Several primary immunodeficiencies are under consideration for gene therapy approaches because of limitations of current standard treatment. Many primary immunodeficiencies are caused by defects in single genes expressed in blood cells; thus addition of a correct copy of the gene to hematopoietic stem cells (HSCs) can generate immune cells with restored function. HSCs can be removed from a patient, treated outside the body, and reinfused. In the last decade, significant improvements have been made in transferring genes by means of retroviruses to HSCs in vitro, and gene therapy trials for patients with X-linked severe combined immunodeficiency (XSCID) and adenosine deaminase–deficient severe combined immunodeficiency have restored immune competence. Gene therapy is actively being pursued in other immunodeficiency disorders, including chronic granulomatous disease and Wiskott-Aldrich syndrome. However, enthusiasm for the correction of XSCID by means of gene therapy has been tempered by the occurrence of 2 cases of leukemia in gene therapy recipients caused by insertion of the retroviral vector in or near the oncogene LMO2. The likelihood of retroviral insertional mutagenesis was estimated to be very low in the past on the basis of theoretic calculations and the absence of observed malignancies in animal studies and early clinical trials. Emerging new findings on retroviral integration both in the patients with XSCID and experimental animals now indicate that the insertion of retroviral sequences into the genome carries significant risk. Understanding the magnitude of risk is now a priority so that safety can be improved for future gene therapy clinical trials.