| OCR Text |
Show Journal ofNeuro- Ophthatmology 20( 2): 67- 72, 2000. i 2000 Lippincott Williams & Wilkins, Inc., Philadelphia Thrombocytosis in Temporal Arteritis Rising Platelet Counts: A Red Flag for Giant Cell Arteritis Norah S. Lincoff, MD, Philip D. Erlich, MD, and Lorna S. Brass, MD Objectives: To determine whether early recognition and detection of thrombocytosis in patients with giant cell arteritis can help secure an earlier diagnosis, and whether it can help differentiate cases of arteritic optic neuropathy from other forms of optic neuropathy. Methods: Medical and ophthalmologic records from 1993 to 1998 of the authors' patients with biopsy- proven temporal arteritis versus the authors' patients with nonarteritic anterior ischemic optic neuropathy and idiopathic demyelinating optic neuritis were prospectively collected. Past and present blood analyses were collected, and platelet counts were compared between patients with giant cell arteritis and control populations. This was done to determine whether thrombocytosis could help with the diagnosis and differentiation of these different disease states. Results: There was a significant difference in the frequency of thrombocytosis in patients with giant cell arteritis ( 13 out of 19 patients), with or without arteritic ischemic optic neuropathy, as compared with those with nonarteritic anterior ischemic optic neuropathy ( zero out of 30 patients), idiopathic optic neuritis ( zero out of 26 patients), and healthy age- matched controls ( one out of 22 control subjects). This difference was especially helpful in patients whose sedimentation rates were within the normal range ( adjusting for age). Also noted was the finding that the rise in the platelet counts was not acute, but rather it was a slow gradual increase for approximately 12 months before the onset of significant systemic or visual symptoms. Conclusion: Thrombocytosis should be considered an important marker in patients being referred for evaluation of ischemic optic neuropathy, diplopia, amaurosis fugax, headache, or even generalized malaise. Westegren sedimentation rates < 50 mm/ hr are often erroneously viewed as nondiagnostic or equivocal in the elderly and just followed. An over- the- phone review of patients' sedimentation rates, complete blood counts, and platelet counts can lead to expedited evaluation and treatment of patients who may be at high risk of visual loss from temporal arteritis. Thrombocytosis should lower a physician's threshold Manuscript received August 19, 1999; accepted March 6, 2000. From the Departments of Neurology and Ophthalmology ( NSL, PDE), State University of New York at Buffalo School of Medicine, Buffalo, New York; and the Department of Otolaryngology and Head and Neck Surgery ( LSB), Kaiser Permanente, Denver, Colorado. This project was partially supported by a grant from the Vascular Disease Prevention Research Center, Kaleida Health, Buffalo, New York. Address correspondence and reprint requests to Norah S. Lincoff, MD, Department of Neurology, Buffalo General Hospital, 100 High Street, Buffalo, NY 14203. to acutely treat patients for possible arteritic ischemic optic neuropathy until the disease is definitely ruled out. Key Words: Giant cell arteritis- Platelet count- Temporal arteritis- Thrombocytosis. Giant cell arteritis ( GCA) is a systemic vasculitis of medium- and large- sized arteries. Focal granulomatous inflammation within the vessel wall leads to structural changes, which in turn reduces blood flow and causes subsequent ischemia of tissues supplied by the artery. The disease usually manifests itself in patients over the age of 60 and in women more often than men ( 1). The classic symptoms include headache or temple and scalp tenderness, jaw claudication, myalgias, visual disturbance, and other systemic features ( 1). Giant cell arteritis is commonly suspected when these classic symptoms are accompanied by certain laboratory values, including a significantly elevated Westegren sedimentation rate. More recently, other lab values, including an elevated c- reactive protein or an elevated fibrinogen level, have been suggested to be indicative of GCA ( 2). Unfortunately, at times the