- Email: [email protected]
Visual and anatomic outcomes associated with posterior segment complications after ganciclovir implant procedures in patients with AIDS and cytomegalovirus retinitis
Visual and Anatomic Outcomes Associated With Posterior Segment Complications After Ganciclovir Implant Procedures in Patients With AIDS and Cytomegalovirus Retinitis JENNIFER I. LIM, MD, RICHARD A. WOLITZ, MD, ANNA H. DOWLING, BS, HOWARD R. BLOOM, MD, ALEX R. IRVINE, MD, AND DANIEL M. SCHWARTZ, MD
● PURPOSE:
To determine anatomic and visual acuity outcomes of posterior segment complications after ganciclovir implant surgery. ● METHODS: We reviewed the medical records of 63 patients with acquired immunodeficiency syndrome who had active cytomegalovirus retinitis in 82 eyes and who underwent 110 consecutive ganciclovir implant procedures. Preoperative and postoperative visual acuity, type of postoperative complication, treatment, and lines of visual acuity change were determined. ● RESULTS: Thirty-eight eyes of 19 patients had bilateral ganciclovir implant procedures, and 25 eyes of 19 patients underwent two or more ganciclovir implant procedures. Thirteen (12%) of 110 ganciclovir implant procedures developed posterior segment complications: rhegmatogenous retinal detachment in six, vitreous hemorrhage in four, endophthalmitis in two, and cystoid macular edema with epiretinal membrane in one. Treatment included pars plana vitrectomy with silicone oil in two cases and without silicone oil in three cases, scleral buckling in one, intravitreal antibiotic injection in two, and laser photocoagulation in two. Overall, median visual acuity was 20/25 preoperatively. Median follow-up was 6 months for all eyes and 7 months for eyes with complications. Postoperative median visual acuity was 20/25 for eyes without complications vs 20/40 at 1 month, 20/60 at 3 and 6 months, and 20/100 at 12 months in eyes with complications (P < .001). More eyes with than without complications lost 2 or more lines of visual acuity (P < .001). Accepted for publication Oct 2, 1998. From the Department of Ophthalmology, Permanente Medical Group, San Francisco, California (Drs Lim, Wolitz, and Bloom and Ms Dowling); and the Department of Ophthalmology, University of California San Francisco, San Francisco, California (Drs Irvine and Schwartz). Reprint requests to Jennifer I. Lim, MD, Department of Ophthalmology, University of Southern California, Doheny Eye Institute, 1450 San Pablo St, #4703, Los Angeles, CA 90033; fax: (323) 442-6460.
288
©
1999 BY
● CONCLUSION: Postoperative complications occurred in
12% of the ganciclovir implant procedures and were associated with decreased visual acuity despite treatment. (Am J Ophthalmol 1999;127:288 –293. © 1999 by Elsevier Science Inc. All rights reserved.)
T
HE USE OF THE GANCICLOVIR IMPLANT TO TREAT
cytomegalovirus (CMV) retinitis has been shown to be effective in a randomly assigned, controlled clinical trial.1 Acute and long-term complications associated with ganciclovir implant surgery include retinal detachment, vitreous hemorrhage, endophthalmitis, and suprachoroidal placement of the implant.1–3 We sought to See also pp. 276 –282, 283–287, 327–328, 329 –339, and 349 –350.
determine the frequency of posterior segment complications after ganciclovir implant procedures for the treatment of CMV retinitis in patients with acquired immunodeficiency syndrome (AIDS). We compared anatomic and visual acuity outcomes of eyes with complications vs outcomes in eyes without these complications.
PATIENTS AND METHODS WE REVIEWED VISUAL AND ANATOMIC RESULTS OF 110
consecutive ganciclovir implant procedures performed in 63 patients at Kaiser Permanente Medical Center, San Francisco, California, and at the University of California San Francisco over a period of 2 and a half years. Patients were not being treated with protease inhibitors when they received the ganciclovir implant. Two patients were subsequently placed on protease inhibitors. Institutional review board approval for placement of the ganciclovir
ELSEVIER SCIENCE INC. ALL
RIGHTS RESERVED.
0002-9394/99/$20.00 PII S0002-9394(98)00443-7
implant was obtained before the procedures from both institutions. Preoperative and postoperative visual acuity, types of postoperative complications, treatment of these complications, and Karnofsky scores were recorded from the medical records. The ganciclovir implants were inserted into the pars plana region by the technique described by Martin and associates.1 An inferonasal or inferotemporal 5-mm pars plana incision was created, prolapsed vitreous was excised with a vitrectomy cutter, and the ganciclovir implant was placed through the incision. The ganciclovir implant was sewn into place with an 8-0 nylon suture placed through the strut of the implant. The wound was closed with an 8-0 nylon suture. The number of implant procedures per eye was recorded. Time intervals were calculated both between the last follow-up visit and the initial implant procedure and between the onset of the complication and the most recent implant procedure. The lines of visual acuity change were calculated 1, 3, 6, and 12 months after the implant procedure. The specific type and outcome of posterior segment complications were recorded from the medical record. Retinal detachments were defined as rhegmatogenous detachments of any size. Vitreous hemorrhage was recorded in this review if the presence of blood in the vitreous cavity caused a measurable decrease of visual acuity from baseline to 20/100 or worse. Endophthalmitis was defined as exogenous endophthalmitis caused by a presumed microbial infection in which an inflammatory response was present and there were intraocular cells, aqueous or vitreous turbid debris with or without hypopyon, ocular discomfort, decreased visual acuity, chemosis, and media opacity. Cultures were required to have been drawn. An epiretinal membrane was defined as a premacular tissue causing surface wrinkling, microaneurysms, or distortion of retinal vessels with or without macular edema. Cystoid macular edema was defined as macular thickening with the characteristic petaloid leakage pattern, as demonstrated on a fluorescein angiogram.
