Endophthalmitis ( post operative, post trauma, bleb-related, endogenous ) and penetrating eye injuries - Guideline for the Management of
|Publication: 31/01/2002 --|
|Last review: 10/11/2020|
|Next review: 10/11/2023|
|Approved By: Improving Antimicrobial Prescribing Group|
|Copyright© Leeds Teaching Hospitals NHS Trust 2020|
This Clinical Guideline is intended for use by healthcare professionals within Leeds unless otherwise stated.
Guideline for the management of endophthalmitis (post operative, post trauma, bleb-related, endogenous) and penetrating eye injuries
Endophthalmitis is an inflammatory reaction occurring as a result of intraocular infection by bacteria or fungi. Parasites rarely cause endophthalmitis. It can occur post surgery or penetrating eye injuries (exogenous) or secondary to bacteraemia or fungaemia (haematogenous, or endogenous). Endophthalmitis has a poor prognosis with severe sight loss in >30% of patients and retinal detachment in 10% of patients. Acute bacterial endophthalmitis and penetrating eye injuries are medical emergencies.
Endophthalmitis is associated with ocular pain and in the majority of cases (>90%) with a reduction in vision. Pain may be severe or relatively mild. The timing and nature of previous eye surgery should be documented. Symptoms of systemic infection are unusual except when endophthalmitis is secondary to an underlying infective process.
Endophthalmitis is characterized by hypopyon, red eye, lid swelling and there can be corneal oedema and involvement of the posterior segment. Decreased medial clarity and poor fundus visualization may also be present.
In patients with systemic symptoms of infection (fevers, chills, sweats, anorexia, lethargy etc) a full examination should be undertaken to identify any potential primary focus for infection.
- If there is no retinal view, a B scan ultrasound is required to diagnose choroidal effusion or retinal detachment before surgery. [Evidence level C]
- An intravitreal tap (and anterior chamber tap where there is significant anterior chamber activity) should be performed within one hour of clinical diagnosis of acute bacterial endophthalmitis to obtain samples for microbiological identification and culture. [During this procedure intra-vitreal antimicrobials are instilled.]
[Evidence level C]
- General anaesthetic (GA) is favoured because the eye is often inflamed and painful, but organising a GA must not delay the procedure >1 hour from diagnosis. Acute bacterial endophthalmitis is a medical emergency.
[Evidence level D]
- For an aqueous tap 0.2mls of aqueous should be removed via the cataract section or a limbal paracentesis using a 27g needle.
[Evidence level C]
- Recommended vitreal biopsy procedure [Evidence level A] :
- Set up the vitrectomy machine for posterior vitrectomy. No infusion line is required, the tubing should be kept dry.
- Attach a 2ml syringe to the aspiration tubing at the 3 way tap.
- Insert a 25G valved trocar into the sclera in the supero-temporal quadrant. Use a double-plane tunnel technique- penetrate the sclera at about 30 degrees, then adjust to a 60-80 degree to penetrate the globe. Remove the blade leaving the port in place.
- Insert the vitrector through the valved port.
- Withdraw 0.2-0.5ml vitreous using cutter at 5000 cuts/minute.
- The space created by the biopsy is sufficient for direct intra-vitreal injection of antimicrobials (see empirical antimicrobial therapy). Inject through the valved port.
- Remove the port using St Martin’s forceps. Massage the wound with a Q tip to ensure closure.
- Record the procedure on Medisoft and ensure a diagnosis of endophthalmitis is entered for audit purposes.
- NB. The microbiology laboratory should be told about the sample in advance and if a fungal infection is suspected this must be communicated clearly to the laboratory.
- Patients requiring admission for systemic antimicrobial therapy should have baseline full blood count and urea and electrolytes measured. [Evidence level D]
- Blood culture (3 sets at different times) should be sent from patients suspected of having endogenous endophthalmitis prior to commencing systemic antimicrobials. N.B. intraocular antimicrobials can be administered prior to blood cultures. [Evidence level D]
- For the purposes of surveillance inform colleagues of the occurrence of an episode of endophthalmitis (or suspected endophthalmitis). E-mail Fiona.Bishop@nhs.net and copy to all ophthalmology consultants & trainees.
