Clinical Spectrum, Prognosis, and Laboratory Abnormalities

CLINICAL REVIEW Linear Scleroderma Clinical Spectrum, Prognosis, and Laboratory Abnormalities VINCENT FALANGA, M.D.; THOMAS A. MEDSGER, Jr., M.D.; MORRIS REICHLIN, M.D.; and GERALD P. RODNANt, M.D.; Pittsburgh, Pennsylvania; and Oklahoma City, Oklahoma The clinical features and natural history of linear scleroderma in 53 patients and the laboratory tests helpful in the management of this disease are described. No patient had Raynaud's phenomenon or signs of systemic connective tissue disease in a mean follow-up of 10 years. Blood eosinophilia (> 300 cells/mm 3 ) was present in half the patients, usually those with clinically active disease rather than inactive disease (p < 0.02). An elevated serum IgG level correlated with the presence of joint contractures (p < 0.02). Antinuclear antibodies, commoner in patients with extensive and prolonged disease, were present in 31% and 46% of patients whose sera were tested on mouse kidney and HEp-2 cells, respectively. Antibodies to single-stranded DNA, present in 50% of patients, were associated with extensive disease, joint contractures (p < 0.001), and active disease of greater than 2 years' duration (p < 0.001). Discordance in immune reactivity indicates that at least three serum autoantibodies exist in these patients: antibodies to single-stranded DNA and antinuclear antibodies with homogeneous and nucleolar immunofluorescence patterns. LOCALIZED SCLERODERMA differs from systemic sclerosis in that localized scleroderma is not characterized by Raynaud's phenomenon, acrosclerosis, or involvement of internal organs (1). The morphologic features of localized scleroderma have led investigators to distinguish three variants: morphea, generalized morphea, and linear scleroderma (2, 3). Patients with morphea have one or few circumscribed sclerotic plaques with variable degrees of hyperpigmentation, hypopigmentation, and a violaceous border. In those with generalized morphea, more individual patches occur, involving more surface area. Patients with linear scleroderma have sclerotic areas of skin in a linear, bandlike distribution, often crossing joint lines and leading to mild, but occasionally severe and disabling, joint contractures. Furthermore, the inflammatory and fibrotic process in linear scleroderma may involve the underlying subcutaneous tissue and muscle (4-6). When linear scleroderma occurs on the face or scalp, it may assume an ivorylike, depressed appearance referred to as en coup de sabre. Localized scleroderma likely has an immunologic basis. The disease has been reported to coexist with systemic sclerosis (3, 7) and systemic lupus erythematosus (8, 9). Also, by immunofluorescence techniques, antinuclear From the Department of Dermatology and the Division of Rheumatology and Clinical Immunology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; and the Arthritis and Immunology Laboratory, Oklahoma Medical Research Foundation, Oklahoma City, Oklahoma. t Dr. Gerald Rodnan is deceased. Annals of Internal Medicine. 1986;104:849-857. antibodies can be detected in 15% (10) to 47% (11) of patients. The frequency of positive antinuclear antibody titers is greatest in patients with the linear form of localized scleroderma, having been as high as 67% when HeLa cells were used as the test substrate (12). Other serologic abnormalities, such as high titers of rheumatoid factor and antibodies to single-stranded DNA, have also been described (13, 14). In this article, we review our experience with the clinical features and prognosis of linear scleroderma. Laboratory abnormalities that may be helpful in judging the activity of the cutaneous lesions and in predicting the extent and course of the cutaneous involvement are also described. Patients and Methods Fifty-three patients with linear scleroderma were followed at the University of Pittsburgh from 1962 to 1983. All had linear scleroderma as denned as linear, sclerotic areas of bound-down skin showing variable degrees of hyperpigmentation or hypopigmentation, or both, and histologic features of dermal sclerosis with entrapment of eccrine glands and blunting of the rete ridge. Each patient had a complete history and physical examination on the initial and