POST-OPERATIVE = ENDOPHTHALMITIS: AN UPDATE

POSTOPERATIVE = ENDOPHTHALMITIS

- = AN=20 UPDATE

DR.MANOJ.S =20 DNB, MNAMS, FRCS(Glasgow)

CONSULTANT VITREO-RETINAL=20 SURGEON

CHAITHANYA EYE HOSPITAL AND = RESEARCH=20 INSTITUTE

KESAVADASAPURAM

TRIVANDRUM

KERALA

PHONE- = 04712447183

mailto:EMAIL-%20soman.mano= j@gmail.com

Post-operative=20 Endophthalmitis: An Update

Settings

Two-thirds of all the cases of endophthalmitis = occur after=20 ocular surgery. Since cataract surgery is the commonest ocular surgery=20 performed, most cases are seen after cataract extraction. Other = intraocular=20 surgeries (eg., glaucoma filtering surgery, vitrectomy, keratoplasty) = have a=20 similar incidence of endophthalmitis (0.1 =96 0.3%); but with a much = worse=20 prognosis. Post-filtering cases have a typically delayed onset with a = classic=20 white-on-red appearance caused by a pus-filled bleb within a congested = eye.=20 However, unlike delayed post cataract endophthalmitis, the causative = organism is=20 usually pneumococcus, and the prognosis is very poor. Post-keratoplasty=20 infections typically have a pain-free onset.

Organisms

The Endophthalmitis Vitrectomy Study (EVS) found = positive=20 cultures for infective organisms in 70% of their patients. All the = infections=20 (within 1 month of cataract surgery) were bacterial, Gram-positive cocci = (90%)=20 being the commonest. The single commonest causative organism in this = study was=20 staphylococcus aureus, which accounted for 70% of all the infections. = Other=20 staphylococci and streptococci constituted 19% of the infections. Gram = =96=20 negative infectious were rare in EVS (4%), fungi or higher bacteria = (eg.,=20 Nocardia) were not isolated.

The Indian spectrum of infective organisms is = however quite=20 different, and must be borne in mind while applying EVS results to the = Indian=20 patients. Three large Indian studies from different parts of the country = (Hyderabad¹, Chennai² and Chandigarh³) = illustrate the Indian scenario. The Chandigarh =96 study was based on = rural=20 eye-camps, and therefore has to be considered separately. The Hyderabad = and=20 Chennai studies showed low culture-positivity (45-54%), a reduced = prevalence of=20 gram-positive cocci (42-47%) an increased number of gram-negative = organisms=20 (26-42%), fungi (17-22%) and polymicrobial infections (13-17%). Gupta et = al=20 showed an even higher number of fungi (58%) in a rural setting, which = had 13%=20 each of gram =96 positive and negative cultures, and a very low culture = =96=20 positivity rate of 38%³.

Overall, Indian studies reveal that the commonest = organisms=20 here (Gram =96 negative bacteria and fungi) are not the ubiquitous = commensals like=20 staphylococcus epidermidis. This implies that the commonest source of = infections=20 in our setting are not ocular adenexa, but other sources in the = operating room=20 like irrigating fluids etc, which might be related to the sub-optimal=20 maintenance of surgical field sterility.

Risk=20 factors

Before a discussion of risk factors, it must be = underscored=20 that such a discussion presumes a perfect asepsis as a basic surgical=20 requirement.

A. =20 Surgical factors

(i) = ; =20 Type of IOL⁴:

It is generally = presumed that bacteria adhere strongly to = hydrophobic=20 IOL surfaces; and making the surface hydrophilic (or less hydrophobic) = repels=20 bacterial adhesion. However, intermediate hydrophobicity (silicone) = attracts=20 bacteria most; and they adhere least to hydrophilic materials like = hydrogel or=20 very hydrophobic ones, = like fluorine=20 =96 coated PMMA. Fluorine appears to be a far superior coating for PMMA = IOLs than=20 heparin, which is reported to allow bacterial adhesion. Lastly, = polypropylene=20 (used to make IOL loops) is probably the worst choice (worse than = silicon) if=20 bacterial growth is feared.

