Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement

Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.

Potential Risk Factors for Lupus Nephritis.

Hojjati3,  Mehrnaz, Ayub4,  Semi, Wang5,  Qi, Behrens2,  Timothy W., Gillespie5,  Emily, Petri1,  Michelle A.

Timonium, MD
Genentech Inc, South San Francisco, CA
Univ of Minnesota, Minneapolis, MN
Univ of Minnesota, Little Canada, MN
Univ of Minnesota


Kidney involvement is a major clinical problem in systemic lupus erythematosus (SLE). Consideration of possible predictive factors for development of lupus nephritis (LN) may help to identify high-risk patients. The purpose of this study was to assess whether certain demographic, clinical and immunological variables can predict the occurrence of lupus nephritis (LN).


We studied 679 SLE patients enrolled from the Hopkins Lupus Cohort via the Autoimmune Biomarkers Collaborative Network. Patients were classified into 2 groups: (1) LN, defined as patients who had either a) renal biopsy, World Health Organization class II-V and/or b) met ACR criteria for LN (persistent proteinuria >=3 or cellular casts attributable to SLE) and/or c) Lupus Activity Index renal subscore >=1. (2) Non-LN (SLE patients with no history of LN and maintaining renal subscore=0, urine protein=0, urine RBC<=5 per high power field and serum creatinine <=1.1mg/dl over the enrollment period. The variables assessed at the beginning of the study were from 2 domains, the demographic/clinical domain and the laboratory/immunologic domain. Chi-square tests were used to examine the association between LN and each variable, and we used t-tests to compare mean age and disease duration between LN and non-LN. Logistic regression was used for prediction of LN from knowledge of the person's demographic, clinical, laboratory and immunologic information at study entry. Stepwise multivariate regression was used to select the best set of predictors after adjusting for SLE disease duration.


Of the 679 patients studied, there were 348 LN patients and 222 non-LN. The LN patients were significantly younger (mean age 40.82, SD=12) compared to non-LN patients (mean age 45.6, SD=12.7; p<0.0001). Mean SLE disease duration was 132 (SD=104) months in the LN group compared to 112.1 (SD=87) months in the non-LN group (p=0.014). Photosensitivity was negatively associated with LN (51.01% in LN group vs 61.3% in non-LN, p=0.02). Variables were included in the stepwise logistic regression model as significant predictors for LN are shown in Table 1.

Interestingly, when we included only demographic and clinical variables in the logistic regression model, smoking was negatively associated with LN (adjusted OR=0.65, p=0.04); however, disease duration (adjusted OR=1.003, p=0.005) and diabetes (adjusted OR=1.7, p=0.02) were positively associated with LN. Anti RNP (adjusted OR=2.1, p=0.004) and low C3 (adjusted OR=2.04, p=0.0009) were also positively associated with LN when we included only lab/immunologic parameters in the logistic regression model.


Our results suggest that certain demographic, clinical and laboratory/immunologic factors may be predictive of LN, including obesity, high cholesterol, diabetes, and disease duration.

Table 1. Variables included in stepwise multiple regression model for LN.

VariableAdjusted odds ratiop-value
High cholesterol1.70.02
Male gender3.30.003
Caucasian ethnicity0.420.0001
Hispanic and Asian ethnicity1.50.08
Low C42.30.0009
dsDNA antibody1.60.05

To cite this abstract, please use the following information:
Hojjati, Mehrnaz, Ayub, Semi, Wang, Qi, Behrens, Timothy W., Gillespie, Emily, Petri, Michelle A.; Potential Risk Factors for Lupus Nephritis. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :473
DOI: 10.1002/art.28242

Abstract Supplement

Meeting Menu