Cancer risk in a large inception SLE cohort: Effects of demographics, smoking, and medications.

OBJECTIVE:To assess cancer risk factors in incident SLE. METHODS:Clinical variables and cancer outcomes were assessed annually among incident SLE patients. Multivariate hazard regression models (over-all risk, and most common cancers) included demographics and time-dependent medications (corticosteroids, antimalarial drugs, immunosuppressants), smoking, and adjusted mean SLE Disease Activity Index-2K. RESULTS:Among 1668 patients (average 9 years follow-up), 65 cancers occurred: 15 breast, 10 non-melanoma skin, seven lung, six hematological, six prostate, five melanoma, three cervical, three renal, two each gastric, head and neck, and thyroid, and one each rectal, sarcoma, thymoma, and uterine cancers. Half of cancers (including all lung cancers) occurred in past/current smokers, versus one-third of patients without cancer. Multivariate analyses indicated over-all cancer risk was related primarily to male sex and older age at SLE diagnosis. In addition, smoking was associated with lung cancer. For breast cancer risk, age was positively and anti-malarial drugs were negatively associated. Anti-malarial drugs and higher disease activity were also negatively associated with non-melanoma skin cancer (NMSC) risk, whereas age and cyclophosphamide were positively associated. Disease activity was associated positively with hematologic and negatively with NMSC risk. CONCLUSIONS:Smoking is a key modifiable risk factor, especially for lung cancer, in SLE. Immunosuppressive medications were not clearly associated with higher risk except for cyclophosphamide and NMSC. Antimalarials were negatively associated with breast cancer and NMSC risk. SLE activity was associated positively with hematologic cancer and negatively with NMSC. Since the absolute number of cancers was small, additional follow-up will help consolidate these findings