Visual Loss after Successful Photocoagulation of Choroidal N eovascularization DENNIS P. HAN, MD, JAMES C. FOLK, MD, ANGELA R. BRATfON, MD

Abstract: A retrospective study was performed on 75 eyes with presumed ocular histoplasmosis syndrome and 59 eyes with aging macular degeneration to determine risk factors for visual loss in the early postoperative period after successful laser photocoagulation of primary and recurrent choroidal neovascular membranes (CNVMs). Despite successful obliteration of neovascularization, visual loss of two lines or greater occurred in 18.8% of eyes with presumed ocular histoplasmosis syndrome and 40.4% of eyes with aging macular degen­eration (mean follow-up, 33.6 weeks). Factors which predicted visual loss after treatment included a short distance of CNVMs from the foveal center (P = 0.002), good preoperative visual acuity (P = 0.003), increased area of CNVM (P = 0.025), and postoperative internal limiting membrane wrinkling (P = 0.017). Eyes with CNVMs greater than 0.8 mm2 in area or less than 0.6 mm from the foveal center had at least twice the frequencies of visual loss of eyes without these characteristics. [Key words: aging macular degeneration, argon laser photocoagulation, choroidal neovascularization, complications of laser treatment, internal limiting membrane wrinkling, krypton laser photocoagulation, presumed ocular histoplasmosis syndrome, visual loss.] Ophthalmology 95: 1380-1384, 1988

Argon laser photocoagulation prevents visual loss from choroidal neovascularization in patients with aging mac­ular degeneration, presumed ocular histoplasmosis, and idiopathic neovascularization. 1-4 Despite successful oblit­eration of the neovascularization, however, a number of these eyes will lose vision because of either a direct effect oflaser treatment, or from the underlying disease process. Visual loss from laser treatment may be from direct dam­age to foveal photoreceptors and retinal pigment epithe­lium or from damage to the nerve fiber layer serving foveal

Originally received: November 24, 1986. Revision accepted: June 20, 1988.

From the Department of Ophthalmology, University of Iowa Hospitals and Clinics, Iowa City.

Presented in part at the ARVO Annual Meeting, Sarasota, April 27-May 3,1986.

Supported in part by an unrestricted grant from Research to Prevent Blind­ness and the Retinal Research Fund.

Reprint requests to Dennis P. Han, MD, The Eye Institute, Medical College of Wisconsin, 8700 W. Wisconsin Ave, Milwaukee, WI 53226.

1380

function. Other complications such as subretinal hem­orrhage, internal limiting membrane wrinkling, or retinal pigment epithelial rips may also contribute to visual 10ss.2,4,5 It is not known what effect factors such as the size and location of the choroidal neovascular membrane (CNVM), presence of neurosensory detachment or sub­retinal hemorrhage, or variables of laser treatment have on the postoperative visual outcome after successful obliteration of choroidal neovascularization. We per­formed a retrospective study to determine whether such factors contribute to visual loss in the early postoperative period after argon or krypton laser photocoagulation of primary and recurrent extrafoveal and juxtafoveal cho­roidal neovascular membranes. We also determined the frequency of early visual loss according to clinically dis­cernible preoperative characteristics which were found to significantly influence the postoperative visual outcome.

PATIENTS AND METHODS

We reviewed the records of71 patients (75 eyes) with presumed ocular histoplasmosis syndrome and 59 patients

HAN et al • CHOROIDAL NEOVASCULARIZA TION

(59 eyes) with aging macular degeneration who were treated successfully with argon or krypton laser photo­coagulation at the University ofIowa for CNVMs between June 1979 and January 1985. Only patients with extra­foveal or juxtafoveal primary or recurrent membranes were eligible for the study . We defined successful treatment as ultimate obliteration of the neovascular membrane without subfoveal extension of neovascularization after either a single or multiple treatment session. Fluorescein angiography demonstrated complete obliteration of neo­vascularization at least 1 month after the last treatment session in all cases. Absence of recurrent neovasculari­zation was confirmed by biomicroscopy or angiographic­ally thereafter, and all cases were followed for at least 2 months after treatment. To allow comparison between eyes treated with argon versus krypton laser, patients who had mUltiple treatments were included only if the same laser type was used for all treatments. Patients were ex­cluded if they had retinal pigment epithelial detachments, mUltiple CNVMs, CNVMs located outside the temporal vascular arcades, recurrent CNVMs extending subfov­eally, pretreatment internal limiting membrane wrinkling, or absent or poor quality pretreatment or posttreatment photographs.