RESULTS IN 63 PATIENTS, 82 EYES UNDERWENT 110 GANCICLOVIR
implant procedures (Vitrasert; mean release rate, 1.40 mg/h; Chiron, Emeryville, California) between May 1994 and December 1996. Nineteen (38 eyes) of the 63 patients had bilateral ganciclovir implant procedures. Twenty-five eyes underwent more than one ganciclovir implant procedure (22 eyes had two implants, three eyes underwent three implant procedures). Eleven of these eyes had the original implant removed at the time of replacement. There were 62 men and one woman; 47 were white, seven Hispanic, seven Asian, and two African-American. Median age was 39 years (range, 28 to 61 years). Median preoperative visual acuity was 20/25 (range, 20/20 to VOL. 127, NO. 3
20/400). Karnofsky scores ranged from 50 to 90 (median, 70). Four patients had a Karnofsky score of 50 and 11 patients had a score of 90. The median Karnofsky scores were the same for patients with and without complications. Follow-up ranged from 2 weeks (three patients; two patients died and one was lost to follow-up) to 3 years; median follow-up was 6 months. Follow-up was less than 3 months for 12 patients (15 [18%] of the 82 eyes) because of death in all but two patients (three eyes) who were lost to follow-up. Six-month follow-up data were not available for 31 (31 eyes in 22 patients) of the 110 procedures (28%) because of death (26 eyes in 18 patients) or loss to follow-up (five eyes of four patients). Twelve-month follow-up data were not available for 48 (44%) of the 110 procedures because of death (33 eyes in 19 patients), loss to follow-up (six eyes in five patients), or limited follow-up (nine eyes in seven patients). Only five (6%) of the 82 eyes had CMV that was resistant to the ganciclovir implant. Four of these eyes underwent replacement of the ganciclovir implant before the scheduled removal and replacement date. One patient (one of the five eyes) was switched to treatment with cidofovir. These eyes did not have posterior segment complications from the implant. Four other (5%) of the 82 eyes showed progression after initial control of the CMV retinitis and underwent replacement of the implant, with subsequent control of the CMV retinitis. Two of these eyes had posterior segment complications (retinal detachments). Thirteen (12%) of the 110 ganciclovir implant procedures were associated with posterior segment complications, which included rhegmatogenous retinal detachment (six cases), severe vitreous hemorrhage obscuring retinal details (four cases), endophthalmitis (two cases), and epiretinal membrane with cystoid macular edema (one case). Median follow-up time was 7 months for eyes with complications. Three of 11 eyes undergoing removal of the original implant at the time of replacement developed complications, including vitreous hemorrhage and retinal detachment. However, in only two of these three eyes was the complication temporally associated with implant replacement and removal. Treatment of retinal detachments included pars plana vitrectomy with silicone oil (two cases), laser photocoagulation to wall off the area of peripheral retinal detachment (two cases), scleral buckling surgery (one case), and observation (one patient refused any surgical treatment). The eyes with severe vitreous hemorrhage were treated with pars plana vitrectomy (two eyes with persistent visual acuity of hand motions at 2 months) and observation (two eyes cleared by 6 weeks). The two eyes with endophthalmitis underwent intravitreal antibiotic injection. The eye with an epiretinal membrane and cystoid macular edema underwent pars plana vitrectomy with epiretinal membrane peeling. Data for this group of eyes with complications are given in the Table. One patient developed endophthalmitis 11 weeks after the second implant procedure. He had visual acuity of no
GANCICLOVIR IMPLANT COMPLICATIONS
289
TABLE. Posterior Segment Complications After Ganciclovir Implant Procedures: Patient Data and Postoperative Visual Acuities Time Interval: Implant to Complication
Age (yrs)
Eye
Complication
51
RE
RRD macula-off
49 days
39 50 45
LE RE LE
RRD RRD RRD RRD
macula-on macula-off macula-on macula-off
36 days 65 days 6 mos 8 mos later 17 mos 5 mos later 46 days 5 mos 1 day 1 day Intraoperative 11 wks 48 days
39
LE
61 28 46 43 48 38 47
LE RE RE LE LE RE LE
RRD macula-on RRD macula-off RRD macula-on Vitreous hemorrhage Vitreous hemorrhage Vitreous hemorrhage Vitreous hemorrhage Endophthalmitis Endophthalmitis
37
RE
ERM, CME
6 mos
Preintervention VA
Intervention
PPVx, silicone oil tamponade Laser Scleral buckle None PPVx, silcone oil tamponade Laser Refused treatment None PPVx None PPVx None Intravitreous antibiotics Intravitreous antifungal agent PPVx
Best Postintervention VA
Final VA
Postintervention Follow-up (mos)
Other
CF
20/70
20/200
6
Cataract
20/30 CF 20/25 CF
20/70 20/200
20/80 20/400
1* 5
— —
20/70
20/400
17*
Cataract
20/30
20/30
20/30
11
—
20/200 HM 20/400 HM 20/200 NLP 20/400
20/200 20/50 20/70 20/20 20/60 NLP 20/400
HM 20/100 20/70 20/25 20/80 NLP 20/400
14 9 2.5* 14 8 1* 4*
20/200
20/80
20/100
4
— Cataract Cataract — — — — —
CF 5 counting fingers; CME 5 cystoid macular edema; ERM 5 epiretinal membrane; HM 5 hand motions; NLP 5 no light perception; PPVx 5 pars plana vitrectomy; RRD 5 rhegmatogenous retinal detachment; VA 5 visual acuity. *Death of patient limited the follow-up data available.
