- Email: [email protected]
Prevalence of a Normal C-Reactive Protein with an Elevated Erythrocyte Sedimentation Rate in Biopsy-Proven Giant Cell Arteritis
Prevalence of a Normal C-Reactive Protein with an Elevated Erythrocyte Sedimentation Rate in Biopsy-Proven Giant Cell Arteritis Mona Parikh, BA,1 Neil R. Miller, MD,1 Andrew G. Lee, MD,2 Peter J. Savino, MD,3 M. Noel Vacarezza, MD,3 Wayne Cornblath, MD,4 Eric Eggenberger, MD,5 Aileen Antonio-Santos, MD,5 Karl Golnik, MD,6 Randy Kardon, MD,2 Michael Wall, MD2 Objective: The erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are laboratory tests that have been said to have a strong correlation with a positive temporal artery biopsy in patients with suspected giant cell arteritis (GCA). Published reports suggest that the CRP is a more sensitive diagnostic indicator of GCA and can be elevated when the ESR is normal. It is also clear that the CRP and ESR can both be normal or both be elevated in patients with biopsy-proven GCA and that the CRP can be elevated when the ESR is normal. The purpose of this study was to ascertain if the CRP can be normal when the ESR is elevated in biopsy-proven GCA. Design: Retrospective, longitudinal, comparative study. Participants: One hundred nineteen patients from 6 major tertiary-care university-affiliated medical centers. Methods: The charts from 119 patients with temporal artery biopsies positive for GCA were reviewed for age, gender, pretreatment ESR, and pretreatment CRP. Main Outcome Measures: The ESR in millimeters per hour Westergren was graded as normal or abnormal based on 2 validated formulas. The CRP was graded as normal or abnormal based on established criteria set forth in the literature as well as at The Johns Hopkins Hematology laboratory. Results: In this study, the ESR had a sensitivity of 76% to 86%, depending on which of 2 formulas were used, whereas an elevated CRP had a sensitivity of 97.5%. The sensitivity of the ESR and CRP together was 99%. Only 1 of the 119 patients (0.8%) presented with a normal ESR and normal CRP (double false negative); 2 patients (1.7%) had a normal CRP despite an elevated ESR according to both formulas. Conclusion: Although most patients with GCA have both an elevated ESR and CRP, there can be nonconcordance of the 2 blood tests. Although such nonconcordance is most often a normal ESR but an elevated CRP, the finding of an elevated ESR and a normal CRP also is consistent with GCA. The use of both tests provides a slightly greater sensitivity for the diagnosis of GCA than the use of either test alone. Ophthalmology 2006;113: 1842–1845 © 2006 by the American Academy of Ophthalmology.
Giant cell arteritis (GCA), also known as temporal or cranial arteritis, is one of the most important ophthalmic emergencies because of its devastating complication of visual loss in 13% to 50% of patients.1– 6 Although a temporal artery biopsy showing evidence of granulomatous inflammation is Originally received: December 4, 2005. Accepted: May 16, 2006. Manuscript no. 2005-1178. 1 Wilmer Eye Institute, Johns Hopkins Hospital, Baltimore, Maryland. 2 Department of Ophthalmology, University of Iowa Medical Center, Iowa City, Iowa. 3 Wills Eye Hospital, Philadelphia, Pennsylvania. 4 Kellogg Eye Center, Ann Arbor, Michigan. 5 Department of Neurology, Michigan State University, East Lansing, Michigan. 6 Cincinnati Eye Institute, Cincinnati, Ohio. The authors have no commercial or proprietary interest in the products or companies mentioned in the article. Correspondence to Neil R. Miller, Wilmer Eye Institute, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287. E-mail: nrmiller@ jhmi.edu.
1842
© 2006 by the American Academy of Ophthalmology Published by Elsevier Inc.
considered the gold standard for diagnosis of GCA,6 –10 there are certain laboratory tests that have a strong correlation with a positive temporal artery biopsy and that can be obtained rapidly in a situation in which delay in diagnosis and treatment may result in unilateral or bilateral blindness, stroke, or heart attack. In particular, the erythrocyte sedimentation rate (ESR) and the C-reactive protein (CRP) assay have been cited as being highly predictive of biopsyproven GCA, with a combined sensitivity of 97%.5,9,11 Although it is clear from published reports that the CRP can be elevated when the ESR is normal and that the CRP and ESR can both be normal or both be elevated in patients with biopsy-proven GCA, it is unclear if the CRP is ever normal when the ESR is elevated.9,11–13 Indeed, it has been stated that the CRP increases in the setting of tissue injury earlier than the ESR. Thus, it is reasonable to postulate that patients with GCA who have an elevated ESR always have an elevated CRP and that a normal CRP in the setting of an elevated ESR indicates a disorder other than GCA. If this were the case, it would be clear that a patient with an elevated ESR but a normal CRP should undergo further ISSN 0161-6420/06/$–see front matter doi:10.1016/j.ophtha.2006.05.020
Parikh et al 䡠 Normal CRP and Elevated ESR in Giant Cell Arteritis diagnostic studies rather than being immediately treated with corticosteroids for GCA pending the results of a temporal artery biopsy. This not only would save the patient from receiving potentially inappropriate and dangerous treatment as well as undergoing an unneeded surgical procedure, but also would result in a more rapid diagnosis of the true underlying condition. To test this hypothesis, we performed a retrospective analysis of the medical records of 119 patients with biopsy-positive GCA from 6 major centers who had both a pretreatment ESR and CRP assay.
abnormal based on this range as well as the established range at the Johns Hopkins Hospital hematology laboratory (0 – 0.5). Finally, the percentages of patients were calculated who had (1) elevated ESR and elevated CRP; (2) elevated ESR and normal CRP; (3) normal ESR and elevated CRP; and (4) normal ESR and normal CRP. The primary outcome variable was the percentage of patients with a positive temporal artery biopsy who also had an elevated ESR but a normal CRP. The secondary outcome variable was the percentage of patients with a positive temporal artery biopsy who had a normal ESR and a normal CRP.
