Abstract

Dear Editor:
In past decades, it has been well recognized that endogenous KP endophthalmitis almost unavoidably brought an extremely poor visual outcome.2–4 As KP endophthalmitis is prevalent in Taiwan, we would like to report our successful experience in achieving good visual outcomes with improved treatment modalities.
A 53-year-old male with underlying diabetes mellitus suffered from liver abscess with KP bacteremia. Endogenous endophthalmitis in the right eye developed 3 days later. Initial corrected visual acuity was counting fingers at 80 cm. Immediate intravitreal injection (IVI) of vancomycin and ceftazidime was given. Hypopyon developed and vision deteriorated to hand motion by the next day. Emergent pars plana vitrectomy (PPV) was done. In addition to systemic administration of ceftriaxone, oral levofloxacin 500 mg once daily was also given for 2 weeks. Repeated IVI of ceftazidime and a second PPV were done 1 and 3 weeks after the first surgical intervention, respectively. Intraocular inflammation completely subsided with only localized subretinal scar remaining. Unfortunately, rhegmatogenous retinal detachment occurred at 5 weeks after initial symptom onset and PPV and silicone oil injection were performed. After silicone oil removal, the patients' vision reached 20/32.
Studies of endogenous KP endophthalmitis have shown that the initial visual acuity at diagnosis and intervention is an important prognostic factor for a better visual outcome.2–4 Thus early diagnosis and intervention are believed to be crucial in treating this devastating disease. The latest case–control study revealed that the presence of hypopyon is a marker for poorer visual outcome. However, it is noteworthy that in our case of typical endogenous KP endophthalmitis with an initial visual acuity of counting fingers and documented hypopyon, a final visual acuity of 20/32 was achieved. This dramatic improvement in visual acuity was very rare compared with previous reports. Prompt surgical intervention and intravitreous administration of antibiotics should predominate, but we believe that administration of an additional systemic antibiotic, levofloxacin, a divergence from the traditional regimen, may be helpful. The use of intravenous ceftriaxone as the first choice in endogenous KP endophthalmitis is according to the result of a single study done by Sharir et al. 5 in 1989, which showed reliable vitreal penetration of ceftriaxone through systemic administration. By contrast, good ocular penetration of fluoroquinolones has been documented in several human studies and animal models. Achieving a therapeutic level of levofloxacin against most gram-negative bacteria has also been observed in oral drug administration.6,7 Vitreous concentrations of 1.6 μg/mL can be achieved 2.5–4 h after oral intake of levofloxacin 500 mg, which is well over the MIC90 of KP (0.13 μg/mL). 6 The use of systemic levofloxacin seems to be supported by more evidence. In their work, George and associates demonstrated that a single oral dose of levofloxacin 750 mg achieved an average concentration of 2.83±0.80 μg/mL about 602 min after administration, which is higher than needed.
The past 2 decades have shown no significant improvement in the visual outcome of endogenous KP endophthalmitis. We postulated that intravenous ceftriaxone as the conventional treatment might not have a high enough intraocular concentration to be effective against KP endophthalmitis. In our case, prompt intervention and the use of systemic levofloxacin produced a good outcome. Although early detection and intervention are of most importance, we also believe modifications to the traditional systemic antibiotic treatments are worth trying.
We congratulate George and associates on their work and for their clarification of the issue. Their application in exogenous postoperative or traumatic endophthalmitis is rational and may become the standard of treatment. We hope our clinical experience also helps advance the treatment of endogenous KP endophthalmitis.
Footnotes
Author Disclosure Statement
The authors have no proprietary or financial interest connected with any product mentioned herein.
