Pars Plana Vitrectomy in the Early Treatment Diabetic Retinopathy Study

Pars Plana Vitrectomy in the Early Treatment Diabetic Retinopathy Study ETDRS Report Number 17 Harry W. Flynn, Jr., MD,l Emily Y. Chew, MD, 2 Brad D. Simons, PhD, 2 Franca B. Barton, MS,3 Nancy A. Remaley, MS,2 Frederick L. Ferris III, MD,2 The Early Treatment Diabetic Retinopathy Study Research Group* Background: The Early Treatment Diabetic Retinopathy Study (ETDRS) enrolled 3711 patients with mild-to-severe non proliferative or early proliferative diabetic retinopathy in both eyes. Patients were randomly assigned to aspirin 650 mg/day or placebo. One eye of each patient was assigned randomly to early photocoagulation and the other to deferral of photocoagulation. Follow-up examinations were scheduled at least every 4 months, and photocoagulation was initiated in eyes assigned to deferral as soon as high-risk proliferative retinopathy was detected. Aspirin was not found to have an effect on retinopathy progression or rates of vitreous hemorrhage. The risk of a combined end point, severe visual loss or vitrectomy, was low in eyes assigned to deferral (6% at 5 years) and was reduced by early photocoagulation (4% at 5 years). Vitrectomy was carried out in 208 patients during the 9 years of the study. This report presents baseline and previtrectomy characteristics and visual outcome in these patients. Methods: Information collected at baseline and during follow-up as part of the ETDRS protocol was supplemented by review of clinic charts for visual acuity and ocular status immediately before vitrectomy. Results: Vitrectomy was performed in 208 (5.6%) of the 3711 patients (243 eyes) enrolled in the ETDRS. The 5-year vitrectomy rates for eyes grouped by their initial photocoagulation assignment were as follows: 2.1 % in the early full scatter photocoagulation group, 2.5% in the early mild scatter group, and 4.0% in the deferral group. The 5-year rates of vitrectomy (in one or both eyes) were 5.4% in patients assigned to aspirin and 5.2% in patients assigned to a placebo. The indications for vitrectomy were either vitreous hemorrhage (53.9%) or retinal detachment with or without vitreous hemorrhage (46.1 %). Before vitrectomy, visual acuity was 5/200 or worse in 66.7% of eyes and better than 20/100 in 6.2%. One year after vitrectomy, the visual acuity was 20/ 100 or better in 47.6% of eyes, including 24.0% with visual acuity of 20/40 or better. Conclusions: With frequent follow-up examinations and timely scatter (pan retinal) photocoagulation, the 5-year cumulative rate of pars plana vitrectomy in ETDRS patients was 5.3%. Aspirin use did not influence the rate of vitrectomy. Ophthalmology 1992;99: 1351-1357

Diabetic retinopathy is a leading cause of blindness in the United States. l The Early Treatment Diabetic Retinopathy Study (ETDRS), a multicenter collaborative clinical Originally received: October 18, 1991. Revision accepted: March 27, 1992. I The Department of Ophthalmology, Bascom Palmer Eye Institute, University of Miami School of Medicine, Miami. 2 National Eye Institute, National Institutes of Health, Bethesda. 3 Maryland Medical Research Institute, Baltimore.

trial, was designed to evaluate photocoagulation and aspirin treatment in the management of patients with nonproliferative or early proliferative diabetic retinopathy,z-6 The ETDRS photocoagulation results have been previously reported.?

* A complete list of ETDRS Investigators appears at the end of ETDRS

report number 7 (Ophthalmology 1991;98:741-756). Reprint requests to Biometry & Epidemiology Program, National Eye Institute, Building 31, Room 6A24, 9000 Rockville Pike, Bethesda, MD 20892.

1351

Ophthalmology

Volume 99, Number 9, September 1992

Visual loss from proliferative diabetic retinopathy is often caused by vitreous hemorrhage and tractional and/ or rhegmatogenous retinal detachment. Pars plana vitrectomy is commonly performed to manage these treatable forms ofvisualloss. 8 -1 5 In the ETDRS, 208 (5.6%) ofthe 3711 patients had pars plana vitrectomy. The purpose of this report is to describe the baseline patient and retinopathy characteristics, initial treatment assignments, indications for vitrectomy, and visual acuity outcomes of ETDRS patients undergoing pars plana vitrectomy.

