Neuroimaging patterns of brain injury in children following near-drowning

Abstract

Background

Near-drowning is a prevalent cause of hypoxic ischemic injury in children and young adults. Despite its clinical significance, there is a lack of studies examining the brain magnetic resonance imaging (MRI) findings associated with near-drowning incidents.

Purpose

The aim of this study is to identify patterns of imaging findings on brain MRI scans of children who have experienced near-drowning.

Methods

This retrospective study included children who experienced near-drowning incidents and had brain MRI scans available for review between November 2000 and September 2023. Abnormal MRI findings were categorized into three patterns: (1) gray matter injury, (2) white matter injury, and (3) combined gray matter and white matter injury. For each category, we distinguished those with MRI scans obtained in the acute setting and those with MRI studies performed in the non-acute setting. When available in the electronic medical records (EMR), collected parameters for patients with acute MRI scans included: sex, age, best estimate of the water temperature, water type, duration of submersion, and whether cardio pulmonary resuscitation (CPR) was performed.

Results

The study included 50 patients (32 males, 18 females) with a median age of 32.9 (interquartile range, 19.9–69.2) months. Of these patients, 28 had acute MRI scans available, while 22 had only non-acute MR imaging. Among the 28 patients with acute MRI, 12 (42%) had primarily cortical and/or deep gray matter injury without visible white matter injury, 8 (29%) had both cortical and/or deep gray and white matter injury, and 8 (29%) were normal. The median age was 26.7 (interquartile range, 16.6–43.6) months in the acute MRI group and 42.9 (interquartile range, 27-130.3) months in the non-acute MRI group. Water temperature information was available in 25/50 cases, all occurring in warm water (9 in a bathtub and 16 in a pool). In patients with isolated gray matter injury, the submersion duration was < 3 min in 7/12 patients, while 5/12 did not have data on submersion duration. CPR was performed in 8 patients, with data unavailable for 4 cases. In patients with gray and white matter injury, submersion duration was < 3 min in 1/8 cases, with data not available for 7 patients. CPR was performed in 5 patients, with data unavailable for 3 cases. In patients with normal findings the submersion duration was < 3 min in 2/8 patients, and CPR was performed in all 8 patients.

Conclusion

Our study suggests that children who suffer near-drowning are likely to have gray matter injury on MRI obtained within the first 7 days after injury. In contrast, delayed white matter lesions, may develop weeks after the initial hypoxic-ischemic event and may be observed in chronic imaging either alone or in combination with gray matter lesions. These imaging patterns appear to resemble those described in neonatal hypoxic-ischemic injuries, although further studies are needed to confirm these associations.

Introduction

Drowning is defined as respiratory impairment resulting from submersion or immersion in any liquid, with outcomes ranging from death to non-fatal injury or no injury [1]. According to the World Health Organization, it is the fourth leading cause of death among children aged 1–4 years and the third leading cause of death among those aged 5–14 years [2]. Non-fatal drowning, commonly referred to as near-drowning, is a common cause of hypoxic-ischemic injury in children, contributing significantly to severe neurological disabilities and mortality [3–5]. Anatomical and functional neuroimaging play a vital role in assessing the severity of cerebral injury, helping management in the acute setting, and providing valuable insights into long-term prognosis. Neurological outcomes are influenced by factors such as the duration of drowning, the need for resuscitation, and the temperature of the water [3]. Additionally, the type and timing of imaging studies, as well as brain maturity, can affect the neuroimaging findings in children who have experienced a drowning event. Magnetic resonance imaging (MRI) is the most valuable imaging modality in the workup of these patients, allowing early detection of the injury and assessment of its extension and severity [6–9]. Despite its clinical importance, there is a scarcity of studies examining acute brain MRI findings in patients following near-drowning incidents. This study aims to investigate injury patterns on brain MRI in a cohort of children who have experienced near-drowning and to explore potential factors influencing these imaging features.

