Comparison Of ACR Classification Criteria For Systemic Lupus Erythematosus

【Objectives】: We compared the sensibility and specificity of the following3criteria：theclassification criteria of American Rheumatism Association revised in1982、1997and2009,and discussed the diagnosis efficiency on each of them.【Methods】：278cases of SLE patients from Union Hospital of Fujian Medical Universityincluding135patients with SLE and143with non SLE used as control,were enrolled in thestudy. Their clinical and laboratory data that coincide with1982,1997and2009ACR-SLEclassifications were filled in each form.The results were analized by using SPSS17.0statisticalsoftware package.【Results】:1. Comparing the patients with SLE and non SLE patients, the difference in theskin lesions、Oral ulcers、Nonscarring alopecia、Serositis、Renal disorder、Hematologic disorder(Leukopenia)、 ANA、 anti-ds-DNA、 anti-Sm、 Antiphospholipid antibodies and Lowcomplement had statistical significance (P<0.05). But the difference in Synovitis\Inflammatory synovitis、Neurologic disorder had no statistical significance (P>0.05).2. The sensitivity of the1982、1997and2009ACR criteria were86.67%、86.67%、97.78%(P=0.002);the specificity were95.80%、95.80%、88.81%respectively (P=0.022),showing thatthe2009criteria had greater sensitivity but lower specificity by contrast while the1997、1982ACR criteria come out to have same sensitivity and specificity.In terms of accuracy, the threeclassifications were91.37%、91.37%、93.17%respectively,showing no statistical significance(P=0.666）.3.When tracing rheumatic diseases (except SLE)as control group(n=113), specificity ofthe three classifications were96.46%、96.46%、91.15%,showing no statistical significancebetween them (P>0.05). The accuracy of the three classifications were91.02%、91.02%、94.69%,with no statistical significance difference (P>0.05).【Conclusions】：The sensitivity of the2009clssification was higher,but specificity was lowerwhich compared to those of ACR-SLE classification revised in1982and in1997.There were no statistical significance in accuracy among three classifications.