diagnosis may be obtuse or missed due to a lack of traditional symptoms, signs, or laboratory parameters. Therefore, patients may be erroneously followed with no diagnosis or an incorrect diagnosis until new or more classic symptoms emerge. For example, patients with visual loss may be followed as if suffering from nonarteritic anterior ischemic optic neuropathy ( NAION), and patients with a vague headache but no visual symptoms may be followed as if suffering from sinus headache or depression. A rise in platelet count is another laboratory parameter that has been associated with GCA ( 3- 10). This study was performed to determine whether early recognition of thrombocytosis in GCA can help to differentiate patients with this disease from those with NAION or other forms of optic neuropathy. Specifically, we evaluated whether monitoring sequential platelet counts, especially in patients whose erythrocyte sedimentation rate is normal or only slightly elevated for age, can be considered significant enough to lower a physician's threshold for the diagnosis of GCA. 67 N. S. L1NCOFF ET AL. PATIENTS AND METHODS We examined the case records of 19 consecutive patients with biopsy- proven GCA that presented to our two offices from 1993 to 1998. In all cases, laboratory values, including a complete blood count with platelet count and sedimentation rate, were obtained on the initial day of evaluation and on subsequent follow- up visits after the initiation of corticosteroid therapy. Each patient's record was reviewed for laboratory results from previous admissions or clinic appointments, and each patient's primary care doctor was contacted so that any previous values ( mostly in the 1 year before GCA symptoms) could be obtained. The normal platelet range used for comparing our groups of patients was 150 to 375 x 103 ( JLL; thrombocytosis is defined as a platelet count ^ 375 x 10 ( JLL. The percent of patients with thrombocytosis in the GCA group was compared with the percent for each of the control groups using the Yates X2 test. We prospectively reviewed these 19 patients with biopsy- proven GCA ( mean age, 75; 12 women) and compared the test results with that of 30 patients with NAION ( mean age, 64; 13 women), 26 patients with acute demyelinating optic neuritis ( mean age, 37; 18 women), and 22 healthy age- matched controls ( mean age, 74; 14 women). Most of the cases of NAION were classic, and the patients did not undergo temporal artery biopsy ( TAB); none of these patients turned out to have temporal arteritis ( TA) or developed systemic symptoms to suggest polymyalgia rheumatica or GCA. Twenty-three of the 26 patients with optic neuritis carried the diagnosis of multiple sclerosis, while three of the 26 patients had idiopathic inflammatory optic neuritis. Of these three patients, none were found to have lesions on magnetic resonance imaging or developed any new neurologic symptoms. No patient had abnormal laboratory parameters or other clinical findings to suggest another autoimmune state. None of the 26 patients were on any form of immunosuppressive treatment for at least 3 months before their acute visual loss. Three case reports of GCA are described to elaborate how detection of escalating thrombocytosis can help in the early diagnosis of certain patients with vague systemic symptoms, who may be harbingers of this serious illness. CASE REPORTS Case 1 ( patient HG) A 70- year- old woman with a 5- month history of feeling unwell sought treatment. She experienced vague symptoms of fatigue and unexplained right ear pain. Her review of systems and work- up was only significant for uncontrolled hypertension, mild iron deficiency anemia, and unexplained thrombocytosis in the presence of a sedimentation rate of 52. From March I, 1996 to December 22, 1997, her platelet rate rose slowly from 484 to 887. Only because of temple tenderness in January of 1998 was a TAB performed and found to be positive. All of the patient's symptoms resolved with steroids, and her platelet count slowly dropped to 279 within the next 6 months. Her sedimentation rate dropped to 12 within 2 months of treatment. The patient never experienced visual loss. Case 2 ( patient LK) A 