- Email: [email protected]
Comparison of Histopathologic Features, Clinical Symptoms, and Erythrocyte Sedimentation Rates in Biopsy-Positive Temporal Arteritis
Comparison of Histopathologic Features, Clinical Symptoms, and Erythrocyte Sedimentation Rates in Biopsy-Positive Temporal Arteritis Vicente A. Diaz, MD, MBA,1 Brian M. DeBroff, MD,2 John Sinard, MD, PhD2,3 Purpose: To compare the histopathologic stage of temporal arteritis observed in biopsy-positive specimens with clinical symptoms and erythrocyte sedimentation rates (ESR). Also, to compare the degree of involvement of biopsies from the right versus left side in patients receiving bilateral biopsies. Design: Retrospective case series. Participants: Thirty-seven patients, 30 of whom received unilateral biopsies and 7 of whom had bilateral biopsies. Methods: Biopsy specimens were reviewed and assigned a severity score on the basis of the presence of histopathologic features known to be associated with temporal arteritis. The charts for these patients were then reviewed, and patients’ clinical symptoms were classified by use of an ordinal scale of severity. The first level of severity entailed headache or superficial pain without any other symptoms. The second level included constitutional symptoms such as weight loss, anorexia, fever, and fatigue. The third level of severity included the presence of visual manifestations. The ESR values were also recorded from patients’ charts and examined as a continuous variable. Results: There is a statistically significant correlation between histopathologic stage of disease observed in biopsy specimens and clinical presentation. (P⬍0.0001) The biopsy severity did not correlate significantly with ESR values (P ⫽ 0.09). There was significant variation of ESR values among the 3 levels of clinical severity (analysis of variance test yielded P ⫽ 0.04). By use of the independent samples t test, we found that mean ESR values from patients with constitutional symptoms and visual symptoms were not statistically different (P nonsignificant), yet patients with only headache/superficial tenderness had significantly lower ESR values than the rest of our patient population (P ⫽ 0.009). Last, we found that the mean difference in biopsy severity was significantly above zero (P⬍0.001) when comparing simultaneous bilateral biopsy specimens. Conclusions: Because of the strong correlation between biopsy severity and clinical presentation, this study supports the use of temporal artery biopsy to identify patients’ risk for complications caused by temporal arteritis. Furthermore, this study indicates that ESR may be more elevated in patients with constitutional or visual symptoms than in patients with headache or scalp pain. Finally, the pathologic grade of the disease may differ significantly from 1 side to the contralateral side. Ophthalmology 2005;112:1293–1298 © 2005 by the American Academy of Ophthalmology.
Temporal arteritis is a chronic vasculitis that typically affects cranial branches of vessels originating from the aortic arch, although involvement of larger vessels has also been reported. The disease usually occurs in individuals older than the age of 50. In this age group, the prevalence is approximately 200/100 000 persons.1
Symptoms of temporal arteritis range from abrupt to gradual and from mild to severe. Some of the milder symptoms associated with the disease are face pain, headaches, (particularly over the temporal areas), and tender temporal or occipital arteries; this pain may become progressively worse but may also be transient in some cases. Systemic
Originally received: December 16, 2003. Accepted: February 2, 2005. Manuscript no. 230859. 1 Yale University School of Medicine, New Haven, Connecticut. 2 Department of Ophthalmology and Visual Science, Yale University School of Medicine, New Haven, Connecticut. 3 Department of Pathology, Yale University School of Medicine, New Haven, Connecticut.
Presented at: Association for Research in Vision and Ophthalmology Annual Meeting, May, 2002; Fort Lauderdale, Florida. The authors have no proprietary or commercial interest in the marketing of this material and have not received any financial support of any kind. Correspondence and reprint requests to Vicente A. Diaz, Yale Eye Center, 330 Cedar Street, P.O. Box 208061, New Haven, CT 06520-8061. E-mail: [email protected].
© 2005 by the American Academy of Ophthalmology Published by Elsevier Inc.
ISSN 0161-6420/05/$–see front matter doi:10.1016/j.ophtha.2005.02.016
1293
Ophthalmology Volume 112, Number 7, July 2005 Table 1. Biopsy Severity Score Histologic Feature
Point Breakdown
Necrosis Depth of inflammation Circumference Involved Small branch vessel involvement Medial and intimal fibrosis Disruption of internal elastic lamina Giant cells
1 1 1 1 1 1 1
Present, 0 not Media, .5 adventitia, 0 intima Complete, .5 ⬎50%, 0 ⬍50% Present, 0 not Present, 0 not Complete, .5 partial, 0 not present Extensive, .5 present, 0 not
Weight breakdown High ⫽ ⫻ 5 Medium ⫽ ⫻ 3 Low ⫽ ⫻ 1
Total possible score: 21
Weight High High High Medium Low Low Low
Table outlines point allocation system for assessment of severity of giant cell arteritis. High, medium, and low refer to weight given to characteristic to reflect relative contribution to overall severity.
symptoms are often present and include fever, fatigue, and weight loss.2 Visual manifestations of this disease can occur when the vasculitis progresses to the ophthalmic arteries, causing ischemic optic neuropathy. Atypical symptoms of temporal arteritis such as respiratory symptoms, tongue infarctions, and extremity claudication have also been described as a result of inflammation in other vessels throughout the body.3– 6 Although it is difficult to make the diagnosis on the basis of symptoms alone, the absence of symptoms can be helpful in ruling out temporal arteritis.7–9 A temporal artery biopsy should be considered for any patient older than the age of 50 with the aforementioned symptoms.2,4,7–9 A temporal artery biopsy may be performed unilaterally, sequential bilaterally, or simultaneous bilaterally. Recent data suggest that performing a unilateral biopsy can be sufficient to exclude temporal arteritis when clinical suspicion is low.10 However, there may be up to a 5% increase in efficacy when performing a simultaneous bilateral biopsy compared with a unilateral biopsy.11 A proposed mechanism for these findings is that “skip lesions” may be present and, thus, prevent proper diagnosis of disease. Given the grave visual consequences of temporal arteritis, it is critical that physicians have a clear understanding of how this disease progresses. It is the goal of this article to examine the relationship between stage of disease, as quantified by examining temporal artery biopsies, clinical presentation, erythrocyte sedimentation rate (ESR), and age. Also, we examine the severity of disease on 1 side versus the contralateral side in patients having undergone simultaneous bilateral biopsies.
