Skin ulcers in a patient with Sjögren's syndrome



History of the present illness

A 49-year-old woman with Sjögren's syndrome presented for evaluation of nonhealing lower extremity ulcers. Four years prior to the current presentation, the patient developed hyperpigmented, nonpruritic, nonpainful macular lesions on her legs that recurred intermittently. These lesions would begin as painless macules colored blue or black, and according to the patient, over the course of a few weeks would evolve into flat, reddish-brownish macules that resolved spontaneously.

Four months before presentation, the patient developed new skin ulcerations on her legs that were similar to earlier exacerbations. However, instead of resolving, these new lesions ulcerated, became quite painful, and did not heal. A skin biopsy sample demonstrated leukocytoclastic vasculitis (Figure 1). Laboratory testing results included positive tests for rheumatoid factor (529 IU/ml, normal value <20), antinuclear antibodies, and antibodies to Ro/SSA, La/SSB, and RNP antigens. The serum C3 level was 103 mg/dl (normal range 82–120), but the serum C4 level was undetectable (normal range 16–70 mg/dl). Serologic tests for antibodies to hepatitis B and C viruses and antineutrophil cytoplasmic antibodies (ANCAs) were negative, as was testing for cryoglobulins. Repeat tests for hepatitis C, ANCAs, and cryoglobulins were negative.

Figure 1.

A, A skin biopsy sample demonstrating an interstitial neutrophilic infiltrate with extravasated erythrocytes and occlusion of small blood vessels in the papillary dermis (hematoxylin and eosin stained; original magnification × 40). Additionally, periodic acid–Schiff (PAS) stain reveals foci of leukocytoclastic vasculitis as evidenced by inflammatory cell infiltration of the vessel wall (B) and PAS-positive material (staining pink) occluding small blood vessels (C).

The patient declined treatment with glucocorticoids for her vasculitis. Over the subsequent 3 months, new skin lesions developed frequently, lower extremity edema developed, and she had increasing pain. The patient was treated with oxycodone and gabapentin without significant pain relief. She was then referred for evaluation and management, and admitted to the hospital because her leg pain interfered markedly with ambulation and activities of daily living.

Past medical history

More than 10 years previously, the patient was diagnosed with Sjögren's syndrome based on manifestations of keratoconjunctivitis sicca, xerostomia, and parotid enlargement. Her symptoms were mild, and she was treated only with over-the-counter moisturizing eye drops. The patient had undergone a hysterectomy for uterine fibroids.

Medications and allergies

The patient's only medication on admission was oxycodone 10 mg every 4 hours, which she took for her leg pain. She did not ingest herbal or other alternative therapies. She had no known drug allergies.

Social history

The patient was employed as an administrative assistant in an office, but had been unable to work because of her leg symptoms. She was not sexually active.


She drank alcohol a few times a year on special occasions, did not smoke, and had never used injectable or other illegal drugs.

Family history

The patient was not aware of any rheumatic or immunologic diseases in her extended family.


The patient reported a 5-pound weight loss over the 4 weeks preceding admission, and she attributed the weight loss to poor appetite. She also reported dry eyes, dry mouth, sharp pain at the skin ulceration sites, and a diffuse burning pain in the lower extremities. She had been having intermittent mild blood-tinged nasal discharge for the past 6 months. She had not experienced fevers, seizures, hearing loss, visual changes, alopecia, dyspnea, chest pain, abdominal pain, diarrhea, oral or vaginal ulcerations, vaginal discharge, dysuria, myalgias, arthralgias, or symptoms of Raynaud's phenomenon.

Physical examination

The patient was uncomfortable. She was afebrile, her pulse rate was 90/minute, and her blood pressure was 140/82 mm Hg. Her sclerae were not injected and both her ocular tear film and salivary pooling were normal by visual inspection. She had neither oral ulcerations nor lymphadenopathy. The pulmonary, cardiac, abdominal, and joint examinations were unremarkable. Her carotid, radial, femoral, dorsalis pedis, and posterior tibial pulses were palpable and symmetric. Deep tendon reflexes were present and symmetric and strength was intact. Vibration, position, and light touch examinations were normal, but hyperesthesia was present in the feet and legs. Her legs appeared brawny. There was a diffuse, light brown macular rash on her legs, along with scattered grey and black macules that ranged in diameter from 2 mm to 2 cm. In addition, numerous ulcerations of varying diameter and depth with pink flat borders were present (Figure 2). The largest ulcer measured 8 cm in diameter and 0.5 cm in depth. Her legs were diffusely tender, and pitting edema of her feet and legs extended to her patellae.

Figure 2.