Materials and methods

We conducted a retrospective review of our electronic medical records (EMR) to identify children who experienced near-drowning incidents and had brain MRI scans available for review between November 2000 and September 2023. MRI scans were performed as part of the clinical evaluation. Our standard departmental protocol includes multiplanar T1-weighted images (WI) as well as T2WI, Fluid-Attenuated Inversion Recovery (FLAIR), Susceptibility-Weighted Imaging (SWI), Diffusion-Weighted Imaging (DWI) and Apparent Diffusion Coefficient (ADC) maps. Examinations were performed on both 1.5T and 3T MRI scanners. During the study period, 3T MRI units were progressively installed and used alongside preexisting 1.5T scanners. Overall, 11 of the 50 MRIs in our cohort were performed on 3T units, while the remaining 39 were acquired on 1.5T scanners. The search was conducted by cross-referencing the keywords “near drowning” and “MRI”, and yielded 69 records. Inclusion was based on the availability of brain MRI scans related to near-drowning incidents. Exclusion criteria included poor imaging quality (n = 1); follow-up MRI scans from patients with multiple MRIs over time (n = 11), as only the first MRI per patient was included in the analysis; and patients with preexisting or alternative neurologic conditions (n = 7). Preexisting neurologic conditions were defined as any previously diagnosed disorder that could potentially impact the observed MRI findings, regardless of whether the condition was well controlled. Specifically, excluded cases included children with generalized epilepsy (n = 4), xeroderma pigmentosum (n = 1), and Rett syndrome (n = 1). MRI scans were reviewed by a research scientist (LM) and a pediatric neuroradiologist with >30 years of experience (TAGMH) in consensus. We classified abnormal MRI findings into three patterns: 1) gray matter injury (central and/or cortical gray matter), 2) white matter injury, and 3) combined gray matter and white matter injury. For each category, we differentiated between those with acute MRI following the incident and those with only non-acute (beyond one week of the injury) MRI studies. The 7-day cutoff was a natural division based on our available data, rather than a predefined threshold. For those with available acute MRI scans (obtained within a week of the incident), the following parameters were collected from the EMR, when available: sex, age, best estimate of water temperature (warm vs. cold), water type (salty, sweet, or contaminated by soap), duration of submersion, and whether CPR was performed. Water temperature was classified as ‘warm’ if the event occurred in a bathtub or pool during the spring, summer, and autumn months. In the state where this study was conducted, the climate is generally warm for most of the year (40–100 degrees Fahrenheit). When available, submersion duration data were reported. Children for whom the EMR did not report data on the duration of submersion, or who were left unattended for more than 30 min with no specific submersion duration data available, were classified as ‘unavailable’. In one case, the patient was left unattended for less than 3 min, and it was classified in the 1–3 min group.

The study was approved by the Institutional Review Board for Human Subject Research at Baylor College of Medicine and Affiliated Hospitals (Approval Code: H-40231, Approval Date: 04/18/2023).

Statistical analysis

Descriptive statistics were used to summarize the study cohort. Continuous variables are reported as median with interquartile range (IQR), and categorical variables as counts and percentages.

Results

The study included 50 patients (32 males, 18 females) with median age of 32.9 (IQR, 19.9–69.2) months. Of these patients, 28 had acute MRI scans available, while 22 had only non-acute MR imaging (Table 1). Among the 22 patients with only non-acute MRI scans, 4 had injuries limited to the white matter, 7 exhibited a pattern of injury involving gray and white matter, and 11 presented normal findings. Among the 28 patients with acute MRI (Table 2), 12 (42%) had primarily cortical and/or deep gray matter injury without visible white matter injury (Fig. 1), 8 (29%) had both cortical and/or deep gray and white matter injury (Fig. 2), and 8 (29%) were normal. The mean median age was 26.7 (IQR, 16.6–43.6) months in the acute MRI group and 42.9 (IQR, range 27-130.3) months in the non-acute MRI group. Among the incidents where the water temperature information was available (n = 25), all occurred in warm water: 9 incidents in a bathtub and 16 in a pool during the warm months. In patients with isolated gray matter injury, the submersion duration was < 3 min in 7 of 12 cases, and no reliable data of submersion duration were available for the remaining 5 patients. CPR was performed in 8 patients, 5 of whom submerged for < 3 min. CPR data were unavailable for 4 patients. In patients with gray and white matter injury, submersion duration was less than 3 min in 1/8 cases, with data unavailable for the remaining 7 patients. CPR was performed in 5 patients, including the one who submerged for < 3 min, and data were unavailable for 3 patients. In patients with normal findings the submersion duration was less than 3 min in 2/8 patients, and CPR was performed in all 8 patients.

