OCR Text |
Show Tournai ofClinirnl Neuro-ophtiulll1lology 8(2): 121-125,1988. '£I 1988 Raven Press, Ltd., New York Specimen Length in Temporal Artery Biopsies Wiley A. Chambers, M.D., and Vitaliano B. Bernardino, M.D. The pathology files of 319 consecutive temporal artery biopsies performed at the Wills Eye Hospital during an ll-year period were reviewed. Fifty-three (17%) of the biopsies were shown to have histological evidence of arteritis. Histopathologic disease was found predominately in women and only in individuals over the age of 60 years. There was no correlation between the length of the biopsy specimen and histological finding in all specimens 4 mm or longer. Sampling probability provided a method of calculating the likelihood of detecting small lesions. With proper sampling techniques, the likelihood of not detecting arteritis due to "skip lesions" in a specimen may be set under 1%. Key Words: Histopathologic disease-Sampling probability- Skip lesions-Specimen length-Temporal arteritis- Temporal artery biopsy, From the Departments of Computer Medicine and Ophthalmology (W.A.C.), George Washington University Medical Center, Washington, D.C., and the Department of Eye Pathology (V.B.B.), Wills Eye Hospital, Philadelphia, Pennsylvania. Address correspondence and reprint requests to Wiley A. Chambers, M.D., Department of Computer Medicine, George Washington University Medical Center, 2300 K Street, N.W., Washington, D.C. 20037, U.S.A. /2/ Temporal arteritis was first described by Hutchinson (1) in 1890, when he identified an 80-yearold gentleman who was unable to wear his hat due to inflamed temporal arteries. The cause of the ailment was ascribed to the hat. As more cases were reported and our understanding of the disease expanded, numerous synonyms became attached to this disease entity including "cranial arteritis," "giant cell arteritis," "granulomatous arteritis," and "arteritis of the aged." Temporal arteritis was later to become the customary term. In 1932, histological evidence of inflammation was reported by Horton et al. (2) and became the definitive basis for diagnosis of the disease. The clinical symptoms of temporal arteritis may include headache, weakness, malaise, jaw claudication, visual loss, anemia, anorexia, insomnia, weight loss, low-grade fever, tenderness over the temporal artery, and elevations in erythrocyte sedimentation rate. These symptoms may occur alone, in combinations, or not at all. The absence of specific clinical symptoms and of any pathognomonic signs has led to ambiguities in diagnosis. Histologic methods have been tried, often with unsatisfactory results (3). In view of good clinical indications to perform biopsies of the temporal artery, the number of positive histological findings of arteritis is variable and often surprisingly low. The percentage of biopsies with positive histological findings has been reported as 12.6% (3), 12.8% (4), 25% (5), 31% (6), and 59% (7). Retrospective reviews of series performed by multiple surgeons report "false-negative" rates of 10% (8), 44% (5), and 61% (9). In these studies, full clinical information, response to therapy, and long-term followup were used to make retrospective diagnoses and establish false-negative rates. The large variability suggests that there are variable factors that influence the histological diagnosis of temporal arteritis. The sampling of arteries is one possible factor. Reports of "skip lesions" have led to suppositions that the i22 W. A. CHAMBERS AND V. B. BERNARDiNO histological findings would be missed if small random segments of the artery were biopsied. The validity of this theory has been questioned (10) and is controversial. If true, the degree of its influence and significance should be examined. To examine the influence of artery specimen size in histological diagnosis of temporal arteritis, a retrospective review was performed of temporal artery biopsies. MATERIALS AND METHODS The surgical pathology files at Wills Eye Hospital in Philadelphia were reviewed for all temporal artery biopsies performed between January 1975 and September 1986. A total of 319 temporal artery biopsies were reviewed on 318 different patients. In only one case was the contralateral side biopsied. Biopsies performed at other institutions were excluded from further consideration because the original sectioning and clinical information was not available. The biopsies were statistically analyzed with respect to age, sex, and length of biopsy specimen. The sampling technique