Materials and Methods
Results
A retrospective, longitudinal, comparative study was performed. Chart review and data collection were conducted in the ophthalmology departments of 6 major tertiary-care university-affiliated medical centers from August 2004 through June 2005. Deidentified data were then sent to the coordinating center at The Johns Hopkins Hospital for final analysis. The study sample consisted of 119 patients with biopsy-proven GCA who had pretreatment ESR and CRP assays drawn between 1994 and 2005, a time during which all patients at all 6 institutions with suspected GCA underwent an assay for both ESR and CRP, followed by a temporal artery biopsy if either the ESR or CRP was elevated or the patient’s symptoms, signs, or both suggested GCA regardless of the ESR or CRP values. All patients had a detailed systemic and ophthalmic history, a complete ophthalmologic evaluation, ESR (Westergren), and CRP assay. Patients with negative temporal artery biopsies were excluded, as were 90 patients with positive biopsies who had already been placed on treatment by the time they were evaluated at 1 of the participating institutions and thus did not have a pretreatment ESR and/or CRP. Data accumulation from patient charts was approved by the institutional review boards of each of the 6 centers. Patients with biopsies consistent with GCA and diagnosed from 1994 to 2005 in the pathology laboratories of the various centers were identified. Hospital records from these patients were reviewed to determine if both a preoperative and pretreatment ESR and CRP were obtained. Patients whose charts included both a preoperative and pretreatment ESR and CRP were considered eligible for inclusion. Patient charts were then reviewed for the following data: age, gender, ESR, and CRP.
Demographic Characteristics
Statistical Data Analysis The ESR in millimeters per hour was graded as normal or abnormal based on 2 validated formulas, 1 from Miller et al14 and 1 from Hayreh et al.11 The Miller et al formula is: Top normal ESR for men: age ⁄ 2; top normal ESR for women: (age ⫹ 10) ⁄ 2.
Of the 119 patients in this study, 87 (73.1%) were female. The mean age at presentation was 75.8, with a range of 52 to 97 years.
Erythrocyte Sedimentation Rate and C-Reactive Protein Levels In this study, 99 of the 119 patients (83.2%) presented with both an elevated pretreatment ESR and CRP. As shown in Table 1, the ESR was elevated in 88 of these patients using both formulas. The other 11 patients had an elevated ESR by the Hayreh et al formula but not the Miller et al formula. Seventeen patients (14.3% of the original 119 patients) had a normal pretreatment ESR using both formulas but an elevated pretreatment CRP. One patient (0.8%) had a normal pretreatment ESR according to both formulas and a normal pretreatment CRP. The remaining 2 patients (1.7%) had an elevated pretreatment ESR according to both formulas but a normal pretreatment CRP. In this study, 91 of 119 patients had an elevated ESR according to the formula of Miller et al, resulting in a sensitivity of 76.5%, whereas 102 of 119 patients had an elevated ESR according to the formula recommended by Hayreh et al, for a sensitivity of 85.7%. Of 119 patients, 116 had an elevated CRP, making the sensitivity of this test 97.5%. Using the combination of ESR and CRP as diagnostic criteria resulted in a sensitivity of 99.2%, with 118 of the 119 patients having an elevated ESR, elevated CRP, or both. As shown in Table 2, we performed the McNemar test to determine if the difference in sensitivity of the ESR versus the CRP for the diagnosis of GCA was significant. The results indicate that the difference in sensitivity between the 2 tests is significant (P ⫽ 0.0003).
Table 1. Percentages of Patients in Each Category Using Both Erythrocyte Sedimentation Rate (ESR) and C-Reactive Protein (CRP) as Diagnostic Indicators
The Hayreh et al formula is: Top normal ESR for men: 17.3⫹(0.18 ⫻ age) mm ⁄ hour; top normal ESR for women: 22.1⫹(0.18 ⫻ age) mm ⁄ hour. Unlike estimation of ESR, there is no standardized technique for CRP, and the “normal” levels vary tremendously from laboratory to laboratory; however, Costello et al9 used the criteria of ⱕ0.5 mg/dl as normal in a comparison of CRP in 71 patients with and 103 patients without GCA. This study showed a sensitivity of 98.6% and specificity of 75.7%, with false-negative results in 1.3% and false-positive results in 26.3%.9 Thus, for analysis in our study, the CRP in milligrams per deciliter was graded as normal or
ESR > Cutoff by ESR > Cutoff by Both Formulas*† Only 1 Criterion*† ESR < Cutoff CRP ⬎ cutoff CRP ⱕ cutoff
88 (73.9%) 2 (1.7%)
11 (9.2%) 0 (0%)
17 (14.3%) 1 (0.84%)
*Criteria recommended by Miller A, Green M, Robinson D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J (Clin Res Ed) 1983;286:266. † Criteria recommended by Hayreh SS, Podhalsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96.