cations lost 2 or more lines of visual acuity (x2 5 15.4, P , .001). Twenty-one (26%) of the 82 eyes lost more than 2 lines of visual acuity. Eleven (16%) of the 69 eyes without complications and 10 (77%) of the 13 eyes with complications lost more than 2 lines of visual acuity (x2 5 26, P , .001). Two (1.8%) of the 110 eyes (two patients) gained 2 lines of visual acuity after the implant. The preoperative median Snellen visual acuity for the eyes that developed complications was 20/30. At the time of diagnosis of the posterior segment complication, the median visual acuity was 20/200 to 20/400. After treatment of the complication, the median visual acuity improved to 20/100 (Table). Compared with the eyes without complications, significantly more eyes with complications lost 2 or more lines of visual acuity at the time of last available follow-up (x2 5 15.4, P , .001). Eyes with endophthalmitis had the worst final visual acuity (no light perception, 20/400), whereas eyes with vitreous hemorrhage had the best final visual acuity. Follow-up was less than 6 months in two patients because of death. The median time interval between onset of the complication and the last implant procedure was 49 days (mean, 69 days). The median time interval between the onset of the complication and the first implant procedure was 49 days (mean, 129 days). Of the 25 eyes that underwent more than one implant procedure, complications occurred in four eyes. Of these four eyes, the complications occurred
light perception and severe ocular pain. Conjunctival, aqueous, and vitreous specimens were drawn and sent for culture. Gram stains of the intraocular fluid showed many white blood cells and rare gram-positive cocci. Intravitreal ceftazidime (2.2 mg) and vancomycin (1.0 mg) were administered as well as subconjunctival cefazolin (100 mg). The vitreous culture grew Streptococcus viridans. The patient’s pain resolved 1 day later. His implant extruded 11 days later and required surgical removal. The other patient with endophthalmitis had visual acuity of 20/400, mild ocular discomfort, and floaters 48 days after the ganciclovir implant procedure. The diagnosis was presumed to be fungal endophthalmitis. Intravitreal amphotericin (400 mg in 0.1 ml) was injected. Cultures did not show growth. The patient was not taking protease inhibitors. The discomfort and floaters resolved, and visual acuity remained 20/400. Overall, median postoperative visual acuity was 20/25 at 1, 3, 6, and 12 months. Median visual acuity in eyes with posterior segment complications was 20/40 at 1 month, 20/60 at 3 and 6 months, and 20/100 at 12 months. At the last follow-up visit available for each patient, median visual acuities were 20/25 for the entire group and 20/100 for the group with complications. Thirty (37%) of the 82 eyes lost 2 or more lines of visual acuity when initial visual acuity was compared with final visual acuity. Nineteen (28%) of the 69 eyes without complications and 11 (85%) of the 13 eyes with compli290
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
MARCH 1999
after the first procedure in two eyes and after the second and third procedures in two other eyes (P . .05). All eyes with a retinal detachment (except for one eye with zone 3 lesions only) had CMV retinitis in zones 1, 2, and 3 and involvement of more than 25% of the retina by CMV scar formation. Retinal detachments occurred 36 days to 6 months after the most recent implant procedure (median onset, 65 days to 5 months). Eighteen eyes of 18 patients with unilateral CMV retinitis underwent 39 procedures. The median age was 40 years. Median visual acuity was 20/25 preoperatively and was unchanged at 1, 3, 6, and 12 months postoperatively. Three eyes of these patients had three ganciclovir implant procedures per eye; the remaining 15 eyes had two ganciclovir implant procedures per eye. Four eyes suffered posterior segment complications in this group. Complications included late-onset retinal detachment (two cases) and vitreous hemorrhage (two cases). Twenty-six eyes in 26 patients underwent one ganciclovir implant procedure per eye. The median age was 37 years. Median visual acuity was 20/25 preoperatively and remained unchanged 1, 3, 6, and 12 months postoperatively. Follow-up was limited to less than 6 months in 12 of the 26 patients because of death in nine patients and loss to follow-up in three patients. There were five posterior segment complications, which included endophthalmitis (one case), retinal detachment (one case), vitreous hemorrhage (two cases), and epiretinal membrane with cystoid macular edema (one case). Thirty-eight eyes of 19 patients underwent bilateral ganciclovir implant procedures. Seven eyes (five patients) underwent two ganciclovir implant procedures each, resulting in 45 total ganciclovir implant procedures in this group. The median age was 40 years. Median visual acuity was 20/25 and remained stable 1, 3, 6, and 12 months postoperatively. Follow-up was limited to less than 6 months in 10 of the 19 patients because of death in nine patients and loss to follow-up in one patient. There were four posterior segment complications in this group, including retinal detachment (three cases) and endophthalmitis (one case).
DISCUSSION PRIOR RESEARCH ON THE GANCICLOVIR IMPLANT HAS UN-
covered a range of complications involving both anterior and posterior segments.1–3 Retinal detachment, vitreous hemorrhage, endophthalmitis, and suprachoroidal placement have been reported as posterior segment complications. Our series includes two cases of delayed-onset endophthalmitis associated with the ganciclovir implant procedure, a need for pars plana vitrectomy for nonclearing vitreous hemorrhage, the development of cystoid macular edema with epiretinal membrane after ganciclovir implant VOL. 127, NO. 3
surgery, and a lower rate of retinal detachment formation than previously reported. The rate of retinal detachments has been reported to be 11% to 18% with the implant and 15% to 29% without the implant. Overall, retinal detachments have been reported to occur in 26% of eyes during the natural course of CMV retinitis.4,5 Our series of posterior complications shows a lower rate of retinal detachment formation (six [5.5%] of 110) with similar lengths of follow-up. The recent report from the Chiron study showed 21 cases of retinal detachments (11.9%) occurring among 177 eyes treated with ganciclovir implants vs four (5.1%) of 79 eyes treated with intravenous ganciclovir. Analysis showed no significant difference between those groups (P 5 .23).6 The retinal detachment rate has recently been found to be lower than in previous reports using the ganciclovir implant (D Martin, oral communication Oct 28, 1997). Similar to the results described by Martin, the retinal detachments in our cases occurred within 2 months of the implant procedure. In the series by Anand and associates2 describing 30 eyes that had received ganciclovir implants, early complications (within 30 days) included endophthalmitis (one case), vitreous hemorrhage (one case), and progressive retinitis (two cases). Late complications included retinal detachment in three (11%) of 27 eyes at 35 to 140 days. In the series by Marx and associates3 describing use of ganciclovir implants for recurrent CMV retinitis (91 eyes of 70 patients), 17% developed a retinal detachment; the cumulative risk was 23% at 6 months after insertion. Eleven of the 12 retinal detachments occurred in patients with greater than 25% of the retina involved by CMV retinitis at baseline vs 39 of 58 eyes without retinal detachment. All retinal detachments were treated with pars plana vitrectomy and silicone oil. The natural course of CMV retinitis includes the occurrence of a retinal detachment. The ganciclovir implant procedure may influence the timing of the retinal detachment. It is difficult to prove that the retinal detachment is a direct result of the ganciclovir implant procedure. In the initial ganciclovir implant study,1 of the 39 eyes undergoing ganciclovir implant surgery (39 primary and 12 exchange procedures), the authors reported that seven patients (18%) had retinal detachments (with associated posterior vitreous detachment and tears at the border of healed CMV retinitis), one patient had a retinal tear (at the border of atrophic CMV retinitis), one patient had postoperative hypotony, and several patients had vitreous hemorrhage. In our study, cases of retinal detachment occurred throughout the period of review. More cases of complications were not noted earlier in the study. There was no “learning curve” effect detectable during the review period. Our series differs from previous series of patients with ganciclovir implants in the need for pars plana vitrectomy for removal of severe nonclearing vitreous hemorrhage. In the original implant study,1 no eyes required pars plana