Patients and Methods Between April 1980 and July 1985, the ETDRS research group enrolled 3711 patients with diabetes mellitus meeting the following eligibility criteria: (1) no macular edema and visual acuity of 20/40 or better with moderate or severe nonproliferative diabetic retinopathy or early proliferative diabetic retinopathy, or (2) macular edema and visual acuity of 20/200 or better with mild, moderate, or severe nonproliferative retinopathy or early proliferative diabetic retinopathy. Patients with high-risk proliferative retinopathy as defined by the Diabetic Retinopathy Study were not eligible for entry into the ETDRS. 7 Patients with retinopathy meeting study eligibility requirements in both eyes and with a favorable prognosis for long-term followup were enrolled in the ETDRS and assigned randomly to aspirin, 650 mg per day, or matching placebo. One eye of each patient was assigned randomly to early photocoagulation and the other eye to deferral of photocoagulation. Eyes assigned to early photocoagulation also were assigned randomly to one of two scatter panretinal photocoagulation techniques (full or mild), and eyes with macular edema were assigned to one of two timing strategies for focal photocoagulation (immediate or delayed). The result of this randomization scheme is that four immediate treatment strategies were used. Details of the treatment strategies are outlined in previous ETDRS publications. 2- 7 Best-corrected visual acuity in the ETDRS was measured at baseline and at 4-month intervals after randomization. During each follow-up visit, any ocular surgery (including vitrectomy) that had been performed since the previous follow-up visit was recorded. After study closure in 1989, additional information was collected from the individual clinical centers to determine the patient's visual acuity at the time of vitrectomy and the indications for vitrectomy. The reported indications for vitrectomy were either vitreous hemorrhage or retinal detachment with or without vitreous hemorrhage. For each vitrectomy, operative notes were requested in an effort to ascertain the extent and nature of retinal detachments, iatrogenic complications, use oflaser endophotocoagulation, etc. Because one third of the operative notes were either missing or incomplete, the complete information concerning the details of the macular status and intraoperative complications is not available. Severe visual loss was a major outcome variable in the ETDRS and was defined as a visual acuity of less than

1352

5/200 on 2 consecutive follow-up examinations scheduled at 4-month intervals. 7 The initial ETDRS protocol for management of nonclearing vitreous hemorrhage specified that vitrectomy should generally be postponed until after severe visual loss occurred. After the 1985 report from the Diabetic Retinopathy Vitrectomy Study (DR VS), 10 indicating the benefit of early vitrectomy (1 to 6 months after severe vitreous hemorrhage), the ETDRS vitrectomy recommendations were changed to allow earlier surgery. Because surgery could be performed either 1 month after detection of vitreous hemorrhage or immediately on detection of progressive retinal detachment, many ETDRS patients did not reach the end point of severe visual loss before vitrectomy. For this reason, previous ETDRS reports comparing early photocoagulation with deferral of treatment used a combined end point of severe visual loss or vitrectomy.7 The current report summarizes the visual outcome for those eyes requiring vitrectomy. Comparisons of variables expressed as proportions were made with two-sample Z test of equality of proportions. The Cutler-Ederer actuarial life table method was used to estimate rates offirst occurrence ofvitrectomy. The Mantei-Cox test was used to compare life table curves. Comparisons of continuous variables were based on the twosample t test of equality of means.

