezolid, azithromycin, and atropine sulfate has been reported previously. Clinical comparisons of antifungal and antibacterial therapy have been reported, with a combination of topical linezolid and topical and oral azithromycin being found to be superior to topical natamycin, which might indicate the response of P. insidiosum to antibacterial agents. In vitro susceptibility of P. insidiosum to antibacterials including minocycline, tetracyclines, macrolides, and linezolid, as monotherapy and in combination with antifungal agents has been reported. Moreover, there is in vitro and in vivo evidence of a synergistic antibacterial effect between minocycline and other antibacterial agents when used to treat P. insidiosum,. However, none of the previous studies have described the efficacy of minocycline and chloramphenicol in the management of human subjects. Administration of minocycline together with linezolid and chloramphenicol was effective for P. insidiosum keratitis in our case. Therefore, triple therapy consisting of minocycline, linezolid, and chloramphenicol, which act by inhibiting protein synthesis, may be a promising candidate treatment for P. insidiosum.

Conclusions

We have reported a case of P. insidiosum keratitis that responded to a triple regimen of minocycline, linezolid, and chloramphenicol. Treating clinicians should be aware of P. insidiosum as a potential causative agent when a patient presents with fungal keratitis refractory to antifungal therapies.

Patient consent

The patient who is the subject of this case report provided informed consent for his case to be published.

Funding

No funding or grant support

Conflicts of interest

None of the authors has any financial disclosures to make.

Authorship

All authors attest that they meet the current ICMJE criteria for authorship.