Patients and Methods All patients who had temporal artery biopsies at the Yale New Haven Hospital from January 1990 to July 2001 were included in the study (416 patients). Of these, the biopsy findings were positive in 67 patients. Twenty-six of these patients with unilateral biopsies were not included in our study, because neither ESR values nor qualitative descriptions of symptoms were documented in their medical record. Also, 4 patients were removed from our
1294
study because of treatment with corticosteroids for more than 14 days before biopsy excision. Permission for review of medical records and biopsy specimens for this study was granted by the Human Investigations Committee at Yale University School of Medicine. Because this was a retrospective review, doses of corticosteroids were not standardized. The mean duration of treatment before biopsy excision ranged between 0 and 3 days, with a mean of 0.57 days. Because previous studies have documented that histologic findings persist 2 weeks after starting therapy, it is unlikely that treatment lead time affected the findings in our study.12,13 Part 1 of the study involved correlating the histopathologic severity of the giant cell arteritis, on the basis of the biopsy score, with the severity of clinical symptoms and ESR. For this part of the study we incorporated all biopsy-positive patients for whom patient histories or ESR values were available (35 patients). These specimens were determined to be biopsy positive by the Yale University Department of Pathology. The main criterion used by our Department of Pathology was inflammation within the vessel wall with evidence of injury to the media of the vessel. For our comparison of clinical symptoms versus biopsy severity, we had had symptom information in the medical chart for 35 patients, and so they were included in this analysis. Similarly, when comparing ESR versus biopsy severity, we used values from 28 patients for whom ESR values were documented in the medical record. In addition, 2 patients who had received simultaneous bilateral biopsies were included in Part 2 of the study (described later) but were not included in this part of the study, because they did not have ESR values or symptom information documented in their medical record. A scoring system was created (Table 1) including histologic features known to be associated with giant cell arteritis. Moreover, the criteria were weighted so that features indicative of more active disease received more value.14,15 The highest weighted features were the presence of tissue necrosis, circumference involved, and depth of tissue involved; although tissue necrosis is also found in systemic necrotizing vasculitis, its presence together with media inflammation would suggest severe disease. Moderate weight was assigned to inflammation of the small branch vessels. Small branch vessels in this study were defined as small arteries or arterioles that were present in all biopsy slides included in our study. The presence of giant cells received a lower weight, because although this is an important finding for diagnosis, it is less indicative of disease severity. The presence of medial fibrosis and disruption of the internal elastic lamina received a lower weight, which is
Diaz et al 䡠 Biopsy-Positive Temporal Arteritis consistent with data from Cox and Gilks, who found that these features are seen with aging and are not specific for giant cell arteritis.16,17 For each patient included in our study, we first examined every biopsy specimen available and selected 1 slide with the most severe features of inflammation. The number of slides available depends on the length of the specimen obtained and ranged from 2 to 6 slides per specimen. Then, we randomized each of the selected biopsy specimens and graded them blindly, without knowledge of which patient they belonged to or the patient’s medical history. We graded each selected biopsy specimen according to the aforementioned scoring system. We then compared the resulting biopsy severity score with clinical features and ESR. For patients who received bilateral biopsies, we chose the higher of the 2 biopsy scores for analysis. In addition, we also examined the relationship between each histologic feature evaluated and clinical severity. In analyzing the severity of clinical symptoms, we used an ordinal variable to indicate 3 different levels of severity. The first level was superficial pain or headache without any other symptoms. The second level of severity included manifestations of systemic symptoms, such as fever, anorexia, or weight loss. Finally, the third level of severity included visual symptoms such as diplopia, blurriness, or loss of vision. Symptom severity was compared as an ordinal variable, and the biopsy score was compared as a continuous variable. The correlation between biopsy scores and symptom severity was examined by use of the Spearman’s test of correlation. When examining the relationship between biopsy severity score and ESR, we used all available values and compared both the biopsy severity score and ESR as continuous variables. The correlation between biopsy severity scores and ESR values was tested by use of the Pearson’s correlation test. To examine ESR variation among the 3 levels of symptom severity, we used a 1-way analysis of variance test. To examine the differences between the mean ESR values for different symptom categories, we used the independent samples t test. Part 2 of the study involved comparing the severity of giant cell arteritis in the specimen of 1 artery with that obtained from the contralateral side of the same patient. Only patients who had simultaneous bilateral biopsies were studied (7 patients). On review of the medical records, there was no clear reason documented for performing bilateral versus unilateral biopsy other than surgeon preference. Specifically, increased severity of symptoms was not documented as a reason for performing bilateral biopsies. For each patient included in this part of the study, biopsy specimens were selected and graded as described for Part 1. We then examined the difference in the severity between biopsies on the left side and right side. We then used a 1-sample t test to find out whether the mean difference in biopsy severity significantly differed from zero, the null hypothesis. We performed the same type of analysis on each individual histopathologic characteristic to find out whether they differed significantly. Also, we examined the relationship between the 2 sides by plotting left-sided biopsy scores versus right-sided biopsy scores, using a Bland–Altman plot.
Results We found a strong correlation between the severity of clinical symptoms and the stage of disease observed in biopsy specimens (P⬍0.0001). Figure 1 illustrates a rise in biopsy severity score with each level of clinical symptom severity. Also, there are comparable sample sizes within each of the clinical severity groups. Moreover, all of the individual features we examined were significantly correlated with symptom severity, except for fibrosis
Figure 1. Biopsy severity versus symptom severity. Patients in symptom severity group 1 had scalp pain or headache, patients in group 2 had constitutional symptoms (weight loss, fever, etc.), and patients in group 3 had visual manifestations. Biopsy severity scores increase as symptom severity increases. Asterisks indicate outliers. Spearman’s test, P⬍0.001.