GANCICLOVIR IMPLANT COMPLICATIONS
291
vitrectomy for vitreous hemorrhages. In a follow-up article on implant exchanges, however, one patient did undergo a vitreous washout.7 Martin and associates7 reported that vitreous hemorrhage occurred more commonly with second and third implant procedures and that almost all vitreous hemorrhages cleared spontaneously within 28 days. One patient in their follow-up study required a vitreous and anterior chamber washout procedure. In our study, a greater proportion of vitreous hemorrhages was not associated with a second procedure. All four patients with vitreous hemorrhage had the hemorrhage after their first procedure. This may be because the majority of patients did not have the original implant removed at the time of the second implant procedure. Anand and associates2 also reported spontaneous resolution (time not given) in four patients with mild vitreous hemorrhage occurring after ganciclovir implant surgery. Marx and associates3 reported two eyes with vitreous hemorrhages that cleared spontaneously (time not given). Malley and associates8 reported a child with bilateral ganciclovir implants and bilateral vitreous hemorrhages that spontaneously resolved by 5 months after surgery. Unlike the reported case of endophthalmitis developing 48 hours after ganciclovir implant surgery with loss of light perception,2 our two patients with endophthalmitis had delayed-onset endophthalmitis (11 weeks and 7 weeks after the implant was placed), and one patient had stabilization of Snellen visual acuity after treatment. Martin and associates1 reported that almost all eyes developed a temporary decrease in visual acuity immediately after the implantation of the ganciclovir implant. However, final visual acuity was 20/25 or better in all eyes except those with a retinal detachment. These eyes with a retinal detachment were repaired with pars plana vitrectomy using silicone oil in five eyes and perfluoropropane gas in two eyes. In the series by Anand and associates,2 seven of 21 eyes with initial visual acuity of 20/40 or better lost 2 or more lines of visual acuity. The causes included endophthalmitis (one case), retinal detachment (two cases), macular retinitis (one case), and expressive aphasia (one case). One eye with a retinal detachment had stable visual acuity after repair, and one did not undergo treatment. In the series by Marx and associates,3 12 eyes developed a retinal detachment and all underwent pars plana vitrectomy with silicone oil for repair. No mention of visual acuity outcome for these eyes was made in their report. In our series, four of the six eyes with retinal detachment had good final visual acuity after treatment. However, compared with the initial visual acuity, more than 2 lines of visual acuity were lost. Our findings are similar to the previous series that reported marked visual acuity loss in eyes with complications in contrast to the eyes without complications. The follow-up intervals for the groups with complications were similar to those of the group without complications, and thus follow-up interval 292
AMERICAN JOURNAL
differences cannot have influenced final visual acuity measurements. Cystoid macular edema has been reported to occur in patients with AIDS who have CMV retinitis and who have not undergone intraocular procedures.9 –12 Reports have documented the occurrence of cystoid macular edema in patients with inactive CMV retinitis.9,10 One the patients in the series by Silverstein and associates10 had a juxtafoveal epiretinal membrane that caused the cystoid macular edema to be more severe. Recently, both cystoid macular edema and epiretinal membrane have been associated with an immune recovery vitreitis.11,12 Our patient, however, did not have vitreitis and was not taking protease inhibitors at the time of initial onset of the epiretinal membrane and cystoid macular edema. Rather, the cystoid macular edema occurred before the manifestation of the epiretinal membrane. We felt that the cystoid macular edema resulted as a postoperative effect and also from subsequent traction from the epiretinal membrane, which may have developed as a result of intraocular surgery. Our follow-up was limited because of patient mortality. The initial Karnofsky scores show that most of our patients were able to care for themselves; only four of the 63 patients had a Karnofsky score of 50. This group does not appear to differ greatly from those in previous series of patients with AIDS and CMV retinitis who have been treated with ganciclovir implants. Our study is a small series, which may overestimate the rate of complications. However, our study emphasizes that visual acuity outcomes may be severely decreased when the complications do occur. Our patients with complications did not differ from those without complications by baseline characteristics, rate of reactivation, duration of follow-up, or proportion with multiple implant procedures. Visual loss over time cannot be attributed to different follow-up intervals. Our study did not identify any subgroup at higher risk of a posterior segment complication. Eyes with multiple implants procedures did not have a greater rate of complications. The eyes with retinal detachments had CMV in all three zones and greater than 25% retinal area of involvement, in contrast to the other eyes. The 12% rate of posterior segment complications should be recognized. Final visual acuity is greatly decreased compared with the initial visual acuity for eyes with posterior segment complications compared with eyes without complications. Among patients with complications, final visual acuity after treatment was the worst for eyes with endophthalmitis and the best for eyes with vitreous hemorrhage. Retinal detachment, typically occurring 3 to 6 months after the implant surgery, and endophthalmitis result in decreased final visual acuity. Patients should be told of the rate of these complications associated with ganciclovir implant surgery and the visual acuity outcomes. It is important to note that visual acuity can be permanently decreased even with treatment for some of these complications. OF
OPHTHALMOLOGY
MARCH 1999
REFERENCES 1. Martin DF, Parks DJ, Mellow SD, et al. Treatment of cytomegalovirus retinitis with an intraocular sustained-release ganciclovir implant. Arch Ophthalmol 1994;112:1531– 1539. 2. Anand R, Nightingale SD, Fish RH. Control of cytomegalovirus retinitis using sustained release of intraocular ganciclovir. Arch Ophthalmol 1993;111:223–227. 3. Marx JL, Kapusta MA, Patel SS, et al. Use of the ganciclovir implant in the treatment of recurrent cytomegalovirus retinitis. Arch Ophthalmol 1996;114:815– 820. 4. Jabs DA, Enger C, Haller J, deBustros S. Retinal detachments in patients with cytomegalovirus retinitis. Arch Ophthalmol 1991;104:794 –799. 5. Freeman WR, Henderly DE, Wan WL, et al. Prevalence, pathophysiology, and treatment of rhegmatogenous retinal detachment in treated cytomegalovirus patients. Am J Ophthalmol 1987;103:527–536. 6. Musch DC, Martin DF, Gordon JF, et al. Treatment of cytomegalovirus retinitis with a sustained-release ganciclovir implant. N Engl J Med 1997;337:83–90.
7. Martin DF, Ferris FL, Parks DJ, et al. Ganciclovir implant exchange: timing, surgical procedure, and complications. Arch Ophthalmol 1997;115:1389 –1394. 8. Malley DS, Barone R, Heinemann MH. Treatment of bilateral cytomegalovirus retinitis with sustained-release ganciclovir implants in a child. Arch Ophthalmol 1996;122:731– 732. 9. Weinberg DV, Moorthy RS. Cystoid macular edema due to cytomegalovirus retinitis in a patient with acquired immune deficiency syndrome. Retina 1996;16:343–344. 10. Silverstein BE, Smith JH, Sykes SO, et al. Cystoid macular edema associated with cytomegalovirus retinitis in patients with the acquired immunodeficiency syndrome. Am J Ophthalmol 1998;125:411– 415. 11. Zegans ME, Walton RC, Holland GN, et al. Transient vitreous inflammatory reactions associated with combination antiretroviral therapy in patients with AIDS and cytomegalovirus retinitis. Am J Ophthalmol 1998;125:292–300. 12. Karavellas MP, Lowder CY, Macdonald JC, et al. Immune recovery vitritis associated with inactive cytomegalovirus retinitis. Arch Ophthalmol 1998;116:169 –175.
