Objective: To evaluate the advisability of removing silicone oil from
eyes after surgery for severe (with a classification of at least C-3)
proliferative vitreoretinopathy. Design: Subgroup analysis of the Sili
cone Study, a randomized, multicentered, surgical trial. Setting: Comm
unity- and university-based clinics. Patients: Two hundred twenty-two
eyes with severe proliferative vitreoretinopathy followed up in the Si
licone Study. Interventions: Vitrectomy for proliferative vitreoretino
pathy with silicone oil as the intraocular tamponade. Outcome Measures
: Changes in visual acuity, recurrent retinal detachment, and incidenc
e of complications. Results: Ninety-nine (45%) of 222 eyes had surgery
for silicone oil removal (oil-removed eyes). Compared with the eyes t
hat did not undergo silicone oil removal (oil-retained eyes) evaluated
at a comparable time after oil injection, oil-removed eyes at the exa
mination prior to oil removal were more likely to be attached (85% vs
40%; P<.0001), have a visual acuity of 5/200 or greater (63% vs 35%; P
<.0001), and not be hypotonous (5% vs 22%; P<.001). There was no assoc
iation between the length of oil retention and incidence of recurrent
retinal detachment after oil removal. Eyes with attached retinas at th
e time of oil removal generally improved in visual acuity at the last
follow-up examination (P<.0001), which was not evident in eyes with de
tached retinas at the time of oil removal. In a matched-pair cohort an
alysis comparing both sets of eyes, there was an increased risk for re
current retinal detachment at the last follow-up examination in the oi
l-removed eyes (odds ratio [OR], 2.1; P=.09). However, overall visual
acuity improved for oil-removed eyes in 19 (29%) of 66 pairs and for o
il-retained eyes in one (2%) of 66 pairs (OR, 19.0; P<.0001). Although
nonsignificant, incidence rates of keratopathy (OR, 0.5) and hypotony
(OR, 0.5) were lower in oil-removed eyes. Conclusion: Removal of sili
cone oil in anatomically successful eyes significantly increases the l
ikelihood of improved visual acuity with a slight increase in the like
lihood of recurrent retinal redetachment. There was a trend for a redu
ction in the incidence of complications in the oil-removed eyes.