(P ⫽ 0.273). The histopathologic features with the strongest correlation to clinical symptoms were necrosis (P⬍0.0001), circumference of involvement (P ⫽ 0.0007), and involvement of the small vessels (P ⫽ 0.002). When comparing ESR to biopsy severity, there seemed to be a positive correlation, as illustrated in Figure 2. However, the correlation did not reach statistical significance (P ⫽ 0.09). Also, the ESR did not correlate significantly with symptom severity (P ⫽ 0.1). There was, however, statistically significant variation among the 3 levels of symptom severity (P ⫽ 0.04). Patients with only superficial pain and headache had a lower mean ESR (49.75) than did patients with systemic (98.36) or ocular (96.17) symptoms. We found that there was not a significant difference in ESR values for patients with systemic versus ocular symptoms (P nonsignificant). However, there was a statistically significant difference when comparing patients with superficial pain/headache versus patients with systemic or ocular symptoms (P ⫽ 0.009). We found that the mean difference in biopsy scores for patients who had bilateral biopsies performed was 5.1, which was significantly above zero (P⬍0.001). Furthermore, our data suggest that the 95% confidence interval for the true mean difference between biopsy specimens is between 2.34 and 7 units of our biopsy severity scale. When we examined each of the histopathologic characteristics individually, we found significant differences in disruption of the internal elastic lamina (P ⫽ 0.03) and circumference of artery involved (P ⫽ 0.0082). In Figure 3, we plotted the absolute difference in biopsy severity score from bilateral biopsies in individual patients versus the average of these biopsy severity scores. This figure illustrates the degree of disparity in disease progression from 1 side compared with the contralateral side, because most points are above zero. There seems to be decreased variation between bilateral biopsy severity scores as the mean score increases, but this trend was not statistically significant.
1295
Ophthalmology Volume 112, Number 7, July 2005
Figure 2. Erythrocyte sedimentation rate versus biopsy severity. Biopsy severity units are according to severity scale in Table 1. There seems to be an increasing trend in erythrocyte sedimentation rate as biopsy severity increases; however, this correlation is not statistically significant. Pearson correlation test, P ⫽ 0.09.
Discussion Giant cell arteritis is a serious disease that can have profound visual consequences. Often unilateral, sequential bilateral, or even simultaneous bilateral biopsies are performed to diagnose the condition. In this study, we formulated a means to study the severity of biopsy specimens and clinical severity to correlate these factors with other features associated with temporal arteritis. An important outcome of this study is the finding that clinical severity of symptoms significantly correlates with the grade of severity of biopsy specimens. This is consistent with the current pathophysiologic model for temporal arteritis in which the inflammation leads to narrowing, especially of the cranial vessels, ultimately leading to ischemia
Figure 3. Bland–Altman plot of biopsy severity. This plot represents all bilateral biopsies included in our study. The absolute value of the difference between left and right biopsies is on the y-axis. Average biopsy severity between left and right side biopsies is on the x-axis. The mean difference in biopsy severity scores was 5.1 (standard deviation [SD] ⫽ 2.975) and is significantly above zero, 1-sample t test, P ⫽ 0.004. Increments in both axes are according to biopsy severity score defined in Table 1.
1296
of the optic nerve.3– 6 This may explain why we found that the temporal artery specimens showing the more severe manifestations of disease came from patients exhibiting visual symptoms. This study showed that patients exhibiting systemic features of disease exhibited more severe evidence of disease on their temporal artery biopsies than did patients with just headache or scalp tenderness. Features such as necrosis and inflammation of the small branch vessels, which showed a particularly high correlation with symptom severity, seem to be indicative of more extensive inflammation. Inflammation of branch vessels may correlate to clinical symptoms, because they represent evidence of more disseminated disease, particularly to smaller vessels. Because smaller vessels have narrower lumens, stenosis of these vessels may be more likely to lead to clinical symptoms. It is unclear whether these trends would be found in patients exhibiting extracranial ischemia, because some studies have shown that these patients may even have negative biopsies.17 Elevated ESR values are associated with temporal arteritis, but we were unaware of any study correlating ESR values with pathologic biopsy severity. This study found that when the ESR values were compared with biopsy severity as 2 continuous variables, there was not a significant correlation (P ⫽ 0.09). However, the group of patients with just face or scalp tenderness had significantly lower ESR values than the other 2 groups (P ⫽ 0.009). This finding suggests that systemic and ocular symptoms occur when there is a greater degree of inflammation. It is possible that the 2 more severe levels of clinical severity—level 2, systemic symptoms, and level 3, visual symptoms— have comparable levels of inflammation, but different symptoms are observed depending on the end organs involved. Therefore, increased inflammation of temporal artery specimens may be detected in patients with visual manifestations because of anatomic proximity to the inflamed cranial vessels, leading to visual symptoms. Given the sensitivity of the acute phase reactant C-reactive protein