Results The ETDRS patients were followed for up to 9 years, during which time pars plana vitrectomy was performed Table 1. Baseline Characteristics of ETDRS Patients in Vitrectomy and No Vitrectomy Groups No Vitrectomy (n = 3503)

Vitrectomy (n = 208)

56.9 75.7

49.0 87.5

±SD)

32.7 ± 16.3

22.5 ± 14.9

<0.0001

± SD)

47.6 ± 14.0

38.2 ± 14.6

<0.0001

29.2 38.7 32.1

51.9 35.1 13.0

<0.0001

27.7

39.1

0.0003

12.3 ± 2.5

13.3 ± 2.7

Male(%) White (%) Age at diagnosis (mean Age at entry (mean Diabetes type* Type I (%) Mixed (%) Type II (%) Positive urine protein (%)

Glycosylated hemoglobin (mean ± SD) (n = 2665)

P Value

0.0259 0.0001

<0.0001

SD = standard deviation. • Definitions used in the EDTRS: Type I = age at diagnosis of diabetes ~ 30 years, started on continuous insulin use within 1 year of diagnosis and percent desirable weight < 1ZOOfo; Type II = age at diagnosis of diabetes > 30 years and never or only intermittently used insulin or age at diagnosis of diabetes >40 years and percent desirable weight ~1200f0. The remaining patients are designated as "mixed."

Flynn et al . Pars Plana Vitrectomy in the ETDRS Table 2. Baseline Characteristics of ETDRS Eyes in Vitrectomy and No Vitrectomy Groups

Visual acuity (mean ± SD) (score of 80 = 20/25) Severe NPDR or worse (%) CSME(%)

No Vitrectomy (n = 7179)

Vitrectomy (n = 243)

81.7 ± 11.2

80.5 ± 11.0

32.6 50.2

58.8 50.2

P Value

P < 0.0001

SD = standard deviation; NPDR = nonproliferative diabetic retinopathy; CSME = clinically significant macular edema.

at least once on 243 eyes of 208 patients. Baseline characteristics for ETDRS patients who either did or did not have a vitrectomy are presented in Table 1. The majority of patients undergoing vitrectomy had type I diabetes mellitus. Patients undergoing vitrectomy were younger at onset of diabetes mellitus and at study entry and were more likely to have proteinuria and higher glycosylated hemoglobin. Baseline ocular characteristics for eyes that either did or did not have vitrectomy during follow-up are presented in Table 2. Eyes undergoing vitrectomy were more likely at baseline to have severe non proliferative or worse retinopathy. There were no differences at baseline in the mean visual acuity scores or percentages with clinically significant macular edema. Twenty percent (42 of 208) of vitrectomies occurred within 2 years of patient entering into the study and 78% (163 of 208) occurred within 5 years (Table 3). High-risk proliferative retinopathy was identified at a scheduled ETDRS examination before vitrectomy in 91 % (221 of 243) of these cases. The average interval between declaration of high-risk proliferative retinopathy and vitrectomy was 20.7 months. In the 22 eyes that did not have high-risk characteristics identified before vitrectomy, 5 of the 22 patients missed the scheduled 4-month visit immediately before the vitrectomy, and 3 of these 5 patients missed the 3 consecutive visits (I-year offollow-up) before the vitrectomy.

The 5-year vitrectomy rates were 2.1 % of eyes assigned to early full scatter photocoagulation as the initial treatment, 2.5% of eyes assigned to early mild scatter photocoagulation, and 4.0% of eyes assigned to deferral. All but 2 of the 89 eyes assigned to immediate photocoagulation received some degree of scatter treatment before vitrectomy. These two eyes received focal treatment first for macular edema in eyes with mild-to-moderate non proliferative retinopathy and scatter photocoagulation was delayed until more severe retinopathy was observed. In these two eyes, vitreous hemorrhage occurred before initiating scatter photocoagulation treatment. According to ETDRS protocol, eyes assigned to deferral of photocoagulation were to have full scatter treatment (~1200 burns) initiated at the first observation of highrisk proliferative retinopathy. Scatter photocoagulation treatment was initiated in 128 (83%) of the 154 eyes assigned to deferral, but only 99 eyes (64%) in the deferral group had received full scatter treatment before the pars plana vitrectomy. In the early photocoagulation treatment group, 69 of 89 eyes (78%) received full scatter prior to vitrectomy. Overall, 88% of vitrectomy cases in the ETDRS received prior scatter photocoagulation treatment, including 69.1 % (168 of 243) receiving full scatter treatment. The 5-year vitrectomy rates (vitrectomy in either eye) for patients assigned to aspirin (5.4%) versus those assigned to placebo (5.2%) showed no statistically significant dif-