Non antimicrobial treatment
The Gold standard is a three port pars plana vitrectomy with intravitreal antibiotics once completed. The silver standard is a vitreous biopsy and injection of intravitreal antibiotics. However, the performance of a vitreous biopsy and injection of intravitreal antibiotics MUST NOT be delayed in an attempt to perform a full vitrectomy.
A three port pars plana vitrectomy involves removal of the core vitreous gel followed by administration of intravitreal antimicrobials. Vitrectomy is the initial procedure of choice because the infected vitreous behaves like an abscess, impeding the activity of antimicrobials.
In most cases a vitreo-retinal surgeon will not be available to perform a vitrectomy within an hour of presentation, and a vitreous biopsy and injection should be undertaken.
Intravitreal injections of dexamethasone 0.4mg in 0.1ml at the end of a vitrectomy have been found to produce a more rapid subsidence of the intraocular inflammation but without improving the long-term functional outcome. However the ESCRS recommend its use as cells of dead bacteria especially streptococci are highly inflammatory. The role of oral steroid is not established. [Evidence level C].
Empirical antimicrobial treatment
Acute post surgical and bleb-associated endophthalmitis
Routine empirical systemic antibacterial treatment is not recommended
Endogenous (haematogenous) endophthalmitis immunocompetent patient
If current or recent central venous catheter, injecting drug user or known candidaemia add:
intravitreal amphotericin 0.005mg in 0.1ml
Discuss empirical systemic antibacterials with microbiology
Endophthalmitis in immunocompromised patient
Empirical systemic antibacterial treatment should be discussed with microbiology.
Penetrating eye injury
NB Vancomycin and Ceftazidime must not be mixed in same syringe as physically incompatible.
For out of hours use Vancomycin & Ceftazidime (together with dilution instructions) can be found in the ‘red box’ in Theatre 1, Chancellor's Wing, SJUH. Where necessary these can be transferred to LGI for urgent treatment of a child.
Recommended second line intravitreal antimicrobials. Vancomycin 1mg in 0.1ml and Amikacin 0.4mg/ml.
Review microbiology results regularly. Once the causative organism is known it may be appropriate to commence systemic therapy and subsequent intravitreal antimicrobials can be adjusted accordingly. If no microbial cause is identified, empirical therapy should be continued.
Staphylococcus aureus, Coagulase negative staphylococci
Give further intravitreal dose of Vancomycin at 48 hours.
If poor response to therapy discuss adding oral Linezolid with microbiology.
Give further intravitreal dose of Vancomycin at 48 hours.
If poor response to therapy discuss adding oral Linezolid with microbiology.
Oral Clarithromycin 500mg 12-h where surgery is not possible.1
Late case endophthalmitis is milder and usually caused by Propionibacterium acnes. As it is often found enclosed in the synaechised capsular sac it has a high rate of recurrence. This can be reduced by vitrectomy possibly combined with posterior capsulectomy
Enterobacteriaceae or Pseudomonas aeruginosa
If poor response to therapy discuss adding oral Ciprofloxacin with microbiology.
1st line: intravitreal amphotericin 0.005mg in 0.1ml
Confirmed fungal endophthalmitis requires vitrectomy
Fluconazole sensitive Candida
Oral Fluconazole 400mg once a day (after 1 day of 12-hourly regimen for loading)
Duration: 6-12 weeks
Fluconazole resistant or filamentous fungi
Oral Voriconazole 400mg 12-hourly for two doses then 200mg 12-hourly
Duration up to 5 weeks
Negative culture results
If initial culture results are negative and the patient is improving following empirical intravitreal antimicrobials then the addition of systemic antimicrobial therapy is not recommended.