64- year- old woman with a 7- month history of fatigue, jaw heaviness, and night sweats sought treatment. Because of her age, GCA was not suspected by her primary physician. One year before our evaluation, her platelet count was mildly elevated at 352, and 1 month before her appointment, her platelet count climbed from 548 to 636. Results of a TAB were positive, and steroid treatment was initiated. Three months after treatment with oral prednisone, her platelet count was 254. The patient never experienced visual loss. Case 3 ( patient JH) A 71- year- old woman who suffered from complete visual loss in one eye from TA in 1997 sought treatment. At that time, the only other symptom was vague bitemporal pain for 2 weeks. Her sedimentation rate was 67. After a positive TAB result, she was appropriately treated with prednisone for 1 year. In 1999, 1 year after termination of treatment, she experienced amaurosis fugax in her only seeing eye; she had no other systemic symptoms. The only significant laboratory finding was severe recurrent escalating thrombocytosis ( platelet count, 854) in the absence of an elevated sedimentation rate ( 5 mm/ hr). A repeated biopsy revealed recurrent activity of the disease, and she was restarted on prednisone. The patient continues to maintain good vision in her seeing eye. RESULTS Of the 19 patients diagnosed with biopsy- proven GCA, 7 were men ( mean age, 75; range, 63- 87) and 12 were women ( mean age, 73; range, 60- 78). A year- long escalation of the thrombocytosis in 13 of the 19 patients with GCA was noted ( Table 1). No significant change in platelet count was noted in patients with NAION, optic neuritis, or in the age- matched healthy controls. The percent of patients in the GCA group with thrombocytosis was significantly higher ( P < 0.001) than in any of the three other control groups ( Table 2). We found that 13 of the 19 patients with GCA had evidence of escalating thrombocytosis several months to 1 year before our initial evaluation ( Fig. 1). These 13 patients ( four men and nine women) had a mean platelet count of 571 ( range, 422- 887) at or just before our first examination. The other six patients with GCA had stable platelet counts distributed within the normal laboratory range ( 204- 359; mean, 274), although even these counts rose an average of 45 points ( Table 1). The average platelet count for all 19 patients with GCA at or just before our first examination was 477 ( range, 204- 887). The thirteen patients with GCA with thrombocytosis showed a drop in their platelet count to an average of 275 ( range, 118- 437) after corticosteroid treatment in tapering doses. J Neuro- Ophthalmol, Vol. 20. No. 2. 2000 THROMBOCYTOSIS IN TEMPORAL ARTERITIS 69 TABLE 1. Patient data for escalating thrombocytosis in temporal arteritis Name MG WH LK SS CF WP JZ GR JH GD BF ES CV LP GR* SW Dl DC GS Age/ Sex 73/ F lire 64/ F 86/ M 77/ F 69/ M 66/ M 75/ F 71/ F 77/ F 68/ F 78/ F 77/ F 78/ F 63/ M 60/ F 74/ M 87/ M 83/ M Systemic presentation M ( 5 mo), ST ( 1 mo) HA, JC, F, PMR ( 2 wk) • LVA ( OS)( l d) F ( 8 mo), M, JC ( 1 mo) ST, A ( 3 mo), IVA( OU)( l wk) HA, ST, PMR ( 4 mo) HA, JC ( 3 mo) A, F, amaurosis xl ( 6 mo) HA ( 1 mo) HA, A ( 2 mo), JC ( 1 wk) HA, lVA( OS)( l mo) Recurrence at 2 y: amaurosis x3 ( OD) ( 6 mo) PMR ( 1 y), IVA ( OU)( 2 wk) HA ( 1 mo) HA, JC, iVA ( OD)( l mo) HA, iVA ( OD), D ( 1 d) iVA ( OS) ( 6 mo) iVA ( OD) ( 1 d) HA, F ( l mo) HA, iVA ( OS), ST ( 2 wk) amaurosis xl, A ( 2 wk) HA, F, PMR ( 2 wk) iVA ( OS) ( 4 d) JC, iVA ( 1 wk) ESR initial evaluation Westegren ( mm/ hr) 52 118 115 34 96 104 58 101 67 5 52 125 101 118 40 44 50 60 41 30 Baseline platelet count ( xlOVL) 454 357 377 241 - 266 276 298 315 457 280 290 240 296 311 214 251 240 194 162 Platelet count within 12 mo before Dx 569 349 352/ 548 400 340/ 413 625 284/ 456 380 560/ 762 - 434 - 426 352 - 260 - - 224 Day 0 initial evaluation platelet count 887 703 636 630 610 580 542 507 506 854 492 460 450 422 359 304 282 276 219 204 Post- tx platelet count 533/ 456/ 118 437 254/ 339 220/ 325 492/ 372 318 461/ 280 352/ 242 232 351 393/ 285/ 175 348 319/ 233 375/ 227 292/ 207 287/ 205 312/ 236 208 171 144 ESR post- tx 12 ( 2 mo) 18 ( 1 mo) 7( 1 