Systemic lupus erythematosus (SLE) is uncommon after the age of 50 years, and studies of elderly patients with SLE are scarce. We conducted the current study to analyze characteristics and outcome of patients with late-onset SLE in a French tertiary referral center, and to compare them with those of younger patients with SLE. From 1980 to 2000, 47 patients were identified as having late-onset SLE, defined as SLE diagnosed at or over the age of 50 years. These patients were compared with a group of 114 randomly selected patients aged younger than 50 years at SLE diagnosis. We compared clinical characteristics, laboratory data, therapy, and course. The female to male ratio was smaller in the late-onset SLE group (p = 0.0012). Some manifestations occurred less frequently in late-onset SLE: arthritis (p = 0.009), malar rash (p = 0.013), and nephropathy (p = 0.009). High-dose corticosteroids (p = 0.0016) and immunosuppressive drugs (p = 0.006) were less commonly used in the elderly. Deaths occurred more frequently in late-onset SLE (p = 0.019), with a 10-year survival rate of 71% versus 95% in early-onset SLE (p < 0.01). In patients with late-onset SLE, causes of death were usually unrelated to SLE. Analysis of pooled data from the literature, based on 714 old and 4700 young SLE patients, confirmed that late-onset SLE was characterized by a smaller female to male ratio (4.4:1 vs. 10.6:1; p = 3.10); a higher occurrence of serositis (36.7% vs. 28.6%; p = 7.10) and pulmonary involvement (21.2% vs. 11.3%; p = 6.10); and a lower occurrence of malar rash (31.1% vs. 62.4%; p = 10), photosensitivity (26.2% vs. 38.2%; p = 6.10), purpura/cutaneous vasculitis (13.4% vs. 25.9%; p = 9.10), alopecia/hair loss (24% vs. 44.9%; p = 3.10), Raynaud phenomenon (24.8% vs. 37.2%; p = 3.10), neuropsychiatric manifestations (15.3% vs. 20.2%; p = 0.025), lymphadenopathy (9.1% vs. 19.6%; p = 2.10), nephrotic syndrome (8.1% vs. 24.3%; p = 0.015), and nephritis (28.6% vs. 42.7%; p = 2.10). Regarding laboratory features, rheumatoid factor positivity was more frequent (32.7% vs. 20.1%; p = 3.10), whereas anti-RNP positivity (10.4% vs. 20.9%; p = 9.10), anti-Sm positivity (9.1% vs. 17.1%; p = 0.001), and a low CH50 complement fraction (45% vs. 64.9%; p = 0.002) were less frequent in old compared with young SLE patients. In conclusion, the clinical pattern of late-onset SLE is characterized by a lower disease severity. The reduced survival observed in this group seems to result mainly from the consequences of aging.