Authors Interactivet We encourage questions and comments regarding this article via the Internet on Authors Interactivet at http://www.ajo.com/ Questions, comments, and author responses are posted.
VOL. 127, NO. 3
GANCICLOVIR IMPLANT COMPLICATIONS
293
● PURPOSE:
To determine anatomic and visual acuity outcomes of posterior segment complications after ganciclovir implant surgery. ● METHODS: We reviewed the medical records of 63 patients with acquired immunodeficiency syndrome who had active cytomegalovirus retinitis in 82 eyes and who underwent 110 consecutive ganciclovir implant procedures. Preoperative and postoperative visual acuity, type of postoperative complication, treatment, and lines of visual acuity change were determined. ● RESULTS: Thirty-eight eyes of 19 patients had bilateral ganciclovir implant procedures, and 25 eyes of 19 patients underwent two or more ganciclovir implant procedures. Thirteen (12%) of 110 ganciclovir implant procedures developed posterior segment complications: rhegmatogenous retinal detachment in six, vitreous hemorrhage in four, endophthalmitis in two, and cystoid macular edema with epiretinal membrane in one. Treatment included pars plana vitrectomy with silicone oil in two cases and without silicone oil in three cases, scleral buckling in one, intravitreal antibiotic injection in two, and laser photocoagulation in two. Overall, median visual acuity was 20/25 preoperatively. Median follow-up was 6 months for all eyes and 7 months for eyes with complications. Postoperative median visual acuity was 20/25 for eyes without complications vs 20/40 at 1 month, 20/60 at 3 and 6 months, and 20/100 at 12 months in eyes with complications (P < .001). More eyes with than without complications lost 2 or more lines of visual acuity (P < .001). Accepted for publication Oct 2, 1998. From the Department of Ophthalmology, Permanente Medical Group, San Francisco, California (Drs Lim, Wolitz, and Bloom and Ms Dowling); and the Department of Ophthalmology, University of California San Francisco, San Francisco, California (Drs Irvine and Schwartz). Reprint requests to Jennifer I. Lim, MD, Department of Ophthalmology, University of Southern California, Doheny Eye Institute, 1450 San Pablo St, #4703, Los Angeles, CA 90033; fax: (323) 442-6460.
288
©
1999 BY
● CONCLUSION: Postoperative complications occurred in
12% of the ganciclovir implant procedures and were associated with decreased visual acuity despite treatment. (Am J Ophthalmol 1999;127:288 –293. © 1999 by Elsevier Science Inc. All rights reserved.)
T
HE USE OF THE GANCICLOVIR IMPLANT TO TREAT
cytomegalovirus (CMV) retinitis has been shown to be effective in a randomly assigned, controlled clinical trial.1 Acute and long-term complications associated with ganciclovir implant surgery include retinal detachment, vitreous hemorrhage, endophthalmitis, and suprachoroidal placement of the implant.1–3 We sought to See also pp. 276 –282, 283–287, 327–328, 329 –339, and 349 –350.
determine the frequency of posterior segment complications after ganciclovir implant procedures for the treatment of CMV retinitis in patients with acquired immunodeficiency syndrome (AIDS). We compared anatomic and visual acuity outcomes of eyes with complications vs outcomes in eyes without these complications.
PATIENTS AND METHODS WE REVIEWED VISUAL AND ANATOMIC RESULTS OF 110
consecutive ganciclovir implant procedures performed in 63 patients at Kaiser Permanente Medical Center, San Francisco, California, and at the University of California San Francisco over a period of 2 and a half years. Patients were not being treated with protease inhibitors when they received the ganciclovir implant. Two patients were subsequently placed on protease inhibitors. Institutional review board approval for placement of the ganciclovir
ELSEVIER SCIENCE INC. ALL
RIGHTS RESERVED.
0002-9394/99/$20.00 PII S0002-9394(98)00443-7
implant was obtained before the procedures from both institutions. Preoperative and postoperative visual acuity, types of postoperative complications, treatment of these complications, and Karnofsky scores were recorded from the medical records. The ganciclovir implants were inserted into the pars plana region by the technique described by Martin and associates.1 An inferonasal or inferotemporal 5-mm pars plana incision was created, prolapsed vitreous was excised with a vitrectomy cutter, and the ganciclovir implant was placed through the incision. The ganciclovir implant was sewn into place with an 8-0 nylon suture placed through the strut of the implant. The wound was closed with an 8-0 nylon suture. The number of implant procedures per eye was recorded. Time intervals were calculated both between the last follow-up visit and the initial implant procedure and between the onset of the complication and the most recent implant procedure. The lines of visual acuity change were calculated 1, 3, 6, and 12 months after the implant procedure. The specific type and outcome of posterior segment complications were recorded from the medical record. Retinal detachments were defined as rhegmatogenous detachments of any size. Vitreous hemorrhage was recorded in this review if the presence of blood in the vitreous cavity caused a measurable decrease of visual acuity from baseline to 20/100 or worse. Endophthalmitis was defined as exogenous endophthalmitis caused by a presumed microbial infection in which an inflammatory response was present and there were intraocular cells, aqueous or vitreous turbid debris with or without hypopyon, ocular discomfort, decreased visual acuity, chemosis, and media opacity. Cultures were required to have been drawn. An epiretinal membrane was defined as a premacular tissue causing surface wrinkling, microaneurysms, or distortion of retinal vessels with or without macular edema. Cystoid macular edema was defined as macular thickening with the characteristic petaloid leakage pattern, as demonstrated on a fluorescein angiogram.
