Flowsheet depicting systemic abnormalities in 67 consecutive youngerpatients with central retinal vein occlusion. ACE indicates angiotensin-convertingenzyme.
Recchia FM, Carvalho-Recchia CA, Hassan TS. Clinical Course of Younger Patients With Central Retinal Vein Occlusion. Arch Ophthalmol. 2004;122(3):317-321. doi:10.1001/archopht.122.3.317
To describe the clinical course of patients 55 years and younger withcentral retinal vein occlusion (CRVO).
Design and Methods
Retrospective, noncomparative case series. Medical records of 67 patientswere reviewed for demographic, photographic, clinical, and visual acuity (VA)data. Data from 57 patients with at least 6 months of follow-up (mean, 29.2months) were analyzed statistically.
Main Outcome Measures
Best-corrected visual acuity and incidence of intraocular neovascularization.
Of 67 consecutive patients (55% men; mean age, 45 years), the medianpresenting VA was 20/50. Forty-five patients (67%) were found to have at least1 systemic disease. In 57 patients with at least 6 months of follow-up, thefinal VA was 20/40 or better in 42%, 20/50 to 20/100 in 18%, and 20/200 orworse in 40%. Visual decline was most common within 3 to 6 months of CRVOonset. Visual improvement was uncommon after 12 months. Of the 22 patientswith a presenting VA of 20/40 or better, 36% declined to 20/400 or worse atthe most recent examination. Of the 10 patients with a presenting VA of 20/200to 20/400, 8 improved to 20/60 or better. None of the 6 patients with a presentingVA of counting fingers or worse improved. Intraocular neovascularization wasdiagnosed at 1 to 9 months following CRVO in 10 patients (18%). Neovascularizationof the anterior segment developed in 6 patients (11%), including neovascularglaucoma in 3 (5%). The occurrence of neovascularization appeared to be unrelatedto sex, age, presence of associated disease, duration of symptoms, or presentingVA.
Younger patients with CRVO have a variable clinical course. PresentingVA does not appear to be predictive of visual or anatomic outcome. As a significantnumber of patients with good vision at presentation develop legal blindness,therapeutic intervention during periods of visual decline may be consideredin these patients.
Central retinal vein occlusion (CRVO) is a significant cause of visualimpairment and may afflict persons of any age.1,2 Asnew therapeutic options for patients with CRVO emerge, an understanding ofthe condition's clinical course is essential for appropriate timing of anypotential intervention. Large prospective studies (such as the Central VeinOcclusion Study [CVOS] and the Eye Disease Case-Control Study)1- 3 andretrospective case series4 have delineatedthe natural history of the condition and enumerated systemic risk factors.Overall, poor visual outcome (visual acuity [VA] worse than 20/200) and pooranatomic outcome (intraocular neovascularization [NV]) are correlated withpoor VA at presentation. Most patients in these studies, however, were olderthan 55 years, and the conclusions regarding the clinical course of CRVO maynot apply to younger patients. Several studies focusing exclusively on patientsyounger than 50 years have arrived at conflicting conclusions and are limitedby small sample size, nonconsecutive case selection, or incomplete follow-up.The present study was undertaken specifically to evaluate the visual and anatomicoutcomes of younger patients with CRVO and is, to our knowledge, the largestseries published to date of such patients treated at a single institution.
Following approval from the institutional review board of William BeaumontHospital (Royal Oak, Mich), records of all patients seen at Associated RetinalConsultants P.C. from February 1992 through February 2001 with the diagnosisof CRVO were obtained. Only the records of patients aged 18 to 55 years atthe time of presentation were studied further.In all eyes,the diagnosis of CRVO was made clinically at the time of patients' presentationbased on venous dilation and tortuosity, 4 quadrants of intraretinal hemorrhage,and angiographic evidence of impaired venous return. The diagnosisof CRVO was confirmed for the present study in all cases by review of fluoresceinangiograms or color fundus photographs.
The following data were obtained: age; sex; duration of visual impairment;medical history; results of systemic medical evaluation; best-corrected VA(baseline and at 1, 3, 6, and 12 months, and at the most recent follow-up);occurrence of intraocular NV; and results of treatment. Presumed date of onsetof CRVO was calculated by the duration of patients' symptoms.