Diaz et al 䡠 Biopsy-Positive Temporal Arteritis (CRP) for detecting temporal arteritis,18 one possibility for further study would be to examine the extent to which C-reactive protein follows the pattern of inflammation we have found with ESR. The findings that biopsy severity strongly correlates with severity of symptoms and that patients with systemic or ocular systems have significantly higher ESR values suggest that there is more extensive inflammation with more severe symptoms. There has been some controversy in the literature regarding the degree of inflammation observed in temporal arteritis, particularly in patients with irreversible cranial ischemic complications (ICICs). The article by Cid et al19 indicates that there is a decreased systemic inflammatory response in patients with ICIC. Also, Gonzalez-Gay et al20 report that there may be a decreased risk of visual manifestations in patients with anemia. Nesher and Sonnenblick,21 however, did not find any correlation between inflammatory response and risk for ICIC. Liozon et al22 report that the intensity of the inflammatory response may correlate indirectly with the development of ICICs. Unlike these other studies, however, we have examined inflammation by looking at biopsies directly. This invasive approach illustrates that there is, indeed, increasing inflammation along with severity of symptoms. Also, the fact that ESR was significantly higher in patients with systemic and ocular symptoms further suggests increased inflammatory response in these patients. Last, this study demonstrates that the histopathologic severity of disease can differ significantly from 1 side compared with the contralateral side. Figure 3 illustrates the variation between both sides for patients in our study. If the disease process progressed identically on both sides, we would expect the absolute difference between simultaneous bilateral biopsies to be much lower. Although these features are being used to measure biopsy severity, they are not meant to be diagnostic. For example, the presence of giant cells alone is considered essentially pathognomonic for giant cell arteritis, regardless of whether the other features are present. Therefore, we found a 5% increased yield from performing bilateral biopsies, which is consistent with the work from Boyev et al,11 which also showed a 5% increase in efficacy from performing a contralateral biopsy. Limitations of this study include the fact that patient symptoms are based on retrospective chart review. There exists the possibility of omission or inaccurate depiction of patients’ symptoms as written in charts; the same pathologist initially reviewed all biopsy specimens, however. Another possible limitation is interobserver bias. Because temporal arteritis may be suspected in a variety of settings, including admitted inpatients, emergency room patients, and outpatients, it is possible that physicians across these settings may not assess the patients in exactly the same manner. Also, the use of corticosteroids before biopsy may have influenced our results, because corticosteroids have been found to limit progression of visual loss in temporal arteritis.23 To address this issue, we removed from our study all patients who received corticosteroids for more than 2 weeks before biopsy excision, because several studies have shown that features of temporal arteritis are present after 2
weeks of therapy.11,12 Most of the remaining patients had their biopsy specimens excised within hours of starting corticosteroid therapy (the mean treatment lead time before biopsy excision was 0.57 days), and so it is unlikely that this would have biased our results. To control for all of these limitations, a prospective study would be necessary. In conclusion, our data indicate that there is a strong correlation between severity of clinical symptoms in temporal arteritis and extent of disease observed on histopathologic analysis of temporal artery biopsy specimens. Also, we have found that the ESR values for patients with systemic or ocular symptoms are significantly higher than for patients with superficial pain or headache. Last, we found that the severity of temporal arteritis can differ significantly between simultaneous bilateral temporal artery biopsy specimens from the same patient. Further prospective studies are needed to continue to examine the relationship between the histopathology of temporal arteritis and clinical symptoms.
References 1. Lawrence RC, Helmick CG, Arnett FC, et al. Estimates of the prevalence of arthritis and selected musculoskeletal disorders in the United States. Arthritis Rheum 1998;41:778 –99. 2. 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. 3. Hellmann DB. Temporal arteritis: a cough, toothache, and tongue infarction. JAMA 2002;287:2996 –3000. 4. Smetana GW, Shmerling RH. Does this patient have temporal arteritis? JAMA 2002;287:92–101. 5. Esteban MJ, Font C, Hernandez-Rodriguez J, et al. Smallvessel vasculitis surrounding a spared temporal artery: clinical and pathologic findings in a series of twenty-eight patients. Arthritis Rheum 2001;44:1387–95. 6. Kattah JC, Mejico L, Chrousos GA, et al. Pathologic findings in a steroid-responsive optic nerve infarct in giant-cell arteritis. Neurology 1999;53:177– 80. 7. Gabriel SE, O’Fallon WM, Achkar AA, et al. The use of clinical characteristics to predict the results of temporal artery biopsy among patients with suspected giant cell arteritis. J Rheumatol 1995;22:93– 6. 8. Duhaut P, Pinede L, Bornet H, et al, Groupe de Recherche sur l’Artérite à Cellules Géantes. Biopsy proven and biopsy negative temporal arteritis: differences in clinical spectrum at the onset of the disease. Ann Rheum Dis 1999;58:335– 41. 9. Hunder GG, Bloch DA, Michel BA, et al. The American College of Rheumatology 1990 criteria for the classification of giant cell arteritis. Arthritis Rheum 1990;33:1122– 8. 10. Hall JK, Volpe NJ, Galetta SL, et al. The role of unilateral temporal artery biopsy. Ophthalmology 2003;110:543– 8. 11. Boyev LR, Miller NR, Green WR. Efficacy of unilateral versus bilateral temporal artery biopsies for the diagnosis of giant cell arteritis. Am J Ophthalmol 1999;128:211–5. 12. Achkar AA, Lie JT, Hunder GG, et al. How does previous corticosteroid treatment affect the biopsy findings in giant cell (temporal) arteritis? Ann Intern Med 1994;120:987–92. 13. Gonzalez-Gay MA, Blanco R, Rodriguez-Valverde V, et al. Permanent visual loss and cerebrovascular accidents in giant cell arteritis: predictors and response to treatment. Arthritis Rheum 1998;41:1497–504.
1297
Ophthalmology Volume 112, Number 7, July 2005 14. Lie JT. The classification and diagnosis of vasculitis in large and medium-sized blood vessels. Pathol Annu 1987;22:125– 62. 15. Lie JT, American College of Rheumatology Subcommittee on Classification of Vasculitis. Illustrated histopathologic classification criteria for selected vasculitis syndromes. Arthritis Rheum 1990;33:1074 – 87. 16. Cox M, Gilks B. Healed or quiescent temporal arteritis versus senescent changes in temporal artery biopsy specimens. Pathology 2001;33:163– 6. 17. Brack A, Martinez-Taboada V, Stanson A, et al. Disease pattern in cranial and large-vessel giant cell arteritis. Arthritis Rheum 1999;42:311–7. 18. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96. 19. Cid MC, Font C, Oristrell J, et al. Association between strong inflammatory response and low risk of developing visual loss and other cranial ischemic complications in
1298
20.
