Background/Purpose: The ability of predicting disease flares in systemic lupus erythematosus (SLE) by serial measurements of antibodies to double stranded DNA (anti-dsDNA) remains uncertain. Previous studies have found inconsistent results and clinical guidelines provide vague recommendations regarding this aspect of disease management. The objective of this study is to assess the diagnostic value of rise in anti-dsDNA for prediction of a disease flare in SLE.

Methods: A MEDLINE search without language restrictions was conducted to identify papers on anti-dsDNA and disease flares in SLE. Papers that could not be excluded by a review of titles and abstracts were subject to a detailed review. The primary inclusion criteria were 1) availability of the data to extract the sensitivity and specificity of a rise in anti-dsDNA for a disease flare and 2) the rise in anti-dsDNA preceded the flare. For each study, the positive and negative likelihood ratios were calculated and information extracted on disease manifestations and laboratory methods of anti-dsDNA measurements. Summary estimates with 95% confidence intervals (CI) were calculated with a random effects model and results expressed with forest plots. Between-study heterogeneity was assessed with I², which takes a value from 0 to 1 and refers to the part of the total variability in the meta-analysis that is due to heterogeneity. Meta-regression was used to assess the impact of the i)publication year, ii)gender proportions, iii)proportion of patients who had renal manifestations, iv)method- and v)frequency of anti-dsDNA measurements on summary estimates and heterogeneity. Egger’s test was used to assess for publication bias using p-value of <0.05 as cut-off. The trim-and-fill method was used to adjust the summary estimates for publication bias.

Results:

Literature search identified 1690 papers of which 58 underwent full-text review. Nine studies with a total of 922 patients were included, of them 256 patients had a rise in anti-dsDNA and 243 had flares. Summary estimates for positive and negative likelihood ratios for anti-dsDNA rise and subsequent disease flare were 4.15 (95% CI 2.03-8.47) and 0.47 (95% CI 0.31-0.72) respectively (Figure). There was substantial between-study heterogeneity I²=0.89). Neither disease manifestations nor the methods anti-dsDNA measurements affected the summary estimates or explained heterogeneity. Findings were consistent with presence of publication bias (p=0.021). Adjustment for publication bias substantially attenuated the summary estimates resulting in positive and negative likelihood ratios of 1.90 (95% CI 0.84-4.33) and 0.72 (95% CI 0.46-1.14), respectively.

Conclusion:

A rise in anti-dsDNA has a limited ability for predicting flares in SLE and publication bias may have exaggerated the utility of serial anti-dsDNA measurements. These findings have direct implications for clinical management of SLE.