The impact of perioperative factors on subsequent intelligence quotient deficits in children treated for medulloblastoma/posterior fossa primitive neuroectodermal tumors

Background. Despite success in treating children with medulloblastoma/posterior fossa primitive neuroectodermal tumor (PF PNET), some children survive with significant neurocognitive sequelae. This study was performed to understand better the significance of perioperative factors on subsequent full scale intelligence quotient (FSIQ) deterioration in these children. Methods. Twenty‐eight children who underwent prospective and serial neurocognitive testing were studied. All children underwent surgery followed by radiotherapy with or without chemotherapy between 1983 and 1987 for medulloblastoma/PF PNET and were disease free when this study was conducted. IQ testing was performed before surgery and after the completion of radiation therapy. The clinical courses of the patients were correlated with changes in the corresponding intelligence quotients of each child. Factors correlating with neurocognitive declines were examined by chi‐square or Fisher exact test analysis. Differences in mean IQs were examined by the t test. Factors found to be significant were analyzed by exact logistic regression analysis. Results. The presence of adverse factors such as neurologic deficits, meningitis, or shunt infections, or the need for repeat surgery was correlated significantly with IQ deficits after treatment. Of the subset of children with one or more of these factors, 13 of 16 (81%) sustained decreases in FSIQ; 7 of 16 (43.8%) had decreases of 20 points of more. In contrast, only 3 of 12 (25%) of the children without the factors sustained FSIQ decreases, and no child sustained a decrease of more than 13 points. The mean FSIQ change after treatment in the group with factors was ‐15.7 (95% confidence interval [CI]: ‐24.0, ‐8.4), and the median was ‐18. The mean FSIQ change in the group without factors was 4.8 (95% CI: ‐0.5, 10.1), and the median was 5. The difference in mean FSIQ change between the two groups was significant (P < 0.0001). On univariate analysis, both the presence of adverse factors and an age less than six years correlated with neurocognitive deficit. On regression analysis, only the presence of adverse factors was significant (odds ratio 11.53; 95% CI, 1.65‐116.58; P = 0.009), whereas age was not (P = 0.27). Conclusions. Perioperative events or complications may account for some of the neurocognitive deterioration seen in these children after treatment, especially in the very young. The occurrence of these factors is associated with a significantly great