21.
22.
23.
giant cell [temporal] arteritis. Arthritis Rheum 1998; 41:26 –32. Gonzalez-Gay MA, Garcia-Porrua C, Rodriguez-Ledo P, Llorca J. Trends in the inflammatory response in biopsyproven giant cell arteritis: comment on the article by Cid et al, and the letters by Nesher and Sonnenblick and Liozon et al [letter]. Arthritis Rheum 2000;43:1427– 8. Nesher G, Sonnenblick M. No association between the inflammatory response and the risk of developing irreversible cranial ischemic complications: comment on the article by Cid et al [letter]. Arthritis Rheum 1998;41:2088 –9. Liozon E, Jauberteau MO, Loustaud V, Vidal E. Association between the inflammatory response and the risk of developing irreversible ischemic complications: comment on the article by Cid et al and the letter by Nesher and Sonnenblick [letter]. Arthritis Rheum 1999;42:2256 – 8. Aiello PD, Trautmann JC, McPhee TJ, et al. Visual prognosis in giant cell arteritis. Ophthalmology 1993;100:550 –5.
Temporal arteritis is a chronic vasculitis that typically affects cranial branches of vessels originating from the aortic arch, although involvement of larger vessels has also been reported. The disease usually occurs in individuals older than the age of 50. In this age group, the prevalence is approximately 200/100 000 persons.1
Symptoms of temporal arteritis range from abrupt to gradual and from mild to severe. Some of the milder symptoms associated with the disease are face pain, headaches, (particularly over the temporal areas), and tender temporal or occipital arteries; this pain may become progressively worse but may also be transient in some cases. Systemic
Originally received: December 16, 2003. Accepted: February 2, 2005. Manuscript no. 230859. 1 Yale University School of Medicine, New Haven, Connecticut. 2 Department of Ophthalmology and Visual Science, Yale University School of Medicine, New Haven, Connecticut. 3 Department of Pathology, Yale University School of Medicine, New Haven, Connecticut.
Presented at: Association for Research in Vision and Ophthalmology Annual Meeting, May, 2002; Fort Lauderdale, Florida. The authors have no proprietary or commercial interest in the marketing of this material and have not received any financial support of any kind. Correspondence and reprint requests to Vicente A. Diaz, Yale Eye Center, 330 Cedar Street, P.O. Box 208061, New Haven, CT 06520-8061. E-mail: [email protected].
© 2005 by the American Academy of Ophthalmology Published by Elsevier Inc.
ISSN 0161-6420/05/$–see front matter doi:10.1016/j.ophtha.2005.02.016
1293
Ophthalmology Volume 112, Number 7, July 2005 Table 1. Biopsy Severity Score Histologic Feature
Point Breakdown
Necrosis Depth of inflammation Circumference Involved Small branch vessel involvement Medial and intimal fibrosis Disruption of internal elastic lamina Giant cells
1 1 1 1 1 1 1
Present, 0 not Media, .5 adventitia, 0 intima Complete, .5 ⬎50%, 0 ⬍50% Present, 0 not Present, 0 not Complete, .5 partial, 0 not present Extensive, .5 present, 0 not
Weight breakdown High ⫽ ⫻ 5 Medium ⫽ ⫻ 3 Low ⫽ ⫻ 1
Total possible score: 21
Weight High High High Medium Low Low Low
Table outlines point allocation system for assessment of severity of giant cell arteritis. High, medium, and low refer to weight given to characteristic to reflect relative contribution to overall severity.
symptoms are often present and include fever, fatigue, and weight loss.2 Visual manifestations of this disease can occur when the vasculitis progresses to the ophthalmic arteries, causing ischemic optic neuropathy. Atypical symptoms of temporal arteritis such as respiratory symptoms, tongue infarctions, and extremity claudication have also been described as a result of inflammation in other vessels throughout the body.3– 6 Although it is difficult to make the diagnosis on the basis of symptoms alone, the absence of symptoms can be helpful in ruling out temporal arteritis.7–9 A temporal artery biopsy should be considered for any patient older than the age of 50 with the aforementioned symptoms.2,4,7–9 A temporal artery biopsy may be performed unilaterally, sequential bilaterally, or simultaneous bilaterally. Recent data suggest that performing a unilateral biopsy can be sufficient to exclude temporal arteritis when clinical suspicion is low.10 However, there may be up to a 5% increase in efficacy when performing a simultaneous bilateral biopsy compared with a unilateral biopsy.11 A proposed mechanism for these findings is that “skip lesions” may be present and, thus, prevent proper diagnosis of disease. Given the grave visual consequences of temporal arteritis, it is critical that physicians have a clear understanding of how this disease progresses. It is the goal of this article to examine the relationship between stage of disease, as quantified by examining temporal artery biopsies, clinical presentation, erythrocyte sedimentation rate (ESR), and age. Also, we examine the severity of disease on 1 side versus the contralateral side in patients having undergone simultaneous bilateral biopsies.