follow-up visits. The patients were seen every 6 months, with some patients being seen more frequently at the beginning of their illness. After the disease was considered inactive for at least a year, the patients had follow-up visits at yearly intervals. In most, initial and sequential diagrams were made with precise measurements of the extent and severity of the cutaneous lesions. The disease was considered to be inactive if no new lesions or extension of existing lesions was documented for a period of at least 6 months. The disease was considered active if new lesions had developed or old lesions had grown within the preceding 6 months. The duration of active disease was calculated as the number of years from the time of onset of the first cutaneous lesion to the time that the last active lesion became quiescent. Patients whose courses were characterized by exacerbations and remissions of the cutaneous lesions were said to have recurrent disease. On the initial visit and on many subsequent visits, we determined a complete blood count and platelet count, peripheral blood eosinophil count, serum immunoglobulin level, and the titers of rheumatoid factor and other serum autoantibodies. Peripheral blood eosinophilia was defined as a total eosinophil count of greater than 300 cells/mm 3. Serum levels of IgG, IgM, and IgA were considered increased if the level was greater than the upper limit of normal for the patient's age and sex. A rheumatoid factor titer of 1:20 or greater by the latex agglutination method was considered elevated. Complete blood counts and platelet counts were repeated at least at yearly intervals in 42 patients and were done for all 1986 American College of Physicians 849

Table 1. Clinical and Laboratory Features in 53 Patients with Linear Scleroderma Sites of Disease Age (yrs)/sex Sites of Linear Scleroderma* Contractures Duration and Patient Face Arms L Trunk Elbow Wrist Fingers Knee Ankle. of Disease yrs Face only 1 1/F + Recurrent 2 3/F + Recurrent 3 6/F + Recurrent 4 8/F + Recurrent 5 Q/F + - - - - - - - - 1 6 16/F + i 7 16/F + 4 Face plus morphea 8 8/F + - - - - - - - - 1 Face plus trunk and extremities plus morphea 9 16/F + L,R L L 4 4 4 - - 4 10 16/F + R L,R - + 4- + + + 4 11 3/M + R L - - - - - - 12 15/M 4 L,R R - + + + 4 Trunk and extremities only 13 2 / F - - L - - - - + - 4 14 3 / F - L - - + + - - 1 15 5 / F - - R - - - - + + 4 16 6/F - - L,R - - - - - 2 17 12/F - - L - - - - - 5 18 18/F R _ + _ 3 19 20/F _ R 2 20 28/F R + 3 21 32/F - L - - - - - - 1 22 36/F L - - - - - - 23 46/F _ R _ l 24 51/F _ L,R - - - - - 1 25 67/F - L - - - - - - 26 3/M R R R + 4 + 4 27 4/M _ R + _ l 28 9/M - - R - - - - 2 29 21/M - - - 3 30 24/M _ R _ 31 64/M _ R _ i Trunk and extremities plus morphea 32 2/F - L,R - - - - - - 1 33 2/F - L L R + + + + 3 34 5/F R R _ Recurrent 35 6/F L L,R R + + + + 5 36 7/F - L L,R - + 4- - + 37 7/F - R R - - - - 38 8/F - L L,R - 4 + 4 + 4-4 39 9/F - R R 4 4 4-5 40 10/F - - R - - - - 4 41 12/F L,R L,R - 4 4 4 4 3 42 14/F L,R R - 4 4 4 4 3 43 15/F L,R 4 Recurrent 44 18/F _ R L - - - - - 5 45 20/F - R - - 4 - - - - 46 54/F - - L,R L,R _ Recurrent 47 57/F L, R - _ l 48 2/M - L L - - 4 - + - 5 49 4/M _ - _ 2 50 6/M - L L,R - - - - 4 + 4 51 10/M - L L - + _ + 4 52 14/M L,R L,R L,R - 4 4 4 + 4 53 20/M L - - - - - - * L indicates disease on the left side; R indicates disease on the right side, t h homogeneous; s = speckled; n = nucleolar. 850 June 1986 Annals of Internal Medicine Volume 104 Number 6

Table 1. (Continued) Sites of Disease Duration Activity Eosinophils IgG Rheumatoid Titer and Pattern Single-Stranded and Patient of Follow-up of Disease Factor of Antinuclear Antibodiesf DNA Binding Mouse HEp-2 Substrate Cells yrs n/mm 3 mg/dl % Face only 1 10 + 325 1220 1 2 26 + 144 1150 27 3 64 + 1300 4 10 59 680 4 5 2 + 0 6 2 648 980 2 7 6 181 _ Face plus morphea 8 3 + 511 1430 24 Face plus trunk and extremities plus morphea 9 6 + 1300 640 1620, h 1080, h 13 10 19 + 836 1475 1280 180, h 140, h 59 11 2 + 80 740 180, n 360, n 4 12 19 136 1025 20 20, h 40, h 20 Trunk and extremities only 13 9 + 1020 1250 20 120, n 38 14 2 15 11 + 68 1070 1620, h 360, h 14 16 4 17 13 + 308 800 20, h 33 18 32 350 5 19 14 + 0 880 20 40, h 1 20 5 + 270 1400 15 21 2 + 156 22 1 + 615 1100 320 180, s 360, s 