Table 1.

Distribution of injury patterns on acute and non-acute imaging

Patterns of injury	Acute imaging	Non-acute imaging only	Total
Gray matter	12	0	12
White matter	0	4	4
Gray matter + white matter	8	7	15
Normal	8	11	19
Total	28	22	50

Table 2.

Clinical and demographic characteristics by injury pattern on acute imaging

Patterns of injury	Acute imaging	Median age (in months)	IQR (in months)	Temperature			Kind of water			Duration of submersion		CPR
				Warm	Cold	NA	Bathtub	Pool	NA	< 3 min	NA	Yes	NA
Gray matter	12 (42%)	26.9	22-44.9	10	0	2	3	7	2	7	5	8	4
White matter	0	NA	-	0	0	0	0	0	0	0	0	0	0
Gray matter + white matter	8 (29%)	27.6	16.2–46.1	7	0	1	3	4	1	1	7	5	3
Normal	8 (29%)	26.2	13-43.6	8	0	0	3	5	0	2	6	8	0
Total	28	26.7	16.6–43.6	25	0	3	9	16	3	10	18	21	7

*NA: not available

Fig. 1.

Gray matter pattern. Axial FLAIR (A, D, G), DWI (B, E, H), and ADC map (C, F, I) images in three different patients aged 16, 27, and 23 months, obtained 3, 3, and 1 days after injury, respectively. The images demonstrate diffuse, mild, bilateral swelling of the central gray matter, with DWI hyperintensity and corresponding low ADC values (ADC dark), consistent with cytotoxic edema

Fig. 2.

Gray matter and white matter pattern. 1-year-old patient found unresponsive in the pool. MRI performed 4 days after the injury. Acute axial T2-weighted MR image (A) and DWI (B) show mild bilateral T2-hyperintensity of the basal ganglia, particularly the caudate nuclei (arrows), and thalami (arrowheads). DWI (C) and ADC map (D) demonstrate DWI-hyperintensity with matching low ADC values (ADC dark) of the hemispheric white matter, consistent with cytotoxic edema. One-year follow-up axial T2-weighted MR image (E) shows chronic volume loss and T2-hyperintense gliosis of the basal ganglia (arrows) and thalami (arrowhead), confirming basal ganglia involvement that was not apparent on the acute DWI at day 4. Follow-up also demonstrates mild diffuse white matter volume loss (F) with ex vacuo widening of the ventricles and hemispheric sulci

Discussion

Our study suggests that children who suffer near-drowning are likely to have gray matter injury on MRI obtained within the first 7 days after injury. In contrast, delayed white matter lesions, may develop weeks after the initial hypoxic-ischemic event and may be observed in chronic imaging either alone or in combination with gray matter lesions. These imaging patterns appear to resemble those described in neonatal hypoxic-ischemic injuries, although further studies are needed to confirm these associations.

Near-drowning is a frequent cause of hypoxic-ischemic injury of the brain in children, and imaging findings may vary significantly. These variations are influenced by factors such as the mechanism of injury (hypoxia, with or without ischemia), the patient’s age and related brain maturity, the severity and duration of the event, as well as the type and timing of the imaging studies [4]. MRI scans should be performed within 3–7 days after the injury [10], and MRI protocols should include at least T1WI, T2WI, DWI sequences, and ADC map. If the MRI scan performed in the first 24 h is negative, a second examination should be obtained at 2–4 days to rule out delayed injury [4–6]. Although T1- and T2-WI may demonstrate the lesions, DWI is the earliest imaging sequence to detect abnormalities [4,6]. However, DWI often underestimates the extent of the injury in the initial 6–12 h and again after approximately 8 days. Abnormal DWI signal typically peaks at 3–5 days and then “pseudonormalizes” by the end of the first week, despite the evolution of the underlying damage (note that decreased ADC values may persist into the second week) [6, 8, 11]. T1 and T2WI are most helpful at this stage when DWI has pseudonormalized [6, 8, 11].