The patients' clinical and photographic records were reviewed and examined for the following characteristics to determine their predictive effect on the postoperative change in visual acuity: disease category, preoperative vi­sual acuity, visual acuity at last follow-up visit, distance ofCNVM from foveal center, area (mm2) ofCNVM, area (mm2

) of subretinal blood, presence of neurosensory ret­inal detachment, use of argon versus krypton laser, single versus multiple treatments, peripapillary versus nonper­ipapillary location of CNVM, and the presence of post­operative internal limiting membrane wrinkling. Krypton laser was used for CNVMs which extended within 200 ~m of the fovea. Krypton also was usually used in lieu of argon when the CNVM was surrounded by a moderate or severe amount of subretinal hemorrhage.

Fundus photographs and fluorescein angiograms had been obtained using a Zeiss (Thornwood, NY) No.3, 30° fundus camera (X2.5 linear magnification of the emme­tropic fundus) in all cases. Area (mm2) ofthe CNVM and of subretinal hemorrhage was determined by XI 0 mag­nification of the fluorescein angiograms and color pho­tographs, followed by calculation of the areas using a computerized digitizer pad and dividing by 625 (X2.5 mUltiplied by X 10).2 No attempt was made to correct for the patients' refractive errors. Linear distances were ob­tained using the same method of magnification except that the measured linear distance was divided by 25 (X2.5 multiplied by X 10). The presence of neurosensory retinal detachment was determined from elevation of the retina on stereo photographs. One examiner who was masked to the preoperative patient characteristics determined whether postoperative internal limiting membrane wrin­kling was present using magnified stereo photographs. Preoperative and postoperative visual acuities consisted of the best distance measurements obtained using the pa-

tient's most recent spectacle correction with pinhole im­provement. Visual acuity measurements were converted to 10gIO of the minimal angle of resolution (log MAR), and the postoperative change in visual acuity was deter­mined by subtracting the preoperative value from the postoperative value. Final postoperative visual acuity was obtained at the patient's last recorded clinic visit.

Statistical analysis was performed using multiple step­wise linear regression to determine the predictive effect of the preoperative characteristics on the postoperative change in visual acuity (final postoperative log MAR mi­nus preoperative log MAR). P values greater than 0.05 were considered not significant. To more accurately assess the prevalence of visual loss in the early postoperative period, we excluded all cases with greater than 12 weeks between initial and final laser treatments from subsequent analyses. The chi-square test and Fisher's exact test were performed to compare the overall prevalence of visual loss between eyes in the presumed ocular histoplasmosis syndrome and aging macular degeneration subgroups. Because measurements of the visual change after treat­ment were not distributed normally, the Wilcoxon rank sum test was used to compare the severity of visual loss between the two groups.

RESULTS

The age (mean ± standard deviation) at the time of treatment for patients with presumed ocular histoplas­mosis syndrome was 41.9 ± 13.4 years, and in patients with aging macular degeneration was 72.6 ± 9.1 years. The mean duration of follow-up for all patients was 33.6 weeks (median, 10.4 weeks). The change (mean ± stan­dard deviation) in visual acuity for all patients was 0.102 ± 0.398 log units (equivalent to one line of visual acuity loss). Thirty-one eyes were treated with krypton laser and 103 eyes were treated with argon laser. Nine eyes with CNVMs located within 200 ~m of the foveal center were treated with krypton laser. One hundred eight eyes re­ceived a single treatment; 26 eyes received two or more treatments.