RESULTS IN 63 PATIENTS, 82 EYES UNDERWENT 110 GANCICLOVIR
implant procedures (Vitrasert; mean release rate, 1.40 mg/h; Chiron, Emeryville, California) between May 1994 and December 1996. Nineteen (38 eyes) of the 63 patients had bilateral ganciclovir implant procedures. Twenty-five eyes underwent more than one ganciclovir implant procedure (22 eyes had two implants, three eyes underwent three implant procedures). Eleven of these eyes had the original implant removed at the time of replacement. There were 62 men and one woman; 47 were white, seven Hispanic, seven Asian, and two African-American. Median age was 39 years (range, 28 to 61 years). Median preoperative visual acuity was 20/25 (range, 20/20 to VOL. 127, NO. 3
20/400). Karnofsky scores ranged from 50 to 90 (median, 70). Four patients had a Karnofsky score of 50 and 11 patients had a score of 90. The median Karnofsky scores were the same for patients with and without complications. Follow-up ranged from 2 weeks (three patients; two patients died and one was lost to follow-up) to 3 years; median follow-up was 6 months. Follow-up was less than 3 months for 12 patients (15 [18%] of the 82 eyes) because of death in all but two patients (three eyes) who were lost to follow-up. Six-month follow-up data were not available for 31 (31 eyes in 22 patients) of the 110 procedures (28%) because of death (26 eyes in 18 patients) or loss to follow-up (five eyes of four patients). Twelve-month follow-up data were not available for 48 (44%) of the 110 procedures because of death (33 eyes in 19 patients), loss to follow-up (six eyes in five patients), or limited follow-up (nine eyes in seven patients). Only five (6%) of the 82 eyes had CMV that was resistant to the ganciclovir implant. Four of these eyes underwent replacement of the ganciclovir implant before the scheduled removal and replacement date. One patient (one of the five eyes) was switched to treatment with cidofovir. These eyes did not have posterior segment complications from the implant. Four other (5%) of the 82 eyes showed progression after initial control of the CMV retinitis and underwent replacement of the implant, with subsequent control of the CMV retinitis. Two of these eyes had posterior segment complications (retinal detachments). Thirteen (12%) of the 110 ganciclovir implant procedures were associated with posterior segment complications, which included rhegmatogenous retinal detachment (six cases), severe vitreous hemorrhage obscuring retinal details (four cases), endophthalmitis (two cases), and epiretinal membrane with cystoid macular edema (one case). Median follow-up time was 7 months for eyes with complications. Three of 11 eyes undergoing removal of the original implant at the time of replacement developed complications, including vitreous hemorrhage and retinal detachment. However, in only two of these three eyes was the complication temporally associated with implant replacement and removal. Treatment of retinal detachments included pars plana vitrectomy with silicone oil (two cases), laser photocoagulation to wall off the area of peripheral retinal detachment (two cases), scleral buckling surgery (one case), and observation (one patient refused any surgical treatment). The eyes with severe vitreous hemorrhage were treated with pars plana vitrectomy (two eyes with persistent visual acuity of hand motions at 2 months) and observation (two eyes cleared by 6 weeks). The two eyes with endophthalmitis underwent intravitreal antibiotic injection. The eye with an epiretinal membrane and cystoid macular edema underwent pars plana vitrectomy with epiretinal membrane peeling. Data for this group of eyes with complications are given in the Table. One patient developed endophthalmitis 11 weeks after the second implant procedure. He had visual acuity of no
GANCICLOVIR IMPLANT COMPLICATIONS
289
TABLE. Posterior Segment Complications After Ganciclovir Implant Procedures: Patient Data and Postoperative Visual Acuities Time Interval: Implant to Complication
Age (yrs)
Eye
Complication
51
RE
RRD macula-off
49 days
39 50 45
LE RE LE
RRD RRD RRD RRD
macula-on macula-off macula-on macula-off
36 days 65 days 6 mos 8 mos later 17 mos 5 mos later 46 days 5 mos 1 day 1 day Intraoperative 11 wks 48 days
39
LE
61 28 46 43 48 38 47
LE RE RE LE LE RE LE
RRD macula-on RRD macula-off RRD macula-on Vitreous hemorrhage Vitreous hemorrhage Vitreous hemorrhage Vitreous hemorrhage Endophthalmitis Endophthalmitis
37
RE
ERM, CME
6 mos
Preintervention VA
Intervention
PPVx, silicone oil tamponade Laser Scleral buckle None PPVx, silcone oil tamponade Laser Refused treatment None PPVx None PPVx None Intravitreous antibiotics Intravitreous antifungal agent PPVx
Best Postintervention VA
Final VA
Postintervention Follow-up (mos)
Other
CF
20/70
20/200
6
Cataract
20/30 CF 20/25 CF
20/70 20/200
20/80 20/400
1* 5
— —
20/70
20/400
17*
Cataract
20/30
20/30
20/30
11
—
20/200 HM 20/400 HM 20/200 NLP 20/400
20/200 20/50 20/70 20/20 20/60 NLP 20/400
HM 20/100 20/70 20/25 20/80 NLP 20/400
14 9 2.5* 14 8 1* 4*
20/200
20/80
20/100
4
— Cataract Cataract — — — — —
CF 5 counting fingers; CME 5 cystoid macular edema; ERM 5 epiretinal membrane; HM 5 hand motions; NLP 5 no light perception; PPVx 5 pars plana vitrectomy; RRD 5 rhegmatogenous retinal detachment; VA 5 visual acuity. *Death of patient limited the follow-up data available.
cations lost 2 or more lines of visual acuity (x2 5 15.4, P , .001). Twenty-one (26%) of the 82 eyes lost more than 2 lines of visual acuity. Eleven (16%) of the 69 eyes without complications and 10 (77%) of the 13 eyes with complications lost more than 2 lines of visual acuity (x2 5 26, P , .001). Two (1.8%) of the 110 eyes (two patients) gained 2 lines of visual acuity after the implant. The preoperative median Snellen visual acuity for the eyes that developed complications was 20/30. At the time of diagnosis of the posterior segment complication, the median visual acuity was 20/200 to 20/400. After treatment of the complication, the median visual acuity improved to 20/100 (Table). Compared with the eyes without complications, significantly more eyes with complications lost 2 or more lines of visual acuity at the time of last available follow-up (x2 5 15.4, P , .001). Eyes with endophthalmitis had the worst final visual acuity (no light perception, 20/400), whereas eyes with vitreous hemorrhage had the best final visual acuity. Follow-up was less than 6 months in two patients because of death. The median time interval between onset of the complication and the last implant procedure was 49 days (mean, 69 days). The median time interval between the onset of the complication and the first implant procedure was 49 days (mean, 129 days). Of the 25 eyes that underwent more than one implant procedure, complications occurred in four eyes. Of these four eyes, the complications occurred
