Traumatic Microbleeds in Mild Traumatic Brain Injury: Stability,Distribution,and Association with Other Injuries

Abstract

This study aimed to investigate whether the amount of traumatic microbleeds (TMBs) detected on magnetic resonance imaging (MRI) changes in different brain regions after a 3-month follow-up in adult patients with mild traumatic brain injury (mTBI). A 3 T MRI of the brain was conducted at baseline (3–17 days after trauma) and at 3 months after patients were diagnosed with mTBI according to the World Health Organization criteria. The TMBs were evaluated, counted, and assigned locations by a neuroradiologist according to the Common Data Elements for Neuroimaging of Traumatic Brain Injury (CDE-TBI). At baseline, 22 out of 113 (19%) patients had at least one TMB. After initial imaging, 22 patients dropped out, resulting in 88 patients, of whom 21 (24%) had at least one TMB. Across these patients, a total of 129 TMB lesions were detected at baseline, all of which were visible at follow-up imaging. No new TMB lesions that were not detected at baseline were seen at follow-up imaging. The most common CDE-TBI regions for TMBs were the frontal (77/129, 60%), temporal (33/129, 26%), and parietal (8/129, 6%) subcortical white matter. For each patient, the TMB lesions were located in one region in 10% (9/88), two regions in 3% (3/88), three regions in 3% (3/88), and four or more regions in 7% (6/88) of the CDE-TBI regions. Twenty-five out of 88 (28%) patients also had other trauma-related intracranial abnormalities: 10/25 (40%) of which had at least one TMB, which was true only for 17% (11/63) for those without (p < 0.05). In conclusion, initial TMBs can still be reliably found with conventional MRI after 3 months in patients with mTBI: delayed imaging does not necessarily compromise the detection of these lesions. Other radiological abnormalities are linked with higher prevalence of TMBs, possibly suggesting increased injury burden within the mTBI population.

Keywords

Common Data Elements for Neuroimaging magnetic resonance imaging mild traumatic brain injury prospective cohort study traumatic microbleeds TMB

Introduction

Traumatic microbleeds (TMBs), possibly resulting from a vascular injury,¹ are commonly found with conventional magnetic resonance imaging (MRI) of the brain in patients with mild traumatic brain injury (mTBI)^2,3 and resulting from low-energy trauma.⁴ However, more and smaller TMBs are also detected in patients with mTBI and in normal population controls when high-field or ultra-high-field MRI scanners are used,^2,5 making it more challenging to determine clinically relevant lesions in mTBI.³ In relation, the presence of TMBs has been reported to be associated with varying patient-related outcomes and post-traumatic complaints in mTBI.^2,3,6 Patients with mTBI, neuroradiologically accompanied by TMBs or other trauma-related intracranial abnormalities (such as subdural hematoma [SDH], intracerebral hemorrhage [ICH], brain contusions, epidural hematoma [EDH], or traumatic subarachnoid hemorrhage [tSAH]), have also been called “complicated” mTBI.^3,7

In clinical practice, there is occasional radiological concern regarding time or region-sensitive TMBs (Table 1) in patients with mTBI and potential TMBs but without any other radiological trauma-related intracranial findings. This may potentially lead to misinterpretation of the traumatic brain injury (TBI) severity and to medicolegal challenges.

Table 1.

Results of the Literature Search

Search words used	[traumatic microbleeds] and [MRI] and [traumatic brain injury]
Articles found from MEDLINE	89
Articles included^a	8