Data for multiple linear regression analysis of visual change versus preoperative characteristics were missing for 7 of the 134 eyes. Factors which predicted a decrease in visual acuity at the last follow-up examination after treatment in the remaining 127 eyes included (1) short distance of CNVM from the foveal center (P = 0.002), (2) good preoperative visual acuity (P = 0.003), (3) in­creased area ofCNVM (P = 0.025), and (4) the presence of postoperative internal limiting membrane wrinkling (P = 0.017). Eyes with aging macular degeneration were as­sociated with a greater postoperative decrease in visual acuity compared with eyes with presumed ocular histo­plasmosis syndrome (P = 0.002). Factors which were not predictive of a postoperative change in visual acuity in­cluded area of subretinal blood, presence of a neurosen­sory retinal detachment, use of argon versus krypton laser, peripapillary location of the CNVM, and single versus

1381

OPHTHALMOLOGY • OCTOBER 1988 • VOLUME 95 • NUMBER 10

multiple treatments. The regression equation predictive of the postoperative change of visual acuity is described as follows:

Postoperative log MAR -Preoperative log MAR = -0.077 + 0.222 (X) + 0.180 (Y) + 0.042 (area of CNVM in mm2) -0.329 (preoperative log MAR) -0.185 (distance ofCNVM from foveal cen­ter in mm)

where X = 1 for presumed ocular histoplasmosis syn­drome and X = 2 for aging macular degeneration, and Y = 1 or 0 if postoperative internal limiting membrane wrinkling was or was not present, respectively. The R2 value for the regression equation incorporating the above five variables was 0.170, indicating that the equation was poorly predictive of the final visual acuity for any given patient.

The frequency of visual loss at the last follow-up ex­amination was determined according to the preoperative characteristics which were found by multiple regression analysis to significantly influence the postoperative visual outcome. We defined visual loss as two or more lines of visual acuity decrease between the preoperative exami­nation and final follow-up examination. Study eyes in each disease category were divided into subgroups ac­cording to the median values, rounded to the nearest tenth, of the distance ofCNVM to foveal center (0.6 mm), area of CNVM (0.8 mm2), and preoperative visual acuity (20/50, not rounded). The median distance to the foveal center and median area of CNVM were approximately equivalent to one-third the disc diameter and one-third the disc area, respectively. Visual loss frequencies at the final follow-up examination for each disease category are presented in Table 1. Eyes with CNVMs greater than 0.8 mm2 in area or less than 0.6 mm from the foveal center were associated with two- or threefold increases, respec­tively, in the frequency of visual acuity loss in both disease categories. In this analysis, in which multiple linear regression was not used, preoperative visual acuity of20/ 50 or greater was associated with an increased frequency of visual loss in only the aging macular degeneration cat­egory.

Fourteen eyes with peripapillary CNVMs were included in the study. Peripapillary CNVMs were significantly larger in area than those of the group as a whole. The area (mean ± standard error of the mean) ofCNVM was 4.61 ± 0.69 mm2 for the peripapillary group versus 1.52 ± 0.14 mm2 for the entire group (P < 0.0005, unpaired t test). However, whether the CNVM was peripapillary in loca­tion or not did not appear to influence the risk of visual loss (multiple regression analysis).

Eighteen eyes with greater than 12 weeks between initial and final laser treatments were excluded from subsequent analysis to more accurately assess the prevalence of visual loss during specified intervals in the early postoperative period. The prevalence of visual loss or gain of two or more lines of visual acuity according to diagnosis and

1382

Table 1. Frequency of Visual Loss' at Last Follow-up Visit According to Selected Preoperative Characteristics

Distance of CNVM to foveal center (JLm)

<600:j: ~600

Preoperative visual acuity ~20/50 <20/50

Area of CNVM (mm2) >0.8:j: :s:0.8

No. of POHS Eyes

(%)

9/28 (32.1) 5/41 (12.1)

10/51 (19.6) 4/18 (22.2)

8/28 (28.5) 6/41 (14.6)

No. of AMD Eyes

(%)

16/26 (61.5) 7/32 (21.8)

14/31 (45.1) 9/27 (33.3)

17/35 (48.5) 6/23 (26.0)

Pt

0.002

0.003

0.025

POHS = presumed ocular histoplasmosis; AMD = aging macular de­generation; CNVM = choroidal neovascular membrane.

