Reconsidering Pathophysiological pathways in abusive infant shaking: Questions beyond the triad

Dear Editor,

We read with great interest the recent article by Thiblin et al., “Pathophysiological hypotheses of the triad in abusive infant shaking: A systematic review and analysis of corroborated cases” (Forensic Sci Int Synergy. 2025; 11:100618) [1]. The authors should be commended for their rigorous attempt to disentangle the mechanistic hypotheses underlying subdural hemorrhage (SDH), retinal hemorrhage (RH), and encephalopathy in infants subjected to shaking. However, after careful appraisal, several aspects merit further discussion and clarification.

First, while the review highlights the limitations of relying solely on witnessed or admitted cases, the assumption that confessions are inherently valid remains problematic. As the authors acknowledge, legal pressures may influence confessions [2]. Future studies should prioritize objective biomechanical and clinical correlates over confessional data, possibly through the use of advanced neuroimaging biomarkers or finite element modeling of infant cranial dynamics. However, even the most advanced simulations including recent wavelet-based biomechanical models remain inherently limited, as they rely on artificial datasets that cannot fully reproduce the physiological variability of living infants. Could the authors clarify whether any biomechanical simulations were considered to triangulate the clinical findings?

Second, the reliance on case reports and case series, while unavoidable, raises concerns about selective reporting and lack of standardization. The review suggests that hypoxia may mediate the triad, yet the absence of detailed documentation on respiratory parameters, blood gases, or cerebral oximetry weakens this inference. However, hospital-measured hypoxia values may reflect secondary physiological deterioration, and therefore cannot answer questions regarding primary hypoxic mechanisms proposed in alternative models. We also acknowledge that such detailed physiologic data are rarely available in real-world AHT assessments, where events are often unwitnessed and presentation is delayed. Recent advances in multimodal monitoring in pediatric intensive care could provide a more precise understanding of hypoxic cascades in suspected abusive head trauma (AHT) [3]. Would it not be valuable for future forensic assessments to integrate such physiologic data systematically?

Third, the assumption that mixed-density SDH reflects re-bleeding in chronic collections may oversimplify neuropathological realities. Contemporary neuroradiological literature suggests that mixed attenuation can arise from clot retraction, CSF admixture, or venous congestion unrelated to trauma [4]. Standardized radiologic reporting, perhaps using structured templates, could improve reproducibility and reduce interpretive variability.

Finally, the review's conclusion that absence of extracranial trauma does not exclude shaking is prudent. Yet, this also raises the forensic dilemma of specificity. If hypoxia and re-bleeding pathways can mimic traumatic findings, how do we safeguard against overdiagnosis of abuse in infants with predisposing vulnerabilities such as benign external hydrocephalus (BEH) [5]? Incorporating population-based registries and comparative cohorts may offer greater clarity for example, prospective registries of infants with benign enlargement of the subarachnoid space (BESS) enrolled before any concern for trauma, or registries of infants with primary hypoxic events such as drowning, age-delimited and designed to exclude mechanical injury. Clear operational criteria for determining abuse through multidisciplinary assessment or exclusion of trauma markers would further strengthen such analyses.

In sum, the review provides an important synthesis but also underscores the epistemic uncertainty surrounding AHT diagnostics. Greater integration of biomechanical science, standardized radiological descriptors, and physiological monitoring data could strengthen the evidentiary foundation in both clinical and forensic practice.

CRediT authorship contribution statement

Parth Aphale: Writing – review & editing, Writing – original draft, Supervision, Conceptualization. Himanshu Shekhar: Writing – review & editing, Writing – original draft, Visualization. Shashank Dokania: Writing – review & editing.

Declaration of generative AI

The authors declare that AI-based language assistance was used solely for grammar refinement. All intellectual content, interpretation, and conclusions are the sole responsibility of the authors.

Funding and disclosures

None.

Declaration of competing interest

None.