Author(s) year, PubMed identifier	Country	Study type	Number of patients	Severity of TBI	Method of quantifying TMBs	Tesla	Average time from trauma to MRI (days)	Average follow-up time (months)	Results
van der Eerden AW et al., 2022, 34719725	Netherlands	Prospective cohort	31	Moderate or severe	Neuroradiologist	3 T	∼21.0	∼6.0	69% of the possible microbleeds at baseline are classified differently at follow-up. Only 16% of them represent a definite microbleed at follow-up.
Nael K, 2020, 33033046	United States	Prospective cohort	10	Mild	Neuroradiologist	3 T	∼2.2	∼1.5	No evolution reported, MAGPI-SWI was able tp detect CMBs on four patients that were initially SWI/CMB negative.
Rizk T et al., 2020, 32228304	United States	Prospective cohort	30	7 mild, 19 moderate, 4 severe	Neuroradiologist	1.5 and 3 T	∼1.0	∼42.0	83% CMBs remained visible on MRI. Only three (17%) patients with TMBs identified during initial scan showed resolution at 1 year. In two patients, new MBs not present at the initial acute scan were observed during 90-day and 1-year follow-up.
Einarsen CE et al., 2019, 31280698	Norway	Prospective cohort	152 (11 with TAI)	Mild	Neuroradiologist	3 T	∼2.4	∼12.0	All 11/11 (100%) TAI-SWI-lesions at baseline were detectable at 12-month timepoint, but tended to be smaller and uniformly less hypotense at follow-up imaging.
Studerus-Germann AM, 2018, 30245171	Switzerland	Prospective cohort	30 (4 with TMBs)	Mild	Neuroradiologist	3 T	∼2.7	∼12.3	4/4 (100%) of TMBs were observable at the end of the radiological follow-up.
Lawrence TP, 2018, 28663054	United Kingdom	Prospective cohort	13 (6 with TMBs)	Mild, moderate, or severe	Neuroradiologist	3 T	∼0.4	∼0.2	The TMB volumes did not change significantly between the two time points. TMB counts were not reported.
Watanabe J, 2016, 27106841	Japan	Case series	3	Moderate	Radiologist	1.5 T	∼0.6	∼1.7	The TMBs were observable at initial imaging but became unobservable at subacute phase imaging. Same lesions were observable at later MRI
Toth A, 2016, 26912192	Hungary	Prospective cohort	5 (4 with TMBs)	Moderate or severe	Neuroradiologist	3 T	∼0.6	∼0.3	In three out of four (75%) patients, TMBs displayed a change during the follow-up. For two out of four patients, the total TMB count decreased due to TMB expansion. No new TMB formation at a previously unaffected site or the disappearance of an initially observable TMB was detected.

The inclusion criteria were as follows: (1) cohort with an MRI follow-up, (2) traumatic microbleeds were evaluated at baseline and at follow-up, (3) not a review or an animal model.

MRI, magnetic resonance imaging; TMB, traumatic microbleed; TBI, traumatic brain injury.

Eight studies have conducted longitudinal evaluations of TMBs in TBI patients using serial brain MRIs (Table 1).^8–15 There is some heterogeneity among these studies in terms of the MRI scanners used (approximately 70% using 3 T), the timing of initial brain scans (ranging from 0.4 to 21 days), the duration of radiological follow-up (from 0.2 to 42 months), and the reporting of lesion anatomical locations (Table 1). Although Griffin et al.¹ had a large mixed-severity TBI cohort in which 94 patients with TMBs at baseline completed 90-day radiological follow-up, they did not report TMB counts at follow-up for comparison. Even when these prospective study populations are combined, only 93 individuals with mild, moderate, or severe TBI and at least one TMB lesion have been radiologically reevaluated for TMB counts. For 28 out of 99 patients (30%), the TMB count remained unchanged during the 1.5- to 12.3-month follow-up period.^9,11,12,14 A change in TMB count was observed in 65 out of 93 patients (70%) during the 0.3- to 42-month follow-up period.^8,10,15 The change in TMB counts is thought to be multifactorial. One potential factor is the inherent uncertainty in classifying lesions at baseline,^8,11 as a proportion of suspected TMBs are later determined not to be true TMBs upon reexamination,⁸ or possible misclassification of lesions of nontraumatic origin.¹⁰ Imaging the brain very shortly after the injury might capture time-dependent signals that affect visibility^11,14,15 or quantifying the number of lesions due to lesion expansion or fusion.^13,15 It has been suggested that the structural and cellular differences in macro- and microanatomical regions may alter the temporal evolution of TMBs.^8,13 Additionally, a small percentage of radiologically observable lesions seen in susceptibility-weighted imaging (SWI) might turn out to be false positives when histologically examined.¹⁶ Individual TMB lesions may persist for years, but the total count may decrease in longer follow-up.¹⁷

Generalizing these findings to a mTBI population is challenging, as they are derived from populations with more severe or mixed-severity TBI. However, the existing small studies performed on mTBI patients^9,11,12 suggest TMB stability from 1.5 months up to a year. Building on this, we aim to determine whether TMBs persist 3 months after mTBI in the generally healthy working-age population using repeated conventional MRI and systematic counting and macroanatomical allocation of TMBs. The timing of imaging was set to reflect a non-acute follow-up in a TBI outpatient clinic. In addition, we examine whether any new TMBs are later detected in patients initially not showing any, potentially indicating time-dependent detection. In addition to the conventional variables to classify TBI severity, we radiologically grouped mTBI patients based on the presence of other trauma-related intracranial abnormalities to capture those with potentially higher trauma energy.