was reviewed by examining the number of segments, number of slides, and total number of sections of the artery. All sectioning consisted of multiple serial sections. Random sampling probability methods were used to calculate the likelihood of detecting focal lesions. Normality testing and Student's t test criteria were used to compare the groups of positive and negative biopsies. Cross-tabs of individual variables and their X2 values calculated (11). Correlation coefficients were examined with respect to histological findings, age, and length of biopsy. A level of p =0 0.05 was used to determine statistical significance. RESULTS . During th~ 11-year period, 319 temporal artery blOPSY speCImens were received by the Department of Eye Pathology of the Wills Eye Hospital. Due to the nature of the hospital and its use as a primar:Y center for ocular problems, the majority of patients presented with some form of ocular compla.int. The most common complaints recorded Included decreased or blurred vision, temporal tenderness, and headaches. The mean age of the respective patients was 70.5 ± 8.2 years (± standard deviation). There were 229 (72%) females and 89 (28%) males, and 163 (51 %) right eyes and 156 (49%) left eyes. The mean length of biopsy specimens was 9.9 ± 5.0 mm. "rl. 1-. - . J' . J i . '1 .,~.. '..'prp '"'."(1", int0 two groups on the basis of histological findings of arteritis (Table 1), Fifty-three of the biopsies (17%) had histological confirmation of arteritis (positive group). This percentage is analogous to previous studies (3-5). Within this group, the mean age of the patients was 73.8 ± 6.6 years, and there were 42 (79%) females and 11 (21 %) males, and 24 (45%) right eyes and 29 (55%) left eyes, The mean length of biopsy specimens was 10,7 ± 4,1 mm. The mean age of the histologically positive group (74 years old) was significantly (p < 0,001) older than the histological negative patient group (70 years old). In the positive group, there are no patients under the age of 61 years (Fig. 1). The ratio of female to male patients was approximately 4:1 in the positive biopsy group. While there is a preponderance of women in the positive group, this ratio is not significantly different from the overall group. The choice of biopsy site, right versus left, was approximately equal. There was no apparent preference by surgeons and no clinical conditions favoring the performance of biopsies of the right or left side, This resulted in an approximately equal num~er of left and right temporal arteries being blOpsied. There was no significant difference in the rate of positive histological findings between left and right temporal artery biopsies. In the single case where a bilateral biopsy was performed, neither side showed histological evidence of arteritis. . The .longer average length of the biopsy specImen In the histologically positive group (10,7 mm) was not significantly different from the histologicall~ .negative group (9.8 mm). The frequency of arten~s showed, no correlation with the length of the bIOPSY speCImen, In the cases reviewed in our institution, variability in the degree of inflammation within specimens was noted, but not quantified . DISCUSSION Temporal ar~eritis is a disease found primarily in older populations as suggested by its synonym TABLE 1. Histologically positive and negative groups Positive Negative group group Significance Number 53 266 Age (years) 73.8 ± 6.6 69.8 ± 8.3 p < 0.001 % Female 79 70 NS % Right eye 45 52 NS Specimen length (mm) 10.7 ± 4.1 9.8 ± 5.1 NS NS, not significant. TEMPORAL ARTERY BIOPSIES 123 70 POSITIVE BIOPSY 60 50 a: UJ :c;o; 40 z=' 30 20 10 cases and a reversal with male predominance in the presumptive group. In the present study and most reports reviewed (6,7), the percentage of women in biopsy-proven cases is proportionate to the percentage in the total number of biopsy cases. It is therefore not a factor in predicting histological findings. The longer mean length of biopsy specimens in the positive group is not statistically (p > 0.05) different from the negative group. The differences between the groups arise from the lower tail of the specimens (Fig. 2). There were no biopsies under 4 mm in length that were found to have histological findings of arteritis. If the percentage of positive findings had been evenly distributed throughout the lower tail, three specimens would have been expected. The distribution may be by chance, or may be a function of the sampling of the segment. 