mo) 15( 1 mo) 8 ( 1 mo) 30 ( 3 mo) 11 ( 2 mo) 38 ( 3 mo) 15 ( 5 mo) 1 ( 1 mo) 5 ( 3 mo) 34( 1 mo) 15( 1 mo) 30 ( 1 mo) 20( 1 mo) 35 ( ly) 1 ( 1 mo) 9 ( 2 mo) 11 ( 1 mo) 14 ( 2 mo) 3 ( 1 mo) A, anorexia; D, diplopia; F, fever; HA, headache; JC, jaw claudication; M, malaise; PMR, polymyalgia rheumatica; SOB, shortness of breath; VA, visual acuity; ST, scalp tenderness; 1, decreased; ESR, erythrocyte sedimentation rate; Dx, diagnosis; tx, treatment. Of the 30 patients with NAION ( mean platelet count, 215) and the 26 patients with optic neuritis ( mean platelet count, 250), none were noted to have a significant escalation in platelet count before initial evaluation, and all patients had platelet counts that were within the normal laboratory range before and after their episode of acute optic neuropathy. The mean platelet count in the healthy age- matched control group was 243 ( Table 2). Only one of the 22 age- matched healthy controls was found to have a platelet count above the normal platelet range ( 380; normal range, 150- 375 x 103 \ xlS). In Table 1, nine of the 19 patients with TA are shown to have had Westegren sedimentation rates at or below 52 mm/ hr before treatment with corticosteroids, which TABLE 2, Thrombocytosis in GCA group compared with control groups at time of diagnosis ( time 0) Patient group GCA ( n = 19) NAION ( n = 30) Demyelinating ION ( n = 26) Healthy controls ( n = 22) Thrombocytosis Percent of cases 13/ 19( 68.4) 0/ 30 ( 0) 0/ 26 ( 0) 1/ 22 ( 4.5) no./ (%) Yates corrected x2* - 24.54 21.80 15.77 P value - < 0.001 < 0.001 < 0.001 * Compared with the GCA group. GCA, giant cell arteritis; NAION, nonarteritic anterior ischemic optic neuropathy; ION, ischemic optic neuropathy. may be viewed by some as nonpathologic ( range, 5- 52 mm/ hr). Six of these nine patients experienced severe visual loss in one or both eyes. Only one of these six patients had escalating thrombocytosis ( patient SS). Eight of the 19 patients had sedimentation rates at presentation greater than 90 mm/ hr. Two of these eight patients experienced severe visual loss in one eye. Three of these eight patients had evidence of thrombocytosis. In the NAION group, the mean sedimentation rate was 13 ( range, 1- 35 mm/ hr), and in the healthy age- matched control group, the mean sedimentation rate was 20 ( range, 1^ 46 mm/ hr). Most of our patients with optic neuritis did not have a sedimentation rate drawn; of the few patients who had a sedimentation rate drawn, it was under 20 mm/ hr. Eight of the 19 patients in the GCA group had mild anemia on initial evaluation ( hemoglobin/ hematocrit < 12/ 35). Only one of the 19 patients with GCA had severe anemia ( hemoglobin/ hematocrit = 9.2/ 29.3). Eight of these patients with anemia had evidence of thrombocytosis. Six of the patients in the GCA group had sedimentation rates greater than 100 mm/ hr at presentation. Only two patients had all three abnormal laboratory findings at or before presentation ( i. e., thrombocytosis and anemia in the presence of a markedly elevated sedimentation rate; patients WB and WP). All 19 patients had normalization of their laboratory parameters after corticosteroid treatment; both their sedimentation rates and platelet counts fell as expected, and their hemoglobin J Neuro- Ophlhalmol. Vol. 20, No. 2, 2000 70 N. S. LINCOFF ET AL. 1000 Normal Platelet Range ( 150- 375x103ul) oo- 1 yr. - 1 mo ~+ 1 mo 1 yr oo Period of Acute Presentation Time ~ 2 Normal platelet range ( encompassing 99% of controls ( 77/ 78) and 6/ 19 patients with GCA) • Patients with GCA Fig. 1. Escalating thrombocytosis in patients with GCA in comparison with control subjects. and hematocrit values rose. Most patients' values normalized within 6 months ( 8). DISCUSSION Giant cell arteritis is a systemic disease that is often diagnosed at a time when permanent visual loss occurs in one or both eyes. An elevated platelet count must be viewed as a red flag in patients over the age of 50 who are referred for visual loss, diplopia, or headache. We found that 13 out of 19 patients diagnosed with biopsy-proven TA had evidence of escalating thrombocytosis several months to 1 year before our initial evaluation and examination. Review