(ii) =20 Posterior capsule tear^5,6 ;

Aqueous has strong = antibacterial=20 activity. Anterior chamber has been shown to be contaminated in 14-18% = cataract=20 surgeries; but the rate of endophthalmitis remains at a low 0.1%. = However,=20 vitreous is a good culture medium for bacteria. Therefore, a tear in = posterior=20 capsule increased the risk of endophthalmitis about 10-fold!

(iii) = Surgical=20 incision^7,8

Phacoemulsification = incisions for=20 foldable lenses are frequently made temporally (to counter = against-the-rule=20 astigmatism) and in clear cornea (for a quieter eye). However, this = situation=20 deprives the wound from the protection of conjunctiva and lids, and = exposes the=20 wound to trauma, tear debris, and aqueous leak. It has been reported = that=20 temporal clear corneal incisions carry 3-4 fold risk of endophthalmitis = compared=20 to superior scleral incisions.

B. =20 Non-surgical factors⁹

It has been recently reported that endophthalmitis = is more=20 common in patients more than 30 years of age; when the patients are = discharged=20 the same day; and when the surgery is carried out in a small private = clinic as=20 opposed to a public hospital or institute. The probable causes may be = lowered=20 patient immunity, less compliance with postoperative instructions; and=20 non-standard levels of perioperative asepsis.

Modes of=20 onset

a) Acute: Endophthalmitis occurred within a week of = surgery=20 in 60% eyes in the EVS study. Such a fulminant infection is usually due = to=20 staph. aureus, pneumococcus, or gram negative organisms.

b) Subacute: Almost all cases of bacterial = endophthalmitis=20 occur within 6 weeks of intraocular surgery. When the latent period of = the=20 infection is a week or more, the organisms are generally less virulent: = the=20 commonest, Staph. epidermidis.

c) Delayed / chronic: Infections occurring after 6 = weeks of=20 surgery are typically caused by Propionibacterium acnes, Staph. = epidermidis or=20 less commonly, nocardia or fungi. The latter are however more common in = Indian=20 settings as previously mentioned.

Clinical = features=20 (EVS)

The universal complaint is decrease in vision, = which may=20 range from 6/18 to perception of light. Pain is also very common, but = EVS=20 reported that 25% of their acute cases had a painless onset. Lid edema = is seen=20 in a third of cases, =91red eye=92 is near universal (82%). Cornea is = clear or=20 slightly edematous in most cases, actual corneal infiltrate occur in a = minority=20 (5%). A hypopyon is almost always present (86%) along with a reduced red = reflex;=20 in most of the EVS cases (80%), no view of the retinal vessels could be=20 obtained¹⁰. As is obvious from this description, no feature = is=20 pathognomonic and therefore when in doubt, one must presume the worst, = and=20 suspect endophthalmitis.

Etiological=20 diagnosis

According to EVS, there were certain clinical = features which=20 indicated gram negative/ Staph. aureus / Pneumococcus infection, and = therefore a=20 poor prognosis: onset of symptoms within 2 days, presenting vision PL, = relative=20 afferent pupillary defect, cataract wound abnormalities (like suture=20 infiltration, leaking wound etc), corneal infiltrate and absence of red = reflex.=20 In fact, they found that if any retinal vessels were visible, a gram- = negative=20 infection was unlikely. However, eye pain was not discriminatory.

Even though the above features are suggestive, they = do not=20 pinpoint the etiological diagnosis; and therefore cannot be used to = choose=20 antibiotics, which must be broad- spectrum.

