Acute promyelocytic leukemia (APL) is characterized by a specific gene rear
rangement and the generation of the PML-RAR alpha fusion transcript which r
esults from a translocation between chromosomes 15 and 17. Targeted therapy
with all-trans retinoic acid (ATRA) and anthracycline-based chemotherapy r
esults in an apparent cure in 70-80% of patients. Both allogeneic (ALLO) an
d autologous (AUTO) hematopoietic stem cell transplantation (HSCT) are effe
ctive in acute myeloid leukemia (AML), but their role in APL is not clear g
iven the excellent outcome with ATRA and chemotherapy. Several retrospectiv
e studies have analyzed the outcome of patients undergoing AUTO or ALLO-HSC
T in first (CRI) or second (CR2) complete remission. Most of these studies
have shown significant transplant-related mortality (TRM) with ALLO-HSCT, b
ut a reduction in relapse rate compared with AUTO-HSCT. The high TRM with A
LLO-HSCT and the excellent outcome with ATRA and chemotherapy do not justif
y recommending this procedure for the majority of patients in CRI. The role
of AUTO-HSCT in CRI also is unclear. A small subset of patients at high ri
sk of relapse, possibly identifiable by a high white blood cell count at pr
esentation may benefit from HSCT. Most patients with relapsed disease achie
ve CR2 with ATRA, arsenic trioxide, or combination therapy. However, it is
not known if these responses are sustained or if consolidation with HSCT ha
s a place in this setting. The outcome of AUTO-HSCT in CR2 using stem cells
that are negative for PML-RAR alpha is excellent. It is unclear whether AL
LO-HSCT from an HLA-identical sibling is superior to AUTO-HSCT with PML-RAR
alpha -negative cells in CR2 since the former would be associated with gra
ft-versus-leukemia effects and the latter with lower TRM. Alternatively, ar
senic trioxide or re-treatment with ATRA, followed by intensive chemotherap
y may also be effective. A randomized prospective clinical trial, or a retr
ospective analysis of the available data would be useful in answering this
critical question.