Ulcers of varying ages and stages on the patient's left leg, as well as hyperpigmentation in the location of previous lesions.

Laboratory evaluation

Laboratory data from the time of admission are shown in Table 1. These included a platelet count of 1,117,000/μl (normal range 150–400 1,000/μl), aspartate aminotransferase level of 249 units/liter (normal range 5–40), alanine aminotransferase level of 186 units/liter (normal range 5–31), gamma glutamyl transferase level of 45 units/liter (normal range 7–33), rheumatoid factor of 529 IU/ml (normal value <20), and undetectable C4.

Table 1. Laboratory values (combination of preadmission serologies and tests at the time of the hospital admission)
TestValueNormal value
White cell count, 1,000/μl4.24–11
Hemoglobin, gm/dl11.412–16
Hematocrit, %31.638–47
Platelet count, 1,000/μl1,117150–400
Aspartate aminotransferase, units/liter2495–40
Alanine aminotransferase, units/liter1865–31
Alkaline phosphatase, units/liter10925–100
Total bilirubin, mg/dl0.40.3–1.2
Gamma glutamyl transpeptidase, units/liter457–33
Creatinine, mg/dl0.60.5–1.1
Rheumatoid factor, IU/ml529<20
Antinuclear antibodiesPositiveNegative
Ro/SSA antibodies, units153<5
La/SSB antibodies, units162<5
RNP/Sm antibodies, units14<5
Double-stranded DNA antibodies, IU/ml<5<5
Anticardiolipin antibodies  
 IgM, IgM phospholipid units<12.5<12.5
 IgG, IgG phospholipid units<15<15
Lupus anticoagulantNot detectedNot detected
C3, mg/dl10382–120
C4, mg/dlUndetectable16–70
Antineutrophil cytoplasmic antibodiesNegativeNegative
Viscosity, centipoise1.51.4–1.8


The patient is a 49-year-old woman with a history of mildly symptomatic Sjögren's syndrome who presented with leg ulcerations and a skin biopsy sample revealing leukocytoclastic vasculitis.


There are numerous causes for a small-vessel vasculitis, including infections, paraneoplastic syndromes, toxin or drug exposure, and a long list of idiopathic autoimmune vasculitides. In this patient, the etiologies strongly considered included connective tissue disease–associated vasculitis, cryoglobulinemia, ANCA-associated vasculitis (Wegener's granulomatosis or microscopic polyangiitis), paraneoplasia, myeloproliferative syndrome, and infections with Mycobacterium tuberculosis, hepatitis C virus, or hepatitis B virus.

Connective tissue diseases such as systemic lupus erythematosus, rheumatoid arthritis, mixed connective tissue disease, and Sjögren's syndrome may be associated with a small-vessel vasculitis. This patient's history, physical examination, and laboratory findings are most consistent with primary Sjögren's syndrome. If Sjögren's syndrome was the underlying etiology of her small-vessel vasculitis, it developed at a time when glandular involvement was minimally active.

The history of 6 months of epistaxis, years of an intermittent rash that resulted in postinflammatory hyperpigmentation, and paresthesias were suggestive of an ANCA-associated vasculitis. However, repeated ANCA tests were negative and computed tomography scans of her sinuses and chest were unremarkable. The lack of radiologic findings combined with the absence of additional signs or symptoms common in ANCA-associated vasculitis made this diagnosis unlikely.

Except for mild weight loss and a thrombocytosis, she did not have any other signs or symptoms suggestive of a malignancy. She did not have a family history of cancer, she had never smoked, and she had a recent normal mammogram. Infections such as hepatitis B and C viruses are often included in the differential diagnosis for patients with rheumatic and dermatologic symptoms. Hepatitis viral infection was considered in this woman with a transaminitis and a cutaneous small-vessel vasculitis. However, repeat testing for hepatitis B virus and hepatitis C virus serologies, as well as hepatitis C viral load, were negative.


The most likely etiology of the patient's condition was Sjögren's syndrome–associated small-vessel vasculitis with or without cryoglobulinemia. Despite the 2 prior negative tests for cryoglobulins, testing was repeated because serum samples are often collected and handled improperly, resulting in false-negative tests for the detection of cryoglobulins. Repeat testing was positive for the presence of serum cryoglobulins, with a cryocrit of 61% (Figure 3). A second sample also showed a remarkably high cryocrit (60%).

Figure 3.

The process for diagnosing cryoglobulins is based on the identification of immunoglobulins or complexes of immunoglobulins and complement precipitating in serum at temperatures below 37°C. This tube contains the precipitated cryoglobulins from the patient in this report.