1843
Ophthalmology Volume 113, Number 10, October 2006 Table 2. Statistical Analysis of the Sensitivities of Erythrocyte Sedimentation Rate (ESR) versus C-Reactive Protein (CRP) as Diagnostic Indicators of Giant Cell Arteritis
ESR (⫹) ESR (⫺) Total
CRP (ⴙ)
CRP (ⴚ)
Total
99 17 116
2 1 3
101 18 119
Sensitivity of ESR: 102/119 ⫽ 85.7%; sensitivity of CRP: 116/119 ⫽ 97.5%; P ⫽ 0.0003.
Discussion Visual loss is the most common devastating and irreversible complication of GCA and can be prevented with treatment. The first step in management, therefore, is early and accurate diagnosis. It is clear from previous studies that both the ESR and the CRP are highly sensitive tests for the diagnosis of GCA and that their combination increases the sensitivity further.5,9,11–13 Because the CRP is an acute-phase plasma protein that rises before ESR in most disease states, we hypothesized that an elevated ESR in the setting of a normal CRP would not be consistent with the diagnosis of GCA. Thus, the primary objective in this study was to determine the prevalence of a normal CRP in patients with GCA who had an elevated ESR. Because all the patients in this study had positive temporal artery biopsies, we could not calculate a false-positive rate for the ESR, the CRP, or both combined.
Demographics Giant cell arteritis is universally accepted as a disease of the elderly that affects women more than men. In the Western hemisphere, where GCA is common, reported annual incidence rates vary significantly from 0.49 to 23.3 per 100 000 persons aged 50 years and older, and the incidence rate dramatically increases with advancing age.15,16 Our findings were consistent with these figures. Of the 119 patients in this study, 87 (73.1%) were female; mean age at presentation was 75.8 years, with a range of 52 to 97 years.
Erythrocyte Sedimentation Rate (Westergren) An elevated ESR is a generally accepted and well-established aid in the diagnosis of GCA. There is, however, a good amount of ambiguity about what ESR level should be considered a diagnostic indicator for GCA. Many studies in the past have attempted to answer the question of what is the normal and abnormal ESR in persons 50 years and older, and there are several proposed formulas for calculating normal values.17,18 In a study of 27 912 individuals aged 20 to 65, Miller et al derived an empiric formula for ESR that included 98% of healthy persons: age in years divided by 2 for men, and age in years plus 10, the sum divided by 2 for women.14 In a more recent study of 363 patients with suspected GCA, 106 of whom were temporal artery biopsy positive, Hayreh et al derived a formula from regression of age on
1844
ESR: 17.3⫹(0.18⫻age) for men, and 22.1⫹(0.18⫻age) for women.11 The formula used by Hayreh et al is more sensitive in predicting biopsy-proven GCA than the formula of Miller et al. Indeed, in our study, 11 patients with biopsy-proven GCA who had a normal ESR according to the formula of Miller et al had an elevated ESR according to the formula of Hayreh et al. In our study, the formula of Miller et al had a 76.5% sensitivity for GCA, whereas the Hayreh et al formula had an 85.7% sensitivity. Although the ESR clearly has a high sensitivity for diagnosing GCA, it is not 100%. Numerous studies have described patients with biopsy-proven GCA who have had ESR values ranging from 5 to 30 mm/hour.19 –24
C-Reactive Protein The CRP is an acute-phase plasma protein of hepatic origin that is also often elevated in GCA. Studies show that in the diagnosis of GCA, the use of CRP has several distinct advantages over the use of ESR, including higher sensitivity and specificity (98.6% and 75.7%, respectively),9 more rapid elevation than ESR in most disease states, and relative insensitivity to age, gender, and other hematologic parameters. In our study, all 6 centers tested the CRP by the clinical laboratory method that is typically used to measure the acute inflammation that develops in the settings of trauma, operations, bacterial infections, and arthritis rather than the highly sensitive CRP assay that is most often used to measure the low-grade, chronic inflammation that occurs in patients with known or suspected coronary artery disease.25 The levels were graded as normal or abnormal based on the range set forth in the study by Costello et al,9 which is also the established range of normal at The Johns Hopkins Hospital hematology laboratory (0 – 0.5 mg/dl). As shown in Table 2, our results are consistent with prior studies that indicate that the difference in sensitivity between the ESR and the CRP (85.7% versus 97.5%) is significant (P ⫽ 0.0003).