Methods

Study design and participants

The study protocol was reviewed and approved by the research ethics committee of Helsinki University Hospital (Study Code: 105/13/03/01/2014). The research was conducted in accordance with the Declaration of Helsinki, and all patients provided written informed consent. The study was conducted in the Traumatic Brain Injury Outpatient Clinic of Helsinki University Hospital, where adult patients with TBI are referred for non-acute follow-up from emergency departments and wards located in the Helsinki Metropolitan area. From referred patients, study participants were prospectively recruited from 2015 to 2018.

This article is part of a larger nested longitudinal study,⁶ where patient and TBI-related factors influencing returning to work were studied. This article presents an extended radiological follow-up of 113 adults with mTBI. Here, we present a brief summary of methods from the original nested study,⁶ followed by a detailed description of the extended radiological analysis.

Initial eligibility evaluation was done at the emergency departments (Fig. 1), where computed tomography (CT) imaging, Glasgow Coma Scale (GCS), loss of consciousness (LOC), and presence and length of post-traumatic amnesia (PTA) were examined and documented by the attending physicians. Exclusion criteria were as follows: (1) persons who were nonworking age (age <18 years or >68 years) or who were not employed during the time of injury, (2) patients with more severe form of TBI than a mild one, defined by the World Health Organization,¹⁸ (3) persons with other native language than Finnish, (4) alcohol or drug dependence defined according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria, (5) patients diagnosed with visual or auditory disability, schizophrenia, or schizoaffective disorder, (6) contraindications for MRI, and (7) patients who underwent MRI later than 17 days after injury. The criteria 7 was based on a clinical expert opinion, the rationale being to capture potential time-bound signals. Persistence of symptoms following mTBI was not a requirement for participation in the study.

FIG. 1.

Flowchart of the study population. CT, computed tomography; MRI, magnetic resonance imaging; mTBI, mild traumatic brain injury; TMB, traumatic microbleed.

Consented participants were scheduled to undergo prompt baseline brain MRIs. One-month post-injury, patients were clinically reexamined, structurally interviewed, and health records examined at the study site. This resulted in a total of 113 patients with mTBI (Figs. 1 and 2), 22/113 (19%) of whom had at least one TMB on their baseline brain MRI. All patients were scheduled for follow-up brain imaging at approximately 3 months. A total of 25 patients dropped out from the radiological follow-up, and reasons for missingness are reported in Figure 1. For other variables, no missing data were detected. No significant tendencies for missingness were observed (Little’s missing completely at random test; p = 0.528).

FIG. 2.

TMB lesion count heatmap. Each column represents one patient. NA, not applicable; TMB, traumatic microbleed.

MRI data acquisition

A board-certified neuroradiologist systematically evaluated all MRI scans, and TMB presence and counts were recorded using Common Data Elements (CDE) for TBI neuroimaging¹⁹ (Fig. 2). The follow-up imaging was assessed by a neuroradiologist with knowledge of the baseline radiological findings. Imaging was conducted using a 3 T Siemens Magnetom Verio scanner equipped with a 32-channel head coil. The imaging protocol included a fast T1 gradient-echo localizer in three orthogonal directions (0.9 × 0.8 mm in-plane resolution, 8 mm slice, TR = 9 ms, TE = 4 ms, FA = 20°), T1 gradient-echo, sagittal localizer (1.0 × 0.8 mm in-plane resolution, 5 mm slices, 20 slices, 1.5 mm gap, TR = 250 ms, TE = 2.48 ms, FA = 90°), axial FLAIR (0.9 × 0.9 mm in-plane resolution, 4 mm slice, 35 slices, TR = 900 ms, TE = 91 ms, TI = 2500 ms, FA = 150°), coronal T2 (0.7 × 0.7 mm in-plane resolution, 4 mm slice, 40 slices, TR = 4000 ms, TE = 96 ms, FA = 150°), 3D T2 SPACE (1 mm isotropic voxel, TR = 3200 ms, TE = 416 ms), 3D T1 MPRAGE (1 mm isotropic voxel, TR = 1800 ms, RE = 2.47 ms, FA = 9°), and 3D gradient-echo SWI sequence (1.0 × 0.9 mm in-plane resolution, 1.8 mm slice resolution, TR = 27 ms, TE = 20 ms, FA = 15°). TMBs were identified as small hemorrhagic lesions in the white matter or gray–white interface, detectable via the SWI sequence. Nonhemorrhagic diffuse axonal injury lesions were not observed.

