Arthritis & Rheumatism, Volume 63,
November 2011 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Chicago, Illinois November 4-9, 2011.

Immunoglobulin Concentrations and Infection Risk Among Patients with ANCA-Associated Vasculitis Treated with Rituximab or Cyclophosphamide.

Specks1,  Ulrich, Merkel2,  Peter A., Seo3,  Philip, Spiera4,  Robert F., Langford5,  Carol A., Hoffman5,  Gary S., Kallenberg6,  Cees G.M.

Mayo Clinic, Rochester, MN
Immune Tolerance Network, Bethesda, MD
Univ of Alabama-Birmingham, Birmingham, AL
National Institute of Allergy and Infectious Diseases
NIAID, Bethesda, MD
Johns Hopkins University, Baltimore, MD
Rho, Chapel Hill, NC
Food & Drug Administration, Bethesda, MD
Boston University School of Medicine, Boston, MA
Johns Hopkins Vasculitis Center, Baltimore, MD
Hospital for Special Surgery, New York, NY
Cleveland Clinic, Cleveland, OH
University Medical Center Groningen, Groningen, Netherlands
Duke University Medical Center, Durham
Genentech, Inc., South San Francisco, CA
Genentech, So San Francisco, CA


Patients (pts) with ANCA-associated vasculitis (AAV) are treated with immunosuppression and are at increased risk of infection. The incidence of low immunoglobulin (Ig) concentrations observed with rituximab (RTX) and cyclophosphamide (CYC) is therefore a potential concern. Limited data exist on the incidence and risk for infection of low Ig concentrations in AAV. We evaluated changes in Ig concentrations among patients (pts) treated with either RTX or CYC followed by azathioprine (CYC/AZA) in the RAVE trial (NEJM 2010), and investigated the impact of these changes on infection.


Post hoc analysis of RAVE, a trial of 197 pts with severe AAV randomized and treated with RTX followed by placebo (n=99) or CYC followed by AZA (n=98). Quantitative assays for IgM, IgG, and IgA were performed at baseline, 6 months (M), and 18M. Proportions of pts with low Ig (<lower limit of normal (LLN)) at each time point and the median changes in Ig at 6M and 18M were assessed, along with the rates of overall and serious infections.


Of the 56 pts who had low Ig of any isotype at baseline, 36 (64%) entered with relapsing disease. Pts in both treatment groups had low Ig at baseline and the proportions with low Ig increased at 6M and 18M (Table 1). Median changes from baseline in Ig concentrations were similar in the RTX and CYC/AZA groups. In the RTX group, pts with low Ig levels at any time had received numerically higher quantities of glucocorticoids (GC) at baseline compared to those who did not have low Ig (IgM: 1.51 g prednisone or equivalent vs 1.03; IgG: 1.31 vs 1.27; IgA 1.73 vs 1.11; p=0.07, 0.45, 0.32, respectively). Rates of overall and serious infections were similar in pts with low Ig at any time compared to those with normal Ig levels (Table 2).

Table 1. Change from Baseline and Proportion of Patients with Low Ig (< LLN&) by Visit

Baseline (n*)9994
  IgM < LLN16 (16.2%)6 (6.4%)
  IgG < LLN18 (18.2%)20 (21.3%)
  IgA < LLN5 (5.1%)10 (10.6%)
6M (n*)8377
    Median Change from baseline (mg/dL)-42.0-42.5
    < LLN49 (59.0%)36 (46.8%)
    Median change from baseline (mg/dL)-259-252
    < LLN54 (65.1%)43 (55.8%)
    Median change from baseline (mg/dL)-48-59
    < LLN25 (30.1%)24 (31.2%)
18M (n*)6960
    Median Change from baseline (mg/dL)-38.0-23.5
    < LLN27 (39.1%)17 (28.3%)
    Median Change from baseline (mg/dL)-185-131
    < LLN31 (44.9%)19 (31.7%)
    Median Change from baseline (mg/dL)-44-42
    < LLN19 (27.5%)16 (26.7%)
& LLN: IgM 23 mg/dL (16 – 19 yrs), 40 mg/dL (>=20 yrs); IgG 549 mg/dL, 700 mg/dL; IgA 61 mg/dL, 70 mg/dL
* n indicates number of subjects with available data at the corresponding visits.

Table 2. Rates of overall and serious infectious events (SIE) in patients with low and normal Ig levels

Normal IgMIgM < LLN at any timeNormal IgMIgM < LLN at any time 
N (pt-year)45 (51.5)54 (74.7)57 (66.5)39 (51.2)
rate of infection per patient year1.611.281.081.56
rate of SIE per patient year0.
 Normal IgGIgG < LLN at anytimeNormal IgGIgG < LLN at anytime
N (pt-year)36 (37.1)63 (89.1)43 (48.4)53 (69.3)
rate of infection per patient year1.861.231.161.38
rate of SIE per patient year0.
 Normal IgAIgA < LLN at anytimeNormal IgAIgA < LLN at any time
N (pt-year)70 (86.3)29 (39.9)66 (77.4)30 (40.3)
rate of infection per patient year1.301.681.241.39
rate of SIE per patient year0.


A substantial proportion of pts with AAV receiving either RTX or CYC/AZA had low Ig concentrations at baseline, perhaps resulting from immunosuppression prior to trial entry. The proportions of pts with low Ig increased at 6M and 18M in both groups, and median changes in Ig levels were similar in both groups. Low Ig was not associated with increased rates of infections.

To cite this abstract, please use the following information:
Specks, Ulrich, Merkel, Peter A., Seo, Philip, Spiera, Robert F., Langford, Carol A., Hoffman, Gary S., et al; Immunoglobulin Concentrations and Infection Risk Among Patients with ANCA-Associated Vasculitis Treated with Rituximab or Cyclophosphamide. [abstract]. Arthritis Rheum 2011;63 Suppl 10 :789

Abstract Supplement

Meeting Menu