Combination of Erythrocyte Sedimentation Rate and C-Reactive Protein in the Diagnosis of Giant Cell Arteritis Although other investigators have emphasized an increased sensitivity for diagnosis of GCA when using the combination of ESR and CRP,5,9,11 the results of our study indicate that performing both an ESR and a CRP assay and pursuing a diagnosis based on abnormal results of either test has only a slightly greater diagnostic yield than performing CRP alone. In our study, an ESR alone had a 76% to 86% sensitivity depending on whether the criteria of Miller et al or Hayreh et al were used; however, an elevated CRP had a sensitivity of 97.5%. When employing criteria of an abnormal ESR or CRP, the sensitivity was 99.2%, as 118 of 119 patients had elevation of the ESR, CRP, or both. Although an increase in sensitivity from 97.5% to 99.2% would seem minimal, GCA has such potential catastrophic complications that any increase in sensitivity is important. It must also be emphasized that because we reviewed only cases of
Parikh et al 䡠 Normal CRP and Elevated ESR in Giant Cell Arteritis biopsy-proven GCA, we could not calculate a false-positive rate. Thus, we do not know how that rate would be affected if one relied only on the CRP to diagnose GCA as opposed to using either the ESR alone or both values together. We believe, however, that the false-positive rate is not as important as the false-negative rate; the potential risks of a temporal artery biopsy in a patient subsequently found not to have GCA are far fewer than the risks of a devastating visual or systemic complication of undiagnosed and untreated GCA. Only 1 of the 119 patients (0.8%) presented with a normal ESR and normal CRP (double false negative). This patient, a 77-year-old white female with a long-standing history of depression and dementia, presented with rightsided temporal tenderness and acute painless loss of vision in her right eye. She had no headaches, jaw claudication, or other systemic signs of GCA. Although 0.8% represents a small proportion of total GCA patients, these individuals are the most worrisome because diagnosis and treatment may more often be delayed. Given the fact that the CRP rises in response to tissue injury sooner than the ESR and that most studies (including our own) indicate that the CRP is a more sensitive diagnostic test for GCA than the ESR, our hypothesis was that it was unlikely that patients with GCA would have a normal CRP in the setting of an elevated ESR. In our study, however, 2 patients (1.7%) with biopsy-proven GCA had a normal pretreatment CRP despite an elevated pretreatment ESR according to both Miller et al and Hayreh et al criteria. One of these patients, a 73-year-old white female, presented with left-sided loss of vision, but no headaches, jaw pain, or other systemic symptoms of GCA. The other patient, a 66-year-old white female, presented with new-onset leftsided headaches, but no change in vision, jaw pain, or other systemic symptoms of GCA. Based on these findings, we conclude that the combination of an elevated ESR and a normal CRP, although rare, is not inconsistent with a diagnosis of GCA and that this combination should not be taken to indicate that another diagnosis is responsible. We agree with Hayreh et al and Costello et al that the use of both tests together identifies the vast majority of patients with GCA with a higher sensitivity than the use of either test alone. Acknowledgments. The authors are indebted to Dr Susan Vitale, Johns Hopkins Hospital, for her assistance with the statistical analysis for the manuscript.
References 1. Wagener HP, Hollenhorst RW. The ocular lesions of temporal arteritis. Am J Ophthalmol 1958;45:617–30. 2. Goodman BW Jr. Temporal arteritis. Am J Med 1979;67: 839 –52. 3. Keltner JL. Giant-cell arteritis: signs and symptoms. Ophthalmology 1982;89:1101–10. 4. Spiera R, Spiera H. Inflammatory disease in older adults: cranial arteritis. Geriatrics 2004;59:25–9.
5. Azhar SS, Tang RA, Dorotheo EU. Giant cell arteritis: diagnosing and treating inflammatory disease in older adults. Geriatrics 2005;60:26 –30. 6. Rahman W, Rahman FZ. Giant cell (temporal) arteritis: an overview and update. Surv Ophthalmol 2005;50:415–28. 7. Bhatti MT, Tabendeh H. Giant cell arteritis: diagnosis and management. Curr Opin Ophthalmol 2001;12:393–9. 8. Riordan-Eva P, Landau K, O’Day J. Temporal artery biopsy in the management of giant cell arteritis with neuro-ophthalmic complications. Br J Ophthalmol 2001;85:1248 –51. 9. Costello F, Zimmerman MB, Podhajsky PA, Hayreh SS. Role of thrombocytosis in diagnosis of giant cell arteritis and differentiation of arteritic from non-arteritic anterior ischemic optic neuropathy. Eur J Ophthalmol 2004;14:245–57. 10. Nordborg E, Nordborg C. Giant cell arteritis: strategies in diagnosis and treatment. Curr Opin Rheumatol 2004;16: 25–30. 11. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96. 12. Myklebust G, Gran JT. A prospective study of 287 patients with polymyalgia rheumatica and temporal arteritis: clinical and laboratory manifestations at onset of disease and at the time of diagnosis. Br J Rheumatol 1996;35:1161– 8. 13. Poole TR, Graham EM, Lucas SB. Giant cell arteritis with a normal ESR and CRP [letter]. Eye 2003;17:92–3. 14. Miller A, Green M, Robinson D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J (Clin Res Ed) 1983;286:266. 15. Hayreh SS, Zimmerman B. Management of giant cell arteritis. Our 27-year clinical study: new light on old controversies. Ophthalmologica 2003;217:239 –59. 16. Boesen P, Sorenson SF. Giant cell arteritis, temporal arteritis, and polymyalgia rheumatica in a Danish county: a prospective investigation, 1982-1985. Arthritis Rheum 1987;30:294 –9. 17. Gilbertsen VA. Erythrocyte sedimentation rates in older patients: a study of 4,341 cases. Postgrad Med 1965;38(5): A44 –52. 18. Bottiger LE, Svedberg CA. Normal erythrocyte sedimentation rate and age. Br Med J 1967;2(544):85–7. 19. Wong RL, Korn JH. Temporal arteritis without an elevated sedimentation rate: case report and review of the literature. Am J Med 1986;80:959 – 64. 20. Grodum E, Petersen HA. Temporal arteritis with normal erythrocyte sedimentation rate. J Intern Med 1990;227: 279 – 80. 21. Jundt JW, Mock D. Temporal arteritis with normal erythrocyte sedimentation rates presenting as occipital neuralgia. Arthritis Rheum 1991;34:217–9. 22. Wise CM, Agudelo CA, Chmelewski WL, McKnight KM. Temporal arteritis with low erythrocyte sedimentation rate: a review of five cases. Arthritis Rheum 2001;34:1571– 4. 23. Neish PR, Sergent JS. Giant cell arteritis: a case with unusual neurologic manifestations and normal sedimentation rate. Arch Intern Med 1991;151:378 – 80. 24. Zweegman S, Makkink B, Stehouwer CD. Giant-cell arteritis with normal erythrocyte sedimentation rate: case report and review of the literature. Netherlands J Med 1993;42:128 –31. 25. Rifai N, Ridker PM. High-sensitivity C-reactive protein: a novel and promising marker of coronary artery disease. Clin Chem 2001;47:403–11.