Table 3. ETDRS Vitrectomy Events during Annual Follow-up after Study Entry

Interval

Number at Risk

No. of Patients Undergoing First Vitrectomy (percent)

Baseline 1 year 2 years 3 years 4 years 5 years 6 years 7 years >7 years

3711 3696 3560 3398 3203 2630 1882 1155 537

0 10 (4.8) 32 (15.4) 45 (21.6) 49 (23.6) 27 (13.0) 22 (10.6) 15 (7.2) 8 (3.8)

Event Rate (percent)

Cumulative Event Rate (percent)

0.0 0.3 0.9 1.4 1.7 1.2 1.4 1.8 2.9

0.0 0.3 1.2 2.5 4.2 5.3 6.7 8.3 11.0

1353

Ophthalmology

Volume 99, Number 9, September 1992

Table 4. ETDRS Eyes Undergoing Vitrectomy: Immediate Preoperative Visual Acuity Visual Acuity

Number jPercent

Better than 20/100 20/100 to 20/200 <20/200 but >5/200 5/200 or less Not available

15 (6.2) 25 (10.3) 36 (14.8) 162 (66.7) 5 (2.1) 243 (100)

Note: 5/200 or less visual acuity was the eligibility criteria for the DRVS.lO

ference (P = 0.91). In the subgroup of eyes with vitreous hemorrhage alone as the indication for vitrectomy, there also was no statistically significant difference in the 5-year vitrectomy rates for aspirin (2.6%) and (3.1 %) placebo treatment groups (P = 0.35). The indication for vitrectomy in the 243 eyes was vitreous hemorrhage alone in 53.9% of cases and retinal detachment, with or without hemorrhage, in 46.1 %. Retinal detachment with or without vitreous hemorrhage was the indication for surgery in 72 (55.0%) of the 131 eyes of type I diabetic patients compared with 9 (30.0%) of 30 eyes of type II diabetic patients (P = 0.024). Thirty-one (37.8%) of82 eyes of patients in the mixed diabetes group had retinal detachment as the indication for surgery. The preoperative visual acuities in the vitrectomy group are shown in Table 4. Two thirds of cases (162 of 243) had 5/200 or less visual acuity, but 6.2% had visual acuity better than 20/100. Fifty-three eyes had severe visual loss before the vitrectomy. Severe visual loss was defined in the ETDRS as visual acuity less than 5/200 for 2 consecutive follow-up examinations 4 months apart. One third of vitrectomies (81 of 243) were performed before the 1985 DR VS report and two thirds (162 of 243) after. The postoperative visual acuities at 1, 2, and 3 years after vitrectomy are listed in Table 5. Approximately half of the eyes (47.6%) achieved 20/100 or better visual acuity (including 24.0% with 20/40 or better) 1 year after surgery, and these levels of visual acuity remained essentially constant during the 3 years of follow-up. The percentages of eyes with visual acuity of 20/400 or better were 69.3% (year 1), 67.3% (year 2), and 68.8% (year 3). The percentage of eyes with less than 5/200 visual acuity after vitrectomy were 20.2% (year 1),24.7% (year 2), and 27.7% (year 3). Three years of follow-up after surgery are available for fewer than 50% of the vitrectomy patients. After vitrectomy was performed, 57 eyes (23.5%) met the ETDRS definition at least once for severe visual loss, including 22 eyes that had severe visual loss both before and after vitrectomy. The presence or absence of retinal detachment at the time ofvitrectomy did not alter visual acuity outcomes (Tables 6A and 6B). The postvitrectomy visual acuity results comparing eyes that received either less than full scatter « 1200 burns) or no photocoagulation with eyes that received full scatter panretinal photocoagulation before vitrectomy

