Dexamethasone and Neurocognitive Outcomes in Survivors of Childhood Acute Lymphoblastic Leukemia

Altered brain development and neurocognitive impairment in survivors of childhood acute lymphoblastic leukemia (ALL) is the result of multiple factors that interact with one another and evolve over the lifespan. Within the first 10 years of survival, direct effects of neurotoxic therapies (i.e., cranial irradiation, intrathecal chemotherapy) appear most pronounced, though over the lifespan the impact of other therapies can begin to have indirect effects on brain integrity and function. For example, thoracic radiation increases risk for cardiopulmonary dysfunction,¹ and this dysfunction can impact brain integrity and neurocognitive function in adult survivors.² Both the direct and indirect effects of ALL and therapy can be modified by health behaviors, such as physical fitness and sleep quality.^3,4

In the manuscript by Phillips, et. al. within this issue, evidence is presented that suggests the direct effects of chemotherapy on brain function are not uniform, but rather may have differential impact on brain regions based on the sex of the survivors and the density of receptors for the specific neurotoxic chemotherapy agent within that region. As reported by the authors, glucocorticoid receptors are not uniformly distributed throughout the brain. Rather, the cerebellum and hippocampi seem to have higher concentrations of these receptors and when the brain is exposed to glucocorticoids, like dexamethasone, these regions with higher concentrations may be more effected than other regions. This principal does not necessarily apply just to glucocorticoids but may explain some of the variability in response to other chemotherapy agents and may be further modified by sex.

Phillips et al demonstrated sex differences in brain structures, including smaller subcortical volumes and thinner cortices associated with higher dexamethasone exposure. All differences were adjusted for intracranial volume, which itself did not differ between survivors and same-sexed controls. The fact that intracranial volume was similar but regional volumes were smaller in survivors compared to controls suggests overall brain growth was not suppressed, since brain growth determines cranium size, but rather development of specific regions was selectively altered. Although not presented in the brief report, smaller volume in the subcortical hippocampi and anterior thalamic regions and thinner cortices in medial and lateral frontal regions were also associated with poorer neurocognitive outcomes (p’s<0.05). However, these associations appear influenced by sex, since female survivors with smaller hippocampi demonstrated poorer performance on measures of inhibition, cognitive flexibility and processing speed, while male survivors did not demonstrate this pattern.

We have previously reported that higher methotrexate exposure was associated with poorer performance on measures of executive function and processing speed.⁵ There is now ample evidence that glucocorticoids are also associated with neurocognitive problems. This has been established in children with chronic disease,^6–8 as well as a large epidemiological study in long-term survivors of childhood cancer.⁹ However, controversy remains regarding the type and dose of glucocorticoid associated with neurocognitive impairment. Survivors of the Dana-Farber Cancer Institute 91–01 protocol were treated with either 6 mg/m²/day or 18 mg/m²/day dexamethasone and, although no major difference were noted between these doses, both were found to demonstrate poorer performance on neurocognitive measures compared to children treated with prednisone.¹⁰ A subsequent study examining outcomes in children randomized to dexamethasone or prednisone during induction on the Children’s Cancer Group 1922 protocol found that children who received 6 mg/m²/day dexamethasone demonstrated marginally lower scores on word reading compared to those treated with prednisone, though no differences in neurocognitive or other academic measures were noted.¹¹ The survivors included in the Phillips et al study were treated on the Total XV protocol at St. Jude Children’s Research Hospital and received either 8 mg/m²/day or 12 mg/m²/day of dexamethasone during continuation therapy.¹² This higher dose of dexamethasone over a longer period of time was associated with the neuroanatomical and neurocognitive problems described above.

Although associations between neurocognitive outcomes with methotrexate do not appear to differ by sex, associations with dexamethasone do differ. Poorer performance on measures of attention, working memory, processing speed, and executive function were associated with higher exposure to dexamethasone, but only among female survivors. Male survivors demonstrated only an association between dexamethasone and measures of attention. This is consistent with other studies that demonstrate female survivors on average complete fewer years of schooling and have higher unemployment rates compared to siblings.^13,14

The differences in neuroanatomical and neurocognitive associations between males and females suggests that dexamethasone exposure impacts male and female brains differently. Consistent with the model proposed by Phillips et al, estradiol could protect neurons from oxidative stress induced cell death,¹⁵ and the changes in estradiol over development may account for the fact that younger females appear to be more affected than older females.¹⁶ However, the current study did not include collection of estradiol levels during therapy and subsequently was not able to evaluate this hypothesis.

The unique impacts from dexamethasone outlined above have important implications for current and future clinical trials. As outlined by Phillips et al, oxidative injury to the central nervous system might be limited by use of a N-methyl-D-aspartate receptor antagonist during chemotherapy, such as memantine, though pre-clinical trials are required to ensure such neuroprotection does not reduce efficacy of dexamethasone in treatment of primary disease. Additionally, transcranial photobiomodulation shows promise in limiting oxidative injury in preclinical models and may have future application in prospective clinical trials.¹⁷ As with the use of cranial irradiation in prior ALL protocols, future attention should also be paid to determining minimal doses of chemotherapies required to maintain therapeutic efficacy while limiting neurotoxicity.