INFECTION AFTER RETINAL DETACHMENT SURGERY

Australian and New Zealand Journal of Ophthalmology 1986; 14: 69-73 INFECTION AFTER RETINAL DETACHMENT SURGERY OSMOND BRUCE HADDEN FRACO Auckland Hospifal. Auckland, New Zealand Abstract: In 250 consecutive retinal detachment operations performed by the author, there were 14 cases (5 6%) of infection of the scleral buckle The commonest infecting organism was Staphylococcus aureus The surgery in these infected cases took longer than average, and utilized more than the usual amount of silicone sponge, a higher proportion were reoperations After an average follow-up of 22 months, only six of the 14 had vision of 6/36 or better and of these, two had persisting inferior traction detachments On the basis of this study and others, the preferred management of infected scleral buckles is to remove the sponge as soon as the diagnosis is made The risk of redetachment is a lesser evil than the sequelae of prolonged inflammation which include traction retinal detachment, massive perirebnal proliferabon, and premacular fibrosis Key words Retinal detachment - infection after surgery, scleral buckle infections, silicone sponge infections The purpose of this paper is to describe the frequency, nature, and management of infection after retinal detachment surgery. A previous paper by the author described the postoperative complications in 135 consecutive retinal detachment operations. ' In the course of the past four years, during which the series has expanded to 250 cases, infection has remained a recurring problem. This paper compares the infected with the noninfected cases, in an attempt to identify significant differences, the recognition of which may reduce the rate of this potentially avoidable complication. METHODS The records of 250 consecutive retinal detachment operations performed by the author between January 1976 and January 1984 were analysed. The 14 cases which developed postoperative infection were recalled and reexa- mined if the records were incomplete or if the follow-up period was less than six months. The follow-up time ranged from six months to eight years, with a mean of 22 months. Comparisons with the 236 non-infected cases were made of surgical time, amount of sponge used, rate of reoperation, rate of successful reattachment, and visual outcome. The results of this and previously published series were used to formulate a plan to reduce the incidence of infection and to improve the management of established infection. RESULTS Prevalence Fourteen of the 250 cases (5.6%) became infected. Folio W-up Follow-up time ranged from six months to 93 months, with a mean of 22 months. Reprint requesfs: Dr 0. B. Hadden, Eye Department, Auckland Hospital, Auckland, I. New Zealand. INFECTION AFTER RETINAL DETACHMENT SURGERY 69

TABLE 1 Comparisons Between Infected and Non-infected Cases Feature Number (Yo) of cases Statistical Infected Non-infected significance (14 cases) (236 cases) Attached Vision 6/36 or 7 (50%) 201 (85%) p<0.001 better 6 (43%) 170 (72%) p=0.02 Encircling sponge 4 (28%) 38 (16%) * Reoperations 4 (28%) 35 (15%) * Duration of surgery 140 min 125 min t *Numbers insufficient for analysis.?not statistically significant. min =minutes. Time of Presentation Nine of the 14 infections became apparent within the first week after surgery. The longest interval between surgery and manifest infection was eight weeks. Retinal Reattachment Only seven of the 14 had attached retinas when last seen. Three of those seven had poor vision due to premacular fibrosis. Visual Acuity Only six of the 14 had vision of 6/36 or better. Those six included the four with attached retinas without premacular fibrosis, and two with persisting inferior traction detachments, but with the maculas attached. Of the other five patients with poor vision, two had traction detachments involving the macula and three had massive periretinal proliferaton (MPP). Comparison with Non-in fected Cases Table 1 gives comparisons between the infected and non-infected cases. Sponge Removal All 14 cases had the sponge removed at intervals varying from the first postoperative day to 10 weeks after the diagnosis of infection. The two cases in which the sponge was removed the day after the diagnosis of infection both did well, with vision of 6/18 and 6/24 respectively. The three cases in which the sponge was removed later than eight weeks after the diagnosis of infection had poor vision. Two had attached retinas but premacular fibrosis, and one had traction detachment. Of the other nine in whom the sponge was removed between two and six weeks after the diagnosis of infection, three had good results and six had poor results. Three of the six had MPP, one had premacular fibrosis, and two had traction retinal detachment. Redetachment in relation to sponge removal Seven cases had detached retinas when last seen. In two, the detachment was present before the sponge was removed. In one of those, there was a localized balloon detachment first seen two days after surgery; six days later a subretinal abscess was visible under the balloon detachment. In the other case, a peripheral traction detachment developed during a prolonged period of inflammation, before the sponge was finally removed eight weeks after infection was diagnosed. In the other five cases with detachment, the retina detached at periods ranging from five days to eight weeks after removal of the sponge. No sponge was removed sooner than three weeks after the initial retinal surgery. Numbers were insufficient to correlate the chance of detachment with time when the sponge was removed. However, premacular fibrosis was significantly related to the length of time between manifestation of infection and removal of the sponge. In the three cases in which premacular fibrosis occurred, the times between manifest infection and removal of the sponge were four, nine and 10 weeks respectively; it did not occur in the seven cases in which the sponge was removed within three weeks of the infection becoming manifest. Because of the small numbers, the significance of this timing could not be statistically validated. Infecting Organism Staphylococcus aureus was the culprit in eight cases, staphylococcus epidermidis in two cases, beta-haemolytic Streptococcus in one, and Pseudomonas in one case. The records were missing in two cases. Of the 10 cases infected with Staphylococcus (aureus or epidermidis), 70 AUSTRALIAN AND NEW ZEALAND JOURNAL OF OPHTHALMOLOGY