Patients and Methods All patients who had temporal artery biopsies at the Yale New Haven Hospital from January 1990 to July 2001 were included in the study (416 patients). Of these, the biopsy findings were positive in 67 patients. Twenty-six of these patients with unilateral biopsies were not included in our study, because neither ESR values nor qualitative descriptions of symptoms were documented in their medical record. Also, 4 patients were removed from our
1294
study because of treatment with corticosteroids for more than 14 days before biopsy excision. Permission for review of medical records and biopsy specimens for this study was granted by the Human Investigations Committee at Yale University School of Medicine. Because this was a retrospective review, doses of corticosteroids were not standardized. The mean duration of treatment before biopsy excision ranged between 0 and 3 days, with a mean of 0.57 days. Because previous studies have documented that histologic findings persist 2 weeks after starting therapy, it is unlikely that treatment lead time affected the findings in our study.12,13 Part 1 of the study involved correlating the histopathologic severity of the giant cell arteritis, on the basis of the biopsy score, with the severity of clinical symptoms and ESR. For this part of the study we incorporated all biopsy-positive patients for whom patient histories or ESR values were available (35 patients). These specimens were determined to be biopsy positive by the Yale University Department of Pathology. The main criterion used by our Department of Pathology was inflammation within the vessel wall with evidence of injury to the media of the vessel. For our comparison of clinical symptoms versus biopsy severity, we had had symptom information in the medical chart for 35 patients, and so they were included in this analysis. Similarly, when comparing ESR versus biopsy severity, we used values from 28 patients for whom ESR values were documented in the medical record. In addition, 2 patients who had received simultaneous bilateral biopsies were included in Part 2 of the study (described later) but were not included in this part of the study, because they did not have ESR values or symptom information documented in their medical record. A scoring system was created (Table 1) including histologic features known to be associated with giant cell arteritis. Moreover, the criteria were weighted so that features indicative of more active disease received more value.14,15 The highest weighted features were the presence of tissue necrosis, circumference involved, and depth of tissue involved; although tissue necrosis is also found in systemic necrotizing vasculitis, its presence together with media inflammation would suggest severe disease. Moderate weight was assigned to inflammation of the small branch vessels. Small branch vessels in this study were defined as small arteries or arterioles that were present in all biopsy slides included in our study. The presence of giant cells received a lower weight, because although this is an important finding for diagnosis, it is less indicative of disease severity. The presence of medial fibrosis and disruption of the internal elastic lamina received a lower weight, which is
Diaz et al 䡠 Biopsy-Positive Temporal Arteritis consistent with data from Cox and Gilks, who found that these features are seen with aging and are not specific for giant cell arteritis.16,17 For each patient included in our study, we first examined every biopsy specimen available and selected 1 slide with the most severe features of inflammation. The number of slides available depends on the length of the specimen obtained and ranged from 2 to 6 slides per specimen. Then, we randomized each of the selected biopsy specimens and graded them blindly, without knowledge of which patient they belonged to or the patient’s medical history. We graded each selected biopsy specimen according to the aforementioned scoring system. We then compared the resulting biopsy severity score with clinical features and ESR. For patients who received bilateral biopsies, we chose the higher of the 2 biopsy scores for analysis. In addition, we also examined the relationship between each histologic feature evaluated and clinical severity. In analyzing the severity of clinical symptoms, we used an ordinal variable to indicate 3 different levels of severity. The first level was superficial pain or headache without any other symptoms. The second level of severity included manifestations of systemic symptoms, such as fever, anorexia, or weight loss. Finally, the third level of severity included visual symptoms such as diplopia, blurriness, or loss of vision. Symptom severity was compared as an ordinal variable, and the biopsy score was compared as a continuous variable. The correlation between biopsy scores and symptom severity was examined by use of the Spearman’s test of correlation. When examining the relationship between biopsy severity score and ESR, we used all available values and compared both the biopsy severity score and ESR as continuous variables. The correlation between biopsy severity scores and ESR values was tested by use of the Pearson’s correlation test. To examine ESR variation among the 3 levels of symptom severity, we used a 1-way analysis of variance test. To examine the differences between the mean ESR values for different symptom categories, we used the independent samples t test. Part 2 of the study involved comparing the severity of giant cell arteritis in the specimen of 1 artery with that obtained from the contralateral side of the same patient. Only patients who had simultaneous bilateral biopsies were studied (7 patients). On review of the medical records, there was no clear reason documented for performing bilateral versus unilateral biopsy other than surgeon preference. Specifically, increased severity of symptoms was not documented as a reason for performing bilateral biopsies. For each patient included in this part of the study, biopsy specimens were selected and graded as described for Part 1. We then examined the difference in the severity between biopsies on the left side and right side. We then used a 1-sample t test to find out whether the mean difference in biopsy severity significantly differed from zero, the null hypothesis. We performed the same type of analysis on each individual histopathologic characteristic to find out whether they differed significantly. Also, we examined the relationship between the 2 sides by plotting left-sided biopsy scores versus right-sided biopsy scores, using a Bland–Altman plot.
Results We found a strong correlation between the severity of clinical symptoms and the stage of disease observed in biopsy specimens (P⬍0.0001). Figure 1 illustrates a rise in biopsy severity score with each level of clinical symptom severity. Also, there are comparable sample sizes within each of the clinical severity groups. Moreover, all of the individual features we examined were significantly correlated with symptom severity, except for fibrosis
Figure 1. Biopsy severity versus symptom severity. Patients in symptom severity group 1 had scalp pain or headache, patients in group 2 had constitutional symptoms (weight loss, fever, etc.), and patients in group 3 had visual manifestations. Biopsy severity scores increase as symptom severity increases. Asterisks indicate outliers. Spearman’s test, P⬍0.001.
(P ⫽ 0.273). The histopathologic features with the strongest correlation to clinical symptoms were necrosis (P⬍0.0001), circumference of involvement (P ⫽ 0.0007), and involvement of the small vessels (P ⫽ 0.002). When comparing ESR to biopsy severity, there seemed to be a positive correlation, as illustrated in Figure 2. However, the correlation did not reach statistical significance (P ⫽ 0.09). Also, the ESR did not correlate significantly with symptom severity (P ⫽ 0.1). There was, however, statistically significant variation among the 3 levels of symptom severity (P ⫽ 0.04). Patients with only superficial pain and headache had a lower mean ESR (49.75) than did patients with systemic (98.36) or ocular (96.17) symptoms. We found that there was not a significant difference in ESR values for patients with systemic versus ocular symptoms (P nonsignificant). However, there was a statistically significant difference when comparing patients with superficial pain/headache versus patients with systemic or ocular symptoms (P ⫽ 0.009). We found that the mean difference in biopsy scores for patients who had bilateral biopsies performed was 5.1, which was significantly above zero (P⬍0.001). Furthermore, our data suggest that the 95% confidence interval for the true mean difference between biopsy specimens is between 2.34 and 7 units of our biopsy severity scale. When we examined each of the histopathologic characteristics individually, we found significant differences in disruption of the internal elastic lamina (P ⫽ 0.03) and circumference of artery involved (P ⫽ 0.0082). In Figure 3, we plotted the absolute difference in biopsy severity score from bilateral biopsies in individual patients versus the average of these biopsy severity scores. This figure illustrates the degree of disparity in disease progression from 1 side compared with the contralateral side, because most points are above zero. There seems to be decreased variation between bilateral biopsy severity scores as the mean score increases, but this trend was not statistically significant.
