High titer of serum antiphospholipid antibody levels in adult Henoch-Schönlein purpura and cutaneous leukocytoclastic angiitis




To investigate a possible role of antiphospholipid (aPL) antibodies in adult Henoch-Schönlein purpura (HSP) and cutaneous leukocytoclastic angiitis (CLA).


We reviewed the records of 30 HSP and 8 CLA adults with an initial cutaneous manifestation of palpable purpura on their lower extremities between 2003 and 2007. Eight microscopic polyangiitis (MPA) patients and 30 healthy persons were recruited as controls. Serum anticardiolipin (aCL), anti–phosphatidylserine-prothrombin complex (anti-PS/PT), and anti–β2-glycoprotein I (anti-β2GPI) antibody levels in HSP, CLA, MPA patients, and healthy controls were measured by enzyme-linked immunosorbent assay.


Twenty-two HSP patients (73%) were positive for serum IgA aCL antibodies. Nineteen (63%) had IgA anti-PS/PT antibodies and 4 (13%) had IgA anti-β2GPI antibodies. IgA aCL and anti-PS/PT antibodies showed a significant correlation (P = 0.007). Twenty (67%) HSP patients had IgM anti-PS/PT antibodies and 6 (20%) had IgG anti-PS/PT antibodies. Six (75%) CLA patients had IgM anti-PS/PT antibodies and 2 (25%) had IgG anti-PS/PT antibodies. In contrast, aPL antibodies were not found in any MPA patients or normal controls. Serum IgA aCL antibody levels in HSP patients showed a significant correlation with serum IgA and C-reactive protein (CRP) levels (P = 0.030 and 0.039, respectively). A positive correlation between CRP and serum IgA anti-PS/PT antibody levels was observed in HSP patients (P = 0.023). Serum IgA aCL antibody levels were also significantly associated with proteinuria according to urinalysis (P = 0.024).


Serum levels of IgA aCL and anti-PS/PT antibodies were elevated in adult HSP, suggesting that serum IgA antibodies may play some role in adult HSP. IgA aCL and/or anti-PS/PT antibodies could serve as markers for adult HSP and should be monitored as an indicator of adult HSP activity. Small-vessel vasculitis could be dependently associated with the presence of IgM anti-PS/PT antibodies. These findings suggest that aPL antibodies are closely related to the pathogenic factors that trigger the development of vasculitis.


Henoch-Schönlein purpura (HSP) is characterized by palpable nonthrombocytopenic purpura over the lower extremities, arthritis, abdominal pain with or without gastrointestinal hemorrhage, and, less commonly, glomerulonephritis. Histologically, HSP reveals leukocytoclastic vasculitis, and has been well regarded as a specific clinicopathologic entity based on the vascular deposition of IgA-dominant immune complexes (ICs) (1, 2). HSP in adulthood differs from the pediatric form of the disease. Adults have more pronounced skin lesions, and renal involvement is more frequent than in childhood (3, 4). The pathogenesis of HSP remains poorly understood. The similarity of the histologic findings in HSP to those of Arthus reaction suggests an IgA IC–mediated disorder (5). Arthus reaction results from the deposition of ICs in the small blood vessels of the skin, glomeruli, and gastrointestinal tract and the subsequent activation of inflammatory responses to these complexes. This reaction has served as the basis for dissecting the cellular and molecular events that are triggered by IC deposition and serves as the basis for our understanding of the pathophysiology of IC-mediated diseases. HSP has been associated with a history of preceding infection, particularly upper respiratory tract infection (URI). IgA is the principal antibody in the respiratory system for defense against microbial agents. These findings raise the possibility of infection as a direct cause or a potential trigger of HSP. It has been postulated that various antigenic stimuli cause an elevation of circulating IgA and that complement activation leads to vasculitis. However, no definitive pathogen has been isolated from the majority of patients with HSP.