, Pertains to two or more lines of visual acuity loss. t Significance of preoperative characteristics according to multiple

stepwise linear regression, POHS and AMD groups combined. :j: Equal to one-third disc diameter.

Table 2. Prevalence of Visual Loss or Gain' by Diagnosis and Length of Follow-up

Prevalence Prevalence of Visual of Visual

Loss Gain

Follow-up No. of eyes No. of Eyes (wks)t Diagnosis (%) (%)

3-8 POHS 13/60 (21.7) 9/60 (15.0) AMD 16/44 (36.4) 9/44 (20.4)

13-26 POHS 6/39 (15.4) 7/39 (17.9) AMD 7/23 (30.4) 3/23 (13.0)

27-52 POHS 5/32 (15.6) 3/32 (9.4) AMD 7/19 (36.8) 2/19 (10.5)

>52 POHS:j: 7/35 (20.0) 10/35 (28.6) AMD:j: 12/24 (50.0) 4/24 (16.7)

POHS = presumed ocular histoplasmosis; AMD = aging macular de­generation.

, Pertains to two or more lines of visual acuity change. t Includes all eyes for which data were available within the deSignated

time period. :j: P value of difference between POHS and AMD prevalence of visual

loss was statistically significant (P = 0.015, chi-square test).

follow-up interval are outlined in Table 2. Although aging macular degeneration eyes had a consistently higher prevalence of visual loss than presumed ocular histoplas­mosis syndrome eyes throughout the study, even as early as the 3- to 8-week interval visit, this difference was sta­tistically significant only in the patients followed for longer than 1 year (P = 0.015, chi-square test). The overall fre-

HAN et al • CHOROIDAL NEOVASCULARIZATION

quency of visual loss of two lines or greater at the last follow-up visit for eyes with presumed ocular histoplas­mosis syndrome was 18.8% (12 of64 eyes), and for eyes with aging macular degeneration was 40.4% (21 of 52 eyes). This difference in visual loss rates between the two groups was statistically significant (P = 0.010, chi-square test). There was no significant difference between the prevalence of visual improvement after laser treatment between the presumed ocular histoplasmosis syndrome and aging macular degeneration categories (Table 2).

Visual loss was also greater in magnitude in the eyes with aging macular degeneration compared with those with presumed ocular histoplasmosis syndrome. The mean and median changes in log MAR in eyes with pre­sumed ocular histoplasmosis syndrome was -0.006 (slight visual gain) and zero (no change), respectively, compared with mean and median changes in eyes with aging macular degeneration of +0.246 (>2-line visual acuity loss) and +0.062 (0.5-line visual loss), respectively. The difference in severity of visual loss between the two groups was sta­tistically significant (P = 0.019, Wilcoxon rank sum test).

DISCUSSION

Our results pertain to visual loss in the early postop­erative period after laser treatment since our follow-up for most cases was relatively short (median, 10.4 weeks). In this study, we found a highly significant relationship between the loss of visual acuity and close proximity of the CNVM to the foveal center. This could best be ex­plained by laser-induced destruction of foveal retina or of the nerve fiber layer subserving it, or relatively acute progression of the underlying disease affecting the fovea. Eyes with large membranes were also associated with a greater risk of visual loss. These eyes may lose vision be­cause the large treatment scotoma makes it difficult to fixate effectively or because there is simply less central or paracentral retina remaining to use for fixation. We found that eyes with CNVMs greater than 0.8 mm2 (approxi­mately one-third disc area) in size or less than 0.6 mm (approximately one-third disc diameter) from the foveal center suffered a two- or threefold increase, respectively, in the frequency of visual acuity loss of two or more lines.

Our data indicate that, in the early postoperative period, eyes with aging macular degeneration are twice as likely to sustain visual loss of two or more lines of visual acuity than eyes with presumed ocular histoplasmosis, despite successful obliteration of the choroidal neovascularization (Table 2). The higher prevalence of visual loss in eyes with aging macular degeneration, which occurred even in the earliest 3- to 8-week interval visit, suggests either that eyes with aging macular degeneration are more susceptible to visual loss induced directly by laser treatment than eyes with presumed ocular histoplasmosis, or that eyes with aging macular degeneration undergo a more rapid deterioration in visual function inherent to the underlying disease. Either of these two events may occur to a greater

degree in eyes with aging macular degeneration compared with those with presumed ocular histoplasmosis syn­drome. From our data, it was not possible to determine the relative contribution of either the underlying disease or the treatment itself to early visual loss. Over a longer duration of follow-up, however, the course of the under­lying disease process most likely contributes to an in­creasing degree to the poorer visual prognosis seen in eyes with aging macular degeneration compared with those with presumed ocular histoplasmosis syndrome.