Statistics

The data were analyzed using R language and environment for statistical computing (R 4.4.2 for Windows; R Development Core Team, R Foundation for Statistical Computing, Vienna, Austria) and the Statistical Package for Social Sciences (SPSS version 29.0.0.0 for Windows, IBM Corp., Armonk, NY). The Wilcoxon signed-rank test was applied to test differences in the means or the ranks of the repeated measurements. Fisher’s exact test or the Monte Carlo method for chi-square test was used in multiple comparisons to estimate group differences in nominal variables, and the Mann–Whitney U test for continuous variables of non-normal distribution, respectively. All tests for significance were two-sided, with probabilities of <0.05 accepted as statistically significant.

Results

TMB persistence

Characteristics of the study population are presented in Table 2. A total of 88/113 patients with mTBI completed the full radiological follow-up. The median time from TBI to baseline MRI was 10 days (25th and 75th percentile: 7, 12), and the median time to follow-up imaging was 105 days (25th and 75th percentile: 85.25, 146.25). Patients with TMBs went to the baseline MRI slightly later than their counterparts (Table 2). Patients with TMBs displayed a higher prevalence of ICH (14% vs. 0%, p < 0.05, Table 2) compared to those without. There were no significant differences between the groups regarding the injury mechanism, sex, age, tobacco use, GCS, duration of LOC, or PTA (Table 2). Shared between these patients, a total of 129 TMB lesions were detected at baseline, all of which were visible at follow-up imaging (Table 3). Twenty-one out of 88 (24%) patients had at least one TMB (Table 2, Figs. 1 and 2): 8% (7/88) had one, 5% (4/88) had two, and 11% (10/88) had ≥3 TMB lesions. No new TMB lesions that were not detected at baseline were seen at follow-up imaging. Accordingly, the median sum of TMB lesions remained unchanged from baseline to 3 months (Wilcoxon signed-rank test, p = 1.000).

Table 2.

Characteristics of the Study Population

Variable	67 mTBI patients, no TMBs	21 mTBI patients, TMBs	Comparisons
Radiological follow-up
Median time (days) from trauma to baseline MRI (25th and 75th percentile)	10 (6.0, 11.5)	11 (9.0, 13.0)	p = 0.042^a (Mann–Whitney U: 910,500)
Median time (days) from trauma to follow-up MRI (25th and 75th percentile)	105 (86.5, 161.0)	94 (85, 123.0)	p = 0.493^a (Mann–Whitney U: 633,500)
Traumatic brain injury
Type of injury
Traumatic microbleeds
◦ Median (25th and 75th percentile) microbleed count at baseline	0	2 (1, 11)
◦ Median (25th and 75th percentile) microbleed count at follow-up	0	2 (1, 11)
Traumatic subarachnoid hemorrhage	8 (12%)	5 (24%)	p = 0.287^b
Intracerebral hemorrhage	0 (0%)	3 (14%)	p = 0.012 ^b
Brain contusion	6 (9%)	4 (19%)	p = 0.241^b
Subdural hematoma	11 (16%)	4 (19%)	p = 0.749^b
Epidural hematoma	1 (1%)	0 (0%)	p = 1.000^b
Loss of consciousness
Witnessed	49/67 (73%)	17/21 (81%)	p = 0.314^c
◦ No loss of consciousness	1	0
◦ ≤1 min	21	9
◦ 2–5 min	12	8
◦ 5–29 min	5	0
Clinical interview (witnesses, patient) indicates loss of consciousness, exact time cannot be determined	5/67 (7%)	2/21 (9%)
Loss of consciousness cannot be determined	13 (19%)	2/21 (9%)
Post-traumatic amnesia (yes)	64/67 (96%)	19/21 (90%)	p = 0.729^a (Mann–Whitney U: 735,500)
Median duration (25th and 75th percentile)	1 h 45 min (0 h 30 min, 4 h 15 min)	1 h 10 min (0 h 16 min, 2 h 15 min)
GCS measured by the emergency services	42/67 (62%)	10/21 (48%)	p = 0.773^c
GCS 15	29	8
GCS 14	11	2
GCS 13	1	0
Injury event			p = 0.489^c
Sports	20 (30%)	9 (43%)
Cycling	21 (31%)	4 (19%)
Falling	17 (25%)	5 (24%)
Motor vehicle	4 (6%)	3 (14%)
Other	4 (6%)	0 (0%)
Unknown	1 (1%)	0 (0%)
Injury mechanism			p = 0.093^c
Direct	64 (96%)	18 (86%)
Undirect	0 (0%)	1 (5%)
Combination (direct and undirect)	1 (1%)	2 (10%)
Unknown	2 (3%)	0 (0%)
Any neurosurgical procedure due to the TBI (yes)	0 (0%)	0 (0%)
Patient-related factors
Age, mean (SD)	40.3 (12.0)	37 (12.1)	p = 0.290^a (Mann–Whitney U: 595,500)
Sex (female %)	38 (56%)	15 (71%)	p = 0.309^b
Smoker	7 (10%)	3 (14%)	p = 0.814^b
Anticoagulation or antithrombotic medication	2 (3%)	0 (0%)	p = 0.293^b
Previous history of traumatic brain injury or an intracranial procedure	9 (13%)	0 (0%)	p = 0.107^b

Statistically significant (probabilities of <0.05) were bolded to enhance legibility.