During a drowning event, neurological damage may occur from reduced blood oxygenation (hypoxia) and/or decreased cerebral blood flow (ischemia), depending on the severity and duration of the injury and possible cardiac arrest [3, 9]. Severe and/or prolonged hypoxemia leads to cardiac hypoxia and diminished cardiac output, ultimately resulting in brain ischemia [12]. Thus, ischemic lesions superimposed on hypoxia typically occur in patients with delayed CPR intervention or prolonged submersion, leading to a more severe injury [12, 13]. This pathophysiology is similar to neonatal HIE, where perinatal asphyxia results in systemic hypoxemia and decreased cerebral perfusion. In both conditions, the combined effects of hypoxia and ischemia lead to comparable patterns of brain injury, which explains the overlap in imaging findings.

The specific involvement of the gray matter and/or white matter also depends on the different cellular and biochemical composition of these brain structures, which influences their vulnerability to injury. Areas of the brain with higher concentrations of excitatory amino acid receptors, particularly glutamate and its corresponding NMDA receptors, are more susceptible to hypoxic-ischemic injury. Activation of these receptors leads to an excessive influx of calcium ions, triggering a cascade of harmful processes, including the generation of free fatty acids and free radicals, and the activation of lipases and proteases. These processes damage cell membranes and internal organelles, particularly the mitochondria, further depleting cellular energy stores and ultimately leading to neuronal death [4, 6, 14].

A predominant gray matter pattern, typically involving structures such as the basal ganglia, cerebral cortical gray matter, and hippocampus, is commonly observed in acute, profound hypoxic-ischemic injuries in neonates (parturitional injury), as in our study cohort. These regions are characterized by high concentrations of glutamate and NMDA receptors, making them particularly vulnerable to injury. Additionally, gray matter is more metabolically active than white matter due to its higher synaptic activity, further contributing to its susceptibility. These factors explain why gray matter, particularly in these key structures, is more commonly affected in severe hypoxic-ischemic events. In contrast, a gray matter and white matter pattern is typically observed in chronic and less profound hypoxic-ischemic injuries [9]. The white matter areas are typically less metabolically active and have lower energy demand compared to the central and cortical gray matter, making them less susceptible to early injury [4, 6].

Our study has several limitations. First, the study population is relatively small. Additionally, due to the retrospective nature of this study, several data points were missing in the EMR, preventing us from performing statistical analyses. In particular, we were unable to assess the time interval between the MRI scan and the near-drowning event, as the specific times of these events were not consistently available in the EMR. Moreover we were unable to evaluate the impact of water temperature on imaging findings, as none of the near-drowning incidents occurred in cold water. Submersion in cold water may possibly limit the degree of brain injury. Previous studies in pediatric cardiac arrest [15] have examined the association between brain MRI findings and patient outcomes. While these investigations provide valuable insights into prognostication, our study did not assess outcome correlations due to the limited and heterogeneous follow-up data available for our cohort. Further prospective studies are needed to address these gaps and provide more reliable insights.

Conclusion

Our study suggests that children who suffer near-drowning are likely to have gray matter injury on MRI obtained within the first 7 days after the injury. In contrast, delayed white matter lesions, may develop weeks after the initial hypoxic-ischemic event and may be observed in chronic imaging either alone or in combination with gray matter lesions. These imaging patterns appear to resemble those described in neonatal hypoxic-ischemic injuries, although further studies are needed to confirm these associations.

Abbreviations

ADC map

Apparent Diffusion Coefficient map

CPR

Cardio-respiratory resuscitation

DWI

Diffusion-Weighted Imaging

EMR

Electronic medical record

FLAIR

Fluid-Attenuated Inversion Recovery

MRI

Magnetic Resonance Imaging

NMDA

N-methyl-D-aspartate

SWI

Susceptibility-weighted imaging

T1WI

T1-wighted images

T2WI

T2-weighted images

Author contributions

T.H. contributed to the conception and design of the work. L.M. and T.H. were responsible for data acquisition, analysis, and interpretation. L.M drafted the main manuscript and L.M. and T.H. prepared the figures. All co-authors reviewed and edited the final draft.

Funding

The authors declare no funding.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to patient confidentiality, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Review Board for Human Subject Research at Baylor College of Medicine and Affiliated Hospitals (Approval Code: H-40231, Approval Date: 04/18/2023). The requirement for informed consent was waived by the IRB as there was no secondary patient interaction. All methods were performed in accordance with relevant guidelines and regulations. The study has been conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.