1354

Table 5. ETDRS Eyes Undergoing Vitrectomy: Postoperative Visual Acuities Given as Percentage at 1, 2, and 3 Years I·Year Expected* Deaths Missed visit Examined 20/20 or better <20/20-20/40 <20/40-20/100 <20/100-20/400 <20/400-5/200 <5/200-LP NLP

= 218) = 8) = 2) = 208) 6.3 17.8 23.6 21.6 10.6 13.9 6.3

(n (n (n (n

2·Year (n = (n = (n = (n = 4.7 23.3 20.0 19.3 8.0 12.0 12.7

168) 14) 4) 150)

3·Year (n = 130) (n= 17) (n = 1) (n = 112) 3.6 17.9 25.0 22.3 3.6 10.7 17.0

LP = light perception; NLP = no light perception. • Expected = number of patients for whom this length of follow-up was possible prior to the study end.

showed no statistically significant difference. Comparison of postvitrectomy visual acuity results between eyes assigned to deferral of photocoagulation and eyes assigned to early photocoagulation also showed no statistically significant difference (data not shown). Neovascular glaucoma (NVG), defined as intraocular pressure greater than or equal to 40 mmHg with new vessels in the anterior segment, occurred in 110 (1.5%) of 7179 eyes in the nonvitrectomy group and in 31 (12.8%) of the 243 eyes undergoing vitrectomy. Four eyes had NVG before vitrectomy and persistent NVG after vitrectomy. Of these 31 eyes, 23 lost light perception during Table 6A. ETDRS Eyes Undergoing Vitrectomy for the Indication Vitreous Hemorrhage: Postoperative Visual Acuities Given as Percentage at 1, 2, and 3 Years I·Year Expected" Deaths Missed visit Examined 20/20 or better <20/20-20/40 <20/40-20/100 <20/100-20/400 <20/400-5/200 <5/200-LP NLP LP

=

(n = (n = (n = (n = 7.0 21.1 21.9 18.4 11.4 13.2 7.0

light perception; NLP

=

119) 4) 1) 114)

2·Year

3·Year

(n = (n = (n = (n = 5.1 27.8 16.5 20.3 6.3 11.4 12.7

(n = (n = (n = (n = 3.8 18.9 26.4 20.8 3.8 7.5 18.9

90) 8) 3) 79)

63) 10) 0) 53)

no light perception.

• Expected = number of patients for whom this length of follow-up was possible prior to the study end.

Flynn et al . Pars Plana Vitrectomy in the ETDRS Table 6B. ETDRS Eyes Undergoing Vitrectomy for the Indication Retinal Detachment: Postoperative Visual Acuities Given as Percentage at 1, 2, and 3 Years I-Year Expected" Deaths Missed Visit Examined 20/20 or better <20/20-20/40 <20/40-20/100 <20/100-20/400 <20/400-5/200 <5/200-LP NLP

(n (n (n (n

= 99) = 4) = 1) = 94)

5.3 13.B 25.5 25.5 9.6 14.9 5.3

2-Year (n (n (n (n

= 7B) = 6) = 1) = 71)

4.2 IB.3 23.9 1B.3 9.9 12.7 12.7

3-Year (n (n (n (n

= 67) = 7) = 1)

= 59)

3.4 16.9 23.7 23.7 3.4 13.6 15.3

LP = light perception; NLP = no light perception. • Expected = number of patients for whom this length follow-up was possible before the study end.

follow-up. After vitrectomy in the ETDRS, 21.3% (10 of 47) of aphakic eyes developed NVG while 11% (21 of 196) ofphakic eyes developed NVG (P = 0.09). The proportion of eyes with observed retinal detachment at the time of vitrectomy that developed NVG postvitrectomy was 11.6% (13 of 112). This was similar to the proportion of eyes without retinal detachment at the time of vitrectomy that developed NVG postvitrectomy (13.7% or 18 of 131). Of eight eyes with persistent retinal detachment after vitrectomy, none developed this definition ofNVG during follow-up.