Mann–Whitney U test.

Fisher’s exact test.

Monte Carlo method for chi-square test.

GCS, Glasgow Coma Scale; MRI, magnetic resonance imaging; mTBI, mild traumatic brain injury.

Table 3.

Traumatic Microbleed Counts by Region According to the Common Data Elements Traumatic Brain Injury Neuroimaging

Anatomical location	TMB count at baseline (% from total)	Right/Left	TMB count at follow-up (% from total)	Right/Left
Corpus callosum genu	0 (0%)	0/0	0 (0%)	0/0
Corpus callosum body	0 (0%)	0/0	0 (0%)	0/0
Corpus callosum splenium	7 (5.4%)	0/7	7 (5.4%)	0/7
Subcortical white matter frontal	77 (59.7%)	42/35	77 (59.7%)	42/35
Subcortical white matter parietal	8 (6.2%)	5/3	8 (6.2%)	5/3
Subcortical white matter temporal	33 (25.6%)	20/13	33 (25.6%)	20/13
Subcortical white matter occipital	1 (0.8%)	0/1	1 (0.8%)	0/1
Internal capsule anterior limb	0 (0%)	0/0	0 (0%)	0/0
Internal capsule posterior limb	0 (0%)	0/0	0 (0%)	0/0
Brainstem dorsolateral rostral	1 (0.8%)	0/1	1 (0.8%)	0/1
Brainstem other	2 (1.6%)	2/0	2 (1.6%)	2/0
Cerebellar peduncles	0 (0%)	0/0	0 (0%)	0/0
Other	0 (0%)	0/0	0 (0%)	0/0
Total	129 (100%)	69/60	129 (100%)	69/60

TBI, traumatic brain injury; TMB, traumatic microbleed.

Most common CDE-TBI regions for individual TMBs were frontal (77/129, 60%), temporal (33/129, 26%), and parietal (8/129, 6%) subcortical white matter (Table 3). For each patient, the TMB lesions were located in one region in 10% (9/88), two regions in 3% (3/88), three regions in 3% (3/88), and four or more regions in 7% (6/88) of the CDE-TBI regions.

TMBs relation to other radiological findings

In total, 25/88 (28%) patients also had other trauma-related intracranial abnormalities, most common being SDH (15/88, 17%) (Table 2). Patients with TMBs displayed a higher prevalence of ICH (14% vs. 0%, p < 0.05, Table 2) compared to those without. There were 14/88 (16%) patients with two or more different lesion types (such as tSAH + SDH), 11/88 (13%) with one, and 63/88 (70%) with none (Table 4). When grouped based on other trauma-related intracranial abnormalities, 10 out of 25 (40%) had at least one TMB, which was true only for 17% (11/63) for those without, a statistically significant difference (p < 0.05) (Fig. 3, Table 4). These patients were also older in comparison (p < 0.05) (Table 4). While some injury events (falls) were more prevalent in patients with trauma-related intracranial abnormalities, there were no significant differences between the two groups when all injury events were considered. There were no significant differences between the groups regarding sex, tobacco use, GCS, duration of LOC, or PTA (Table 4).

FIG. 3.

TMB lesion count heatmap grouped by other trauma−related intracranial abnormalities. Each column represents one patient. NA, not applicable; TBI, traumatic brain injury; TMB, traumatic microbleed.

Table 4.

