This is the first study describing specifically biopsy-proven kidney involvement of HUV, with detailed pathological data and long term follow-up. We show that HUV can be associated with different glomerular lesions (mainly membrano-proliferative GN, but also crescentic GN, with or without immune glomerular deposits), lesions of renal arterioles (both acute and chronic, sometimes with C1q and immune complex deposits), and tubulo-interstitial inflammation (sometimes with tubular C1q and immune complex deposits). Notably, at the time of kidney biopsy, the interstitial fibrosis was often absent or mild, suggesting a rapid and recent occurrence of renal lesions. This kidney involvement can be life-threatening, with AKI requiring emergency dialysis, and can lead to ESKD. Yet, even with a severe initial presentation, a complete renal recovery can be obtained with corticosteroids, and possibly immunosuppressive therapy, in a majority of patients.
HUV is an immune-complex mediated systemic vasculitis affecting small vessels (1). Immune complex deposits in vessel walls are usually documented on skin biopsies of urticarial lesions, and we show here that these deposits, comprising C1q, can also be encountered in renal arterioles.
Yet, immune deposits are not systematic in the kidney of patients with HUV, especially in glomeruli, were pauci-immune forms of crescentic GN can be encountered. An overlap between HUV and ANCA-associated vasculitis (AAV) could be discussed in these patients. Here, among the 6 patients with positive ANCA antibodies, only 4 had specific ANCA, among whom 3 displayed membrano-proliferative GN (classically not encountered in AAV) and only 1 displayed pauci-immune CGN. This last patient, indeed, both responded to HUV classification criteria (chronic urticaria, complement consumption, arthralgia, glomerulonephritis, and anti-C1q antibodies) and displayed features of AAV (arthralgia, glomerulonephritis and anti-MPO antibodies).
Similarly, the possible overlap between HUV and systemic lupus erythematosus can be discussed in patients with positive ANA. Among the 7 patients with positive, only 1 had anti-dsDNA, and displayed isolated interstitial nephritis. Only 1 patient with positive ANA displayed a membrano-proliferative GN with full-house immune deposits. Apart from this kidney involvement, this patient had no specific feature of lupus but displayed: biopsy-proven cutaneous leucocytoclastic vasculitis, angioedema, and recurrent bronchitis with chronic respiratory failure.
HUV is a rare disease, but its exact incidence is unknown (6). Kidney involvement of HUV is even more rare: in the largest retrospective cohort of HUV, comprising 57 cases gathered in France over 20 years, 10 (18%) patients displayed kidney involvement of HUV (5). In an American cohort comprising 18 patients with HUV, 9 (50%) patients displayed kidney involvement (4). More recently, a Swedish cohort of 16 patients reported 5 patients with kidney involvement, among whom 2 reached ESKD (6). Membrano-proliferative GN, membranous GN, minimal change disease, focal glomerulopathy and anti-glomerular basement membrane nephritis have been described in the literature (32). We confirm in this work the heterogeneity of clinical presentation and kidney pathological lesions in patients with HUV.
AntiC1q antibodies are IgG antibodies directed against the collagen-like region of C1q. They are encountered in HUV and in other diseases associated with immune complex deposition, such as mixed connective tissue disorder, systemic lupus erythematosus and post-infectious vasculitis (5). The presence of anti-C1q antibodies has been associated with kidney involvement in lupus, and more specifically with crescentic lesions (33). The presence of anti-C1q antibodies was only a minor criterion in Schwartz et al (2), but was considered more discriminant in the revised international Chapel Hill consensus conference of 2012 (1). However, the specificity and sensitivity of anti-C1q antibodies are unknown. Their presence is neither sufficient nor mandatory to diagnose HUV: 67% of patients from the present cohort, and 38% of patients from the literature had positive anti-C1q antibodies. In the French retrospective cohort by Jachiet et al, 55% of HUV patients had positive anti-C1q antibodies and 90% had a low C1q, and 88% from the American cohort by Wisnieski et al had positive anti-C1q antibodies. The prognostic value of anti-C1q antibodies in HUV is controversial: it was associated with renal involvement in the French cohort by Jachiet et al, but showed no prognostic value for the global outcome (ESKD or death) in the present cohort and in the literature review of patients with kidney involvement.
The treatment of HUV and of its kidney involvement is far from established. While corticosteroids were used in all patients from the present cohort, Rituximab was used mostly as a 2nd or 3rd line therapy, but could be proposed as a corticoid-sparing agent in relapsing forms of HUV.
We underline the value of repeat kidney biopsy in these patients, which can allow the differentiation of active versus chronic renal lesions and guide the need for immunosuppressive regimen.