30 50 70 90 60 POSITIVE BIOPSY 50 a: 40 LU <0 ~ :::> 30 z 20 10 0 30 50 70 90 YEARS 60 NEGATIVE BIOPSY 50 a: 40 LU co :~::> 30 z 20 10 0 YEARS FIG. 1. Distribution of patients' age with histological findings of arteritis (positive biopsy) compared to patients without significant histological findings (negative biopsy). o 2 5 8 11 14 17 20 23 26 29 32 35 38 41 LENGTH IN MILLIMETERS "arteritis of the aged" (12). In this review, there were no patients under the age of 61 years who showed positive histological findings. Studies in which age is compared between groups of patients with temporal arteritis show a higher age of patients with positive histological findings (8,9). In each of these studies, all patients in the positive groups were above the age of 60 years old. The mean age of 74 years is remarkably similar in each study (4,8,9), with most patients between the ages of 70 and 85 years. While cases of temporal arteritis do occur below the age of 60 years, rarely do biopsies of these patients show evidence of arteritis. Female predominance was evident in both the positive and negative groups. When the sex distribution of the older age group involved is taken into account, previous studies vary in their amount of female predisposition (6-8). Roth (9) showed a female predilection in biopsy-proven 70 NEGATIVE BIOPSY 60 50 a: UJ :c;o; 40 z=' 30 20 10 o o 2 5 8 11 14 17 20 23 26 29 32 35 38 41 LENGTH IN MILLIMETERS FIG. 2. Distribution of total specimen lengths submitted for pathologic review in cases with histological findings of arteritis (positive biopsy) compared to specimen lengths without significant histological findings (negative biopsy). J Clin Neuro-ophthahnol, Vol. 8. No.2. 1988 124 W. A. CHAMBERS AND V. B. BERNARD/NO Statistically, the element of chance cannot be excluded. As a specimen decreases in size, it is more difficult to divide into multiple segments. As a result, less sections are often made and reviewed. If less sections are reviewed, the "miss rate" due to skip lesions will increase. The sampling primarily affects the smaller length biopsies and should be the basis for sampling size recommendations. The absence or presence of "skip lesions" in temporal arteritis has been a heavily debated topic during the past decade. Studies in 1974 by Cohen and Smith (10) showed no evidence of "skip lesions" in serially sectioned biopsies. In their article, they made reference to the high variability in the degree of inflammation from one layer to another and from one segment to the next. Historically, it is this variability that has led to speculation that there must be areas free from inflammation. Cohen and Smith argued that while there was variability, there were no areas totally free of inflammation. They emphasized the lack of any documented proof of an area totally free of inflammation surrounded by areas of involvement. In 1976, however, Klein (13) and Albert (14) each identified foci of inflammation in otherwise normal biopsies. In Klein's and Albert's studies, focal regions were noted in 28% and 18% of the cases reviewed, respectively. Both studies noted focal lesions with lengths of approximately 0.3 mm. While the existence of "skip lesions" may provide an explanation for some negative biopsies in cases of clinically diagnosed arteritis, statistically it is unlikely to account for all negative results. Even if the assumption is made that "skip lesions" occur in as high as 30% of the cases, and that the biopsy is always reported as negative in cases with "skip lesions," then a 17% positive biopsy rate can only be expected to increase to a maximum of 40%. It is unlikely, however, that all "skip lesions" would be reported as negative. The probability of taking random sections through an artery and finding a focal area may be calculated through the use of the following formula (15) Probability of finding a lesion 1 - probability of not finding area ( affected )n -1- 1------- segment length where n = number of sections. In this formula, the affected length (0.29 mm) is the smallest length of a focal area reported (13,14) and the percentage of arteries with focal arteritis (28%) is the largest percentage of arteries reported to have skip lesions (13). The number of sections is equal to the number of slides per segment. The probability of finding a focal area of inflammation is based on an exponential function (Fig. 3). As the number of sections increases, its additional effect on the probability decreases. The probability of finding small lesions is significantly affected by the segment size (Table 2). Smaller 1.0 A A c:.