Neurologic toxicity of intrathecal liposomal cytarabine when used for CNS prophylaxis in conjunction with the hyper-CVAD regimen

Response:

In randomized trials using liposomal cytarabine for carcinomatous meningitis, toxicity profile was similar to standard intrathecal chemotherapy (free cytarabine for lymphomas; methotrexate for solid tumors).^1,2  A higher frequency of headaches and chemical arachnoiditis occurred with the liposomal preparation. Glantz et al²  commented that attributing neurotoxicity to either tumor or side effects of intrathecal chemotherapy was often difficult. Notably, systemic chemotherapy was permitted for disease outside the meninges (not high-dose methotrexate [> 500 mg/m²/d] or cytarabine [> 2 mg/m²/d] for lymphomas). Neurotoxicity was not delineated by systemic therapy (43% of lymphoma; 20% of solid tumor participants). In the lymphoma trial, liposomal cytarabine was interrupted for hydrocephalus and hearing loss in 1 patient each, although toxicities were not directly attributed to the agent.² 

In their letter, Chamberlain and Glantz comment on the neurologic syndromes we reported and question attribution to liposomal cytarabine. Our attributions were based on 15-year experience in nearly 500 adults with acute lymphoblastic leukemia without central nervous system (CNS) disease, treated with hyper-CVAD (cyclophosphamide, vincristine, doxorubicin, dexamethasone) and standard CNS prophylaxis (methotrexate 12 mg day 2; cytarabine 100 mg day 7 for 3-8 cycles). In this setting, seizures are extremely uncommon (3 of 482 patients [0.006%]). The long half-life of liposomal cytarabine made attribution plausible in patient 1, given absence of discernable metabolic or other etiology. We have not observed hydrocephalus with standard CNS prophylaxis. Obesity and corticosteroids contribute to pseudotumor cerebri. Patient 2 was not obese, and dexamethasone dosing was only slightly increased. Investigators in our Department of Neuro-Oncology have also attributed anecdotal hydrocephalus cases to liposomal cytarabine (Morris D. Groves, M. D. Anderson Cancer Center, written personal communication, March 22, 2007). These data, taken in aggregate with the case in the lymphoma trial, raise the question whether association of hydrocephalus with liposomal cytarabine has been unrecognized or underreported.

Findings for patient 3 are better described as partial cauda equina syndrome or sacral radiculopathy. Although Chamberlain and Glantz suggest any chemotherapy given via lumbar puncture (LP) could cause such neurologic sequelae, we have not observed this after standard CNS prophylaxis given almost exclusively via LP. Decreased perineal sensation and lower-extremity paresthesias in patient 4 were also consistent with sacral radiculopathy. Standard intrathecal chemotherapy was delayed until 4 months later. The morbidity of urinary and/or fecal incontinence associated with these syndromes in 2 of 31 patients was considered unacceptable.

Encephalopathy is more appropriate terminology for patient 5; encephalitis implies an inflammatory process. Encephalopathy related to high-dose systemic methotrexate-cytarabine usually manifests during administration or immediately thereafter, not on day 15, in this case approximately 36 hours after liposomal cytarabine. An exhaustive evaluation was unable to discern any other etiology for the neurotoxicity, which led to the patient's demise. To what degree each agent contributed is an unanswerable question but does not preclude an association.

The character, constellation, and severity of neurotoxicity was concerning in potentially curable patients. We felt it prudent to alert other investigators of the need for monitoring and reporting of neurotoxicity if this agent was used in a similar manner. The absence of similar neurotoxicity in the 56 patients treated with standard CNS prophylaxis since terminating the liposomal cytarabine trial further supports our conclusions.