Visual acuity measurements were converted to logMAR values for statisticalcomparison. Counting fingers (CF) and hand motions (HM) were assigned logMARvalues of 2.0 and 3.0, respectively. Means were compared using the unpaired t test. Frequencies were compared using the χ2 or Fisher exact test as appropriate. Visual acuity trends were analyzedusing the Mann-Whitney U test. A P value of .05 was considered statistically significant.
A total of 67 consecutive patients aged 18 to 55 years were identifiedin the 10-year study period (Table 1).The mean age was 45 years (range, 20-55 years; median, 48 years). Thirty-sevenpatients (55%) were men, and 30 (45%) were women. The mean duration of visualimpairment was 26 days (range, 1 day to 6 months; median, 12 days).
In 36 of the 67 patients (54%), a systemic disease was known at thetime of presentation (Figure 1).All 67 patients were asked specifically about a known diagnosis of hypertension,diabetes mellitus, and hypercholesterolemia: 15 patientshad hypertension; 8 had diabetes mellitus; and 4 had hypercholesterolemia. In the 31 patients with no known systemic disease at presentation,a systemic medical evaluation was performed in 30, and a systemic abnormalitywas discovered in 9 (30%). No patient had proliferative diabetic retinopathy.
Fifty-seven patients (85%) had been followed-up for at least 6 months,and 50 patients (75%) had been followed-up for at least 12 months. There wasno statistical difference in demographics, presenting VA, or incidence ofsystemic disease between the patients with at least 6 months of follow-upand those with shorter follow-up. Data and statistical analyses regardingvisual and anatomic outcomes presented in this study pertain to the groupof 57 patients followed-up for at least 6 months.
The median presenting VA was 20/50. Twenty-two patients (39%) had aVA of 20/40 or better. Nineteen patients (33%) had a VA of 20/50 to 20/100,and 16 (28%) had a VA of 20/200 to light perception. There was no significantdifference in age, sex, or incidence of systemic disease among these 3 groups.Patients with the worst vision tended to be older and female. Among all 57patients analyzed, the final VA was 20/40 or better in 24 patients (42%),20/50 to 20/100 in 10 patients (18%), and 20/200 or worse in 23 patients (40%).No patient had a final VA worse than light perception. Overall, VA improvedby at least 3 lines in 13 (23%) of 57 patients and declined by at least 3lines in 16 (28%). In the cohort of 25 patients with follow-up of more than24 months, vision remained unchanged after the 12-month follow-up visit in17 (68%). Only 2 (8%) experienced subsequent visual improvement (2 or morelines), and 6 (24%) experienced further decline (typically to levels of CFto light perception). The most common causes of visual decline were macularedema and vitreous hemorrhage.
In the cohort of 22 patients with initially good VA (20/40 or better),13 (59%) retained VA in that range at the most recent follow-up (mean, 29.2months), while 8 (36%) had worsened to the level of 20/200 or worse (Table 2). The loss of VA was a minimumof 8 lines. In 5 patients (23%), VA was CF or HM. As presented in Table 3, the most significant decline inVA occurred within 5 months of presumed CRVO onset (or within 3 months ofpresentation). Only 1 of 10 patients who had lost VA by the 3-month follow-upeventually regained vision. (In that patient, the VA was 20/40 at presentation,20/80 three months later, 20/40 at the 12-month follow-up, and 20/30 at 33months.)
In the cohort of 25 patients with intermediate VA at presentation (20/50to 20/200), 11 (55%) retained VA in the same range at 12 months, while 8 (40%)improved to 20/40 or better and 5% worsened to 20/400 or worse (Table 4). Visual acuity improved by at least 3 lines in 26% anddeclined by at least 3 lines in 37%. Four patients had a final VA of CF orHM. While the 3-month mark appeared to be significant in this group as well,the cadence of visual change thereafter was more variable. For example, of6 patients who demonstrated visual improvement at 3 months' follow-up (or3-4 months after the presumed onset of CRVO), 5 ultimately retained a VA of20/15 to 20/30. In 12 patients with worse VA (20/100 or worse) at 3 months'follow-up, 5 ultimately regained VA of 20/50 to 20/80, and 7 declined furtherto levels of 20/200 or worse.