4 23 2 168 950 4 24 5 840 _ 5 25 8 59 1250 360, n 34 26 2 + 1000 24 27 2 + 216 1100 3 28 4 0 _ 2 29 7 + 495 30 1 + 1212 1550 31 4 + 428 660 40, h 2 Trunk and extremities plus mor phea 32 7 + 820 3 33 7 + 1088 1250 180, s 40, h 55 34 52 142 950 2 35 10 + 990 1400 20 20, h 66 36 4 + 469 1950 37 7 + 504 1200 4 38 10 + 67 1850 2560 4860, h 1080, h 42 39 6 132 7 40 7 + 0 930 5 41 16 + 42 10 + 544 1700 160 20, h 1080, h 69 43 14 + 0 1250 40, n 19 44 25 + 672 1150 45 2 + 342 1140 46 7 47 2 204 48 21 0 1250 11 49 7 980 930 40, n 2 50 10 _ 20, h 40, h 34 51 5 + 40, n 5 52 7 + 2486 2900 120, h 16 53 1 + 712 1600 5 Falangaetal. Linear Scleroderma 851

Figure 1. The distribution of the age of onset of linear scleroderma. Open bars indicate females, and hatched bars indicate males. patients at the end of our follow-up period. A urinalysis was done initially on all patients and was repeated yearly in 37 patients and monthly in 12 of these patients who had very extensive and active disease treated with ^-penicillamine. A urinalysis was done at the end of our follow-up in 46 patients. Roentgenograms of the chest and affected extremities were done on the initial visit in 29 and 13 patients, respectively. Roentgenograms of the chest were done once during the followup in another 15 patients. In 12 patients with extensive and widespread linear lesions, but who were otherwise asymptomatic, barium contrast studies of the esophagus and upper gastrointestinal tract were also done. ANTINUCLEAR ANTIBODIES All antinuclear antibody tests were done simultaneously on serum samples that had been stored at 70 C. Indirect immunofluorescence was done on mouse kidney sections that were fixed for 1 minute in cold acetone. The rest of the procedure was done as previously described, and sera were screened at a 1:20 dilution (15). HEp-2 cells were also used as substrate, and the serum specimens were screened for antibody at a 1:40 dilution. Fluorescein-treated anti-human antibody conjugate was polyvalent for IgG, IgM, and IgA. All sections were viewed by one of us with a microscope equipped with epillumination optics. Antibodies to native DNA were sought by using a Crithidia luciliae substrate and screening the serum sample at a 1:10 dilution. SINGLE-STRANDED DNA ANTIBODIES We prepared single-stranded DNA by heating Escherichia coli native DNA (Sigma Chemical Company, St. Louis, Missouri) at 100 C for 10 minutes and then plunging it into ice water. Radioactive labeling with 125 I was done by the method of Commerford (16). Specific activity of the 125 I-single-stranded DNA was approximately 1 Ci/u.g of DNA in phosphatebuffered saline with 1% bovine serum albumin. For each experiment, 10 \x L of radiolabeled DNA was diluted in borate-saline buffer, ph 7.8, to a specific activity of about 2000 counts/min- 75 u,l. The working buffer in all radioimmunoassays was 0.03- M borate and 0.15-M saline, ph 7.8. Triton-X 100 (Sigma Chemical Company) was present in all radioimmunoassay tubes at a final concentration of 0.1% (vol/vol). The Farr assay is a modification of the procedure described by Wold and colleagues (17). Twenty microlitres of serum were incubated overnight with 5 JLIL of buffer acid and 75 ux of diluted l25i_dna at 4 C to give a final volume of 100 JLLL. On the second day, 50 ux of saturated ammonium sulfate was added to each tube, and the tubes were then incubated for 1 hour at 4 C. After centrifugation, each precipitate was washed with 33% saturated ammonium sulfate and centrifuged for 15 minutes twice. The results in tubes were counted as above. Results of the radioimmunoassays are expressed as percent binding of the radiolabeled DNA and were calculated as follows: percent binding = [(p B)/(I B)] X 100, where p is counts per 10 minutes for precipitates, B is counts per 10 minutes for blank tubes, and / is counts per 10 minutes for 75 u.l of diluted 125 I-single-stranded DNA. Statistical analysis was done by chi-square analysis, using the appropriate correction when indicated. Results CLINICAL FEATURES None of our 53 patients had Raynaud's phenomenon or symptoms or signs of systemic lupus erythematosus or systemic sclerosis after a mean follow-up of 10 years. Specifically, none developed photosensitivity, dysphagia, shortness of breath, pleuritic