Early post-operative = Pneumococcal=20

endophthalmitis with = ring abscess=20

Early post-operative Pseudomonas

endophthalmitis with =20 infilteration and hypopyon

&n= bsp; &nb= sp; &nbs= p; = ; =20

Late post-operative=20 Propionibacterium =20 endophthalmitis with equatorial capsular plaque

&n= bsp; &nb= sp; &nbs= p; = ; = &= nbsp; &n= bsp; =20

Late post-operative Nocardia endophthalmitis

with anterior chamber and iris nodules

Additional=20 assessment

Mode of onset may help in diagnosis of infection. = When=20 inflammation occurs immediately after surgery, it is likely to be = iatrogenic,=20 ie., caused by surgical trauma. Even severe infections usually have a = latent=20 period of at least a day. Similarly, the infection following suture = removal is=20 commonly due to pneumococcus, and therefore warrants more aggressive=20 management.

In malnourished, debilitated, immunocompromised and = very old=20 patients, fungal and nocardia infections are more common and also have a = more=20 virulent presentation.

It is important to diagnose and treat / refer the = patient=20 immediately. When the patient presents more than 2 days after the onset, = the=20 efficacy of antibiotic treatment is precipitously = reduced⁶.

Systemic evaluation of immune status and diseases = like TB,=20 diabetes and HIV infections is important, as these diseases worsen the = prognosis=20 after treatment.

Management of acute = endophthalmitis (EVS)

Since the most important prognostic factor is the = vision at=20 presentation, EVS recommended treatment according to presenting vision. = The=20 patients presenting with a vision better than PL did equally well with = vitreous=20 tap (+ intravitreal antibiotic injections) as with primary core = vitrectomy.=20 Intravitreal antibiotics were preferred because No PL vision was more = common=20 after vitrectomy. The patients presenting with PL vision however were = better off=20 with more aggressive treatment (primary vitrectomy with intravitreal=20 antibiotics). Vitrectomy removes the media opacities, reduces infection = load and=20 increases antibiotic penetration in these cases. As an adjuvant to this = primary=20 regimen, EVS recommended topical and subconjunctival antibiotics and = steroids,=20 and also oral steroids in all the cases. They did not find any utility = of=20 systemic antibiotics in endophthalmitis.

Choice of = antibiotics=20 (EVS)

The EVS recommended Vancomycin (1 mg / 0.1 ml) as = the number=20 one choice for intravitreal injection due to excellent coverage of gram = positive=20 organisms. Their recommendation for gram-negative coverage, Amikacin = (0.4 mg /=20 0.1 mL) however was controversial. They tried both Ceftazidime (1.5 =96 = 2 mg / kg=20 tid I/V) and Amikacin (6 mg/kg bid I/V) as systemic antibiotics; = replacing=20 Ceftazidime with oral Ciprofloxacin (750 mg bid) in penicillin =96 = allergic=20 patients. Systemic antibiotics did not affect course of disease in EVS = study.=20 They also used subconjunctival and topical antibiotics for all patients. =

Steroids = in EVS=20 study

Contrary to popular belief, EVS group extensively = used=20 corticosteroids by topical (Pred Acetate 1%), subconjunctival = (Dexamethasone, 6=20 mg / 0.25 mL) and oral (Prednisone 60 mg, 5-10 days) routes. They = however,=20 avoided intravitreal steroid injections. The EVS did not report any = adverse=20 effect / worsening due to steroids in their 400-plus cases.

Vitreous = tap /=20 biopsy

Since both tap (using a hypodermic syringe) and = biopsy (using=20 a vitreous cutter) had the same anatomical and visual results, the more=20 economical and easily available option of vitreous tap is preferred. A = 25-27=20 gauge syringe is recommended. Aspiration of vitreous is done 3.5 mm from = the=20 limbus, withdrawing a 0.1 =96 0.3 mL sample. Biopsy with a cutter is = recommended=20 if the tap is dry. A 0.1 mL sample of aqueous is also taken from = anterior=20 chamber (AC) because in a small number of cases, only AC tap gives a = positive=20 culture of micro organisms. The intravitreal injection of antibiotics = must be=20 done separately (Do not change the syringe with needle in place), two=20 antibiotics should not be mixed. It is mandatory to visualize the needle = in the=20 vitreous cavity and inject with the bevel of the needle pointing = upwards.