1845
Giant cell arteritis (GCA), also known as temporal or cranial arteritis, is one of the most important ophthalmic emergencies because of its devastating complication of visual loss in 13% to 50% of patients.1– 6 Although a temporal artery biopsy showing evidence of granulomatous inflammation is Originally received: December 4, 2005. Accepted: May 16, 2006. Manuscript no. 2005-1178. 1 Wilmer Eye Institute, Johns Hopkins Hospital, Baltimore, Maryland. 2 Department of Ophthalmology, University of Iowa Medical Center, Iowa City, Iowa. 3 Wills Eye Hospital, Philadelphia, Pennsylvania. 4 Kellogg Eye Center, Ann Arbor, Michigan. 5 Department of Neurology, Michigan State University, East Lansing, Michigan. 6 Cincinnati Eye Institute, Cincinnati, Ohio. The authors have no commercial or proprietary interest in the products or companies mentioned in the article. Correspondence to Neil R. Miller, Wilmer Eye Institute, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287. E-mail: nrmiller@ jhmi.edu.
1842
© 2006 by the American Academy of Ophthalmology Published by Elsevier Inc.
considered the gold standard for diagnosis of GCA,6 –10 there are certain laboratory tests that have a strong correlation with a positive temporal artery biopsy and that can be obtained rapidly in a situation in which delay in diagnosis and treatment may result in unilateral or bilateral blindness, stroke, or heart attack. In particular, the erythrocyte sedimentation rate (ESR) and the C-reactive protein (CRP) assay have been cited as being highly predictive of biopsyproven GCA, with a combined sensitivity of 97%.5,9,11 Although it is clear from published reports that the CRP can be elevated when the ESR is normal and that the CRP and ESR can both be normal or both be elevated in patients with biopsy-proven GCA, it is unclear if the CRP is ever normal when the ESR is elevated.9,11–13 Indeed, it has been stated that the CRP increases in the setting of tissue injury earlier than the ESR. Thus, it is reasonable to postulate that patients with GCA who have an elevated ESR always have an elevated CRP and that a normal CRP in the setting of an elevated ESR indicates a disorder other than GCA. If this were the case, it would be clear that a patient with an elevated ESR but a normal CRP should undergo further ISSN 0161-6420/06/$–see front matter doi:10.1016/j.ophtha.2006.05.020
Parikh et al 䡠 Normal CRP and Elevated ESR in Giant Cell Arteritis diagnostic studies rather than being immediately treated with corticosteroids for GCA pending the results of a temporal artery biopsy. This not only would save the patient from receiving potentially inappropriate and dangerous treatment as well as undergoing an unneeded surgical procedure, but also would result in a more rapid diagnosis of the true underlying condition. To test this hypothesis, we performed a retrospective analysis of the medical records of 119 patients with biopsy-positive GCA from 6 major centers who had both a pretreatment ESR and CRP assay.
abnormal based on this range as well as the established range at the Johns Hopkins Hospital hematology laboratory (0 – 0.5). Finally, the percentages of patients were calculated who had (1) elevated ESR and elevated CRP; (2) elevated ESR and normal CRP; (3) normal ESR and elevated CRP; and (4) normal ESR and normal CRP. The primary outcome variable was the percentage of patients with a positive temporal artery biopsy who also had an elevated ESR but a normal CRP. The secondary outcome variable was the percentage of patients with a positive temporal artery biopsy who had a normal ESR and a normal CRP.