light perception and severe ocular pain. Conjunctival, aqueous, and vitreous specimens were drawn and sent for culture. Gram stains of the intraocular fluid showed many white blood cells and rare gram-positive cocci. Intravitreal ceftazidime (2.2 mg) and vancomycin (1.0 mg) were administered as well as subconjunctival cefazolin (100 mg). The vitreous culture grew Streptococcus viridans. The patient’s pain resolved 1 day later. His implant extruded 11 days later and required surgical removal. The other patient with endophthalmitis had visual acuity of 20/400, mild ocular discomfort, and floaters 48 days after the ganciclovir implant procedure. The diagnosis was presumed to be fungal endophthalmitis. Intravitreal amphotericin (400 mg in 0.1 ml) was injected. Cultures did not show growth. The patient was not taking protease inhibitors. The discomfort and floaters resolved, and visual acuity remained 20/400. Overall, median postoperative visual acuity was 20/25 at 1, 3, 6, and 12 months. Median visual acuity in eyes with posterior segment complications was 20/40 at 1 month, 20/60 at 3 and 6 months, and 20/100 at 12 months. At the last follow-up visit available for each patient, median visual acuities were 20/25 for the entire group and 20/100 for the group with complications. Thirty (37%) of the 82 eyes lost 2 or more lines of visual acuity when initial visual acuity was compared with final visual acuity. Nineteen (28%) of the 69 eyes without complications and 11 (85%) of the 13 eyes with compli290
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
MARCH 1999
after the first procedure in two eyes and after the second and third procedures in two other eyes (P . .05). All eyes with a retinal detachment (except for one eye with zone 3 lesions only) had CMV retinitis in zones 1, 2, and 3 and involvement of more than 25% of the retina by CMV scar formation. Retinal detachments occurred 36 days to 6 months after the most recent implant procedure (median onset, 65 days to 5 months). Eighteen eyes of 18 patients with unilateral CMV retinitis underwent 39 procedures. The median age was 40 years. Median visual acuity was 20/25 preoperatively and was unchanged at 1, 3, 6, and 12 months postoperatively. Three eyes of these patients had three ganciclovir implant procedures per eye; the remaining 15 eyes had two ganciclovir implant procedures per eye. Four eyes suffered posterior segment complications in this group. Complications included late-onset retinal detachment (two cases) and vitreous hemorrhage (two cases). Twenty-six eyes in 26 patients underwent one ganciclovir implant procedure per eye. The median age was 37 years. Median visual acuity was 20/25 preoperatively and remained unchanged 1, 3, 6, and 12 months postoperatively. Follow-up was limited to less than 6 months in 12 of the 26 patients because of death in nine patients and loss to follow-up in three patients. There were five posterior segment complications, which included endophthalmitis (one case), retinal detachment (one case), vitreous hemorrhage (two cases), and epiretinal membrane with cystoid macular edema (one case). Thirty-eight eyes of 19 patients underwent bilateral ganciclovir implant procedures. Seven eyes (five patients) underwent two ganciclovir implant procedures each, resulting in 45 total ganciclovir implant procedures in this group. The median age was 40 years. Median visual acuity was 20/25 and remained stable 1, 3, 6, and 12 months postoperatively. Follow-up was limited to less than 6 months in 10 of the 19 patients because of death in nine patients and loss to follow-up in one patient. There were four posterior segment complications in this group, including retinal detachment (three cases) and endophthalmitis (one case).
DISCUSSION PRIOR RESEARCH ON THE GANCICLOVIR IMPLANT HAS UN-
covered a range of complications involving both anterior and posterior segments.1–3 Retinal detachment, vitreous hemorrhage, endophthalmitis, and suprachoroidal placement have been reported as posterior segment complications. Our series includes two cases of delayed-onset endophthalmitis associated with the ganciclovir implant procedure, a need for pars plana vitrectomy for nonclearing vitreous hemorrhage, the development of cystoid macular edema with epiretinal membrane after ganciclovir implant VOL. 127, NO. 3
surgery, and a lower rate of retinal detachment formation than previously reported. The rate of retinal detachments has been reported to be 11% to 18% with the implant and 15% to 29% without the implant. Overall, retinal detachments have been reported to occur in 26% of eyes during the natural course of CMV retinitis.4,5 Our series of posterior complications shows a lower rate of retinal detachment formation (six [5.5%] of 110) with similar lengths of follow-up. The recent report from the Chiron study showed 21 cases of retinal detachments (11.9%) occurring among 177 eyes treated with ganciclovir implants vs four (5.1%) of 79 eyes treated with intravenous ganciclovir. Analysis showed no significant difference between those groups (P 5 .23).6 The retinal detachment rate has recently been found to be lower than in previous reports using the ganciclovir implant (D Martin, oral communication Oct 28, 1997). Similar to the results described by Martin, the retinal detachments in our cases occurred within 2 months of the implant procedure. In the series by Anand and associates2 describing 30 eyes that had received ganciclovir implants, early complications (within 30 days) included endophthalmitis (one case), vitreous hemorrhage (one case), and progressive retinitis (two cases). Late complications included retinal detachment in three (11%) of 27 eyes at 35 to 140 days. In the series by Marx and associates3 describing use of ganciclovir implants for recurrent CMV retinitis (91 eyes of 70 patients), 17% developed a retinal detachment; the cumulative risk was 23% at 6 months after insertion. Eleven of the 12 retinal detachments occurred in patients with greater than 25% of the retina involved by CMV retinitis at baseline vs 39 of 58 eyes without retinal detachment. All retinal detachments were treated with pars plana vitrectomy and silicone oil. The natural course of CMV retinitis includes the occurrence of a retinal detachment. The ganciclovir implant procedure may influence the timing of the retinal detachment. It is difficult to prove that the retinal detachment is a direct result of the ganciclovir implant procedure. In the initial ganciclovir implant study,1 of the 39 eyes undergoing ganciclovir implant surgery (39 primary and 12 exchange procedures), the authors reported that seven patients (18%) had retinal detachments (with associated posterior vitreous detachment and tears at the border of healed CMV retinitis), one patient had a retinal tear (at the border of atrophic CMV retinitis), one patient had postoperative hypotony, and several patients had vitreous hemorrhage. In our study, cases of retinal detachment occurred throughout the period of review. More cases of complications were not noted earlier in the study. There was no “learning curve” effect detectable during the review period. Our series differs from previous series of patients with ganciclovir implants in the need for pars plana vitrectomy for removal of severe nonclearing vitreous hemorrhage. In the original implant study,1 no eyes required pars plana
GANCICLOVIR IMPLANT COMPLICATIONS
291