The = Indian=20 perspective

A recent report by Verma et al¹¹ = documented the=20 course of 37 patients referred with endophthalmitis. Only 10 had = received an=20 intravitreal injection of antibiotics before referral, the rest were on = systemic=20 antibiotics. Of the previously tapped patients, 60% improved to 6/60 or = more. In=20 the group on only systemic antibiotics, only 32% improved to the same = level.=20 Also, 25% of untapped patients ended up with a vision of no PL. This = report=20 highlights the importance of early primary treatment with intravitreal=20 antibiotics in improving the visual outcome.

Core=20 Vitrectomy

Core vitrectomy is recommended as the primary = treatment when=20 the vision is PL; and as a second-rung procedure when primary = intravitreal=20 injection fails to improve the condition. An ultrasound scan is required = to look=20 for the extent of vitreous opacities, choroidal thickening / detachment; = and=20 most importantly, the retinal status. A vitreous sample should be = obtained=20 before starting infusion. The goal of the procedure is to remove about = 50% of=20 the vitreous volume. Anterior chamber exudates frequently block view and = have to=20 be washed out before vitrectomy. Antibiotics are injected at the end of = the=20 procedure.

Additional=20 procedures

EVS recommended a repeat vitrectomy if after 36-60 = hours of=20 the primary procedure, the vision was worse (but not no PL); red reflex = was=20 reduced further; hypopyon increased; corneal infiltrates developed; pain = increased; or culture showed an organism not covered by previously = injected=20 antibiotics.

Visual = results=20 (EVS)

When the patients enrolled for EVS, the baseline = vision of=20 86% of patients was <2/60. At the conclusion of the study, only 11% = remained=20 with a vision of <2/60; 75% were >6/36; and 55% achieved an = excellent=20 vision of 6/12 or better. However, it would be unrealistic to expect a = similar=20 outcome in the Indian scenario due to a more virulent organism-spectrum = (see=20 above).

Objections to=20 EVS

The focus of criticism of this study was the use of = systemic=20 antibiotics. Many authors believed that EVS could not generalize on the = futility=20 of systemic antibiotics because they used intravenous antibiotics with a = primary=20 gram =96 negative spectrum coverage (Amikacin and Ceftazidime, see = above), even=20 though most cases in EVS had gram positive infections. Instead, they = proposed=20 that using intravenous Vancomycin would be a better alternative and = could have=20 altered the results of the study.

Campochiaro et al¹² had already = reported, the=20 possibility of macular infarction with amikacin in a survey of retina=20 specialists; and discouraged its use. The EVS investigators however gave = compelling arguments in response to this criticism. They claimed than=20 notwithstanding the spectrum of antibiotic coverage, >90% of the gram = positive cocci were susceptible to one of the two antibiotics used. The = eyes=20 which had (the remaining 10%) resistant organisms actually had a better = visual=20 outcome than those with susceptible organisms. This conclusively proved = that the=20 systemic antibiotics, even when effective against intravitreal = organisms, did=20 not provide additional benefit beyond direct intravitreal = injections.

In favour of Amikacin, EVS quoted synergism with = vancomycin,=20 better coverage than other alternatives (especially Ceftazidime); less=20 resistance, less inoculum effect (ie., decreased potency with increased=20 bacterial load) as compared to Ceftazidime =85.and no definite proof of=20 toxicity.

Limitations of=20 EVS

In spite of the elegant study design and strong = arguments=20 against objectors, the EVS still had limitations, inherent in any study. = Because=20 of the focus of the study on post-cataract infections, its results could = not be=20 applied to traumatic, chronic and bleb related endophthalmitis, where = early=20 vitrectomy is probably more urgent.

Among the diabetics with >PL vision in EVS, = recovery to=20 6/12 vision was seen in only 39% eyes (compared to 55% non-diabetics) = with=20 intravitreal injections alone; and therefore they are probably = candidates for=20 primary vitrectomy rather than vitreous tap.