Materials and Methods
Results
A retrospective, longitudinal, comparative study was performed. Chart review and data collection were conducted in the ophthalmology departments of 6 major tertiary-care university-affiliated medical centers from August 2004 through June 2005. Deidentified data were then sent to the coordinating center at The Johns Hopkins Hospital for final analysis. The study sample consisted of 119 patients with biopsy-proven GCA who had pretreatment ESR and CRP assays drawn between 1994 and 2005, a time during which all patients at all 6 institutions with suspected GCA underwent an assay for both ESR and CRP, followed by a temporal artery biopsy if either the ESR or CRP was elevated or the patient’s symptoms, signs, or both suggested GCA regardless of the ESR or CRP values. All patients had a detailed systemic and ophthalmic history, a complete ophthalmologic evaluation, ESR (Westergren), and CRP assay. Patients with negative temporal artery biopsies were excluded, as were 90 patients with positive biopsies who had already been placed on treatment by the time they were evaluated at 1 of the participating institutions and thus did not have a pretreatment ESR and/or CRP. Data accumulation from patient charts was approved by the institutional review boards of each of the 6 centers. Patients with biopsies consistent with GCA and diagnosed from 1994 to 2005 in the pathology laboratories of the various centers were identified. Hospital records from these patients were reviewed to determine if both a preoperative and pretreatment ESR and CRP were obtained. Patients whose charts included both a preoperative and pretreatment ESR and CRP were considered eligible for inclusion. Patient charts were then reviewed for the following data: age, gender, ESR, and CRP.
Demographic Characteristics
Statistical Data Analysis The ESR in millimeters per hour was graded as normal or abnormal based on 2 validated formulas, 1 from Miller et al14 and 1 from Hayreh et al.11 The Miller et al formula is: Top normal ESR for men: age ⁄ 2; top normal ESR for women: (age ⫹ 10) ⁄ 2.
Of the 119 patients in this study, 87 (73.1%) were female. The mean age at presentation was 75.8, with a range of 52 to 97 years.
Erythrocyte Sedimentation Rate and C-Reactive Protein Levels In this study, 99 of the 119 patients (83.2%) presented with both an elevated pretreatment ESR and CRP. As shown in Table 1, the ESR was elevated in 88 of these patients using both formulas. The other 11 patients had an elevated ESR by the Hayreh et al formula but not the Miller et al formula. Seventeen patients (14.3% of the original 119 patients) had a normal pretreatment ESR using both formulas but an elevated pretreatment CRP. One patient (0.8%) had a normal pretreatment ESR according to both formulas and a normal pretreatment CRP. The remaining 2 patients (1.7%) had an elevated pretreatment ESR according to both formulas but a normal pretreatment CRP. In this study, 91 of 119 patients had an elevated ESR according to the formula of Miller et al, resulting in a sensitivity of 76.5%, whereas 102 of 119 patients had an elevated ESR according to the formula recommended by Hayreh et al, for a sensitivity of 85.7%. Of 119 patients, 116 had an elevated CRP, making the sensitivity of this test 97.5%. Using the combination of ESR and CRP as diagnostic criteria resulted in a sensitivity of 99.2%, with 118 of the 119 patients having an elevated ESR, elevated CRP, or both. As shown in Table 2, we performed the McNemar test to determine if the difference in sensitivity of the ESR versus the CRP for the diagnosis of GCA was significant. The results indicate that the difference in sensitivity between the 2 tests is significant (P ⫽ 0.0003).
Table 1. Percentages of Patients in Each Category Using Both Erythrocyte Sedimentation Rate (ESR) and C-Reactive Protein (CRP) as Diagnostic Indicators
The Hayreh et al formula is: Top normal ESR for men: 17.3⫹(0.18 ⫻ age) mm ⁄ hour; top normal ESR for women: 22.1⫹(0.18 ⫻ age) mm ⁄ hour. Unlike estimation of ESR, there is no standardized technique for CRP, and the “normal” levels vary tremendously from laboratory to laboratory; however, Costello et al9 used the criteria of ⱕ0.5 mg/dl as normal in a comparison of CRP in 71 patients with and 103 patients without GCA. This study showed a sensitivity of 98.6% and specificity of 75.7%, with false-negative results in 1.3% and false-positive results in 26.3%.9 Thus, for analysis in our study, the CRP in milligrams per deciliter was graded as normal or
ESR > Cutoff by ESR > Cutoff by Both Formulas*† Only 1 Criterion*† ESR < Cutoff CRP ⬎ cutoff CRP ⱕ cutoff
88 (73.9%) 2 (1.7%)
11 (9.2%) 0 (0%)
17 (14.3%) 1 (0.84%)
*Criteria recommended by Miller A, Green M, Robinson D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J (Clin Res Ed) 1983;286:266. † Criteria recommended by Hayreh SS, Podhalsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96.
1843
Ophthalmology Volume 113, Number 10, October 2006 Table 2. Statistical Analysis of the Sensitivities of Erythrocyte Sedimentation Rate (ESR) versus C-Reactive Protein (CRP) as Diagnostic Indicators of Giant Cell Arteritis
ESR (⫹) ESR (⫺) Total
CRP (ⴙ)
CRP (ⴚ)
Total
99 17 116
2 1 3
101 18 119
Sensitivity of ESR: 102/119 ⫽ 85.7%; sensitivity of CRP: 116/119 ⫽ 97.5%; P ⫽ 0.0003.
Discussion Visual loss is the most common devastating and irreversible complication of GCA and can be prevented with treatment. The first step in management, therefore, is early and accurate diagnosis. It is clear from previous studies that both the ESR and the CRP are highly sensitive tests for the diagnosis of GCA and that their combination increases the sensitivity further.5,9,11–13 Because the CRP is an acute-phase plasma protein that rises before ESR in most disease states, we hypothesized that an elevated ESR in the setting of a normal CRP would not be consistent with the diagnosis of GCA. Thus, the primary objective in this study was to determine the prevalence of a normal CRP in patients with GCA who had an elevated ESR. Because all the patients in this study had positive temporal artery biopsies, we could not calculate a false-positive rate for the ESR, the CRP, or both combined.