of their previous laboratory values at the time of referral helped alert us to the possible diagnosis of GCA; those with an elevated platelet count, with or without an elevated sedimentation rate, were treated immediately with corticosteroids while awaiting biopsy. Thirteen out of 19 patients were treated before any permanent visual loss. A few patients experienced subjective symptoms of blurriness or amaurosis fugax, and two patients had cotton wool spots in the absence of optic nerve swelling. The patients with severe visual loss had evidence of pallid optic nerve swelling. Although an association between thrombocytosis and GCA has been reported in the literature ( 1- 10), it is not always recognized as a positive finding in the medical community. Anemia, on the other hand, is a well-accepted finding ( 11). The few reports in the literature concerning thrombocytosis in GCA discuss only the platelet count of patients at the time of initial evaluation, as if it was only an acute phase reaction. No report has ever remarked on the history or chronicity of the thrombocytosis. The average venous platelet count in the population ranges from 214 to 229 ( 3). Age is known to have little to no effect on platelet counts ( 12). The normal laboratory reference value in the literature for platelets is 150 to 375 x 103 ( xL. In 1963, Olhagen ( 5) drew attention to the occurrence of thrombocytosis in patients with GCA; this observation was subsequently confirmed by Hamrin ( 4), who noted that platelet counts were 90,000 higher in patients with polymyalgia rheumatica as compared with control subjects. Another report showed that patients J Neuro- Ophthalmol, Vol. 20, No. 2, 2000 THROMBOCYTOSIS IN TEMPORAL ARTERITIS 71 with TA have a platelet production rate 1.7 times higher than observed in healthy subjects ( 13,14). Reactive thrombocytosis has been described in various other inflammatory diseases, including inflammatory bowel disease, rheumatoid arthritis, Wegener's granulomatosis, and ankylosing spondylitis. It has also been described accompanying malignancy, carcinomatosis, acute blood loss, iron deficiency anemia, tuberculosis, and the postsplenectomy state ( 15,16). None of our patients were known to be suffering from any of these disease states. As in GCA, a normal sedimentation rate does not rule out activity of these diseases, and therefore thrombocytosis also can assist in the early diagnosis of these disease states. The exact pathophysiology that leads to escalating thrombocytosis remains to be elucidated. Several in vivo and in vitro studies suggest a cytokine- mediated process as the underlying mechanism ( Fig. 2). In particular, the proinflammatory cytokine interleukin- 6 has been shown to be present in increased concentrations in patients with GCA ( 17). Interleukin- 6 production originates from the tissue infiltrating macrophages and fibroblasts found within the arterial granulomas of patients with GCA ( 18). Interleukin- 6 and other cytokines are thought to mediate the release of acute phase proteins from hepatocytes ( 19), which cause megakaryocytopoesis in vitro. An association has been shown between thrombocytosis and the ocular and cerebral complications in GCA ( 10), perhaps through an acute thrombotic event. The elevated platelet production rate in patients with GCA has been shown to return to normal after the institution of corticosteroid therapy ( 8). Correlated with this is the finding that interleukin- 6 production is rapidly suppressed in vivo by corticosteroids ( 17,18). Thrombocytosis should be considered an important objective finding in patients referred for evaluation regarding GCA or idiopathic ischemic optic neuropathy. When middle- aged or elderly patients experience vague constitutional symptoms, a complete blood count and Arterial Granulomas/ Macrophage Cells - iProduce IL6 ( Pleotropic Cytokine) 4- Causes Release Acute phase proteins from hepatocytes - » T SED rate ( Proteins, a1, haptoglobin, ceruloplasmin, fibrinogen, C- reactive protein) lActs on Megakaryocytopoiesis in vitro Increases Platelet Count in vivo =? a reactive thrombocytosis iCauses Thrombotic event Fig. 2. Presumed mechanism of thrombocytosis in GCA ( 