In the cohort of 10 patients with initially poor VA (20/400 to HM),30% regained at least 3 lines of VA by 12 months (Table 5). Two patients attained a VA of 20/40 or better, and another2 patients improved to the 20/50 to 20/200 range. In the remaining 6 patients,VA did not improve beyond 20/400. Six patients presented with a VA of CF orHM, and all 6 remained at that level. By contrast, of the 10 patients witha VA of 20/200 or 20/400, 8 (80%) improved to 20/60 or better. Visual acuityimproved during the interval of 1 to 9 months after CRVO onset and was unrelatedto age, sex, or presence of a systemic medical disorder.
Overall, 10 patients (18%) developed intraocular NV, diagnosed 1 to9 months (mean, 5.9 months, median, 6.5 months) following the presumed onsetof CRVO (Table 6). Neovascularizationof the anterior segment developed in 6 patients (11%), including neovascularglaucoma in 3 (5%). Neovascularization of the posterior segment developedin 5 patients (9%). Nine of the 10 patients with NV were treated with panretinalphotocoagulation. Two patients underwent subsequent pars plana vitrectomyfor vitreous hemorrhage. One patient, who developed 1 clock-hour of NV ofthe iris was observed without complication.
Of the 10 patients who developed NV, 4 had presented with an initialVA of 20/40 or better; 2, with an initial VA between 20/50 and 20/100; and4, with an initial VA of CF or worse. One half of the patients were men, and7 (70%) had an associated systemic disease. The final VA in all but 1 patientwas 20/400 or worse.
Table 7 presents a comparisonof the demographic and clinical parameters between the cohort of 10 patientswho developed NV and the cohort of 47 patients who did not. No significantdifference in age, sex, presence of associated disease, duration of visualsymptoms, or presenting VA was detectable.
Central retinal vein occlusion remains a cause of significant visualmorbidity. Although it is diagnosed much more frequently in older individuals,CRVO can affect patients of all ages, with devastating effects on vision.1- 4 Hayreh5 suggested the term optic disc vasculitis to encompass the clinical spectrum of disc swelling, venous abnormalities,intraretinal hemorrhage, and exudates occurring in younger patients. In typeI optic disc vasculitis (similar to the so-called papillophlebits reportedby previous authors6,7), markeddisc edema was the predominant finding, and the visual prognosis was excellent.In type II, more typical clinical and angiographic features of CRVO were observed,and the clinical course was more variable. Only patients with clinical andangiographic findings consistent with CRVO were included in the present study.
Several authors in the last 2 decades have reported on the visual andanatomic outcomes in younger patients with CRVO. Priluck et al8 describedthe long-term course of 42 patients drawn from a cohort of 63 patients, 40years and younger, seen at the Mayo Clinic from 1949 to 1974. However, finalVA was unavailable in 8 patients, and for 28 patients, the most recent ocularexamination was performed elsewhere. Walters and Spalton9 reviewedthe cases of 17 patients aged 40 years or younger, but final visual acuitieswere available in only 11 (65%). Fong et al10 combined39 of their nondiabetic CRVO patients younger than 50 years with 64 such casesfrom other retinal specialists. They did not, however, differentiate visualand anatomic outcomes for patients with different presenting levels of VA.
In the present study, 67 consecutive patients aged 18 to 55 years withCRVO followed-up at a single institution were studied. The presenting VA was20/40 or better in 42% and 20/200 or worse in 28%. These percentages are consistentwith those reported by Walters and Spalton9 andby Fong et al10 in their aggregate of 103 cases.Priluck et al8 reported a higher incidenceof poor presenting VA (50%).