1295
Ophthalmology Volume 112, Number 7, July 2005
Figure 2. Erythrocyte sedimentation rate versus biopsy severity. Biopsy severity units are according to severity scale in Table 1. There seems to be an increasing trend in erythrocyte sedimentation rate as biopsy severity increases; however, this correlation is not statistically significant. Pearson correlation test, P ⫽ 0.09.
Discussion Giant cell arteritis is a serious disease that can have profound visual consequences. Often unilateral, sequential bilateral, or even simultaneous bilateral biopsies are performed to diagnose the condition. In this study, we formulated a means to study the severity of biopsy specimens and clinical severity to correlate these factors with other features associated with temporal arteritis. An important outcome of this study is the finding that clinical severity of symptoms significantly correlates with the grade of severity of biopsy specimens. This is consistent with the current pathophysiologic model for temporal arteritis in which the inflammation leads to narrowing, especially of the cranial vessels, ultimately leading to ischemia
Figure 3. Bland–Altman plot of biopsy severity. This plot represents all bilateral biopsies included in our study. The absolute value of the difference between left and right biopsies is on the y-axis. Average biopsy severity between left and right side biopsies is on the x-axis. The mean difference in biopsy severity scores was 5.1 (standard deviation [SD] ⫽ 2.975) and is significantly above zero, 1-sample t test, P ⫽ 0.004. Increments in both axes are according to biopsy severity score defined in Table 1.
1296
of the optic nerve.3– 6 This may explain why we found that the temporal artery specimens showing the more severe manifestations of disease came from patients exhibiting visual symptoms. This study showed that patients exhibiting systemic features of disease exhibited more severe evidence of disease on their temporal artery biopsies than did patients with just headache or scalp tenderness. Features such as necrosis and inflammation of the small branch vessels, which showed a particularly high correlation with symptom severity, seem to be indicative of more extensive inflammation. Inflammation of branch vessels may correlate to clinical symptoms, because they represent evidence of more disseminated disease, particularly to smaller vessels. Because smaller vessels have narrower lumens, stenosis of these vessels may be more likely to lead to clinical symptoms. It is unclear whether these trends would be found in patients exhibiting extracranial ischemia, because some studies have shown that these patients may even have negative biopsies.17 Elevated ESR values are associated with temporal arteritis, but we were unaware of any study correlating ESR values with pathologic biopsy severity. This study found that when the ESR values were compared with biopsy severity as 2 continuous variables, there was not a significant correlation (P ⫽ 0.09). However, the group of patients with just face or scalp tenderness had significantly lower ESR values than the other 2 groups (P ⫽ 0.009). This finding suggests that systemic and ocular symptoms occur when there is a greater degree of inflammation. It is possible that the 2 more severe levels of clinical severity—level 2, systemic symptoms, and level 3, visual symptoms— have comparable levels of inflammation, but different symptoms are observed depending on the end organs involved. Therefore, increased inflammation of temporal artery specimens may be detected in patients with visual manifestations because of anatomic proximity to the inflamed cranial vessels, leading to visual symptoms. Given the sensitivity of the acute phase reactant C-reactive protein
Diaz et al 䡠 Biopsy-Positive Temporal Arteritis (CRP) for detecting temporal arteritis,18 one possibility for further study would be to examine the extent to which C-reactive protein follows the pattern of inflammation we have found with ESR. The findings that biopsy severity strongly correlates with severity of symptoms and that patients with systemic or ocular systems have significantly higher ESR values suggest that there is more extensive inflammation with more severe symptoms. There has been some controversy in the literature regarding the degree of inflammation observed in temporal arteritis, particularly in patients with irreversible cranial ischemic complications (ICICs). The article by Cid et al19 indicates that there is a decreased systemic inflammatory response in patients with ICIC. Also, Gonzalez-Gay et al20 report that there may be a decreased risk of visual manifestations in patients with anemia. Nesher and Sonnenblick,21 however, did not find any correlation between inflammatory response and risk for ICIC. Liozon et al22 report that the intensity of the inflammatory response may correlate indirectly with the development of ICICs. Unlike these other studies, however, we have examined inflammation by looking at biopsies directly. This invasive approach illustrates that there is, indeed, increasing inflammation along with severity of symptoms. Also, the fact that ESR was significantly higher in patients with systemic and ocular symptoms further suggests increased inflammatory response in these patients. Last, this study demonstrates that the histopathologic severity of disease can differ significantly from 1 side compared with the contralateral side. Figure 3 illustrates the variation between both sides for patients in our study. If the disease process progressed identically on both sides, we would expect the absolute difference between simultaneous bilateral biopsies to be much lower. Although these features are being used to measure biopsy severity, they are not meant to be diagnostic. For example, the presence of giant cells alone is considered essentially pathognomonic for giant cell arteritis, regardless of whether the other features are present. Therefore, we found a 5% increased yield from performing bilateral biopsies, which is consistent with the work from Boyev et al,11 which also showed a 5% increase in efficacy from performing a contralateral biopsy. Limitations of this study include the fact that patient symptoms are based on retrospective chart review. There exists the possibility of omission or inaccurate depiction of patients’ symptoms as written in charts; the same pathologist initially reviewed all biopsy specimens, however. Another possible limitation is interobserver bias. Because temporal arteritis may be suspected in a variety of settings, including admitted inpatients, emergency room patients, and outpatients, it is possible that physicians across these settings may not assess the patients in exactly the same manner. Also, the use of corticosteroids before biopsy may have influenced our results, because corticosteroids have been found to limit progression of visual loss in temporal arteritis.23 To address this issue, we removed from our study all patients who received corticosteroids for more than 2 weeks before biopsy excision, because several studies have shown that features of temporal arteritis are present after 2
weeks of therapy.11,12 Most of the remaining patients had their biopsy specimens excised within hours of starting corticosteroid therapy (the mean treatment lead time before biopsy excision was 0.57 days), and so it is unlikely that this would have biased our results. To control for all of these limitations, a prospective study would be necessary. In conclusion, our data indicate that there is a strong correlation between severity of clinical symptoms in temporal arteritis and extent of disease observed on histopathologic analysis of temporal artery biopsy specimens. Also, we have found that the ESR values for patients with systemic or ocular symptoms are significantly higher than for patients with superficial pain or headache. Last, we found that the severity of temporal arteritis can differ significantly between simultaneous bilateral temporal artery biopsy specimens from the same patient. Further prospective studies are needed to continue to examine the relationship between the histopathology of temporal arteritis and clinical symptoms.
