Clinical trial design in systemic lupus erythematosus: lessons learned and future directions

Designing successful randomized controlled trials (RCTs) in systemic lupus erythematosus (SLE) poses many challenges. It remains difficult to correlate alterations in biologic markers with clinical outcome, especially when signs and symptoms are intermittent and broadly variable between patients. Disease activity indices were not designed specifically as outcome measures in RCTs, as they were developed in the context of longitudinal observational studies. Although all disease activity indices have been validated against each other and demonstrated to show change, organ system manifestations are variably weighted; fatigue and autoantibody titers are scored in some and not in others. Due to the variability of the underlying disease course an assessment of disease activity may most accurately be portrayed as change over time, such as an area under the curve analysis. Another lesson learned is that 'responder indices' proposed in the absence of prospective validation in RCTs do not function well. The argument can always be made that any response criteria will work if the treatment is effective; but without the precedent of a product specifically approved for use in SLE, this is hard to prove. The ACR/Systemic Lupus International Cooperating Clinics (SLICC) damage index was designed to score irreversible manifestations of disease or consequences of its treatment, provided they had been present for at least six months. The damage index may best be utilized to stratify patients or balance randomization at baseline. It may also be incorporated into an endpoint analysis, to ensure that treatment or disease associated deterioration in organ system function (that may be overlooked in scoring disease activity alone) has not occurred. Patient cohort data have demonstrated that the medical outcomes survey short form-36 (SF-36) reflected the effects of SLE better than other patient reported measures. Worsening SF-36 domain scores best correlate with higher disease activity, increased glucocorticoid doses and use of cytotoxic agents. It has been shown sensitive to change in RCTs and observational cohorts, and reflects the impact of treatment with high dose glucocorticoids and immunosuppressive agents, as well as end stage renal disease. There is now a body of data derived from RCTs in SLE. Albeit limited, yet to result in an approved therapy, evidence is accumulating that indicate 'early markers' of response can be defined which may correlate with longer term clinical outcomes. This should inform us in our ongoing efforts to clinically test a broad variety of promising interventions.