Given the extremely elevated cryocrit and the known association between Sjögren's syndrome and a markedly increased risk of lymphoma (1–4), there was concern for an underlying malignancy, and a hematologic evaluation was undertaken. Serum and urine protein electrophoresis, immunofixation electrophoresis, and a bone marrow biopsy were all performed. Serum protein electrophoresis revealed a small gamma band and a decrease in beta globulin (0.78 gm/dl compared with a lower limit of normal of 0.9 gm/dl), and urine protein electrophoresis demonstrated a faint gamma band. Immunofixation electrophoresis of the serum demonstrated a monoclonal IgM kappa band, as well as polyclonal IgG (Figure 4). The bone marrow biopsy sample demonstrated a mildly hypercellular marrow (40% fat) and a mild increase in the number of megakaryocytes and reticulin with normal proportions and maturation of erythroid and myeloid bone marrow elements. The bone marrow aspirate did not show features of lymphoma by flow cytometry. Plasma cells constituted less than 1% of the total cells identified, making multiple myeloma unlikely.

Figure 4.

A, Serum protein electrophoresis (SPEP) and immunofixation electrophoresis (IFE). There is mild banding in the gamma globulin region of the SPEP. IFE demonstrates discrete corresponding monoclonal IgM and kappa bands and polyclonal IgG (no discrete band). B, SPEP of the cryoprecipitate reveals one single band in the gamma globulin region as compared with the control. C, IFE of the cryoprecipitate reveals a monoclonal IgM band with a corresponding kappa band and polyclonal IgG. There is no banding in the lambda and IgA regions, supporting the assumption that the previous faint bands identified in the serum IEP were artifactual.

Additional blood tests revealed normal blood viscosity and negative tests for anticardiolipin antibodies (IgG and IgM), lupus anticoagulant, anti–double-stranded DNA antibodies, and human immunodeficiency virus. Subsequent reevaluation of the original skin biopsy sample revealed terminal vessel occlusion by periodic acid–Schiff-positive material as seen in cryoglobulinemia (Figure 1).


Circulating cryoglobulins and the clinical processes they are associated with have fascinated clinicians for decades and have been the focus of even greater attention since the link between hepatitis C infection and cryoglobulinemia was firmly established. Cryoglobulins are immunoglobulins or complexes of immunoglobulins and complement that precipitate in serum at temperatures below 37°C. Cryoglobulins have been associated with many chronic disease processes, including hematologic abnormalities, infections, and connective tissue diseases (5–9). The cryocrit is a measure of the percentage of cryoprecipitate per volume of serum (9, 10). It is determined by dividing the volume of cryoprecipitate by the volume of serum multiplied by 100 to yield the cryocrit percentage. Cryocrit levels typically do not correlate with disease severity (9).

In 1974, Brouet et al (11) described 3 classes of cryoglobulin classes: I, II, and III, and this schema remains in use today. Cryoglobulin typing is important to help determine the underlying pathogenesis. Type I cryoglobulins are composed of monoclonal immunoglobulins, usually IgM and less commonly IgG or IgA, and are associated with hematologic abnormalities such as lymphoma, Waldenström's macroglobulinemia, and multiple myeloma (5, 8–10). Type I cryoglobulins account for approximately 10% of cases of clinically apparent cryoglobulinemia (9, 10). Types II and III cryoglobulins are collectively labeled as “mixed” cryoglobulins, and are comprised of immunoglobulins with rheumatoid factor activity (type II with monoclonal IgM and type III with polyclonal IgM) that recognize polyclonal IgG. Type II is the most common type of cryoglobulinemia, accounting for approximately 60% of cases; type III accounts for 30% of cases (9, 10). Mixed cryoglobulins are associated with chronic inflammation and infection, including hepatitis B and C viral infections, systemic lupus erythematosus, rheumatoid arthritis, and Sjögren's syndrome (5–10, 12). Cryoglobulins may also be present in the sera of patients with chronic inflammatory conditions without clinical significance. However, when cryoglobulins deposit in small and medium-sized vessels in the skin, kidneys, joints, and nerves, tissue injury may occur.