The selection bias in EVS mandated that any = enrolled eye=20 should have been suitable for either tap or core PPV. Since the later = procedure=20 requires a clear cornea, the study automatically excluded severe = endophthalmitis=20 with gross corneal edema or NPL vision. This was one reason why they had = only 6%=20 gram-negative organisms. This also explains the excellent visual outcome = of EVS,=20 the commonest organism was Staph. epidermidis, a low-virulence = infection.

EVS also assumed a fence-sitting posture on use of = steroids=20 in endophthalmitis while they administered steroids by oral, topical and = subconjunctival routes, and reported no fungi or resistant bacteria; = they=20 abstained from intravitreal steroids. Subsequent studies^13,14 = reported contrasting experiences with intravitreal steroids. Since their = primary=20 use is controversial, it appears reasonable to administer steroids after = KOH=20 smear to rule out fungi under broad spectrum antibiotic cover.

Post-EVS=20 era

The EVS argument against systemic antibiotics was = poor ocular=20 penetration. Recent studies report a new drug, = Gatifloxacin¹⁵, which=20 achieves MIC 90 for most bacteria in vitreous within 3 hours. However, = it does=20 not cover pseudomonas and enterococcus, and may cross-react with some = cardiac=20 medications. It may be useful for traumatic and bleb-related = endophthalmitis=20 which were not covered in the EVS.

Since Campochiaro showed macular infarction with = Amikacin,=20 many surgeons and institutions have shifted to the safer ceftazidime=20 notwithstanding EVS insistence on merits and safety of Amikacin. It may=20 definitely be used; but probably it is wise not to repeat it quickly, = especially=20 with vitrectomy, which increases the drug activity on retina.

Even though EVS desisted from injecting steroids = into the=20 eye; the modern controversy on use of steroids in endophthalmitis does = not=20 include the route of administration. That is, when the use of steroid is = agreed=20 on, every one prefers to push it into the vitreous directly rather than = mouth or=20 conjunctiva.

Though EVS advocated only 50% vitrectomy for = endophthalmitis,=20 many surgeons currently prefer a total vitrectomy along with scleral = buckling,=20 endolaser and silicon oil injection¹⁶. This option probably = has a=20 better anatomical outcome; but functional vessels are not significantly=20 better.

Delayed-onset=20 endophthalmitis

As already mentioned, late onset infections are = mostly due to=20 P. acne, S. epidermidis and fungi. The classical clinical picture is of = mild=20 recurrent inflammation, which transiently responds to topical steroids, = but=20 frequently recurs. Media are relatively clear. A classic finding in P. = acnes=20 infection is an equatorial capsular plaque which can be seen only after = dilating=20 the pupil. It is highly suggestive but not pathognomonic of P. acnes = infection=20 (such plaque may also be seen in fungal infections). The most common=20 presentation of fungal infection is cotton ball exudates in the = vitreous.

Vitrectomy is recommended early in spite of the = benign=20 appearance. In P. acnes endophthalmitis, definitive treatment is removal = of the=20 entire capsular bag along with vitrectomy. Vancomycin remains the drug = of choice=20 for intraocular injection. Fungal endophthalmitis is treated by = vitrectomy with=20 intravitreal Amphotericin B

(5 microgm/0.1 mL); along with topical natamycin = 5%, oral=20 itraconazole 100 mg

/ fluconazole 200 mg twice a day.

Prophylaxis

The single most important drug in perioperative = prophylaxis=20 is povidone =96 iodine. This versatile antiseptic is antifungal, = antiviral and=20 antiprotozoal, besides being antibacterial. This is in fact, the only=20 evidence-based prophylactic measure for intraocular = surgery¹⁷. It is=20 used both for painting the eyelid and periorbital skin (10% solution) = and=20 conjunctival application (5% solution). It can also be used before = postoperative=20 limbal suture removal, and in place of postoperative subconjunctival=20 antibiotics.