Discussion The ETDRS provides long-term follow-up on a large group of patients with mild-to-severe non proliferative or early proliferative diabetic retinopathy at study entry. The ETDRS recommendations for photocoagulation and aspirin treatment have been previously published. 2 - 7 The results of early photocoagulation for diabetic retinopathy were reviewed in ETDRS report number 9. ETDRS investigators concluded that provided careful follow-up can be maintained, scatter photocoagulation treatment was not recommended for mild or moderate nonproliferative diabetic retinopathy. When retinopathy is more severe, scatter photocoagulation treatment should be considered and usually should not be delayed if the eye has reached the high-risk proliferative stage. It is stated in ETDRS report number 9 for eyes with more severe retinopathy and macular edema that the 5-year rate for the combined endpoint of severe visual loss or occurrence of vitrectomy was higher (l 0.3%) in eyes assigned to deferral of photocoagulation and was lower (5.6% in full scatter treated eyes to 6.9% in mild scatter treated eyes) in the groups assigned to early scatter photocoagulation treatment.

The current ETDRS report summarizes the patient characteristics and follow-up data on the small percentage of patients undergoing vitrectomy (5.6%). ETDRS patients undergoing vitrectomy were generally white, and were more likely to have type I diabetes, were younger at onset of diabetes mellitus and at study entry, had proteinuria, and had higher glycosylated hemoglobin. This report also reviews the specific indications for and visual acuity outcomes of vitrectomy in ETDRS patients. There were no differences in visual acuity outcomes between those vitrectomy patients who underwent early scatter photocoagulation treatment compared with those who had deferral of scatter photocoagulation. The results of this report do not change the previous recommendations for scatter photocoagulation treatment. The risk of NVG after vitrectomy for complications of proliferative diabetic retinopathy is reportedly reduced by panretinal photocoagulation, by achieving complete retinal reattachment, and by preserving the crystalline " . the Iens. 8 ' 16-18 N earIy aII eyes und ergomg vltrectomy 10 ETDRS received panretinal photocoagulation treatment (either preoperatively, intraoperatively, or both). The number of eyes with persistent postvitrectomy retinal detachment was small in this series, and this occurrence did not appear to increase the risk ofNVG. There was a higher percentage of NVG in aphakic eyes compared with phakic eyes in this series, although the difference was not statistically significant. Because of case selection and methods of recording visual acuity, comparison of this ETDRS vitrectomy series to other reported series is difficult. Table 7 shows a comparison of the outcomes of ETDRS patients undergoing Table 7. Patient Characteristics, Indications for Surgery, and Visual Acuity Outcomes 2 Years after Pars Plana Vitrectomy

Diabetes Types (Percent of Patients) Type I Mixed Type II Indication for Surgery (%) Vitreous hemorrhage Retinal detachment Two-Year Postoperative Visual Acuity (%) ~20/4O

<20/40 to 20/100 <20/100 to 20/400 <20/400 to 5/200 <5/ 200 DRVS

=

DRVS (Early Group)

ETDRS (All Eyes)

(n = 30B)

(n = 20B)

37.7 27.6 34.7

5l.9 35.1 13.0

(n = 30B)

(n = 243)

60,9 39.1

53.9 46.1

(n = 253)

(n = 150)

24.5 17.1 IB,9 3.6 36.0

2B.0 20.0 19.3 8.0 24.7

Diabetic Retinopathy Vitrectomy Study.