chest pain, or symptoms of pericarditis. During the follow-up period, we did not note periungual telangiectasia, cutaneous lesions of discoid or systemic lupus erythematosus, or widespread and symmetric induration of the face, fingers, and proximal forearms suggestive of systemic sclerosis. No patient developed leukopenia or thrombocytopenia. Urinalysis results were initially normal in all patients. Proteinuria developed in 3 patients taking ^-penicillamine, but it resolved after therapy with the drug was stopped. The roentgenograms of the chest were normal in all but 3 older patients, who had radiographic evidence of chronic obstructive pulmonary disease. The radiographic studies of the esophagus were all normal. The clinical features and laboratory results in our 53 patients are shown in Table 1. The age distribution by sex of all patients at the time of onset of linear scleroderma is shown in Figure 1. Although older patients were seen, 44 (83%) of our patients were under age 25 when the disease began. The female-to-male ratio was 4:1. One patient (Patient 18) had two sons with isolated Raynaud's phenomenon and a female cousin with systemic sclerosis. A history of trauma to the area subsequently involved with linear scleroderma was present in 12 patients (23%). In 1 patient (Patient 2), trauma clearly led to an exacerba- 852 June 1986 Annals of Internal Medicine Volume 104 Number 6

Table 2. Laboratory Abnormalities in Patients with Linear Scleroderma Total Blood Eosinophils Elevated Rheumatoid Antinuclear Antibodies Anti-ssDNA* Patients > 300/mm3 Serum IgG Factor Mouse Kidney HEp-2 Cells Level Substrate Substrate n < n positive/patients tested (%) - All patients 53 21/42 (50) 20/39 (51) 9/34 (26) 12/39 (31) 18/39 (46) 20/39 (51) Patient classification! E only 7 2/6 (33) 2/5 (40) 0/4 0/4 0/4 1/4 (25) E plus M 1 1/1 (100) 1/1 (100) 0/1 0/1 0/1 1/1 (100) E plus M plus LS 4 1/3 (33) 2/4 (50) 3/4 (75) 4/4 (100) 4/4 (100) 3/4 (75) LSTE plus M 22 10/18 (55) 11/17 (65) 3/13 (23) 5/17 (29) 8/17 (47) 9/17 (53) LSTE only 19 7/14 (50) 4/12 (30) 3/12 (25) 3/13 (23) 6/13 (46) 6/13 (46) Disease activity Active 35 18/29 (62)% 18/30 (60) 8/23 (35) 10/26 (38) 14/26 (54) 15/26 (58) Inactive 18 3/13 (23)J 2/9 (22) 1/11 (9) 2/13 (15) 3/13 (23) 5/13 (38) Number of linear scleroderma lesions Two or more 26 8/16 (50) 11/18 (61) 6/18 (33) 9/19 (47) 11/19 (58) 13/19 (68) One 27 13/26 (50) 9/21 (43) 3/16(19) 3/20(15) 7/20 (35) 7/20 (35) Contractures Present 23 9/17 (53) 13/17 (76)t 7/16 (44) 9/19 (47) 12/19 (63) 16/19 (84) Absent 30 12/25 (48) 7/22 (32) J 2/18 (11) 3/20(15) 6/20 (30) 4/20 (20) Disease duration > 2 years 22 10/18 (55) 10/15 (67) 7/15 (47) 10/18 (55) 11/18 (61) 15/18 (83) < 2 years 15 4/11 (36) 2/9 (22) 1/9 (11) 0/10 3/10 (30) 1/10 (10) * Anti-ssDNA = antibodies to single-stranded DNA. te = en coup de sabre; M = morphea; LS = linear scleroderma; LSTE = linear scleroderma of trunk and extremities. X p < 0.02. p< 0.001. tion of the cutaneous lesions. The onset of linear scleroderma was usually slow and insidious, but in several patients it was rapid. Edema and erythema with or without pain appeared early in 5 patients. Cutaneous ulcerations of the involved extremity occurred in 1. In 1 patient, fever and regional adenopathy preceded the onset of linear scleroderma. Lesions were most frequently seen on the trunk and extremities (45 patients), and 22 of these patients also had patches of morphea present. Most linear lesions were located on the extremities, and frequently, two or more lesions were present (61%), often bilaterally (46%). Lesions on the trunk would occasionally cross the midline. The en coup de sabre variety of linear scleroderma was present in 12 (23%) patients, and 4 patients had morphea and en coup de sabre with linear lesions on the extremities or trunk. Five patients with lesions of en coup de sabre also had facial hemiatrophy. Generalized anthralgia, symmetric and involving large joints, was present at the onset or before the development of linear scleroderma in 7 patients (13%); in only 2 of these patients was an elevated titer of rheumatoid factor present. The arthralgia disappeared as the cutaneous disease became more prominent. Roentgenograms of affected joints in the patients with arthralgia