Cut-off values of immunological tests to identify patients at high risk of severe lupus nephritis

ABSTRACT

Cut-off values for anti-dsDNA, anti-nucleosome and anti-C1q antibodies tests and for complement-mediated hemolytic activity (CH50) were explored to identify patients with high risk of developing severe lupus nephritis (LN). Forty-one patients with confirmed systemic lupus erythematosus (SLE) were identified; their levels for the three antibodies and complement had been measured on a same serum sample. These patients were classified based on the presence of renal involvem ent; sixteen had active proliferative LN. With the cut-off values accepted in the laboratory for SLE diagnosis (anti-dsDNA > 100 UI/ml, anti-nucleosome > 50 U/ ml or CH50 < 190 UCH50%) no significant differences were found between patients with and without LN. Anti-C1q > 40 U/ml showed a statistically significant association with LN and had 80% of specificity. Cut-off values for LN identified by Receiver Operating Characteristic curves (ROC) were higher for anti-dsDNA (> 455 IU/ml) and anti-nucleosome (>107 U/ml), lower for CH50 (< 150 UCH50%) and, for anti-C1q (> 41 U/ml) coincided with the cut-off values accepted for SLE. Anti-C1q > 134 U/ml had a 92% of specificity, 56% of sensibility and was associated with a fifteen-fold increased risk of LN. The simultaneous presence of anti-nucleosome > 107 U/ml and anti-C1q > 134 U/ml was associated with a 27-fold higher probability for LN. According to these results, the cut-off values used to detect SLE activity could be inadequate to identify patients at high risk of severe LN.

RESUMEN

Se exploraron valores de corte para los ensayos de anti-ADNdc, anti-nucleosoma, anti-C1q y complemento hemolítico total (CH50) capaces de identificar los casos con mayor riesgo de nefritis lúpica (NL) grave. Se seleccionaron 41 pacientes ≥ 16 años con lupus eritematoso sistémico (LES) confirmado que tenían titulados los niveles de los tres anticuerpos y CH50, en una misma muestra de suero. Fueron clasificados según presencia de compromiso renal; 16 presentaron formas proliferativas de NL activa. Con los valores de corte aceptados por el laboratorio para el diagnóstico de LES (anti-ADNdc > 100 UI/ml, anti-nucleosoma > 50 U/ml o un CH50 < 190 UCH50%) no se encontraron diferencias significativas entre casos con y sin NL. Un anti-C1q > 40 U/ml tuvo una especificidad del 80% y mostró una asociación estadísticamente significativa con NL. Al aplicar curvas Receiver Operating Characteristic (ROC) para NL, se identificaron valores de corte más altos para anti-ADNdc (> 455 IU/ml) y anti-nucleosoma (> 107 U/ml), más bajo para CH50 (< 150 UCH50%) y para el anti-C1q (> 41 U/ml) coincidió con el aceptado para diagnóstico de LES. Un anti-C1q > 134 U/ml presentó una sensibilidad del 56%, una especificidad del 92% y se asoció con quince veces más riesgo de NL. La presencia simultánea de anti-C1q > 134 U/ml y anti-nucleosoma > 107 U/ml se asoció 27 veces más riesgo de NL. De acuerdo a estos resultados los valores de corte empleados para actividad en pacientes con LES podrían resultar inadecuados para identificar pacientes con mayor riesgo de NL grave.