Reactivation of herpes zoster occurs in patients treated with immunosuppressive drugs but is generally not considered an infection associated with either acute myeloid leukemia (AML) or with its intense myelosuppressive treatment. Our institution has been using arsenic trioxide (ATO) since 1997 in all relapsed APL patients and more recently has included it in our front line APL regimens. Here we present a retrospective analysis investigating if ATO treatment is associated with herpes zoster reactivation and whether acyclovir can prevent this in APL patients as well as those with other malignancies on clinical trials.

Methods: We evaluated records of patients treated at Memorial Sloan Kettering Cancer Center dating back to 1970, and selected patients with electronic medical records who received ATO treatment either as conventional therapy or as part of a clinical trial (n=213). We also selected records of any patient with a diagnosis of AML over the same period who had not received ATO for use as a control group (n=3366).

Results: Among 213 patients who received ATO therapy, 27 (12.7%) were diagnosed with herpes zoster a median of 23 days after the initiation of therapy. This cohort included 76 APL patients, 11 (14.5%) of who were diagnosed with herpes zoster a median of 15 days after the start of therapy (1 days - 5 years). Among the 3366 patients in our AML cohort, 122 (3.6%) were diagnosed with herpes zoster a median of 230 days after the initiation of chemotherapy (2 days - 11 years).

In univariate analysis we found a significant association between the likelihood of herpes zoster and ATO treatment (p < 0.01). Patients receiving ATO treatment had a greater than 3 fold odds of herpes zoster compared to non-arsenic treated patients, with a trend toward accelerated development of zoster at cumulative dosages > 1000mg (likelihood ratio test p = 0.28). Acyclovir was strongly protective in both younger and older patients, with an overall relative hazard of 0.07 (likelihood ratio p < 0.01). In a multivariate logistic regression analysis, we found that only arsenic use and older age at diagnosis contributed significantly to the likelihood of herpes zoster diagnosis (p = 0.01), but a diagnosis of APL did not (p=0.89).

Conclusion: ATO is now part of the standard of care for treatment of patients with APL. We observed that the incidence of herpes zoster was more common among patients receiving ATO, with a trend toward greater risk at higher cumulative doses. There is evidence to suggest that ATO results in T-cell dysfunction (Morzadec, Bouezzedine, Macoch, Fardel, & Vernhet, 2012; Thomas-Schoemann et al., 2012), which is consistent with the observation that T-cell depleting diseases such as HIV and T-cell directed immunosuppressives are also associated with higher rates of herpes zoster. We therefore suggest including acyclovir prophylaxis in all patients receiving ATO, and continue prophylactic treatment up to 6 months after completion of therapy.


Douer:Gilead: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.