The use of topical antibiotics is more = controversial. Started=20 3 days preoperatively, topical eye drops (Gatifloxacin / Ofloxacin) have = been=20 reported to prevent endophthalmitis¹⁸. It is important to = avoid=20 prolonged preoperative use however, which might lead to infection with = resistant=20 virulent organisms. In fact, there has been a 10-fold increase in = resistance to=20 ciprofloxacin due to its rampant abuse in perioperative = care¹⁹ .

Postoperative subconjunctival antibiotics do not = show a clear=20 benefit and should be substituted by topical povidone iodine instead. = For=20 topical use, ofloxacin has been reported as superior to ciprofloxacin = due to=20 higher penetration and less resistance (ARVO 2002).

Some aseptic measures have become a standard of = care even=20 though their prophylactic value is uncertain. The surgical masks, for = example,=20 have been shown to reduce the bacterial load in the surgical field, and = are=20 therefore mandatory; even though a causal link with endophthalmitis has = not been=20 established ²⁰(Alwitry, Br J Ophthalmol 2002).

A new factor, which may be linked with a current = increase in=20 incidence of endophthalmitis in some series, is the use of clear corneal = incision. A temporal corneal incision, which carries 3-4 fold the risk, = should=20 be sutured even in presence of a trivial leak. Hydration of corneal = value keeps=20 it closed for less than half an hour. Superior scleral incisions covered = by=20 conjunctiva and lids are the safest²¹ (Nagaki, JCRS, 2003).=20 Similarly, use of silicone IOL or polypropylene haptics should be = avoided in=20 high risk/ one eyed patients due to increased bacterial adhesion (See = above).=20

Meibomian glands harbour most of the pathogenic = organisms for=20 endophthalmitis. Therefore meibomitis, even in absence of frank = blepharitis,=20 should be treated before surgery. Also, eyelash =96 trimming and = conjunctival=20 irrigation should be avoided. Instead, the drape should cover the lid = margins=20 adequately (by cutting it on an open and not closed eye).

Applying = EVS to=20 clinical practice

Having reviewed the recent reports and the Indian=20 microbiological profile vis-=E0-vis EVS, it is obvious that the EVS=20 recommendations cannot be unconditionally applied to our patients with=20 postoperative endophthalmitis. While broadly conforming to the EVS=20 recommendations, following special situations demand individualization = of=20 treatment.

Gram-negative bacteria, fungi,and nocardia should = be expected=20 even in acute and subacute infections, with subsequent change in = treatment and=20 expected prognosis (see above).

Primary vitrectomy is probably recommended for the = above=20 infections, chronic and bleb-related infections, and in diabetic = patients, due=20 to rapid progression and worse prognosis. It is advisable to = aggressively treat=20 conjunctivitis in a post-filter glaucomatous eye, preferably with oral=20 gatifloxacin. This drug may also be a useful adjunct in selected severe = cases of=20 endophthalmitis (see above).

As per the current knowledge, vancomycin is the = single most=20 suitable drug for acute, subacute, and selected cases of chronic=20 endophthalmitis. While vancomycin and ceftazidime / amikacin is the = preferred=20 combination in acute endophthalmitis; even cefazoline and gentamicin = (100 mg,=20 not 400 mg in 0.1 mL) will do, when nothing else is available. However, = if=20 repeat injections are required gentamicin / amikacin should be = avoided.

When a steroid-cover is considered for = endophthalmitis,=20 intravitreal is the current route of choice. It is prudent however, to = rule out=20 fungal infection, before pushing steroids.

Povidone =96 Iodine is the single best agent for = prophylaxis,=20 but it is equally important to take utmost care in surgical field = sterility and=20 irrigating fluids, which appear to be implicated in the Indian patients. = As=20 cataract surgery is being done earlier, and in less opaque lenses with = better=20 pre-operative vision,; it is important to avoid any complications which = severe=20 compromise visual recovery. With these precautions in place, it may be = possible=20 to minimize the risk of this devastating complication of cataract = surgery.