TABLE 2 Sensitivities of Infecting Staphylococcal Species Antibiotic Number Number Number tested sensitive resistant Methicillin 8 7 1 Erythromycin 6 6 0 Tetracycline 3 3 0 Penicillin 6 2 4 Ampicillin 3 0 3 Gentamicin 3 3 0 eight of the 10 were tested against methicillin and seven were sensitive; six were tested against erythromycin and all were sensitive; three were tested against both tetracycline and gentamicin, and all were sensitive; six were tested against penicillin and only two were sensitive; three were tested against ampicillin and none were sensitive. These results are summarized in Table 2. DISCUSSION The frequency of infection of silicone sponges has been reported as ranging from 0.1Vo2 to 24.4%. Hilton and Wallyn4 listed the infection rates published by five other authors, those rates having a mean of 3.3%. Thus the infection rate of 5.6% in the present series is above average. The extremely high frequency of 24.4%, reported by Russo and Ruiz, relates to inclusion of patients in whom sponges were removed up to four years after surgery, because of mechanical problems such as discomfort, mild inflammation, and subconjunctival haemorrhage, and in which organisms were cultured from the removed sponges. If those chronic cases were excluded, their frequency would reduce to 7.1%, which is still moderately high. The use of silicone sponges, as designed by Lincoff et al,s is often criticized because of the higher frequency of infection as compared with solid silicone materials. All cases in the present series utilized silicone sponges, the only solid silicone being an encircling band (strap), which was never used without a local sponge. Hilton and Wall~n,~ in 600 consecutive cases, had seven infections in the 385 cases in which sponge was used, but no infections in the 215 cases in which only solid silicone was used. In 14 infected cases, INFECTION AFTER RETINAL DETACHMENT SURGERY Russo and Ruiz removed the sponges but retained the solid silicone bands, and all healed well. However, silicone sponges are easy to use, create high localized buckles, and do not damage the underlying sclera. Even in infected cases, the underlying sclera usually remains healthy. This series demonstrates the following points. 1. The more sililcone sponge that is used, the higher the chance of infection. Russo and Ruiz and Hahn et ai6 found likewise. 2. Reoperations are more likely to become infected. Ulrich and Burton found likewise. 3. The commonest infecting species is the staphylococcus.2-6therefore the most useful prophylactic antibiotics are methicillin, erythromycin, tetracycline and the aminoglycosides (neomycin, gentamicin, tobramycin). 4. The sponge should be removed as soon as infection is diagnosed, to reduce the chance of premacular fibrosis. Premacular fibrosis has not been mentioned in the literature as a problem associated with sponge infection. In this series it accounted for very poor vision in three cases with attached retinas. It is a recognised cause of poor vision after uncomplicated detachment surgery, but occurs also in vitreous inflammatory disease. * Thus it was not unexpected that the three cases in which it occurred had had the sponge in situ for four to nine weeks after infection had become manifest, although this time-fibrosis relationship could not be statistically validated. Nevertheless it seems reasonable to advise early removal of infected sponges. Early removal of infected sponges may also reduce the chance of massive periretinal proliferation (MPP). Ulrich and Burton removed only 54% of their infected sponges inside three months. Thirteen of the 37 (35%) developed MPP before the sponge was removed. In the present series all the sponges were removed inside 10 weeks, and only three of 14 (21%) developed MPP. Earlier removal does increase the chance of redetachment, but lessening the risk of 71