Other Trauma-Related Intracranial Abnormalities

Variable	No additional radiological findings at baselinen = 63	At least one different trauma-related intracranial abnormality at baselinen = 25	Comparisons
Radiological follow-up
Median time (days) from trauma to baseline MRI (25th and 75th percentile)	9 (6.5, 11.5)	11 (9, 13)	p = 0.062^a (Mann–Whitney U: 586,500)
Median time (days) from trauma to follow-up MRI (25th and 75th percentile)	106 (87.5, 145.5)	94 (83,123)	p = 0.207^a (Mann–Whitney U: 924,000)
Traumatic brain injury
Type of injury
Traumatic microbleeds (yes)	11/63 (17%)	10/25 (40%)	p = 0.0495 ^b
◦ Median (25th and 75th percentile) microbleed count at baseline	0 (0, 0)	0 (0, 2)
◦ Median (25th and 75th percentile) microbleed count at follow-up	0 (0, 0)	0 (0, 2)
tSAH	0	13 (52%)
ICH	0	3 (12%)
CON	0	10 (40%)
SDH	0	15 (60%)
EDH	0	1 (4%)
Injury combinations
tSAH	0	3 (12%)
tSAH + ICH	0	2 (8%)
tSAH + SDH	0	3 (12%)
tSAH + CON	0	3 (12%)
tSAH + CON + SDH	0	2 (8%)
ICH	0	1 (4%)
CON	0	1 (4%)
CON + SDH	0	3 (12%)
CON + SDH + EDH	0	1 (4%)
SDH	0	6 (24%)
Loss of consciousness
Witnessed	51/63 (81%)	15/25 (60%)	p = 0.388^c
◦ No loss of consciousness	0	1
◦ ≤1 min	24	6
◦ 2–5 min	15	5
◦ 5–29 min	4	1
Clinical interview (witnesses, patient) indicates loss of consciousness, exact time cannot be determined	8 (6%)	3 (12%)
Loss of consciousness cannot be determined	8 (13%)	7 (28%)
Post-traumatic amnesia (yes)	61/63 (97%)	22/25 (88%)	p = 0.109^a (Mann–Whitney U: 857,000)
Median duration (25th and 75th percentile)	01 h 30 min (00 h 30 min, 04 h 00 min)	01 h 43 min (0h 20 min, 03 h 00 min)
GCS measured by the emergency services	39/63 (62%)	12/25 (48%)	p = 1.000^c
GCS 15	28	9
GCS 14	10	3
GCS 13	1	0
Injury event			p = 0.084^c
Sports	23 (37%)	6 (24%)
Cycling	21 (33%)	4 (16%)
Falling	12 (19%)	10 (40%)
Motor vehicle	4 (6%)	3 (12%)
Other	3 (5%)	1 (4%)
Unknown	0	1 (4%)
Injury mechanism
Direct	58 (92%)	24 (96%)	p = 0.701^c
Undirect	1 (2%)	0
Combination (direct and indirect)	3 (5%)	0
Unknown	0	1 (4%)
Any neurosurgical procedure due to the TBI (yes)	0	0
Patient-related factors
Age (median, 25th and 75th percentile)	36.0 (29.0, 49.0)	46.0 (36.5, 56.0)	p = 0.008^a (Mann–Whitney U: 502,500)
Sex (female %)	35 (56%)	18 (72%)	p = 0.227^a
Smoker	5 (8%)	5 (20%)	p = 0.139^a
Anticoagulation or antithrombotic medication	1 (2%)	1 (4%)	p = 0.490^a
Previous history of traumatic brain injury or an intracranial procedure	7 (11%)	2 (8%)	p = 1.000^a

Statistically significant (probabilities of <0.05) were bolded to enhance legibility.

Mann–Whitney U test.

Fisher’s exact test.

Monte Carlo method for chi-square test.

CON, brain contusions; EDH, epidural hematoma; GCS, Glasgow Coma Scale; ICH, intracerebral hemorrhage; MRI, magnetic resonance imaging; SD, standard deviation; SDH, subdural hematoma; tSAH, traumatic subarachnoid hemorrhage.

Discussion

Limitations and generalizability

A general population matched control group with radiological follow-up would have given insight into the radiological evolution of possible untraumatic/cerebral microbleed for comparison. The depth or the size of the lesions was not recorded. Higher field MRI scanners or diffusion tensor imaging could have captured changes not shown by the conventional MRI techniques. The potential differences in TMB persistence by brain region could not be answered by our study, as all TMBs persisted. We did not evaluate the changes in the appearances of the other intracranial abnormalities. A risk for selection bias exists in study recruitment, where mTBI patients with persisting symptoms might be more likely to participate in a study offering investigations that would not be offered until clinically required. Our results are generalizable to a generally healthy working-age population with a mild form of TBI.

Interpretation

Every sixth patient with mTBI portrays some trauma-related intracranial abnormalities imaged by MRI or CT.^1,20 In our study, these patients also had a higher prevalence of TMBs compared to those without other radiological abnormalities, possibly suggesting the amount of TMBs to be part of increased injury burden (or trauma strain) within the mTBI population. However, TMBs are relatively common finding in patients with mTBI caused by low-energy trauma.² In mTBI, TMBs are likely the result of traumatic vascular injury rather than axonal damage, as histopathological studies suggest they may occur in isolation from axonal injury.¹ White matter itself may be particularly vulnerable to trauma.²¹ Current study displayed the frontal and temporal lobes to display higher TMB burden in mTBI. Experimental TBI models suggest that TMB lesions (or their locations) may be predicted by simulating the biomechanical vessel strain.²² Further research is needed to understand whether structural, genetical, or biomechanical factors make the frontal and temporal lobes more prone to TMB formation or are these areas just more likely to be exposed to higher strains in TBI events.