We have previously shown high serum levels of IgA anticardiolipin (aCL) in adult HSP, suggesting that serum IgA aCL antibodies may play some role in the pathogenesis (6). Antiphospholipid (aPL) antibodies are a heterogeneous group of immunoglobulins that bind to several anionic phospholipids including cardiolipin and phosphatidylserine (7, 8). Most aPL antibodies require a serum cofactor, β2-glycoprotein I (β2GPI), for optimal binding (9–12). It has been shown that many aPL antibodies may react to a neoepitope formed on the β2GPI molecule by the interaction between the phospholipid and β2GPI (13, 14). Detection of aPL cofactor antibodies, in addition to the classic aCL antibodies, seems to be of considerable clinical importance. Prothrombin is another possible antigenic target of antiphospholipid syndrome (15). Atsumi et al (16) and Amengual et al (17) suggested that anti–phosphatidylserine-prothrombin complex (anti-PS/PT) antibodies alone, rather than antiprothrombin antibodies, are associated with symptoms of antiphospholipid syndrome.

Small-vessel vasculitides are a group of inflammatory diseases that affect the arterioles, venules, and capillaries. These vasculitides can be divided into pauci-immune vasculitides associated with antineutrophil cytoplasmic antibodies (ANCAs) and immune deposit-associated vasculitides. The ANCA-associated vasculitides include microscopic polyangiitis (MPA), and HSP represents immune deposit-associated small-vessel vasculitides. The Chapel Hill Consensus Conference (CHCC) on the Nomenclature of Systemic Vasculitis defined the term cutaneous leukocytoclastic angiitis (CLA) as isolated cutaneous leukocytoclastic angiitis without systemic vasculitis or glomerulonephritis (2). Thus, for the diagnosis of CLA, systemic involvement must be excluded (18). The diagnosis is one of exclusion because evidence of systemic vasculitic disease apart from arthralgias must be absent (19). In the present study, we diagnosed patients who did not have systemic vasculitis for at least 2 years during followup of CLA.

To investigate the aPL antibodies of patients with HSP, CLA, and MPA, we examined the prevalence of aCL antibodies, anti-PS/PT antibodies, and anti-β2GPI antibodies in 30 patients with HSP, 8 patients with CLA, and 8 patients with MPA. Furthermore, we investigated whether these antibodies are closely correlated with clinical or serologic features, which could indicate that IgA aCL and IgA anti-PS/PT antibodies could play a significant role in the pathogenesis of HSP.


Clinical investigation and tissue samples.

Thirty Japanese patients with HSP (15 men, 15 women, mean ± SD age 56.4 ± 20.1 years), 8 with CLA (6 men, 2 women, mean ± SD age 56.6 ± 19.1 years), and 8 with MPA (4 men, 4 women, mean ± SD age 62.9 ± 17.3 years) seen at the Department of Dermatology, St. Marianna University School of Medicine between 2003 and 2007 were examined. Patients presented with typical characteristics of nonthrombocytopenic symmetric palpable purpura over the lower extremities, and other conditions such as connective tissue diseases and infections had been excluded. Skin biopsy samples were obtained from patients during an active phase of their disease. Two skin biopsy samples from each of the 30 HSP and 8 CLA patients were taken for routine and direct immunofluorescence (DIF) staining according to standard procedures. All biopsy samples were taken from the lower extremities. The tissue section was incubated with commercially prepared fluoresceinated antisera specific to human IgG, IgM, IgA, or C3. Each of the skin biopsy samples showed fibrinoid necrosis, an admixture of neutrophils and lymphocytes in and around the blood vessels, intravascular fibrin thrombi, and nuclear dust characteristics of leukocytoclastic vasculitis in the upper and middle dermis. Based on DIF of the skin biopsy samples, vascular deposits of IgA in the dermis led to the diagnosis of HSP. The following selection criteria for HSP were used in the study: the presence of a leukocytoclastic vasculitis and vascular IgA and the absence of thrombocytopenia or known hematologic or connective tissue disorders. CLA is an isolated vasculitis limited to skin, and systemic involvement must be excluded. Apart from arthralgia, no systemic manifestations were observed after a minimum followup of 2 years. Eight MPA patients were diagnosed according to the CHCC definition (2), Sørensen et al criteria (20), and Japanese criteria (21).