premacular fibrosis and MPP justifies the risk. Schwarz and Pruett9 removed buckles from 152 eyes for various reasons. When buckles were removed inside six months, 39% redetached; when removed between six and 12 months, 21.4% redetached; and when removed later than one year, 5.8% redetached. The overall detachment rate was 14.5%, but when only infected cases were analysed the detachment rate was 28% after sponge removal. The value of preoperative and intraoperative antibiotics has been established. Two rabbit experiments are relevant. Buettner et al sutured sponges to the scleras of rabbits. Half the sponges had been squeeze-soaked in a solution of neomycin, and the other half were soaked in neomycin without squeezing. There was a significantly higher neomycin level in the squeezesoaked sponges when they were later removed from the sclera. Doft et a19 immersed sponges in suspensions of Staphylococcus aureus and then sutured them to the sclera of rabbits. Twentyfour hours later the sponges were removed, and colony counts performed. Bacterial growth was eliminated by preoperative subconjunctival gentamicin. Bacterial growth was reduced but not eliminated by preoperative gentamicin eyedrops, by preoperative systemic gentamicin, and by postoperative systemic gentamicin. Postoperative subconjunctival steroids did not increase the bacterial growth. Hahn et a16 studied a series of 900 consecutive operations in New York and 1000 consecutive operations in Bonn, done during the same period. Both used preoperative local antibiotics and both soaked and squeezed their sponges in antibiotic solutions. The New York patients had no postoperative antibiotics, whereas the Bonn patients had systemic penicillin for five days and antibiotic eyedrops for five weeks. The Bonn patients had a conjunctival smear on the morning of surgery, and if bacteria were present, surgery was delayed. Between the two groups there was no significant difference in the rates of infection, which were 3.4% in the New York patients and 2.7% in the Bonn patients. This large study concluded that the rate of infection was reduced neither by using postoperative prophylactic antibiotics nor by requiring a sterile conjunctiva preoperatively. Ulrich and Burton had 37 cases of infection, 31 of which had had a routine preoperative conjunctival swab. In only five of the 31 cases was the infecting organism the same as the preoperative organism. Those cases in which there had been a positive preoperative swab had a higher chance of infection, but excluding them would have reduced the infection rate by only 0.4%. Thus preoperative conjunctival swabs are worthless. They were not done in the present series. Infected cases had a longer than average operating time in this series. Hilton and Wallyn suggested that a shorter operating time was associated with a reduced frequency of infection. However, operating time is probably not an independent factor, as the more significant factors of increased amounts of silicone sponge and reoperations are both automatically associated with a longer operation time. CONCLUSION Data accumulated from this and previous studies suggest that the following are reasonable guidelines for reducing the incidence of infection and for improving its management. 1. Use preoperative antibiotic drops which are effective against staphylococci, such as methicillin or an aminoglycoside. 2. Squeeze-soak the sponge in a solution of neomycin (or one of the newer aminoglycosides, gentamicin or tobramycin). 3. Use a minimum amount of sponge and, where convenient, use solid silicone rather than sponge. Encircle with bands (straps), not with sponges. 4. Flood the scleral bed with neomycin, or another aminoglycoside, during surgery. 5. Avoid the need for reoperation. 6. When infecton is diagnosed, remove the sponge promptly to minimize the chance of premacular fibrosis and MPP. Other reasonable but untested precautions include: (i) using a limbal incision to ensure a thick continuous covering of the sponge by conjunctiva and Tenon s capsule; (ii) thorough 72 AUSTRALIAN AND NEW ZEALAND JOURNAL OF OPHTHALMOLOGY

preparation of the eyelid margins, including trimming of the lashes and covering the lid margins with a sticky sterile plastic drape. ACKNOWLEDGEMENTS I am indebted to the ophthalmologists who allowed me to operate on their patients, and who provided much of the follow-up information. I thank Mrs Wendy Smith for typing the manuscript. References 1. Hadden OB. Lessons from one hundred and thirty-five consecutive retinal detachment operations. Trans Ophthalmol SOC NZ. 1981; 33: 59-64. 2. Buettner H, Goldstein BG, Anhalt JP. Infection prophylaxis with silastic sponge explants in retinal detachment surgery. Ophthalmol 1981; 2: 71-76. 3. Russo CE, Ruiz RS. Silicone sponge rejection: early and late complications in retinal detachment surgery. Arch Ophthalmol 1971; 85: 647-650. 4. Hilton OF, Wallyn RH. The removal of scleral buckles. Arch Ophthalmol 1978; 96: 2061-2063. 5. Lincoff H, Boras I, McLean JM. Modifications to the Custodis procedure for retinal detachment. Arch Ophthalmol 1965; 73: 160-163. 6. Hahn YS, Lincoff BS, Lincoff H, Kreissig I. Infection after sponge implantation for scleral buckling. Am J Ophthalmol 1979; 87; 180-185. 7. Ulrich RA, Burton TC. Infections following scleral buckling procedures. Arch Ophthalmol 1974; 92: 213-215. 8. Gass JDM. Stereoscopic atlas of macular diseases. 2nd ed. St Louis: Molsby, 1977: 362. 9. Schwarz PL, Pruett RC. Factors influencing retinal redetachment after removal of buckling elements. Arch Ophthalmol 1977; 95: 803-807. 10. Doft BH, Lipkowitz MD, Kowalski BS, Taylor F. An experimental model to assess factors associated with scleral buckle infection. Retina 1983; 3: 212-217. One of the most interesting treatments for essential blepharospasm with and without oromandisular dystonia is subcutaneous injection of purified botulinum toxin. Although this procedure, originally developed by Dr Alan Scott, is of short term benefit, lasting about 3-4 months, it takes only a few minutes to perform, requires no anaesthesia, and appears to be perfectly safe. While the long term effects of repeated botulinum injections are unknown, this treatment may represent the optimum form of therapy currently available for patients with essential blepharospasm and Meige s Syndrome. p. 979. 4th Edition. Walsh & Hoyt s Clinical Neuro-Ophthalmology, 1985. INTtC rlon A1 I1 K KI I INAL UI I \( tl\ll N I \UKOLKY 73