Demographics Giant cell arteritis is universally accepted as a disease of the elderly that affects women more than men. In the Western hemisphere, where GCA is common, reported annual incidence rates vary significantly from 0.49 to 23.3 per 100 000 persons aged 50 years and older, and the incidence rate dramatically increases with advancing age.15,16 Our findings were consistent with these figures. Of the 119 patients in this study, 87 (73.1%) were female; mean age at presentation was 75.8 years, with a range of 52 to 97 years.
Erythrocyte Sedimentation Rate (Westergren) An elevated ESR is a generally accepted and well-established aid in the diagnosis of GCA. There is, however, a good amount of ambiguity about what ESR level should be considered a diagnostic indicator for GCA. Many studies in the past have attempted to answer the question of what is the normal and abnormal ESR in persons 50 years and older, and there are several proposed formulas for calculating normal values.17,18 In a study of 27 912 individuals aged 20 to 65, Miller et al derived an empiric formula for ESR that included 98% of healthy persons: age in years divided by 2 for men, and age in years plus 10, the sum divided by 2 for women.14 In a more recent study of 363 patients with suspected GCA, 106 of whom were temporal artery biopsy positive, Hayreh et al derived a formula from regression of age on
1844
ESR: 17.3⫹(0.18⫻age) for men, and 22.1⫹(0.18⫻age) for women.11 The formula used by Hayreh et al is more sensitive in predicting biopsy-proven GCA than the formula of Miller et al. Indeed, in our study, 11 patients with biopsy-proven GCA who had a normal ESR according to the formula of Miller et al had an elevated ESR according to the formula of Hayreh et al. In our study, the formula of Miller et al had a 76.5% sensitivity for GCA, whereas the Hayreh et al formula had an 85.7% sensitivity. Although the ESR clearly has a high sensitivity for diagnosing GCA, it is not 100%. Numerous studies have described patients with biopsy-proven GCA who have had ESR values ranging from 5 to 30 mm/hour.19 –24
C-Reactive Protein The CRP is an acute-phase plasma protein of hepatic origin that is also often elevated in GCA. Studies show that in the diagnosis of GCA, the use of CRP has several distinct advantages over the use of ESR, including higher sensitivity and specificity (98.6% and 75.7%, respectively),9 more rapid elevation than ESR in most disease states, and relative insensitivity to age, gender, and other hematologic parameters. In our study, all 6 centers tested the CRP by the clinical laboratory method that is typically used to measure the acute inflammation that develops in the settings of trauma, operations, bacterial infections, and arthritis rather than the highly sensitive CRP assay that is most often used to measure the low-grade, chronic inflammation that occurs in patients with known or suspected coronary artery disease.25 The levels were graded as normal or abnormal based on the range set forth in the study by Costello et al,9 which is also the established range of normal at The Johns Hopkins Hospital hematology laboratory (0 – 0.5 mg/dl). As shown in Table 2, our results are consistent with prior studies that indicate that the difference in sensitivity between the ESR and the CRP (85.7% versus 97.5%) is significant (P ⫽ 0.0003).
Combination of Erythrocyte Sedimentation Rate and C-Reactive Protein in the Diagnosis of Giant Cell Arteritis Although other investigators have emphasized an increased sensitivity for diagnosis of GCA when using the combination of ESR and CRP,5,9,11 the results of our study indicate that performing both an ESR and a CRP assay and pursuing a diagnosis based on abnormal results of either test has only a slightly greater diagnostic yield than performing CRP alone. In our study, an ESR alone had a 76% to 86% sensitivity depending on whether the criteria of Miller et al or Hayreh et al were used; however, an elevated CRP had a sensitivity of 97.5%. When employing criteria of an abnormal ESR or CRP, the sensitivity was 99.2%, as 118 of 119 patients had elevation of the ESR, CRP, or both. Although an increase in sensitivity from 97.5% to 99.2% would seem minimal, GCA has such potential catastrophic complications that any increase in sensitivity is important. It must also be emphasized that because we reviewed only cases of
Parikh et al 䡠 Normal CRP and Elevated ESR in Giant Cell Arteritis biopsy-proven GCA, we could not calculate a false-positive rate. Thus, we do not know how that rate would be affected if one relied only on the CRP to diagnose GCA as opposed to using either the ESR alone or both values together. We believe, however, that the false-positive rate is not as important as the false-negative rate; the potential risks of a temporal artery biopsy in a patient subsequently found not to have GCA are far fewer than the risks of a devastating visual or systemic complication of undiagnosed and untreated GCA. Only 1 of the 119 patients (0.8%) presented with a normal ESR and normal CRP (double false negative). This patient, a 77-year-old white female with a long-standing history of depression and dementia, presented with rightsided temporal tenderness and acute painless loss of vision in her right eye. She had no headaches, jaw claudication, or other systemic signs of GCA. Although 0.8% represents a small proportion of total GCA patients, these individuals are the most worrisome because diagnosis and treatment may more often be delayed. Given the fact that the CRP rises in response to tissue injury sooner than the ESR and that most studies (including our own) indicate that the CRP is a more sensitive diagnostic test for GCA than the ESR, our hypothesis was that it was unlikely that patients with GCA would have a normal CRP in the setting of an elevated ESR. In our study, however, 2 patients (1.7%) with biopsy-proven GCA had a normal pretreatment CRP despite an elevated pretreatment ESR according to both Miller et al and Hayreh et al criteria. One of these patients, a 73-year-old white female, presented with left-sided loss of vision, but no headaches, jaw pain, or other systemic symptoms of GCA. The other patient, a 66-year-old white female, presented with new-onset leftsided headaches, but no change in vision, jaw pain, or other systemic symptoms of GCA. Based on these findings, we conclude that the combination of an elevated ESR and a normal CRP, although rare, is not inconsistent with a diagnosis of GCA and that this combination should not be taken to indicate that another diagnosis is responsible. We agree with Hayreh et al and Costello et al that the use of both tests together identifies the vast majority of patients with GCA with a higher sensitivity than the use of either test alone. Acknowledgments. The authors are indebted to Dr Susan Vitale, Johns Hopkins Hospital, for her assistance with the statistical analysis for the manuscript.