None of the patients had been given corticosteroids, immunosuppressants, or vasodilators at the time of serum sampling. Furthermore, no patient demonstrated any evidence of coexisting malignancy, other autoimmune diseases, or viral hepatitis, nor were any of the patients positive for mixed cryoglobulinemia. Thirty healthy persons with comparable sex and age distributions were recruited as normal controls. The following tests were negative or within the normal range in the HSP and CLA patients: antinuclear antibodies, ANCAs, and virus serology including hepatitis A, B, and C.

Serologic studies.

The first immunologic assessments were performed on the serum collected at the same time as the skin biopsy samples. IgG, IgM, and IgA isotypes of anti-PS/PT antibodies and IgG, IgM, and IgA aCL antibodies were measured with a specific enzyme-linked immunosorbent assay (ELISA; Medical & Biological Laboratories, Nagoya, Japan) according to the manufacturer's protocol. Briefly, the serum samples were diluted 1:101, added to 96-well plates coated with PS/PT or aCL, and incubated for 1 hour at 20°C. Polyclonal gout anti-human IgG, IgM, and IgA antibodies labeled with horseradish peroxidase were used as conjugate solutions to recognize the 3 isotypes of PS/PT or aCL antibodies. Color was developed with 3,3′,5,5′-tetramethylbenzidine and H2O2 and the plates were read at 450 nm. IgA anti-β2GPI antibodies were measured with a specific ELISA (Bio-Rad Laboratories, Hercules, CA) according to the manufacturer's protocol. IgG and IgM anti-β2GPI antibodies were determined according to the standardized anti-β2GPI antibody ELISA (Diagnostica Stago, Asnieres, France). The cutoff points were established by the mean +5 SD of 50 controls. The following cutoff values were used: 12 units/ml, 10 units/ml, and 10 units/ml for IgG, IgM, and IgA anti-PS/PT antibodies, respectively; 10 IgG phospholipid units, 10 IgM phospholipid units, and 10 IgA phospholipid units (APL units) for IgG, IgM, and IgA aCL antibodies, respectively; and 10 units/ml, 10 units/ml, and 10 units/ml for IgG, IgM, and IgA anti-β2GPI antibodies, respectively. When ANCA was detected, specificity was characterized by ELISA for reactivity with myeloperoxidase or proteinase 3, using the technique recommended by the European ANCA Assay Standardization Group (22). All serum samples were stored at −70°C prior to assay. Serum IgA and C-reactive protein (CRP) levels were assayed with laser nephelometry (Nihon Kohden, Tokyo, Japan).

Statistical analysis.

Kolmogorov-Smirnov test with Lilliefors significance level was used to evaluate age, serum aPL antibodies, serum CRP level, and serum IgA. Statistical analyses were performed using the Mann-Whitney U test for comparisons of serum aPL antibody levels. The Mann-Whitney U test was also used to compare blood parameter levels between positive and negative clinical symptoms (proteinuria, arthralgia, abdominal pain, and URI); the level of significance was set at P < 0.05 in all cases. The correlation between serum IgA or CRP level and variation of the blood parameters was assessed by Spearman's rank correlation test. All data are expressed as the mean ± SD.

The experimental protocol was approved by St. Marianna University, and informed consent was obtained from all patients (No. 1117).