[Evidence level D]
If initial culture results are negative and the patient is deteriorating following empirical intravitreal a second sample should be collected at the time of repeated intravitreal injection.
[Evidence level D]
Intravitreal injection can be repeated according to the clinical picture at intervals of 48-72 hours. If systemic therapy is commenced, continue for 10 days in total, except fungal endophthalmitis which requires 6-12 weeks. Systemic antimicrobials if required are usually given orally.
Repeat doses of intravitreal antimicrobials are recommended if there is not significant resolution of inflammatory signs: vitreous opacification, hypoyon, pain and conjunctival injection, 48 hours after the initial treatment.
Endophthalmitis can be categorized into different groups, which are relevant because the microbiology, surgery and prognosis differs between the groups.
Poor visual outcomes are associated with the more virulent pathogens including Staphylococcus aureus, streptococci, enterococci and Gram negative organisms.1
Acute post operative endophthalmitis
Acute post cataract endophthalmitis is the most common form of endophthalmitis but is an uncommon complication occurring in 0.1-0.2% of operations. The source of infection is usually the patients conjunctival or eyelid flora. Small inocula of bacteria can be cleared from the aqueous by normal immune mechanisms but the vitreous is less resistant to infection. Symptoms usually develop with a week of the operation.
An acute form of endophthalmitis occurs 2-4 days postoperatively and this is mainly due to coagulase negative staphylococci (CNS), Staphylococcus aureus or Gram negative organisms. A delayed form occurs 5-7 days post operatively and is often due to coagulase negative staphylococci, streptococci, enterococci or Gram negatives and rarely fungal species.
Bleb associated endophalmitis
A bleb is a surgically created opening in the sclera, covered only by conjunctiva that is used to treat glaucoma when medical interventions fail. Endophthalmitis may follow “blebitis” and is usually caused by organisms that either colonise or contaminate the conjunctiva, including Streptococcus pneumoniae, oral streptococci, coagulase negative staphylococci, Staphylococcus aureus, Moraxella catarrhalis and occasionally Gram negatives such as Serratia. Infection usually occurs months to years after surgery. Bleb associated endophthalmitis has a poorer prognosis than post cataract endophthalmitis and is more prevalent if antifibrotic agents e.g. 5 fluorouracil or mitomycin c have been used in producing a filtering bleb.1
Endogenous (haematogenous) endophthalmitis
Endopthalmitis occasionally occurs in patients without a history of eye surgery or trauma and is due to haematogenous seeding following a bacteraemia or fungaemia. An underlying cause (e.g. endocarditis, intravascular catheter related infection, intravenous drug use) is present in the majority of cases and the symptoms/signs of the underlying problem may result in a delayed diagnosis.
The most common fungal pathogen isolated is Candidia albicans. Candida species may cause a retinitis without endophthalmitis, which responds well to systemic antifungal therapy alone but if vitritis is present a vitrectomy and intravitreal antifungal therapy is usually required.
Bacterial causes of haematogenous endophthalmitis include streptococci (beta-haemolytic, Streptococcus pneumoniae, oral streptococci), Staphylococcus aureus and Gram negative bacilli.
Post traumatic endophthalmitis
The incidence of endophthalmitis after traumatic penetrating eye injury is 3-8%, rising to 11-26% in the presence of a retained intra-ocular foreign body. The latter cases should be treated as for definite endophthalmitis with vitreous biopsy and intra-vitreal antibiotics. In the absence of intra-ocular foreign body, antibiotic prophylaxis should be given at the time of primary repair. Traumatic endophthalmitis has a worse prognosis than post operative endophthalmitis because the infection is usually mixed and more virulent pathogens may be present. Filamentous fungal causes of endophthalmitis (see below) are more common with penetrating eye injury especially if vegetable matter is present (10-15% of cases).