1355

Ophthalmology

Volume 99, Number 9, September 1992

vitrectomy to the early vitrectomy group of the DRVS. Although the ETDRS had a higher proportion of type I patients and had a higher proportion of patients with retinal detachment, the visual acuity outcomes were quite similar. In the DRVS, 41.6% of the early vitrectomy group achieved visual acuity of 20/100 or better 2 years after surgery, including 24.5% at 20/40 or better. The visual acuity outcomes were somewhat more favorable in the ETDRS vitrectomy group, in which 48% achieved visual acuity of 20/ 100 or better at the 2-year postoperative examination, including 28% with visual acuity of 20/40 or better. Overall, at the 2-year follow-up, 67.3% ofETDRS vitrectomy cases achieved 20/400 or better visual acuity compared with 60.5% in the DRVS early vitrectomy cases. A poor visual outcome (visual acuity less than 5/200) occurred in 24.7% of the ETDRS postvitrectomy eyes, compared with 36% ofDRVS eyes at the 2-year postoperative examination. To the extent that there are differences in the visual acuity outcomes of the ETDRS patients compared with other reported series, there are a number of factors to consider. First, many patients in the ETDRS had the advantage of improved vitreoretinal technology and methodology.9 Laser endophotocoagulation technology became available in 1983 (after the DRVS and many of the other published series were completed) and was available to the majority of ETDRS patients at the time of their surgery. The use of intraoperative scatter endophotocoagulation is believed to reduce the incidence of recurrent postoperative vitreous hemorrhage. 19 Second, 88% of eyes undergoing vitrectomy in the ETDRS received prior scatter photocoagulation treatment. Sixty-nine percent of eyes received 1200 or more scatter burns before vitrectomy. This prior treatment is generally not the case in other series. 10-15 Third, the retinal status ofETDRS vitrectomy eyes may not be comparable with other series with respect to the number with macular detachment or pre-existing macular edema. In a previous ETDRS report comparing aspirin and placebo treatment groups, the rates of vitreous hemorrhage alone and follow-up visual acuity scores were similar between the two treatment groupS.6 This previous ETDRS report stated that aspirin, in a dosage of 650 mg daily, had no clinically important harmful effects for diabetic patients with retinopathy.6 In the current study, the 5year vitrectomy rates were similar for one or both eyes of patients in the aspirin (5.4%) and placebo (5.2%) groups. For the vitrectomy indication of vitreous hemorrhage without retinal detachment, the 5-year rates were 2.6% in the aspirin group and 3.1 % in the placebo group. These results confirm the previous conclusions of the ETDRS that there are no ocular contraindications to aspirin when required for cardiovascular disease or other medical problems. The stability of results for diabetic retinopathy complications after pars plana vitrectomy has been previously reported. 2o- 22 Provided there were good anatomic and visual acuity results 6 months after surgery, the visual acuity results tended to remain stable for up to 10 years of followup. The ETDRS results are consistent with these reports,

1356

although the available 3-year follow-up after vitrectomy included less than 50% of operated cases. In the current ETDRS vitrectomy series, there was a trend over the 3year follow-up period for an increase in the percentage of eyes with no light perception. For the most part, these eyes with no light perception come from the group of cases that had visual acuity of less than 20/400 at the 1year postvitrectomy examination. The clinical significance of the current report is to emphasize that pars plana vitrectomy continues to play an important role in the management of complications from proliferative diabetic retinopathy. ETDRS patients were examined at 4-month intervals after study entry, received full scatter photocoagulation at or before the first observation of high-risk proliferative retinopathy, and still developed visually significant complications from proliferative diabetic retinopathy. This report underscores the fact that adverse events from proliferative diabetic retinopathy can occur in diabetic patients despite early photocoagulation treatment and presents estimates of the risks of complications requiring vitrectomy. Also of clinical significance is the fact that aspirin use did not lead t{:) a higher rate of vitrectomy, either in the overall group or in the subgroup of patients with vitreous hemorrhage.

References I. Klein R, Klein BEK. Vision disorders in diabetes. In: National Diabetes Data Group. Diabetes in America: Diabetes Data Compiled 1984. [Bethesda, MD]: U.S. Depart Health and Human Services. 1985; XII-I-XII-36 (DHHS publ. no. (NIH) 85-1468). 2. The Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study Report Number I. Arch Ophthalmol 1985; 103: 1796-1806. 3. The Early Treatment Diabetic Retinopathy Study Research Group. Treatment techniques and clinical guidelines for photocoagulation of diabetic macular edema. Early Treatment Diabetic Retinopathy Study Report Number 2. Ophthalmology 1987;94:761-74. 4. The Early Treatment Diabetic Retinopathy Study Research Group. Techniques for scatter and local photocoagulation treatment of diabetic retinopathy. Early Treatment Diabetic Retinopathy Study Report Number 3. Int Ophthalmol Clin 1987;27:254-64. 5. The Early Treatment Diabetic Retinopathy Study Research Group. Early Treatment Diabetic Retinopathy Study design and baseline patient characteristics. Early Treatment Diabetic Retinopathy Study Report Number 7. Ophthalmology 1991;98:741-56. 6. The Early Treatment Diabetic Retinopathy Study Research Group. Effects of aspirin treatment for diabetic retinopathy. Early Treatment Diabetic Retinopathy Study Report Number 8. Ophthalmology 1991 ;98:757-65. 7. The Early Treatment Diabetic Retinopathy Study Research Group. Early photocoagulation for diabetic retinopathy. Early Treatment Diabetic Retinopathy Study Report Number 9. Ophthalmology 1991 ;98:766-85. 8. Thompson JT, de Bustros S, Michels RG, Rice TA. Results and prognostic factors in vitrectomy for diabetic traction

