Critics Question Statistical Rigor of Elderly Brain Tumor Study

A letter to the editor published in the Egyptian Journal of Neurosurgery on 20 August 2026 raises significant methodological concerns regarding a recent study on morbidity and mortality in elderly brain tumor surgery. The correspondence, authored by Assia Zanat, Nour Abu Shawish, and Mehnaz Ejaz Khokhar, critiques the statistical design of a prospective cohort study by Hussain et al. The authors of the letter argue that the multivariate logistic regression model used in the original study suffers from insufficient statistical power, potentially compromising the reliability of its findings on mortality predictors in geriatric neurosurgery.

The central criticism focuses on the ratio of events to predictor variables in the statistical analysis. This results in fewer than five events per variable, a ratio that falls well below the commonly recommended threshold of ten events per variable. Statistical literature indicates that when the number of events per variable is this low, regression coefficients can become unstable and biased, leading to imprecise estimates. The letter highlights that this statistical limitation is reflected in the wide confidence intervals reported in the original paper, which restricts the precision of the estimates and their generalizability to other patient populations.

A specific example of this imprecision is provided in the critique. For instance, the original study reported an adjusted Odds Ratio of 5.12 for patients with an American Society of Anesthesiologists (ASA) score of 3 or higher. However, the 95% Confidence Interval for this estimate ranged from 1.15 to 22.84. Such a broad range indicates substantial uncertainty in the magnitude of the risk associated with higher ASA scores.

A second major concern raised by the letter concerns the composition of the study cohort. The original analysis combined patients with markedly heterogeneous tumor pathologies, ranging from benign meningiomas to aggressive glioblastoma multiforme, into a single category of “brain tumor surgery.” Prior systematic reviews have demonstrated that outcomes following meningioma resection in elderly patients follow a distinct risk profile compared to surgery for malignant gliomas. By pooling these diverse diagnoses, the original study may have obscured diagnosis-specific risks. The apparent effects of age, ASA score, or Karnofsky Performance Status (KPS) on mortality may therefore partly reflect the specific risks associated with the tumor type rather than a generalizable age-related surgical risk.

The letter also points to selection bias inherent in the study design. As a surgical cohort, all included patients had already been deemed suitable operative candidates by a multidisciplinary team. This pre-selection process may limit the generalizability of the findings to the broader population of elderly patients presenting with brain tumors. This includes patients managed conservatively or with radiosurgery. While the original authors acknowledged this limitation, the letter emphasizes its implications for external validity, suggesting that the results may not apply to the full spectrum of elderly brain tumor patients.

Further methodological issues relate to the choice of the mortality endpoint. The original study utilized a 90-day mortality endpoint, which the letter argues may not adequately capture outcomes for patients with glioblastoma. Even with optimal multimodal therapy, including surgery, radiotherapy, and concomitant temozolomide, median survival in this population typically extends well beyond 90 days. Consequently, the 90-day window may fail to reflect the true long-term surgical benefit or risk for these patients. The letter suggests that diagnosis-stratified analyses or a longer follow-up period would provide a more complete and accurate picture of outcomes across this heterogeneous population.

The correspondence concludes by congratulating the original authors for their contribution to an important and underexplored area of geriatric neurosurgery. The letter acknowledges the value of the prospective cohort design and the detailed analysis of mortality predictors. The authors of the letter suggest that future research should focus on multicenter studies with larger samples. Additionally, they recommend longer follow-up periods and diagnosis-stratified analyses to strengthen risk-prediction models for elderly patients undergoing brain tumor surgery.

Subject of Research: Methodological critique of a prospective cohort study on morbidity and mortality in elderly brain tumor surgery

Article Title: Morbidity and mortality in elderly brain tumor surgery: methodological considerations

Article References: Zanat, A., Abu Shawish, N., & Khokhar, M. E. (2026). Morbidity and mortality in elderly brain tumor surgery: methodological considerations. Egyptian Journal of Neurosurgery, 41(1), Article 145. https://doi.org/10.1186/s41984-026-00651-6

Image Credits: AI Generated

DOI: 10.1186/s41984-026-00651-6

Keywords: geriatric neurosurgery, brain tumor surgery, statistical power, morbidity and mortality, methodological critique, Egyptian Journal of Neurosurgery, logistic regression, clinical research, Morbidity, mortality, elderly, brain