were normal. Joint contractures, frequently involving more than one joint, were a common result of the subcutaneous and deeper extension of cutaneous lesions; these contractures occurred in 23 (56%) of the patients with linear lesions of the trunk and extremities. The contractures involved the fingers (36%), ankles (34%), wrists (32%), elbows (27%), and knees (24%) and, except for Patient 18, were limited to the areas involved with linear scleroderma. Surgical procedures to restore proper joint motion were only partly successful. Amputation was needed in 3 patients because of severely impaired growth of the affected limb (Patients 11 and 27) and because of impaired vascularity (Patient 50). The overall mean ( it SD) duration of active disease was 2.8 it 1.5 years, but recurrent disease was also seen in 13% of the patients, particularly in those with en coup de sabre. One such patient (Patient 3) had frequent exacerbations of her facial lesions throughout her life. No change in the cutaneous involvement was noted in 3 patients who became pregnant, and in 1 patient, pregnancy led to softening of the preexisting Figure 2. The natural history of eosinophil counts in ten patients with active linear scleroderma who were followed longitudinally. Each set of lines represents a patient; open circles indicate active disease, and closed circles indicate inactive disease. Asterisks indicate treatment with systemic corticosteroids. Falanga et al. Linear Scleroderma 853

Figure 3A. Levels of single-stranded DNA binding of serum samples from normal persons, patients with systemic lupus erythematosus (SLE), and patients with linear scleroderma. Figure 3B. Natural history of single-stranded DNA binding of sera from six patients with active linear scleroderma (open circles). Closed circles indicate inactive disease. lesions, followed by the development of new sclerotic lesions after delivery. EOSINOPHILIA Peripheral blood eosinophilia was seen in 21 of 42 patients (Table 2). In none was it attributable to an adverse drug reaction or parasitic infestation. Eosinophilia was commoner in patients with active disease (p < 0.02), but its presence did not correlate with the number of cutaneous lesions. Figure 2 shows the natural history of eosinophil counts in 10 patients with active linear scleroderma who were followed longitudinally. Active disease, as defined earlier, closely paralleled the eosinophil count, which declined to normal levels in all patients during follow-up. In 2 of the patients in Figure 2, the initial decline in eosinophil count may have been due to therapy with systemic corticosteroids. OTHER LABORATORY FINDINGS The level of IgG was increased in patients with active and extensive disease and was more often elevated in patients with joint contractures (p < 0.02, Table 2). Levels 854 June 1986 Annals of Internal Medicine Volume 104 Number 6

of IgM and IgA were elevated in 51% and 47% of patients, respectively, but did not correlate with any clinical features. A serum rheumatoid factor of 1:20 or more was present in 9 of 34 patients (26%), and the titer ranged from 1:20 to 1:2560. Although the presence of rheumatoid factor did not correlate significantly with any clinical findings (Table 2), a high titer was usually associated with severer cutaneous and articular involvement. AUTOANTIBODIES The frequency of antinuclear antibodies was 31% when serum was tested on mouse kidney substrate and 46% when tested on HEp-2 cells. On both substrates, antinuclear antibodies were commoner in patients with severe and extensive disease (Table 2). In all patients with a positive antinuclear antibody titer on mouse kidney substrate, the duration of disease was longer than 2 years. The antinuclear antibody titer ranged from 1:20 to 1:4860 on mouse kidney substrate and 1:20 to 1:3240 on HEp-2 cells. The commonest nuclear immunofluorescence pattern was homogeneous (75%), followed by fine nuclear speckled (17%) and nucleolar (8%). No antibodies to double-stranded DNA were seen in any of the 39 serum samples tested. The percent binding to single-stranded DNA of serum from normal persons, patients with linear scleroderma, and patients with systemic lupus erythematosus is shown in Figure 3A. In persons with active and multilesion