The statistical analyses in the present study apply to the 57 patientswho were followed-up for at least 6 months. Final VA wasreduced to 20/200 or worse in 40%. This overall level of poor finalVA is higher than that in previous reports (range, 9%-32%).8- 10 Theincidence of intraocular NV was 18% in the present series. The rate of neovascularglaucoma (5%) is consistent with previous reports, while the rate of NV ofthe posterior segment (9%) is higher. Intraocular NV wasdiagnosed later than 6 months following the presumed onset of CRVO in 50%of the patients in whom it occurred, irrespective of presenting VA. Thisresult underscores the importance of regular and vigilant examination.
The major aim of the present study was to detail the clinical courseof patients subdivided by presenting VA. Prognostic information availablefrom presenting VA is valuable for patient counseling and treatment recommendations.In the CVOS, for example, among the cohort of patients with a VA of 20/40or better, 65% retained VA in that range, and only 10% declined to levelsworse than 20/200. Patients in the CVOS with a presenting VA of less than20/200 were unlikely to experience visual improvement and were more likelyto develop intraocular NV.3 Our results inyounger patients, however, stand in contrast with those of the CVOS, in that presenting VA does not appear to be predictive of final VA or clinicalcourse.In younger patients with initially good VA (20/40 or better),only 13 (59%) of 22 patients remained at that level while, more strikingly,more than one third worsened to 20/400 or less. Younger patients with initiallypoor VA (20/200 or worse) may not fare as dismally as expected, as one halfimproved, and a full one fourth attained a VA of 20/40 or better. Moreover,the incidence of NV was not significantly different between the cohorts ofpatients with initially good and initially poor VA (18% compared with 25%; P = .70 by the Fisher exact test).
While an increasing number of systemic factors (especially those relatedto thrombophilia)11,12 have beenreported to be associated with CRVO in younger patients, the appropriate extentof systemic evaluation remains unclear. Conclusions regarding the advisabilityof systemic testing based on the current series should be drawn with caution,since testing of matched controls was not performed. Additionally, since practicepatterns evolved with changes in contemporary knowledge and publications overthe 10-year study period, testing was not uniform for all patients. It isnotable, however, that 4 of 30 patients without known prior disease were foundto have common conditions (anemia, hyperglycemia, and hypercholesterolemia)that both have systemic ramifications and are potentially treatable (Figure 1).
Limitations of the present study are mainly those inherent to any retrospectivereview. There is a possibility of selection bias, in that milder cases ofCRVO may not be referred to retina specialists. Long-term data may be biasedtoward more severe cases, as patients with a benign course may not maintainregular examinations. This potential bias was assessed by comparing the groupof 57 patients with at least 6 months of follow-up with the group of 10 patientswith less follow-up; no significant differences in baseline parameters werefound. Standardized fluorescein angiography, as established in the CVOS, wasnot performed in every patient. The quality of angiographic images and therange of funduscopic views provided were not consistently sufficient to allowmeaningful interpretation, thus precluding a rigorous assessment of perfusionor ischemia. However, VA is correlated with extent of perfusion and offersa general indication of perfusion status.1 Thepresent study may thus provide practical information for cases in which fluoresceinangiography may not be readily available.
It is hoped that our results will aid in the counseling of younger patientswith CRVO, especially with respect to the appropriateness and timing of therapeuticintervention. For example, a significant number of younger patients with apresenting VA of 20/200 to 20/400 may recover vision spontaneously, albeitslowly. On the other hand, it appears that a patient with initially good visionwho loses vision after 3 months may be unlikely to recover spontaneously,and visual improvement in any patient is unlikely after 12 months. Thus, interventionsmay be justified at an early stage in such cases.
Corresponding author and reprints: Franco M. Recchia, MD, 8018 MedicalCenter East, Vanderbilt University Medical Center, Nashville, TN 37232 (e-mail: firstname.lastname@example.org).
Submitted for publication May 21, 2003; final revision received October16, 2003; accepted November 6, 2003.
This research was supported in part by the Heed Ophthalmic FoundationCleveland, Ohio, the AOS-Knapp Foundation,Cleveland, and the Ronald MichelsFellowship Foundation, Baltimore, Md (Dr Recchia).
This research was presented in part at the 106th annual meeting of theAmerican Academy of Ophthalmology; October 23, 2002; Orlando, Fla.