REFERENCES

1: = Kunimoto D Y,=20 Das T et al

Microbiology spectrum and susceptibility of = isolates.Part=20 1-postoperative endophthalmitis

Am J Ophthalmol. 1999;128: 240-42.

2; Anand A R, et al

Spectrum of aetiological agents of postoperative=20 endophthalmitis and antibiotic susceptibility of bacterial = isolates.
Indian J=20 Ophthalmol. 2000 Jun;48(2):123-8.

3: Gupta A, et al

Spectrum and clinical profile of post cataract = surgery=20 endophthalmitis in north India.
Indian J Ophthalmol. 2003=20 Jun;51(2):139-45.

4: = Kodgikian L=20 et al.

Invivo studies of bacterial adhesion in 5 types of=20 intraocular lenses.

Invest Ophthalmol Vis Sci. 2002; 43(12):3717-21. =

5: = Wong TY, et=20 al.

The epidemiology of acute endophthalmitis after = cataract=20 surgery in an Asian

population.

Ophthalmology. 2004 Apr;111(4):699-705.

6: Sunaric G et al

Current approach to post-operative = endophthalmitis

BJO 1997, 81:1006-15

7: = Nagaki Y, et=20 al

Bacterial endophthalmitis after small-incision = cataract=20 surgery. effect of

incision placement and intraocular lens type.

J Cataract Refract Surg. 2003 Jan;29(1):20-6.

8: = Cooper BA, et=20 al.

Case-control study of endophthalmitis after = cataract surgery=20 comparing scleral

tunnel and clear corneal wounds.

Am = J Ophthalmol.=20 2003 Aug;136(2):300-5.

9: Li J, et al =

Significant Nonsurgical Risk Factors for = Endophthalmitis=20 after Cataract

Surgery: EPSWA Fourth Report.

Invest Ophthalmol Vis Sci. 2004 = May;45(5):1321-8.

10: Johnson MW, et = al.

The = Endophthalmitis=20 Vitrectomy Study. Relationship between clinical

presentation = and=20 microbiologic spectrum.

Ophthalmology. 1997 Feb;104(2):261-72.

11: Verma L, et al

First contact management of postoperative = endophthalmitis. A=20 retrospective analysis.
Indian J Ophthalmol. 2004 Mar;52(1):65-6.

12: Campochiaro P A, et al

Aminoglycoside toxicity--a survey of retinal = specialists.=20 Implications for ocular use.
Arch Ophthalmol. 1991 = Jul;109(7):946-50.

13: Das T, et al

Intravitreal dexamethasone in exogenous = endophthalmitis-a=20 prospective study.

BJO 1999,83:1050-53.

14. Shah GK, et al

Visual outcome following the use of steroids in = postoperative=20 endophthalmitis.

Ophthalmology 2000,107; 486-89

15: = Hariprasad=20 SM, et al.

Vitreous and aqueous = penetration=20 of orally administered gatifloxacin in humans.

Arch Ophthalmol. 2003 Mar;121(3):345-50.

16: Kaynak S, et al.

Surgical management of postoperative = endophthalmitis:=20 comparison of 2

techniques.

J Cataract Refract Surg. 2003 May;29(5):966-69.

17: Ciulla T A, et al

Bacterial endophthalmitis prophylaxis in cataract=20 surgery.

Ophthalmology 2002,109: 13-26.

18: Ta C N,et al

Prospective randomized comparison of 3day vs 1 hour = preoperative ofloxacin prophylaxis for cataract surgery.

Ophthalmology 2002,109: 2036-39.

19: Busbee B J, et al

Advances in knowledge and treatment: an update on=20 endophthalmitis.

Curr Opin Ophthalmol. 2004 Jun;15(3):232-7.

20: Altwitry A, et al

The use of surgical face masks during cataract = surgery-is it=20 necessary.

BJO = 2002,=20 86:975-77