References 1. Wagener HP, Hollenhorst RW. The ocular lesions of temporal arteritis. Am J Ophthalmol 1958;45:617–30. 2. Goodman BW Jr. Temporal arteritis. Am J Med 1979;67: 839 –52. 3. Keltner JL. Giant-cell arteritis: signs and symptoms. Ophthalmology 1982;89:1101–10. 4. Spiera R, Spiera H. Inflammatory disease in older adults: cranial arteritis. Geriatrics 2004;59:25–9.
5. Azhar SS, Tang RA, Dorotheo EU. Giant cell arteritis: diagnosing and treating inflammatory disease in older adults. Geriatrics 2005;60:26 –30. 6. Rahman W, Rahman FZ. Giant cell (temporal) arteritis: an overview and update. Surv Ophthalmol 2005;50:415–28. 7. Bhatti MT, Tabendeh H. Giant cell arteritis: diagnosis and management. Curr Opin Ophthalmol 2001;12:393–9. 8. Riordan-Eva P, Landau K, O’Day J. Temporal artery biopsy in the management of giant cell arteritis with neuro-ophthalmic complications. Br J Ophthalmol 2001;85:1248 –51. 9. Costello F, Zimmerman MB, Podhajsky PA, Hayreh SS. Role of thrombocytosis in diagnosis of giant cell arteritis and differentiation of arteritic from non-arteritic anterior ischemic optic neuropathy. Eur J Ophthalmol 2004;14:245–57. 10. Nordborg E, Nordborg C. Giant cell arteritis: strategies in diagnosis and treatment. Curr Opin Rheumatol 2004;16: 25–30. 11. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96. 12. Myklebust G, Gran JT. A prospective study of 287 patients with polymyalgia rheumatica and temporal arteritis: clinical and laboratory manifestations at onset of disease and at the time of diagnosis. Br J Rheumatol 1996;35:1161– 8. 13. Poole TR, Graham EM, Lucas SB. Giant cell arteritis with a normal ESR and CRP [letter]. Eye 2003;17:92–3. 14. Miller A, Green M, Robinson D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J (Clin Res Ed) 1983;286:266. 15. Hayreh SS, Zimmerman B. Management of giant cell arteritis. Our 27-year clinical study: new light on old controversies. Ophthalmologica 2003;217:239 –59. 16. Boesen P, Sorenson SF. Giant cell arteritis, temporal arteritis, and polymyalgia rheumatica in a Danish county: a prospective investigation, 1982-1985. Arthritis Rheum 1987;30:294 –9. 17. Gilbertsen VA. Erythrocyte sedimentation rates in older patients: a study of 4,341 cases. Postgrad Med 1965;38(5): A44 –52. 18. Bottiger LE, Svedberg CA. Normal erythrocyte sedimentation rate and age. Br Med J 1967;2(544):85–7. 19. Wong RL, Korn JH. Temporal arteritis without an elevated sedimentation rate: case report and review of the literature. Am J Med 1986;80:959 – 64. 20. Grodum E, Petersen HA. Temporal arteritis with normal erythrocyte sedimentation rate. J Intern Med 1990;227: 279 – 80. 21. Jundt JW, Mock D. Temporal arteritis with normal erythrocyte sedimentation rates presenting as occipital neuralgia. Arthritis Rheum 1991;34:217–9. 22. Wise CM, Agudelo CA, Chmelewski WL, McKnight KM. Temporal arteritis with low erythrocyte sedimentation rate: a review of five cases. Arthritis Rheum 2001;34:1571– 4. 23. Neish PR, Sergent JS. Giant cell arteritis: a case with unusual neurologic manifestations and normal sedimentation rate. Arch Intern Med 1991;151:378 – 80. 24. Zweegman S, Makkink B, Stehouwer CD. Giant-cell arteritis with normal erythrocyte sedimentation rate: case report and review of the literature. Netherlands J Med 1993;42:128 –31. 25. Rifai N, Ridker PM. High-sensitivity C-reactive protein: a novel and promising marker of coronary artery disease. Clin Chem 2001;47:403–11.
1845