Clinicopathological insights into lupus nephritis in South Africans: a study of 251 patients.

There are few published studies on biopsy proven lupus nephritis (LN) from sub-Sahara Africa, mainly due to lack of expertise and pathology back-up for performing and interpreting renal biopsies in many centres. The purpose of this study was to document factors associated with biopsy proven LN and to determine clinical and laboratory models that best predict proliferative LN in South Africans. Of the 251 patients studied, 84.1% were females and 79.3% were of mixed ancestry. There were more observed cases of proliferative LN (63%) than non-proliferative LN. Factors associated with proliferative LN were male gender (p = 0.049), haematuria on dipstix (p < 0.0001), proteinuria on dipstix (p = 0.042), low serum albumin (p = 0.032), low complement C3 (p < 0.0001), low complement C4 (p = 0.009) and positive double-stranded DNA (p = 0.039). Using four models designed from various combinations of the factors associated with proliferative LN, the specificity and positive predictive values were highest for the model that combined gender (male), presence of dipstix haematuria and proteinuria, hypoalbuminaemia, low C3 and low C4 and positive double-stranded DNA (100% respectively). Further study is recommended to identify the value of using these demographic and laboratory parameters in identifying patients with proliferative LN in resource limited centres where the performance of a biopsy is not possible.