To retrospectively evaluate safety and efficacy of long-term treatment with Cyclosporine A (CSA) in patients with systemic lupus erythematosus (SLE) poorly responsive to treatment with corticosteroids (CCS) and/or conventional disease-modifying anti-rheumatic drugs (DMARDs), SLE patients who had received CSA-based induction and maintenance regimens according to disease activity were recorded. Efficacy was assessed using the SLE Disease Activity Index (SLEDAI) and laboratory analyses. Forty SLE patients (including 18 with lupus nephritis, 11 with neurological involvement and 7 with overlap syndromes (4 Sjögren's syndrome, 2 myasthenia gravis and 1 Behçet's disease) were recorded. According to baseline SLEDAI, 30 patients had severe and 10 moderate SLE. Mean SLEDAI scores and relevant laboratory values significantly reduced from baseline (22∓10 vs 5∓6; P < 0.002) during the follow-up period (8∓2 years; range 1-15). Twenty-three (57.5 percent) patients achieved excellent (improvement in the range 70-100 percent) response to treatment (10 of whom were subsequently maintained on CSA monotherapy), 14 (35 percent) had good-fair (improvement in the range 25-69 percent) response and 3 (7.5 percent) had to interrupt therapy (including CSA) for disease worsening. Mild and transient adverse events occurred in 15 (37 percent) patients, including hypertrichosis (17.5 percent), gum hypertrophy (17.5 percent) hypertension (12.5 percent), abdominal pain (7.5 percent), and dyslipidemia (5 percent), but treatment interruption was not required. Low-dose CSA together with other drugs is effective to induce, or as monotherapy to maintain, long-term (at least 2 years) remission, and is generally well tolerated in patients with moderate or severe SLE poorly responsive to CCS and/or conventional DMARDs. Furthermore, the favourable effect of CSA treatment may allow to spare more cytotoxic drugs.