Recent studies have shown that endophthalmitis caused by liver abscess accounts for an increasing percentage of endogenous endophthalmitis cases each year, particularly in East Asia. The main source of infection for endogenous endophthalmitis is liver abscess, and the predominant cause of endogenous endophthalmitis is the pathogenic bacteria, Klebsiella pneumoniae. A study on endogenous endophthalmitis conducted in Korea showed that 25% of primary lesions were liver abscesses, 10 whereas a study in Taiwan found that 53% of primary lesions were liver abscesses, and 61% of cases were caused by K. pneumoniae, and showed the trend of drug resistance. 11 In the present study, microbial cultures of the aqueous humor and vitreous humor, as well as blood cultures were obtained from all patients Bacteria could be identified in 11 (68.75%) cases, 8 (50%) of which were positive for Klebsiella pneumoniae,3 patients showed carbapenems resistance. The ideal condition for performing a bacterial culture is during the absence of antibiotic treatment following disease onset. However, due to the rapid progression of the disease, and lack of a standardized antibiotic treatment strategy for endogenous endophthalmitis, it is difficult to obtain this condition, and the results of the culture may be unreliable. 12 Therefore, for effective management of endogenous endophthalmitis, blood samples should be collected before starting antibiotic treatment. 13
Liver abscess–associated endophthalmitis can significantly impact the visual acuity of patients, and has a poor prognosis. A study conducted in Southern California investigating endophthalmitis caused by K. pneumoniae showed that approximately half of the patients required enucleation. 14 It has been reported that early use of antibiotics combined with timely vitrectomy in patients with endophthalmitis can effectively improve the visual acuity.15 In the present study, we found that while patients with liver abscess–associated endophthalmitis who underwent vitrectomy combined with an intravitreal injection had significantly improved visual acuity, the overall prognosis was poor. Endogenous endophthalmitis caused by liver abscess is a metastatic infection, in which inflammation occurs rapidly and pathogenic bacteria can invade the inner eyes in a short period of time, which makes surgery more difficult. In this study, some patients (43.75%) with endogenous endophthalmitis were initially diagnosed with only a liver abscess and fever. But four patients had NLP vision when they were transferred to ophthalmology, two of the patients infected with Bacillus licheniformis and Serratia marcescens showed binocular vision loss,they are sensitive to quinolones, However, cephalosporin is commonly used in the treatment of liver abscess. Eye symptoms are easily neglected when rescuing critical symptoms and focusing on systemic conditions, thereby missing the opportunity for early diagnosis and effective treatment. Furthermore, the clinical manifestations of liver abscess–associated endophthalmitis are nonspecific, and the condition can easily be misdiagnosed as other diseases, such as uveitis, leading to a delay in diagnosis and treatment. In addition, the high virulence of pathogenic bacteria in liver abscess–associated endophthalmitis can rapidly result in irreversible damage to the retinal function once the eye is infected. Therefore, retinal function cannot be restored despite the eventual control of infection.
Vitrectomy is effective for the treatment of retinal detachment, vitreous hemorrhage, and diabetic retinopathy. It is also the most commonly used surgical method that has demonstrated a definitive efficacy for the treatment of endophthalmitis.16–18 Vitrectomy can remove intravitreal inflammatory lesions, bacteria and toxins, and reduce the damage caused by toxic substances on retinal function.19 It may also rescue vitreous transparency and reduce or avoid tractional retinal detachment.20,21Intravitreal injection enables the rapid and effective delivery of antibiotics, to achieve high local drug concentrations. This allows the antibiotics to target the pathogenic bacteria more effectively, thereby inhibiting bacterial growth and controlling inflammation. Concurrent intravitreal vancomycin ,low-dose ceftazidime and dexamethasone are recommended for patients with infectious endophthalmitis caused by liver abscess. According to the experimental and clinical observation, most clinicians believe that injection of vancomycin, ceftazidime and dexamethasone into the vitreous cavity is safe and dose not lead to retinal toxicity. Studies have shown that intravitreal antibiotic injection for susceptible pathogens within 48 hours may help in retaining the visual acuity in some patients with liver abscess–associated endophthalmitis, and patients who received antibiotic intervention after 48 hours following disease onset ultimately had poorer visual acuity.22 Despite prompt treatment, the visual sequelae of endogenous endophthalmitis caused by liver abscess are significant and outcomes are poor. In the current series, only seven (36.84%) of our patients had visual acuity of CF or better at the 6-month follow-up visit. In addition, for patients after vitrectomy, beasue some patients are accompanied with bacteremia, postoperative infection is still a high risk factor for blindness. 3 patients in this study underwent enucleation because of repeated infection, so it is necessary to inject antibiotics intravenously according to the results of drug sensitivity after vitrectomy.
The present study showed that upon diagnosis of endophthalmitis caused by liver abscess, treatment should be initiated immediately and involve intravitreal injections of antibiotics and vitrectomy. Meanwhile, the active treatment of the primary lesions was critical. The administration of systemic antibiotics should be accompanied by active treatment of surgical abscess drainage, carbapenem can be used when cephalosporin has no effect. There are still many limitations to this study. First, case data may not be perfect enough as a retrospective study. Second, the sample size of this study is small, and requires a larger sample si ze and multicenter approach.