IgA aCL antibodies, IgA anti-PS/PT antibodies, and/or IgA anti-β2GPI antibodies were present in 25 (83%) of the 30 HSP patients; IgA aCL antibodies alone were present in 5 patients, IgA anti-PS/PT antibodies alone were present in 2 patients, and a combination of both types of antibodies was present in 17 patients (Table 1). IgA aCL antibodies were found in 22 (73%) of the 30 HSP patients, IgA anti-PS/PT antibodies were found in 19 (63%), and IgA anti-β2GPI antibodies were found in 4 (13%). Of the 4 HSP patients with IgA anti-β2GPI antibodies, 3 were found to be positive for IgA aCL antibodies and IgA anti-PS/PT antibodies. IgA aCL antibodies, IgA anti-PS/PT antibodies, and IgA anti-β2GPI antibodies were not found in the serum of the CLA patients (Table 2), the MPA patients, or the healthy controls. Twenty (67%) of the HSP patients had positive titers for IgM anti-PS/PT antibodies and 6 (20%) had positive titers for IgG anti-PS/PT antibodies. By contrast, 6 (75%) of the 8 CLA patients were positive for IgM anti-PS/PT antibodies and 2 (25%) were positive for IgG anti-PS/PT antibodies. Serum IgA aCL levels were significantly elevated in all HSP patients (mean ± SD 12.1 ± 5.3 APL) in comparison with those of the CLA patients (mean ± SD 4.8 ± 2.4 APL) (Figure 1A). Serum IgA anti-PS/PT levels were also significantly higher among the HSP patients (mean ± SD 18.5 ± 24.6 units/ml) than the CLA patients (4.4 ± 2.2 units/ml) (Figure 1B). The titers for IgM anti-PS/PT antibodies were similar between the HSP patients (mean ± SD 13.1 ± 8.1 units/ml) and CLA patients (13.6 ± 5.0 units/ml), whereas these values were significantly higher than those of the MPA patients and normal controls (Figure 1C). Levels of IgA anti-β2GPI antibodies (mean ± SD 6.6 ± 10.9 units/ml versus 3.8 ± 2.1 units/ml) and IgG anti-PS/PT antibodies (mean ± SD 12.8 ± 22.7 units/ml versus 5.5 ± 4.7 units/ml) were not significantly different between HSP patients and CLA patients. Laboratory investigations in the HSP patients revealed elevated serum IgA (mean ± SD 486.3 ± 206.7 mg/dl), as well as elevated CRP levels (mean ± SD 4.2 ± 5.2 mg/dl). We found a significant correlation between serum IgA aCL antibodies and IgA anti-PS/PT antibodies in our HSP patients (rs = 0.50, P = 0.007) (Figure 2).

Table 1. Clinical, serologic, and DIF features in 30 patients with Henoch-Schönlein purpura*
Patient no.AgeSexIgA aCL (APL)IgA PS/PT (units/ml)IgA β2GPI (units/ml)IgM PS/PT (units/ml)IgG PS/PT (units/ml)CRP (mg/dl)IgA (mg/dl)DIF with IgA
  • *

    DIF = direct immunofluorescence; aCL = anticardiolipin antibodies; APL = IgA phospholipid units; PS/PT = phosphatidylserine-prothrombin complex; β2GPI = β2-glycoprotein I; CRP = C-reactive protein.

175M181810454.0559IgG, IgM, C3
672M1623291.85427IgM, C3
781F242.0758IgM, C3
1119F1111233.0241IgM, C3
1341M170.7351IgM, C3
1664M1296151208.4337IgM, C3
1738F111224203.14708IgM, C3
1834M12141.6260IgM, C3
1932F13203.3177IgM, C3
2144M13132.14204IgM, C3
2439F18163.6643IgM, C3
2581M1332104.9675IgM, C3
3082F0.33329IgM, C3
Table 2. Clinical, serologic, and DIF features in 8 patients with cutaneous leukocytoclastic angiitis
Patient no.AgeSexIgA aCL (APL)IgA PS/PT (units/ml)IgA β2GPI (units/ml)IgM PS/PT (units/ml)IgG PS/PT (units/ml)CRP (mg/dl)IgA (mg/dl)DIF
  1. * See Table 1 for definitions.

Figure 1.