linear scleroderma, the mean binding level was similar to that seen in patients with systemic lupus erythematosus. Table 2 shows that antibodies to single-stranded DNA were present in 19 of 38 patients with linear scleroderma and were commoner in those with joint contractures (p < 0.001) and disease duration of greater than 2 years (p < 0.001). Figure 3B shows the natural history of antibodies to single-stranded DNA in 6 patients with active linear scleroderma who were followed longitudinally. The activity of the disease paralleled the serum level of single-stranded DNA binding. In 4 patients followed until their disease became inactive, the percentage of singlestranded DNA binding declined towards normal. There was discordance in the immunofluorescence staining patterns and the presence of antinuclear antibodies. In six patients with antibodies to single-stranded DNA, the antinuclear antibody test was negative. The immunofluorescence staining pattern was homogeneous in two patients and nucleolar in three patients having a positive antinuclear antibody titer and a negative titer for antibody to single-stranded DNA. This finding indicates that antinuclear antibodies to single-stranded DNA may occur in the absence of antinuclear antibodies, and vice versa, suggesting that two separate antibody systems are involved. Also, antibodies to nucleoli are independent of antinuclear antibodies with homogeneous pattern and single-stranded DNA antibodies. Therefore, we have identified at least three distinctive serum antibodies in patients with linear scleroderma. The heterogeneity in immune response is clearly seen in Figure 4, which shows the serologic course of a single patient (Patient 13); three Figure 4. Serologic course in a patient with linear scleroderma showing discordance in time of three distinct antibody systems. distinct antibody systems are seen to be discordant in time. Discussion We evaluated our experience in patients with linear scleroderma by focusing on the clinical features and natural history of the disease and on the laboratory abnormalities that could be useful prognosticators of subsequent disease activity. Although some patients have both linear scleroderma and systemic connective tissue diseases, the paucity of reports describing such cases (3, 7-9) speaks for the rarity of this occurrence. Occasionally, the various serologic abnormalities present in patients with linear scleroderma and in those with other forms of localized scleroderma have been equated with true systemic involvement (18). In this series of patients with linear scleroderma, internal organ involvement, as measured by our tests, was absent at the time of presentation and during an extended follow-up. However, we must be cautious in excluding that some of these patients may not develop a connective tissue disorder or internal organ involvement at some later date. In addition, it is possible that subclinical or even transient internal organ involvement may have been missed. We therefore recommend that long-term follow-up be continued in patients with linear scleroderma. Unfortunately, the cutaneous lesions of linear scleroderma may, by themselves, be quite devastating, as is clear from the high frequency and severity of joint contractures in our patients. Hyperpigmentation, hypopigmentation, and cutaneous and muscular atrophy lead to considerable cosmetic, functional, and psychological problems. Admittedly, our series is skewed somewhat by referral bias of patients with more complicated cases. Morphea has previously been documented in patients with linear scleroderma (19), and it was present in about half of our patients. We have seen an occasional patient initially present with morphea alone and subsequently develop linear lesions. Because of the greater morbidity associated with linear scleroderma, patients with morphea should be watched closely for this evolution. The high frequency of joint contractures demands that pa- Falangaetal. Linear Scleroderma 855