Flynn et al . Pars Plana Vitrectomy in the ETDRS

9. 10.

11.

12.

13. 14. 15.

retinal detachment of the macula. Arch Ophthalmol 1987;105:497-502. Aaberg TM, Abrams GW. Changing indications and techniques for vitrectomy in management of complications of diabetic retinopathy. Ophthalmology 1987;94:775-9. The Diabetic Retinopathy Vitrectomy Study Research Group. Early vitrectomy for severe vitreous hemo.rrhage. in diabetic retinopathy. Two-year results ofa randomized tna!. Diabetic Retinopathy Vitrectomy Study Report 2. Arch Ophthalmol 1985;103:1644-52. The Diabetic Retinopathy Vitrectomy Study Research Group. Early vitrectomy for severe proliferative diabetic retinopathy in eyes with useful vision. Results of a randomized trial. Diabetic Retinopathy Vitrectomy Study Report 3. Ophthalmology 1988;95: 1307 -20. The Diabetic Retinopathy Vitrectomy Study Research Group. Early vitrectomy for severe vitreous hemorrhage in diabetic retinopathy. Four-year results ofa randomized tria!. Diabetic Retinopathy Vitrectomy Study Report 5. Arch Ophthalmol 1990; 108:958-64. Michels RG, Rice TA, Rice EF. Vitrectomy for diabetic vitreous hemorrhage. Am J Ophthalmol 1983;95:12-21. Rice TA, Michels RG, Rice EF. Vitrectomy for diabetic traction retinal detachment involving the macula. Am J Ophthalmol 1983;95:22-33. Thompson JT, de Bustros S, Michels RG, et aI. Results of

16.

17. 18.

19.

20.

21. 22.

vitrectomy for proliferative diabetic retinopathy. Ophthalmology 1986;93:1571-4. Wand M, Madigan JC, Gaudio AR, Sorokanich S. Neovascular glaucoma following pars plana vitrectomy for complications of diabetic retinopathy. Ophthalmic Surg 1990;21:113-18. Blankenship G, Cortez R, Machemer R. The lens and pars plana vitrectomy for diabetic retinopathy complications. Arch OphthalmoI1979;97:1263-7. Rice TA, Michels RG, Maguire MC, Rice EF. The effect of lensectomy on the incidence of iris neovascularization and neovascular glaucoma after vitrectomy for diabetic retinopathy. Am J Ophthalmol 1983;95:1-11. Liggett PE, Lean JS, Barlow WE, Ryan SJ. Intraoperative argon endophotocoagulation for recurrent vitreous hemorrhage after vitrectomy for diabetic retinopathy. Am J OphthalmoI1987;103:146-9. Blankenship GW. Stability of pars plana vitrectomy results for diabetic retinopathy complications. A comparison of five-year and six-month postvitrectomy findings. Arch Ophthalmol 1981 ;99: 1009-12. Rice TA, Michels RG. Long-term anatomic and functional results of vitrectomy for diabetic retinopathy. Am J Ophthalmol 1980;90:297-303. Blankenship GW, Machemer R. Long-term diabetic vitrectomy results. Report of 10 year follow-up. Ophthalmology 1985;92:503-6.

1357