-~c:.-tr--e:.-c:.'"""""6 (/) 0.9 A A • • • Z A • • 0 A • • • (/) 0.8 A • w • ~ A • e,:) 0.7 • z A • )( )( )( c::J 0.6 • )( )( z • )( )( FIG. 3. The probability of finding u.. 0.5 c:. )( u.. • )( focal lesions of 0.29 mm (the )( 0 0.4 )( (> (> smallest size focal area reported >- • )( (> (> (> in the literature) is shown for dif- f- )( (> -l 03 )( co • )( (> (> (> (> ferent segment lengths. The hori- « 0.2 lC (> zontal axis represents the number (> (> co )( (> of random sections viewed within 0 aa:.:. 0.1 ()>( (> (> (> each segment. 0 0 5 10 15 20 NUMBER OF SECTIONS (> 1cm SEGMENT LENGTH )( 5mm SEGMENT LENGTH • 2mm SEGMENT LENGTH _", ':-1-:"0 ~E:_"':~I\:r'IT L~"IGTH TEMPORAL ARTERY BIOPSIES 125 TABLE 2. Probability of finding a focal areaa • The size of the focal area is 0.29 mm. Calculations are based on random sampling. 1 3% 6% 14% 29% 2 6% 11% 27% 50% 3 8% 16% 37% 64% 4 11% 21% 47% 75% 5 14% 26% 54% 82% 10 25% 45% 79% 97% 15 36% 59% 90% 99% 20 44% 70% 96% 99% segments require less sections to achieve the same probability. A sampling of the "routine" procedure at our institution revealed an average of 2 mm segments and an average of eight slides per segment. The calculated "miss rate" from "skip lesions" would be expected to be 8%. The calculations presented represent a worst case situation based on literature previously reported (13,14). The calculations assume that skip lesions occur at the highest frequency reported and always occur as single focal areas of 0.29 mm in size. Although a focal area of 0.29 mm has been reported (13), it was not the only area of arteritis present in the specimen, as assumed in the calculations. Because of these assumptions, the actual probabilities of finding some evidence of arteritis may be expected to be higher and the expected rate of missing a lesion would be lower. The effect of "skip lesions" is probably overrated as an explanation for negative biopsies. These "miss rate" calculations do not consider the total length of the specimen as a factor. They only reflect the probability of finding a lesion within the segment obtained. The lack of correlation with the length of the specimen suggests that REFERENCES 1. Hutchinson J. Diseases of the arteries. Arch Surg 1890; 1:323. 2. Horton BT, Magath TB, Brown GE. An undescribed form of arteritis of the temporal vessels. Mayo Clin Proc 1932; 7:700. 3. Ehsan M, Lange RK, Waxman J. Overdiagnosis of temporal arteritis. South Med J 1981;74:1198-200. 4. McDonnell Pj, Moore GW, Miller NR, Hutchins GM, Green WR. Temporal Arteritis: A clinicopathologic study. Ophthalmology 1986;93:518-30. 5. Allsop Cj, Callagher PJ. Temporal artery biopsy in giantcell arteritis: A reappraisal. Am JSurg PathoI1981;5:317-23. 6. Huston KA, Hunder CG, Lie }T, Kennedy BA, Elveback LR. Temporal arteritis: a 25-year epidemiologic, clinical. and pathologic study. Ann Intem Med 1978;88:162-7. 7. Bengtsson BA, Malmvall BE. The epidemiology of giant cell arteritis including temporal arteritis and polymyalgia rheumatica. Arthritis Rheum 1981;24:899-904. 8. Hedges TR, Gieger GL. Albert OM. The clinical value of negative temporal artery biopsy specimens. Arch OphthalmoI1983; 101:1251-4. 9. Roth AM, Milsow L, Keltner JL. The ultimate diagnosis of patients undergoing temporal artery biopsies. Arch OphthalmoI1984; 102:901-3. 10. Cohen DN, Smith TR. Skip areas in temporal arteritis: myth versus fact. TrailS Am Acad Ophthalmol Otolaryngol 1974;78(OP):772-83. 11. Witte RS. Statistics. New York: CBS College Publishing, 1985. 12. Huston KA, Hunder GG. Giant cell (cranial) arteritis: a clinical review. Am Heart 11980;100:99-107. 13. Klein RG, Campbell Rj, Hunder GG. Carney JA. Skip lesions in temporal arteritis. Mayo Clin Proc 1976;51:504-10. 14. Albert DM, Ruchman MC, Keltner JL. Skip areas in temporal arteritis. Arch Ophthalmol 1976;94:2072-7. 15. Raiffa H. Decision analysis. Reading, PA: Addison-Wesley, 1968;7-21. histological evidence of arteritis would not be found in portions of the artery not biopsied. From this study, the following guidelines are recommended: (a) biopsies should be a minimum of 4 mm in length; (b) biopsies should be sectioned into 1 mm segments; and (c) a minimum of nine sections should be reviewed from each segment. These guidelines are expected to provide a 99% probability of detecting any evidence of arteritis in a specimen. 1 mm segment 2mm segment 5mm segment 1 em Number of views segment 1Gill Nellro,ol,h"t11/",ol. Vol. 8. No.2. 1988 |