tients be properly followed for this complication. We have found, as have others (20), that early and continued physical therapy is crucial in maintaining adequate joint motion. We found that the typical patient has active disease for only 2 to 3 years. This observation should help clinicians in advising patients and their families about the natural history of the disease. However, proper follow-up is indicated in all patients, because recurrence is possible. The presence of peripheral blood eosinophilia in patients with localized scleroderma in general, and in those with linear scleroderma in particular, has been documented previously. Giordano and colleagues (21), who also defined a normal eosinophil count as less than 300 cells/mm 3, noted eosinophilia in 2 of 9 patients with localized scleroderma (1 with morphea and 1 with linear scleroderma). Using less than 400 cells/mm 3 as a normal eosinophil count, Fleischmajer and coworkers (22) found eosinophilia in 3 of 9 patients with localized scleroderma (1 with morphea, 1 with linear scleroderma, and 1 with generalized morphea). Rodnan and colleagues (13) noted increased blood eosinophil counts in 9 of 16 patients with active linear scleroderma; these patients are included in the present study. The correlation we have found between blood eosinophilia and disease activity may be of considerable help in the assessment and management of patients with linear scleroderma. As we have already stated, eosinophilia has, on occasion, preceded a disease exacerbation, and its disappearance has heralded disease remission. We have seen this phenomenon in 7 of our patients. It is not surprising that elevations of IgG, IgM, and IgA were found in our patients, especially during active disease. Hypergammaglobulinemia occurs commonly in various inflammatory processes (23). We have found that these abnormalities in serum immunoglobulins persist for variable periods of time despite clinical improvement. Therefore, the determination of immunoglobulin levels has not been prognostically useful, except for the observed correlation between elevated IgG levels and joint contractures. At present, we have no biologically meaningful explanation for this finding. As we have seen, a positive titer of antinuclear antibody was clearly associated with severer cutaneous and joint disease. The presence of antinuclear antibodies warrants more frequent follow-up and close observation. This review is the first comprehensive report of serum antinuclear antibodies and their immunofluorescence patterns in patients with linear scleroderma. The observed immunofluorescence patterns are not specific for linear scleroderma, as they may be found in patients with other connective tissue diseases. However, it is notable that the peripheral pattern was absent and that the nucleolar pattern was seldom seen; these types of immunofluorescence patterns are more specific for systemic lupus erythematosus and systemic sclerosis, respectively (24). Serum antibody heterogeneity in our series of patients is shown by the presence of antibodies to single-stranded DNA and their independence from other observed antinuclear antibodies with nucleolar and homogeneous immunofluorescence. We found no clear correlation between clinical features and a particular antinuclear antibody titer or immunofluorescence pattern. However, patients whose sera had a higher level of single-stranded DNA binding did have more extensive disease. None of the serum samples tested contained antibodies to doublestranded DNA. The use of a pure DNA source, such as that in Crithidia luciliae, is needed to avoid confusion between antibodies to single-stranded and double-stranded DNA (14). The pathogenic significance of these serum autoantibodies in patients with linear scleroderma is unclear. For example, although antibodies to single-stranded DNA occur in patients with systemic lupus erythematosus who have active renal disease (25), no clinical evidence of renal involvement has been recognized in linear scleroderma. One could speculate that these autoantibodies result simply from tissue destruction leading to antigenic stimulation during active disease, but we have found them very early during the course of linear scleroderma and have not seen any serologic conversion in patients who initially had negative studies. Therefore, although these serum antibodies may not be directly involved in the cause of linear scleroderma, they may be closely linked to the pathogenic events leading to the disease. Regardless of the biologic