References 1. Lawrence RC, Helmick CG, Arnett FC, et al. Estimates of the prevalence of arthritis and selected musculoskeletal disorders in the United States. Arthritis Rheum 1998;41:778 –99. 2. 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. 3. Hellmann DB. Temporal arteritis: a cough, toothache, and tongue infarction. JAMA 2002;287:2996 –3000. 4. Smetana GW, Shmerling RH. Does this patient have temporal arteritis? JAMA 2002;287:92–101. 5. Esteban MJ, Font C, Hernandez-Rodriguez J, et al. Smallvessel vasculitis surrounding a spared temporal artery: clinical and pathologic findings in a series of twenty-eight patients. Arthritis Rheum 2001;44:1387–95. 6. Kattah JC, Mejico L, Chrousos GA, et al. Pathologic findings in a steroid-responsive optic nerve infarct in giant-cell arteritis. Neurology 1999;53:177– 80. 7. Gabriel SE, O’Fallon WM, Achkar AA, et al. The use of clinical characteristics to predict the results of temporal artery biopsy among patients with suspected giant cell arteritis. J Rheumatol 1995;22:93– 6. 8. Duhaut P, Pinede L, Bornet H, et al, Groupe de Recherche sur l’Artérite à Cellules Géantes. Biopsy proven and biopsy negative temporal arteritis: differences in clinical spectrum at the onset of the disease. Ann Rheum Dis 1999;58:335– 41. 9. Hunder GG, Bloch DA, Michel BA, et al. The American College of Rheumatology 1990 criteria for the classification of giant cell arteritis. Arthritis Rheum 1990;33:1122– 8. 10. Hall JK, Volpe NJ, Galetta SL, et al. The role of unilateral temporal artery biopsy. Ophthalmology 2003;110:543– 8. 11. Boyev LR, Miller NR, Green WR. Efficacy of unilateral versus bilateral temporal artery biopsies for the diagnosis of giant cell arteritis. Am J Ophthalmol 1999;128:211–5. 12. Achkar AA, Lie JT, Hunder GG, et al. How does previous corticosteroid treatment affect the biopsy findings in giant cell (temporal) arteritis? Ann Intern Med 1994;120:987–92. 13. Gonzalez-Gay MA, Blanco R, Rodriguez-Valverde V, et al. Permanent visual loss and cerebrovascular accidents in giant cell arteritis: predictors and response to treatment. Arthritis Rheum 1998;41:1497–504.
1297
Ophthalmology Volume 112, Number 7, July 2005 14. Lie JT. The classification and diagnosis of vasculitis in large and medium-sized blood vessels. Pathol Annu 1987;22:125– 62. 15. Lie JT, American College of Rheumatology Subcommittee on Classification of Vasculitis. Illustrated histopathologic classification criteria for selected vasculitis syndromes. Arthritis Rheum 1990;33:1074 – 87. 16. Cox M, Gilks B. Healed or quiescent temporal arteritis versus senescent changes in temporal artery biopsy specimens. Pathology 2001;33:163– 6. 17. Brack A, Martinez-Taboada V, Stanson A, et al. Disease pattern in cranial and large-vessel giant cell arteritis. Arthritis Rheum 1999;42:311–7. 18. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Validity and reliability of various diagnostic criteria. Am J Ophthalmol 1997;123:285–96. 19. Cid MC, Font C, Oristrell J, et al. Association between strong inflammatory response and low risk of developing visual loss and other cranial ischemic complications in
1298
20.
21.
22.
23.
giant cell [temporal] arteritis. Arthritis Rheum 1998; 41:26 –32. Gonzalez-Gay MA, Garcia-Porrua C, Rodriguez-Ledo P, Llorca J. Trends in the inflammatory response in biopsyproven giant cell arteritis: comment on the article by Cid et al, and the letters by Nesher and Sonnenblick and Liozon et al [letter]. Arthritis Rheum 2000;43:1427– 8. Nesher G, Sonnenblick M. No association between the inflammatory response and the risk of developing irreversible cranial ischemic complications: comment on the article by Cid et al [letter]. Arthritis Rheum 1998;41:2088 –9. Liozon E, Jauberteau MO, Loustaud V, Vidal E. Association between the inflammatory response and the risk of developing irreversible ischemic complications: comment on the article by Cid et al and the letter by Nesher and Sonnenblick [letter]. Arthritis Rheum 1999;42:2256 – 8. Aiello PD, Trautmann JC, McPhee TJ, et al. Visual prognosis in giant cell arteritis. Ophthalmology 1993;100:550 –5.