In mTBI, the development—or visibility—of TMBs displayed by conventional MRI seems to be stable.^9,11,12,14 In our study, TMB lesions observed in MRI obtained <17 days after the initial trauma were still seen 3 months later. We have included a radiological example of a TMB stability from our cohort (Fig. 4). However, while the TMB lesions can persist for a year¹¹ or more,¹⁰ significantly delaying (over a year) the initial baseline MRI, less TMB lesions are expected.¹⁷

FIG. 4.

Radiological example of traumatic microbleed stability. The figure shows baseline and follow-up MRI susceptibility-weighted images from one patient with four traumatic microbleeds: three clustered in the frontal region and one located just behind them in the frontomesial area, all indicated by arrows. MRI, magnetic resonance imaging.

While this radiological stability helps in trauma detection and classification of patients with complicated mTBI,^3,5 the clinical implications of isolated TMBs in mTBI remain unclear.^2,3,6 In community-dwelling healthy middle-aged adults with a retrospectively evaluated history of TBI accompanied by cerebral microbleeds (or TMBs), poorer cognitive performance and clinical complaints were observed.²³ However, in a prospective setting, TMBs detected in patients with mTBI did not appear to delay return to work compared to those without TMBs.⁶

However, we know from mixed-severity TBI datasets that one can increase the number of TMBs seen in SWI or T2*-weighted gradient echo by increasing the MRI field strength, partly due to the ability to produce thinner slices and thus increasing sensitivity.^2,5,17 Despite the heightened sensitivity, at this moment, there is no evidence to suggest that going beyond the conventional MRI field strengths improves patient care or decision-making in mTBI.^5,17

Conclusions

Initial TMBs can still be reliably found with conventional MRI after 3 months in patients with mTBI: delayed imaging does not necessarily compromise the detection of these lesions. Other radiological abnormalities are linked with higher prevalence of TMBs, possibly suggesting increased injury burden within the mTBI population.

Transparency, Rigor, and Reproducibility Summary

The study protocol was reviewed and approved by the research ethics committee of Helsinki University Hospital (Study Code: 105/13/03/01/2014). The research was conducted in accordance with the Declaration of Helsinki, and all patients provided written informed consent. While the analysis plan was not preregistered, the study design, exclusion criteria, and outcome indicators were set prior to any data analysis. The study design was developed by investigators with decades of experience treating adult TBI patients in an outpatient setting.

This study was not formally registered because it is not a clinical trial. Being a prospective cohort study, no randomization was performed, and investigators were not blinded. Data were labeled using codes not linked to participant-identifying information. Data were collected between 2015 and 2018. De-identified data are not available in a public archive. No external validation studies have been performed or planned to be conducted. Statistical analysis was performed by A.J. Software (and equipment) used to perform the analysis is publicly available. The analytic and visualization code may be available by emailing the corresponding author.

Authors’ Contributions

A.J. and A.K. contributed to acquisition and analysis of data, conceptualization and design of the study, and drafting a significant portion of the article or figures. I.M. and S.M. contributed to methodology, supervision, design of the study, data acquisition and curation, review, and editing. A.H. and K.M. contributed to design of the study, data collection and analysis, review, and editing.

Footnotes

Acknowledgments

The authors thank Taina Nybo (Psych PhD) for insightful comments during the editing phase of the article.

Author Disclosure Statement

The authors have no financial or other conflicts of interest to declare.

Funding Information

This study was sponsored by the Helsinki University Hospital, State Research Funding (VTR Fund grant number: TYH2020227), and Maire Taponen Foundation.

References

Griffin

, Turtzo

, Parikh

, et al. Traumatic microbleeds suggest vascular injury and predict disability in traumatic brain injury. Brain, 2019; 142(11):3550–3564; doi: 10.1093/brain/awz290

Hageman

, Hof

, Nihom

. Susceptibility-weighted MRI and microbleeds in mild traumatic brain injury: Prediction of posttraumatic complaints? Eur Neurol, 2022; 85(3):177–185; doi: 10.1159/000521389

van der Horn

, de Haan

, Spikman

, et al. Clinical relevance of microhemorrhagic lesions in subacute mild traumatic brain injury. Brain Imaging Behav, 2018; 12(3):912–916; doi: 10.1007/s11682-017-9743-6