We found that eyes with good preoperative visual acuity were at greater risk for visual loss than those with poor visual acuity. This finding is similar to that of the Macular Photocoagulation Study Group in the treatment of cho­roidal neovascularization in presumed ocular histoplas­mosis syndrome.4 This observation may be explained by the phenomenon of eccentric fixation, in which patients with initially poor vision have a relatively large area of the macula on which to eccentrically fixate in order to maintain their preoperative level of visual acuity. In con­trast, patients who initially have good foveal function and subsequently lose it cannot maintain their preoperative visual acuity by eccentric fixation.

Absorptive factors influencing the loci of laser energy uptake, such as laser wavelength or the presence of sub­retinal hemorrhage or neurosensory detachment, appear to have little influence on the visual prognosis compared with the previously mentioned factors. Because only the krypton laser was used for CNVMs within 200 Jlm of the foveal center, the effect of wavelength on CNVMs very close to the fovea cannot be assessed by this study. Ab­sorptive factors may also influence the development of postoperative internal limiting membrane wrinkling, which was also associated with a higher risk of visual loss. This complication occurs more frequently in eyes with presumed ocular histoplasmosis syndrome than eyes with aging macular degeneration, in argon-treated versus krypton-treated eyes, and in eyes subject to heavy treat­ment intensities. 5 Because immediate posttreatment pho­tographs documenting the degree of retinal whitening achieved during treatment were not available in a number of cases in our study, treatment intensity was not included as a variable in the multiple regression analysis. Thus, we cannot rule out the possibility that the higher risk of visual loss in eyes with internal limiting membrane wrinkling may be due to heavier treatment intensities rather than the wrinkling itself. Alternatively, these two factors may be interrelated, with heavy treatment resulting in internal limiting membrane wrinkling and subsequent visual loss from distortion of the fovea.

Because of the retrospective nature of our study, factors such as uniformity of length of follow-up, standardization of treatment technique, and randomization to argon ver­sus krypton laser could not be controlled and are limi­tations of the study. Nevertheless, our results indicate that characteristics inherent to both the underlying macular disease and the CNVM itself may influence the risk of visual loss after successful laser treatment. We recommend

1383

OPHTHALMOLOGY • OCTOBER 1988 • VOLUME 95 • NUMBER 10

that patients be counseled about the risk of visual loss even after successful laser treatment, especially if they have aging macular degeneration, good preoperative visual acuity, a large CNVM, or a CNVM close to the center of the fovea.

REFERENCES

1. Macular Photocoagulation Study Group. Argon laser photocoagulation for senile macular degeneration: results of a randomized clinical trial. Arch Ophthalmol1982; 100:912-8.

2. Macular Photocoagulation Study Group. Argon laser photocoagulation

1384

for ocular histoplasmosis: results of a randomized clinical trial. Arch Ophthalmol1983; 101 :1347-57.

3. Macular Photocoagulation Study Group. Argon laser photocoagulation for idiopathic neovascularization: results of a randomized clinical trial. Arch Ophthalmol1983; 101 :1358-61.

4. Macular Photocoagulation Study Group. Krypton laser photocoagu­lation for neovascular lesions of ocular histoplasmosis: results of a randomized clinical trial. Arch Ophthalmol1987; 105:1499-507.

5. Han DP, Folk JC. Internal limiting membrane wrinkling after argon and krypton laser photocoagulation of choroidal neovascularization. Retina 1986; 6:215-9.

6. Krishan NR, Chandra SR, Stevens TS. Diagnosis and pathogenesis of retinal pigment epithelial tears. Am J Ophthalmol 1985; 100:698-707.