vitrectomy for vitreous hemorrhages. In a follow-up article on implant exchanges, however, one patient did undergo a vitreous washout.7 Martin and associates7 reported that vitreous hemorrhage occurred more commonly with second and third implant procedures and that almost all vitreous hemorrhages cleared spontaneously within 28 days. One patient in their follow-up study required a vitreous and anterior chamber washout procedure. In our study, a greater proportion of vitreous hemorrhages was not associated with a second procedure. All four patients with vitreous hemorrhage had the hemorrhage after their first procedure. This may be because the majority of patients did not have the original implant removed at the time of the second implant procedure. Anand and associates2 also reported spontaneous resolution (time not given) in four patients with mild vitreous hemorrhage occurring after ganciclovir implant surgery. Marx and associates3 reported two eyes with vitreous hemorrhages that cleared spontaneously (time not given). Malley and associates8 reported a child with bilateral ganciclovir implants and bilateral vitreous hemorrhages that spontaneously resolved by 5 months after surgery. Unlike the reported case of endophthalmitis developing 48 hours after ganciclovir implant surgery with loss of light perception,2 our two patients with endophthalmitis had delayed-onset endophthalmitis (11 weeks and 7 weeks after the implant was placed), and one patient had stabilization of Snellen visual acuity after treatment. Martin and associates1 reported that almost all eyes developed a temporary decrease in visual acuity immediately after the implantation of the ganciclovir implant. However, final visual acuity was 20/25 or better in all eyes except those with a retinal detachment. These eyes with a retinal detachment were repaired with pars plana vitrectomy using silicone oil in five eyes and perfluoropropane gas in two eyes. In the series by Anand and associates,2 seven of 21 eyes with initial visual acuity of 20/40 or better lost 2 or more lines of visual acuity. The causes included endophthalmitis (one case), retinal detachment (two cases), macular retinitis (one case), and expressive aphasia (one case). One eye with a retinal detachment had stable visual acuity after repair, and one did not undergo treatment. In the series by Marx and associates,3 12 eyes developed a retinal detachment and all underwent pars plana vitrectomy with silicone oil for repair. No mention of visual acuity outcome for these eyes was made in their report. In our series, four of the six eyes with retinal detachment had good final visual acuity after treatment. However, compared with the initial visual acuity, more than 2 lines of visual acuity were lost. Our findings are similar to the previous series that reported marked visual acuity loss in eyes with complications in contrast to the eyes without complications. The follow-up intervals for the groups with complications were similar to those of the group without complications, and thus follow-up interval 292
AMERICAN JOURNAL
differences cannot have influenced final visual acuity measurements. Cystoid macular edema has been reported to occur in patients with AIDS who have CMV retinitis and who have not undergone intraocular procedures.9 –12 Reports have documented the occurrence of cystoid macular edema in patients with inactive CMV retinitis.9,10 One the patients in the series by Silverstein and associates10 had a juxtafoveal epiretinal membrane that caused the cystoid macular edema to be more severe. Recently, both cystoid macular edema and epiretinal membrane have been associated with an immune recovery vitreitis.11,12 Our patient, however, did not have vitreitis and was not taking protease inhibitors at the time of initial onset of the epiretinal membrane and cystoid macular edema. Rather, the cystoid macular edema occurred before the manifestation of the epiretinal membrane. We felt that the cystoid macular edema resulted as a postoperative effect and also from subsequent traction from the epiretinal membrane, which may have developed as a result of intraocular surgery. Our follow-up was limited because of patient mortality. The initial Karnofsky scores show that most of our patients were able to care for themselves; only four of the 63 patients had a Karnofsky score of 50. This group does not appear to differ greatly from those in previous series of patients with AIDS and CMV retinitis who have been treated with ganciclovir implants. Our study is a small series, which may overestimate the rate of complications. However, our study emphasizes that visual acuity outcomes may be severely decreased when the complications do occur. Our patients with complications did not differ from those without complications by baseline characteristics, rate of reactivation, duration of follow-up, or proportion with multiple implant procedures. Visual loss over time cannot be attributed to different follow-up intervals. Our study did not identify any subgroup at higher risk of a posterior segment complication. Eyes with multiple implants procedures did not have a greater rate of complications. The eyes with retinal detachments had CMV in all three zones and greater than 25% retinal area of involvement, in contrast to the other eyes. The 12% rate of posterior segment complications should be recognized. Final visual acuity is greatly decreased compared with the initial visual acuity for eyes with posterior segment complications compared with eyes without complications. Among patients with complications, final visual acuity after treatment was the worst for eyes with endophthalmitis and the best for eyes with vitreous hemorrhage. Retinal detachment, typically occurring 3 to 6 months after the implant surgery, and endophthalmitis result in decreased final visual acuity. Patients should be told of the rate of these complications associated with ganciclovir implant surgery and the visual acuity outcomes. It is important to note that visual acuity can be permanently decreased even with treatment for some of these complications. OF
OPHTHALMOLOGY
MARCH 1999
REFERENCES 1. Martin DF, Parks DJ, Mellow SD, et al. Treatment of cytomegalovirus retinitis with an intraocular sustained-release ganciclovir implant. Arch Ophthalmol 1994;112:1531– 1539. 2. Anand R, Nightingale SD, Fish RH. Control of cytomegalovirus retinitis using sustained release of intraocular ganciclovir. Arch Ophthalmol 1993;111:223–227. 3. Marx JL, Kapusta MA, Patel SS, et al. Use of the ganciclovir implant in the treatment of recurrent cytomegalovirus retinitis. Arch Ophthalmol 1996;114:815– 820. 4. Jabs DA, Enger C, Haller J, deBustros S. Retinal detachments in patients with cytomegalovirus retinitis. Arch Ophthalmol 1991;104:794 –799. 5. Freeman WR, Henderly DE, Wan WL, et al. Prevalence, pathophysiology, and treatment of rhegmatogenous retinal detachment in treated cytomegalovirus patients. Am J Ophthalmol 1987;103:527–536. 6. Musch DC, Martin DF, Gordon JF, et al. Treatment of cytomegalovirus retinitis with a sustained-release ganciclovir implant. N Engl J Med 1997;337:83–90.
7. Martin DF, Ferris FL, Parks DJ, et al. Ganciclovir implant exchange: timing, surgical procedure, and complications. Arch Ophthalmol 1997;115:1389 –1394. 8. Malley DS, Barone R, Heinemann MH. Treatment of bilateral cytomegalovirus retinitis with sustained-release ganciclovir implants in a child. Arch Ophthalmol 1996;122:731– 732. 9. Weinberg DV, Moorthy RS. Cystoid macular edema due to cytomegalovirus retinitis in a patient with acquired immune deficiency syndrome. Retina 1996;16:343–344. 10. Silverstein BE, Smith JH, Sykes SO, et al. Cystoid macular edema associated with cytomegalovirus retinitis in patients with the acquired immunodeficiency syndrome. Am J Ophthalmol 1998;125:411– 415. 11. Zegans ME, Walton RC, Holland GN, et al. Transient vitreous inflammatory reactions associated with combination antiretroviral therapy in patients with AIDS and cytomegalovirus retinitis. Am J Ophthalmol 1998;125:292–300. 12. Karavellas MP, Lowder CY, Macdonald JC, et al. Immune recovery vitritis associated with inactive cytomegalovirus retinitis. Arch Ophthalmol 1998;116:169 –175.
Authors Interactivet We encourage questions and comments regarding this article via the Internet on Authors Interactivet at http://www.ajo.com/ Questions, comments, and author responses are posted.
VOL. 127, NO. 3
GANCICLOVIR IMPLANT COMPLICATIONS
293