The eligible patients were assigned at random to a 2 × 2 factorial design to one of the four treatment groups (Table 32.2).

Treatment was initiated within 6 h of initial examination. In all four groups, undiluted vitreous specimen was collected before performing the assigned procedure, usually with a vitreous cutter in the VIT group and using a needle in the TAP group. Patients assigned to VIT received a three-port pars plana vitrectomy, and no additional attempt was done to separate the posterior vitreous, if not separated already. A vitreous volume of 0.1–0.3 ml was collected in patients assigned to TAP. Two intravitreal antibiotics were injected in all four groups, vancomycin 1.0 mg in 0.1 ml and amikacin 0.4 mg in 0.1 ml. Patients assigned to intravenous (IV) antibiotics received ceftazidime and amikacin.

Undiluted vitreous and vitrectomy effluent, when vitrectomy was done, were cultured. Three media were used—chocolate agar (37 °C in CO₂), thioglycolate broth, and Sabouraud dextrose agar. The vitrectomy effluent was filtered through a sterile 0.45 μm membrane filter; the filter was divided into three pieces under sterile conditions. The three pieces of filter papers were cultured in chocolate agar, in Sabouraud dextrose agar, and thioglycolate broth. Gram stain was used for microscopy. The microbiological results were categorized into “confirmed positive,” “equivocal,” and “negative.” The microbiology and infection category is shown in Table 32.3 [2].

In the EVS there was 69.3% (n = 291) confirmed growth, 12.9% (n = 54) was equivocal growth, and 17.9% (n = 75) was no growth. Gram-positive cocci were more common (94.2%) than gram-negative bacilli (5.9%), and gram-positive coagulase-negative micrococci were the most common isolation (n = 226; 70%).

All gram-positive coagulase-negative micrococci were sensitive in descending order to vancomycin (100%), amikacin (86.1%), and ciprofloxacin (77.9%); gram-negative bacilli were sensitive in descending order to ciprofloxacin (94.7%) and both ceftazidime and amikacin (89.5% each). In nearly all instances, the study drug combination was sensitive to the isolates in the EVS (Table 32.4).

The EVS used medicines through all routes—topical, subconjunctival, intravitreal, and systemic (intravenous or oral). These consisted of antibiotics, vancomycin, ceftazidime, amikacin, and ciprofloxacin, and corticosteroids, dexamethasone and prednisolone. The details and route of delivery are shown in Table 32.5.

In the EVS 420 patients were recruited in 24 study centers. The major results are tabulated in Table 32.6 [3].