Chronic pseudophakic endophthalmitis
Chronic, indolent infections can occur after cataract surgery involving artificial intraocular lenses. The most common cause is Propionibacterium acnes, but coagulase negative staphylococci and diphtheroids are occasionally implicated. A high rate of recurrence can be problematic, this can be reduced by vitrectomy, possibly combined with posterior capsulectomy. In unresponsive cases the capsular bag and lens implant may need to be removed completely. (Evidence level C)
Fungal endophthalmitis may be caused by yeasts (e.g. Candida) or filamentous fungi (e.g. Aspergillus, Fusarium). Endophthalmitis in an immunosuppressed patient (e.g. leukaemic, post stem-cell transplant) should be presumed fungal (yeast or filamentous fungus) until proven otherwise. Endophthalmitis in a patient with a current or recent central venous catheter, an injecting drug user or a patient who is known to be candidaemic should also be presumed caused by Candida until proven otherwise.
If fungal endophthalmitis is suspected this must be communicated clearly to the laboratory, as the sample will need special processing. Organisms isolated from cases of fungal endophthalmitis should be sent for antifungal susceptibility testing.
Empirical therapy is administered before the causative organism is known and should have activity against the most likely pathogens. The initial antimicrobial therapy in endophthalmitis is almost always empirical. Occasionally when endophthalmitis occurs as a result of haematogenous seeding from a primary infection elsewhere (e.g. infective endocarditis) the causative organism is known prior to surgery and directed antimicrobial therapy can be given.
Antimicrobial penetration into the vitreous is generally poor so antimicrobials are administered topically, directly into the infected space.
Endophthalmitis is usually confined to the infected eye so delivery of antimicrobials directly to the site of infection enables high local levels of antimicrobials and avoids the systemic side effects.
Vancomycin is given to provide broad spectrum cover against Gram positive pathogens including most staphylococci, streptococci and enterococci.
Systemic Vancomycin has very poor retinal penetration and must therefore be given by intravitreal injection: Eighteen patients with endophthalmitis following cataract surgery were given 1g of Vancomycin by IV infusion. One to five hours later the patients had a vitreous tap, intravitreal antimicrobials and the Vancomycin levels in the vitreous fluid were measured. Results were obtained for 15 patients. The mean Vancomycin level 3.5 to 5 hours after IV administration was 2.04 +/- 1.33 mg/L (range non-measurable to 4.5 mg/L). These levels were sub therapeutic and the authors conclude that the addition of IV Vancomycin in the treatment of endophthalmitis is not beneficial. ESCRS guidelines recommend intravenous use of the same drugs as given intra vitreally based on the theory that this should help to maintain adequate levels in the vitreous, however, this has not been shown to be beneficial and we feel that the real risks of toxicity and intravascular catheter-related infection outweigh any theoretical benefit.
Ceftazidime has broad spectrum activity against Gram negative bacteria including Pseudomonas aeruginosa, but poor spectrum against the relevant gram positive organisims. It does not penetrate the eye well when given systemically, hence intravitreal administration is required.
Penetration of systemic fluoroquinolones into the vitreous is relatively good compared with other antibacterials but is still poor. When two doses of 750mg Ciprofloxacin were administered orally 12 hours apart, 20 patients who underwent vitrectomy had mean vitreous levels of 0.51 +/- 0.24 mg/L which exceeded the MIC90 of Staphylococcus epidermidis at the time5 These data are unlikely to be relevant to current practice. Ciprofloxacin resistance has become widespread and it can no longer be relied upon for activity against coagulase negative staphylococci or Gram negatives in the empirical setting. Meticillin-resistant coagulase negative staphylococci have an MIC90 >16mg/L and meticillin-susceptible strains (which are a less common cause) have an MIC90 of 2mg/L – well in excess of reported ocular concentrations. Ciprofloxacin has been used to treat endophthalmitis but there is no evidence to support its addition to intraocular antimicrobials. In addition, Ciprofloxacin has become a major driver of Clostridium difficile infection and is associated with an increased risk of MRSA infection.