A, IgA anticardiolipin antibodies (aCL; antiphospholipid units), B, IgA anti–phosphatidylserine-prothrombin complex (anti-PS/PT; units/ml) antibody levels expressed as relative optical density, and C, IgM anti-PS/PT antibody levels (units/ml) were measured by ELISA in serum samples from normal controls and patients with Henoch-Schönlein purpura (HSP), cutaneous leukocytoclastic angiitis (CLA), and microscopic polyangiitis (MPA). The detection limit was 5 units/ml and the cutoff value was 10 units/ml. Horizontal lines indicate the mean value in each group.

Figure 2.

Positive correlation between serum IgA anti–phosphatidylserine-prothrombin complex (anti-PS/PT) antibody and IgA anticardiolipin (aCL) antibody levels in patients with Henoch-Schönlein purpura. Serum IgA anti-PS/PT antibody levels are shown on the ordinate, and serum IgA aCL antibody levels are shown on the abscissa (rs = 0.50, P = 0.007).

There was a significant positive correlation between serum IgA aCL antibodies and IgA levels (rs = 0.40, P = 0.030) (Table 3). A similar correlation was also seen between IgA aCL antibodies and CRP levels (rs = 0.38, P = 0.039). In addition, there was a positive correlation between serum IgA anti-PS/PT antibodies and CRP level (rs = 0.42, P = 0.023). The IgA anti-β2GPI titer was significantly correlated with serum IgA and CRP level (rs = 0.52, P = 0.005 and rs = 0.44, P = 0.020, respectively). IgG anti-PS/PT antibodies were positively correlated with serum IgA (rs = 0.46, P = 0.014). Proteinuria according to urinalysis was noted in 24 (80%) of the 30 HSP patients. Serum IgA aCL antibody; IgG, IgM, and IgA anti-PS/PT antibody; and IgA anti-β2GPI antibody concentrations in relation to the presence or absence of proteinuria in the HSP patients are listed in Table 4. Serum IgA aCL antibody levels differed significantly between the proteinuria-positive and proteinuria-negative HSP patients (P = 0.024). A similar trend was seen with respect to IgG anti-PS/PT antibodies (P = 0.045). Serum IgA aCL antibody, IgA anti-PS/PT antibody, and IgA anti-β2GPI antibody levels did not differ significantly between the HSP patients who were positive or negative for arthralgia, abdominal pain, and URI. There was no significant correlation between IgM anti-PS/PT antibody in our CLA patients and arthralgia and URI.

Table 3. Correlation between antiphospholipid antibodies and serum IgA, CRP*
 IgA aCL antibodyIgA anti-PS/PT antibodyIgA anti-β2GPI antibodyIgM anti-PS/PT antibodyIgG anti-PS/PT antibody
  • *

    See Table 1 for definitions.

Serum IgA     
Table 4. Serum antiphospholipid antibody (aPL) concentrations associated with proteinuria in 30 patients with Henoch-Schönlein purpura*
Serum aPL antibodiesPresence (n = 24)Absence (n = 6)
  • *

    Values are the mean ± SD. See Table 1 for definitions.

  • P = 0.024.

  • P = 0.045.

IgA aCL antibodies, APL12.99 ± 5.278.67 ± 3.92
IgA anti-PS/PT antibodies, units/ml20.58 ± 26.9310.12 ± 6.82
IgA anti-β2GPI antibodies, units/ml7.38 ± 12.073.75 ± 1.83
IgM anti-PS/PT antibodies, units/ml12.88 ± 8.1414.00 ± 8.51
IgG anti-PS/PT antibodies, units/ml14.71 ± 25.035.33 ± 3.61

HSP is defined as deposition of IgA into the dermal vessels as measured by DIF. In addition, DIF revealed C3 within the affected vessels in 25 (83%) of 30 HSP patients, IgM in 14 (47%), and IgG in 2 (6.7%). Patients 1 and 16 of the HSP patients with IgG deposits showed high serum levels of IgG anti-PS/PT antibody. We found no association between IgA alone (with or without C3) and IgG and/or IgM in the DIF testing. In the CLA biopsy samples, we found C3 deposits in 3 (38%) of the 8 patients, IgM in 1 (13%), and/or IgG in 1 (13%).