20). metabolic panel are often drawn before tests, such as a sedimentation rate. Sedimentation rates < 50 are often followed, or viewed as nonpathologic, and can therefore lead to unnecessary morbidity, such as visual loss. Over-the- phone review of patients' complete blood counts and platelet counts can lead to expedited evaluation and care of patients who may be at high risk. Thrombocytosis should be considered a marker of disease, for it may be present before any new symptoms emerge or before any other laboratory parameters become abnormal. This data supports the known fact that GCA is not an acute disease state but rather a chronic progressive inflammatory disease that can culminate in a catastrophic thrombotic event. Though thrombocytosis is seen in other disease states, such as rheumatoid arthritis and cancer, this study demonstrates some clear clinical use of this finding in GCA. Future studies may help elucidate the sensitivity of this finding. In middle- aged and elderly patients with vague systemic symptoms, the finding of thrombocytosis may lead to earlier diagnosis and preemptive treatment of this systemic disease. Acknowledgement: The authors thank Dr. Roger Priore for his statistical analysis and Donald Watkins and Nancy Mac Donald from Art and Photographic Services at the State University of New York at Buffalo. REFERENCES 1. Ghanchi F, Dutton G. Current concepts in giant cell ( temporal) arteritis. Surv Ophthalmol 1997; 42: 99- 123. 2. Hayreh SS, Podhajsky PA, Roman R, et al. Giant cell arteritis: validity and reliability of various diagnostic data. Am J Ophthalmol 1997; 123: 285- 96. 3. Vilen L, Jacobsson S, Wadenvik H, Kutti J. ADP- induced platelet aggregation as a function of age in healthy humans. Thromb Hae-most 1989; 61: 490- 2. 4. Hamrin B. Polymyalgia arteritica. Acta Med Scand Suppl 1972; 533: 100- 30. 5. Olhagen B. Polymyalgia rheumatica: a form of senile arteritis. Acta Rheum Scand 1963; 9: 157- 64. 6. Price N, Clearkin LG. Thrombocytosis and giant cell arteritis ( letter). Lancet 1994; 343: 672. 7. Malmvall BE, Bengtsson BA. Giant cell arteritis. Scand J Rheumatol 1978; 7: 154- 8. 8. Bergstrom AL, Bengtsson BA, Olsson LB, Malmvall BE, Kutti J. Thrombokinetics in giant cell arteritis, with special reference to corticosteroid therapy. Ann Rheum Dis 1979; 38: 244- 7. 9. Krishna R, Kosmorsky G. Implications of thrombocytosis in giant cell arteritis. Am J Ophthalmol 1997; 124: 103. 10. De Keyser J, De Klippel N, Ebinger G. Thrombocytosis and ischemic complications in giant cell arteritis. BMJ 1991; 303: 825. 11. Jacobson D, Slamovits T. Erythrocyte sedimentation rate and its relationship to hematocrit in giant cell arteritis. Arch Ophthalmol 1987; 105: 965- 7. 12. Swaanenburg J, Rutten W, Holdrinet A, van Strik R. The determination of reference values for hematologic parameters using results obtained from patient populations. Am J Clin Pathol 1987; 88: 182- 91. 13. Gibb W, Urry P, Lees A. Giant cell arteritis with spinal cord infarction and basilar artery thrombosis. J Neurol Neurosurg Psychiatry 1985; 48: 945- 8. 14. Bengtsson BA. Haematological observations in giant cell arteritis. Acta Med Scand 1982; 658: 44- 7. J Neuro- Ophthalmol, Vol. 20, No. 2, 2000 72 N. S. LINCOFF ET AL. • JWt :' & 15. Farr M, Scott DL, et al. Thrombocytosis of active rheumatoid 18. Emilie D, Liozon E, Crevon MC, et al. Production of interleukin 6 disease.' Ann Rheum Dis 1983; 42: 545- 9. by granulomas of giant cell. Hum Immunol 1994; 1: 17- 24. 16. Morowitz D, Allen LW, Kirsner JB. Thrombocytosis in chronic 19. Dasgupta B, Panayi GS. Interleukin- 6 in serum of patients with inflammatory bowel disease. Ann Int Med 1968; 68: 1013- 8. polymyalgia rheumatica and giant cell arteritis. Br J Rheum 1990; 17. Roche NE, Fulbright JW, Wagner AD, Hunder GG, Goronzy JJ, 6: 456- 8. Weyand CM. Correlation of interleukin- 6 production and disease 20. Stahl C, Zucker- Franklin D, Evatt B, Winton E. Effects of human activity in polymyalgia rheumatica and giant ceil arteritis. Arthritis interleukin- 6 on megakaryocyte development and thrombocyto- Rheum 1993; 9: 1286- 94. poiesis in primates. Blood 1991; 78: 1467- 75. J Neuro- Ophthalmol, Vol. 20, No. 2. 2000 |