significance of these autoantibodies, their presence indicates more extensive cutaneous and articular involvement and longer duration of disease. The longer duration is particularly evident in patients with antibodies to single-stranded DNA. The picture of linear scleroderma that emerges from this series of patients is that of a cutaneous and subcutaneous fibrosing disorder with frequent joint contractures. Disease progression is associated with elevated blood eosinophil counts, increased serum IgG levels, and the presence of antinuclear and single-stranded DNA antibodies. Requests for reprints should be addressed to Vincent Falanga, M.D.; Department of Dermatology, University of Pittsburgh, 3601 Fifth Avenue; Pittsburgh, PA 15213. References 1. RODNAN GP. When is scleroderma not scleroderma? Bull Rheum Dis. 1981;31:7-10. 2. TUFFANELLI DL, WINKELMANN RK. Systemic scleroderma. Arch Dermatol. 1955;84:359-71. 3. CHRISTIANSON HB, DORSEY CS, O'LEARY PA, KIERLAND RR. Localized scleroderma: a clinical study of two hundred thirty-five cases. Arch Dermatol. 1956;74:629-39. 4. STERN LZ, PAYNE CM, ALVAREZ JT, HANNAPEL LK. Myopathy associated with linear scleroderma: a histochemical and electron microscopic study. Neurology. 1975;25:114-9. 5. DOYLE JA, CONNOLLY SM, WINKELMANN RK. Cutaneous and subcutaneous inflammatory sclerosis syndromes. Arch Dermatol. 1982;118:886-90. 6. PERSON JR, SU WP. Subcutaneous morphoea: a clinical study of sixteen cases. Br J Dermatol. 1979;100:371-80. 7. CURTIS AC, JANSEN TG. The prognosis of localized scleroderma. Arch Dermatol. 1958;78:749-56. 8. MACKEL SE, KOZIN F, RYAN LM, SHETH KJ, JORDON RE. Concurrent linear scleroderma and systemic lupus erythematosus: a report of two cases. J Invest Dermatol. 1979;73:368-72. 9. DUBOIS EL, CHANDOR S, FRIOU GJ, BISCHEL M. Progressive systemic sclerosis (PSS) and localized scleroderma (morphea) with positive LE cell test and unusual systemic manifestations compatible with systemic lupus erythematosus (SLE). Medicine (Baltimore). 1971;50:199-222. 10. ROWELL NR, BECK JS. The diagnostic value of an antinuclear antibody test in clinical dermatology. Arch Dermatol. 1967;96:290-5. 856 June 1986 Annals of Internal Medicine Volume 104 Number 6

11. BARTHELMES H. Das Verhalten der antinuklearen Faktoren bei Sklerodermie. Dermatol Monatsschr. 1975;161:536-44. 12. TAKEHARA K, MOROI Y, NAKABAYASHI Y, ISHIBASHI Y. Antinuclear antibodies in localized scleroderma. Arthritis Rheum. 1983;26:612-6. 13. RODNAN GP, LIPINSKI E, RABIN BS, REICHLIN M. Eosinophilia and serologic abnormalities in linear localized scleroderma [Abstract]. Arthritis Rheum. 1977;20:133. 14. FALANGA V, MEDSGER TA, REICHLIN M. High titers of antibodies to single-stranded DNA in linear scleroderma. Arch Dermatol. 1985;121:345-7. 15. MATTIOLI M, REICHLIN M. Characterization of a soluble nuclear ribonucleoprotein antigen active with SLE sera. J Immunol. 1971; 107:1281-90. 16. COMMERFORD SL. Iodination of nucleic acids in vitro. Biochemistry. 1971;10:1993-2000. 17. WOLD RT, YOUNG FE, TAN EM, FARR RS. Deoxyribonucleic acid antibody: a method to detect its primary interaction with deoxyribonucleic acid. Science. 1968;161:806-7. 18. BOURGEOIS-DROIN C, TOURAINE R. Sclerodermic en plaques: perturbations immunologiques et viscerales. Ann Med Interne (Paris). 1978;129:107-12. 19. JABLONSKA S, RODNAN GP. Localized forms of scleroderma. Clin Rheum Dis. 1979;5:215-41. 20. RUDOLPH RI, LEYDEN JJ. Physiatrics for deforming linear scleroderma. Arch Dermatol. 1976;112:995-7. 21. GIORDANO M, ARA M, VALENTINI G, CHIANESE U, BENCIVENGA T. Presence of eosinophilia in progressive systemic sclerosis and localized scleroderma. Arch Dermatol Res. 1981;271:411-7. 22. FLEISCHMAJER R, JACOTOT AB, SHORE S, BINNICK SA. Scleroderma, eosinophilia, and diffuse fasciitis. Arch Dermatol. 1978;114:1320-5. 23. MCKELVEY EM. Serum protein electrophoresis in laboratory diagnosis. Med Clin North Am. 1969;53:147. 24. SONTHEIMER RD, DENG JS, GILLIAM JN. Antinuclear and anticytoplasmic antibodies: concepts and misconceptions. J Am Acad Dermatol. 1983;9:335-43. 25. LANGE A. Evaluation of the simultaneous estimation of anti-dsdna and anti-ssdna antibodies for clinical purposes. Clin Exp Immunol. 1978;31:472-81. Falanga etal. Linear Scleroderma 857