Hepatitis C infection has been reported in 70–100% of cases of type II cryoglobulinemia (5), including in many patients with underlying connective tissue disorders (6, 7). In a cross-sectional study of 115 unselected patients with primary Sjögren's syndrome, 16% (18 of 115) tested positive for cryoglobulins (7). Forty-four percent (8 of 18) of the patients with cryoglobulins had evidence of a hepatitis C infection compared with only 8% (7 of 84) of the patients who did not have cryoglobulins. In addition, cutaneous vasculitis was significantly more common among the patients with cryoglobulinemia (56%) than among those without cryoglobulinemia (8%). Similarly, in a cross-sectional study of 122 unselected patients with systemic lupus erythematosus, 25% (31 of 122) had detectable cryoglobulins (6). The majority had cryocrit levels less than 1%; fewer than 10% of patients had cryocrit levels greater than 5%. Twenty-one percent (5 of 24) of the patients with cryoglobulins had hepatitis C virus infections compared with only 5% (4 of 75) of the patients without cryoglobulins. These studies demonstrate that cryoglobulins are frequently seen in Sjögren's syndrome and systemic lupus erythematosus, and that although hepatitis C viral infection is more common in the patients with cryoglobulins than those without hepatitis C, the majority of the patients with connective tissue disease and cryoglobulins do not have hepatitis C infection.


Treatment options for small-vessel vasculitis and cryoglobulinemia include glucocorticoids, cyclophosphamide, methotrexate, azathioprine, other immunosuppressive drugs including biologic agents, and plasmapheresis. For patients with coexisting hepatitis C infection, treatment for the viral disease is considered the primary form of therapy. B cells have been identified as a major contributing factor to the pathogenesis of glandular and extraglandular symptoms in Sjögren's syndrome (13–15), including cryoglobulinemia (14), leading to evaluation of rituximab, a chimeric murine–human anti-CD20 (B cell) monoclonal antibody, as treatment for Sjögren's syndrome (16–22). These relatively small studies each demonstrated the efficacy of rituximab in Sjögren's syndrome, including among patients with vasculitis.

The patient in this report was initially prescribed high-dose prednisone, and this treatment was associated with halting the progression of established lesions and the development of new lesions, a mild to moderate reduction in pain, and normalization of her liver function tests. Azathioprine was subsequently added as a glucocorticoid-sparing agent. Cyclophosphamide was not added due to patient preference. Over the next 6 months, her lesions slowly began to heal. Rituximab was then added to her regimen and she received 2 intravenous doses of 1,000 mg each. Her peripheral blood B cell CD19 count became undetectable. In the months following treatment with rituximab, the patient's pain improved dramatically and she no longer required low-dose narcotic analgesia. Prednisone was tapered down to 5 mg/day. Because many of her lesions did not heal completely, the patient underwent skin grafting of 9 lesions, with good results (Figure 5). Interestingly, her cryocrit decreased during treatment to a low of 10%. CD19 B cell levels recovered 7 months after the initial dose of rituximab and, due to concern for recurrence, the patient then received a second course of rituximab. Despite an appropriate B cell response to rituximab (CD19 count of 0), the patient developed worsening neuropathic pain and recurrence of skin lesions. She was then treated with oral cyclophosphamide and high-dose glucocorticoids with resolution of active vasculitis. The patient is now receiving a maintenance regimen of mycophenolic acid (2,000 mg/day) and low-dose prednisone (5 mg/day), has a cryocrit of <1%, and has no clinical evidence of active vasculitis.

Figure 5.

The patient's left leg with healed wounds several weeks following the placement of skin grafts.


This case illustrates some of the challenges involved in establishing a diagnosis of cryoglobulinemia. Incorrect collection, handling, and/or processing of the patient's previous 2 blood samples may have led to false-negative tests for cryoglobulins. Such blood samples need to be collected and kept at 37°C for 1–2 hours prior to clot formation; the serum and clot should be separated in a warm, or at least room temperature, centrifuge. If the temperature of the specimen falls prior to clot formation, cryoglobulins will precipitate out of the serum, leading to a false-negative test result. It was presumed that the previous tests were collected, handled, or processed incorrectly. Cryoglobulins may, at times, be absent or present in low concentration, even in symptomatic patients, and repeat testing is indicated in suspected cases of cryoglobulinemia. Despite the initial absence of cryoglobulins in this patient's serum on 2 tests, moderate clinical suspicion for the diagnosis remained, given the skin and neurologic findings and the multiple laboratory abnormalities suggestive of a diagnosis of cryoglobulinemia, including undetectable C4 in the presence of normal C3 (9), elevated liver enzymes (9), and rheumatoid factor positivity (5, 8–10). Additionally, it is important to note that automated cell counters may falsely identify cryoglobulins as platelets or leukocytes based on size criteria (23, 24), as was the case with this patient. The optimal treatment strategy for patients with cryoglobulinemia remains to be established.


Sjögren's syndrome–associated type II cryoglobulinemic vasculitis.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Merkel had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Chapnick, Merkel.

Acquisition of data. Chapnick, Merkel.

Analysis and interpretation of data. Chapnick, Merkel.


The authors appreciate the assistance of Lynne J. Goldberg, MD, Robert Sands, MD, and Gunnar Tomasson, MD, in the preparation of this report.