Twenty-two (73%) of our 30 Japanese adult HSP patients were positive for serum levels of IgA aCL antibodies. IgA aCL titers in the adult HSP patients were significantly higher than those in the 8 CLA patients, 8 MPA patients, and 30 normal controls. In a previous study, 21 (81%) of 26 Chinese children with HSP had a higher prevalence of IgA aCL than healthy and juvenile rheumatoid arthritis control patients (23). In the present study, there was a significantly greater presence of IgA anti-PS/PT antibodies in the HSP patients compared with the CLA patients, MPA patients, and normal controls. In addition, IgA aCL antibodies and IgA anti-PS/PT antibodies were significantly correlated. We also found that there was a significant positive correlation between serum IgA aCL antibodies and serum IgA level. IgA abnormalities suggest an immunologic basis for the pathogenetic mechanisms underlying HSP (23,24). Elevated serum IgA aCL antibody and/or IgA anti-PS/PT antibody levels may imply involvement of immunologic elements in the pathogenesis of HSP. Furthermore, we found a significant correlation between serum IgA aCL antibodies, IgA anti-PS/PT antibodies, and CRP level in the 30 HSP patients. CRP is an inflammatory marker and an elevated CRP titer likely contributes to the aggressive clinical condition. Based on these findings, serum IgA aCL antibody and/or IgA anti-PS/PT antibody levels could serve as a marker for adult HSP and therefore should be monitored to appropriately gauge disease activity in adult HSP. Of the 30 HSP patients, 24 (80%) were positive for proteinuria according to urinalysis. Patients who subsequently developed proteinuria showed significantly higher IgA aCL antibody levels than those without proteinuria. These findings suggest that monitoring serum IgA aCL antibody levels might be useful to predict renal involvement.

We found IgM anti-PS/PT antibodies in 20 (67%) of the 30 HSP patients, but this finding was not correlated with IgA anti-PS/PT antibodies and/or IgA aCL antibodies. IgM anti-PS/PT antibodies were detected in 6 (75%) of our 8 CLA patients, but neither IgA anti-PS/PT antibodies nor IgA aCL antibodies were found in these patients. Small-vessel vasculitis (leukocytoclastic vasculitis), HSP, and CLA share a common pathologic background that may be related to the increased IgM anti-PS/PT antibody production in HSP and CLA. Phosphatidylserine is a regular constituent of the inner leaflet of the cell membrane, which is only exposed on the outside of the cell membrane during apoptosis or by damaged endothelial cells (25). Some studies have demonstrated that prothrombin binds specifically to the surface of apoptotic cells (26, 27). We believe that prothrombin binds to apoptotic endothelial cells and combines phosphatidylserine in the dermis. The complexes may cause IgM anti-PS/PT antibody production in small-vessel vasculitis, which would probably be locally produced.

In conclusion, we found high titers of serum IgA aCL antibodies and/or IgA anti-PS/PT antibodies in adult HSP patients, and there was a significant association between these 2 markers. The 2 IgA antibodies were also positively correlated with CRP level. These IgA antibodies in HSP patients may reflect disease activity and play some pathogenic role. In contrast, serum IgM anti-PS/PT antibody levels were elevated in HSP and CLA patients, regardless of IgA aPL antibody levels. We speculate that IgM anti-PS/PT antibodies could be implicated in disease susceptibility for leukocytoclastic vasculitis. These findings further suggest that aPL antibodies are closely related to the pathogenic factors that trigger the development of vasculitis. Further studies are required to confirm the pathogenesis of HSP and CLA.


Dr. Kawakami 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 design. Kawakami.

Acquisition of data. Kawakami, Yamazaki.

Analysis and interpretation of data. Kawakami, Yamazaki.

Manuscript preparation. Kawakami, Mizoguchi, Soma.

Statistical analysis. Kawakami.