Moorfields eye hospital use Moxifloxacin 400mg once daily for 10 days in adults and use Ciprofloxacin (no dose given) in children.2 Moxifloxacin has marginally better activity against coagulase negtive staphylococci than Ciprofloxacin , but in a recent study of coagulase negative staphylococcal isolates from endophthalmitis cases in the USA, Moxifloxacin was only active against 48% of isolates, too few to justify empirical use (Harper et al., 2007). In addition there is no evidence of efficacy.
Linezolid has excellent activity against Gram positive bacteria including coagulase negative staphylococci. Coagulase negative staphylococcal isolates from endophthalmitis cases in a recent study form the USA were uniformly susceptible to Linezolid (Harper et al., 2007). Linezolid penetrates the aqueous and achieves potentially therapeutic concentrations but vitreal levels are considerably lower (Ciulla et al., 2005; Prydal et al., 2005). Although there have been reports of Linezolid use for endophalmitis, its efficacy in this setting has not been formally evaluated so its benefit over intraocular Vancomycin is not known. In an animal model, Linezolid was not toxic to the eye when given by intraocular injection (Duke et al., 2009).
Fluconazole is appropriate empiric systemic antifungal therapy for Candida endophthalmitis (suspected or proven), but will need to be changed if the isolate is found to be resistant to Fluconazole . Voriconazole should be used as empiric systemic antifungal therapy for endophthalmitis caused by filamentous fungi (suspected or proven), as Fluconazole has no activity against these organisms, and vorioconazole is also active against Candida spp.
Systemic amphotericin has very poor intra-vitreal penetration and serious systemic toxicity and is not therefore recommended. In animal experiments the vitreal concentration of Fluconazole reached 56% of the plasma concentration at steady state.7
|Target patient group:||Adult patients with endophthalmitis|
|Target professional group(s):||Secondary Care Doctors
- ESCRS (European Society of cataract and refractive surgery) Guidelines on prevention, investigation and management of post-operative endophthalmitis. Version 2 “007. Editors Barry P et al.
- Moorfields eye hospital NHS foundation trust pharmacists handbook. 2006
- Ferencz J, Assia E, Diamantstein L and Rubinstein E. (1999). Vancomycin concentrations in the vitreous after intravenous and intravitreal administration for postoperative endophthalmitis. Arch Ophthalmol 117: 1023-1027.
- Smith A, Pennefather P, Kaye S, Hart C. Fluoroquinolones. Place in Ocular Therapy. (2001) Drugs. 61. 747-761.
- Lesk M, Ammann H, Marcil G, Vinet B, Lamer L, Sebag M (1993). The penetration of oral ciprofloxacin into the aqueous humor, vitreous, and subretinal fluid of humans. Am.J. Ophthalmol, 115: 623-8,
- Current clinical practice.
- Sunaric-Mégevand G and Pournaras C. Current approach to postoperative endophthalmitis. Br J Ophthalmol 97;81:1006-1015.
- Ciulla, T. A., Comer, G. M., Peloquin, C. & Wheeler, J. (2005). Human vitreous distribution of linezolid after a single oral dose. Retina 25, 619-624.
- Duke, S., Kump, L., Yuan, Y., West, W., Sachs, A., Haider, N. & Margalit, E. (2009). The Safety of Intraocular Linezolid in Rabbits. Invest Ophthalmol Vis Sci.
- Harper, T., Miller, D. & Flynn, H. W., Jr. (2007). In vitro efficacy and pharmacodynamic indices for antibiotics against coagulase-negative staphylococcus endophthalmitis isolates. Ophthalmology 114, 871-875.
- Prydal, J. I., Jenkins, D. R., Lovering, A. & Watts, A. (2005). The pharmacokinetics of linezolid in the non-inflamed human eye. Br J Ophthalmol 89, 1418-1419.
Improving Antimicrobial Prescribing Group
LHP version 2.0
Equity and Diversity
The Leeds Teaching Hospitals NHS Trust is committed to ensuring that the way that we provide services and the way we recruit and treat staff reflects individual needs, promotes equality and does not discriminate unfairly against any particular individual or group.