Moe

, Limandvik Myhr

, Moen

, et al. Association of cause of injury and traumatic axonal injury: A clinical MRI study of moderate and severe traumatic brain injury. J Neurosurg, 2020; 133(5):1559–1567; doi: 10.3171/2019.6.JNS191040

Moenninghoff

, Kraff

, Maderwald

, et al. Diffuse axonal injury at ultra-high field MRI. PLoS One, 2015; 10(3):e0122329; doi: 10.1371/journal.pone.0122329

Huovinen

, Marinkovic

, Isokuortti

, et al. Return to work after mild traumatic brain injury: Association with positive CT and MRI findings. Acta Neurochir (Wien), 2022; 164(7):1707–1717; doi: 10.1007/s00701-022-05244-4

Panenka

, Lange

, Bouix

, et al. Neuropsychological outcome and diffusion tensor imaging in complicated versus uncomplicated mild traumatic brain injury. PLoS One, 2015; 10(4):e0122746; doi: 10.1371/journal.pone.0122746

van der Eerden

, van den Heuvel

TLA

, Maas

, et al. The radiological interpretation of possible microbleeds after moderate or severe traumatic brain injury: A longitudinal study. Neuroradiology, 2022; 64(6):1145–1156; doi: 10.1007/s00234-021-02839-z

Nael

, Dagher

, Downs

, et al. Maximum AmbiGuity Distance for Phase Imaging in detection of traumatic cerebral microbleeds: An improvement over current imaging practice. AJNR Am J Neuroradiol, 2020; 41(11):2027–2033; doi: 10.3174/ajnr.A6774

10.

Rizk

, Turtzo

, Cota

, et al. Traumatic microbleeds persist for up to five years following traumatic brain injury despite resolution of other acute findings on MRI. Brain Inj, 2020; 34(6):773–781; doi: 10.1080/02699052.2020.1725835

11.

Einarsen

, Moen

, Håberg

, et al. Patients with mild traumatic brain injury recruited from both hospital and primary care settings: A controlled longitudinal magnetic resonance imaging study. J Neurotrauma, 2019; 36(22):3172–3182; doi: 10.1089/neu.2018.6360

12.

Studerus-Germann

, Gautschi

, Bontempi

, et al. Central nervous system microbleeds in the acute phase are associated with structural integrity by DTI one year after mild traumatic brain injury: A longitudinal study. Neurol Neurochir Pol, 2018; 52(6):710–719; doi: 10.1016/j.pjnns.2018.08.011

13.

Lawrence

, Pretorius

, Ezra

, et al. Early detection of cerebral microbleeds following traumatic brain injury using MRI in the hyper-acute phase. Neurosci Lett, 2017; 655:143–150; doi: 10.1016/j.neulet.2017.06.046

14.

Watanabe

, Maruya

, Kanemaru

, et al. Transient disappearance of microbleeds in the subacute period based on T2*-weighted gradient echo imaging in traumatic brain injury. Acta Neurochir (Wien), 2016; 158(7):1247–1250; doi: 10.1007/s00701-016-2805-5

15.

Toth

, Kovacs

, Tamas

, et al. Microbleeds may expand acutely after traumatic brain injury. Neurosci Lett, 2016; 617:207–212; doi: 10.1016/j.neulet.2016.02.028

16.

Haller

, Montandon

, Lazeyras

, et al. Radiologic-Histopathologic correlation of cerebral microbleeds using pre-mortem and post-mortem MRI. PLoS One, 2016; 11(12):e0167743; doi: 10.1371/journal.pone.0167743

17.

Scheid

, Ott

, Roth

, et al. Comparative magnetic resonance imaging at 1.5 and 3 Tesla for the evaluation of traumatic microbleeds. J Neurotrauma, 2007; 24(12):1811–1816; doi: 10.1089/neu.2007.0382

18.

Carroll

, Cassidy

, Holm

, et al.; WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. Methodological issues and research recommendations for mild traumatic brain injury: The WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury. J Rehabil Med, 2004(Suppl 43):113–125; doi: 10.1080/16501960410023877

19.

Haacke

, Duhaime

, Gean

, et al. Common data elements in radiologic imaging of traumatic brain injury. J Magn Reson Imaging, 2010; 32(3):516–543; doi: 10.1002/jmri.22259

20.

Isokuortti

, Iverson

, Silverberg

, et al. Characterizing the type and location of intracranial abnormalities in mild traumatic brain injury. J Neurosurg, 2018; 129(6):1588–1597; doi: 10.3171/2017.7.JNS17615

21.

Armstrong

, Sullivan

, Perl

, et al. White matter damage and degeneration in traumatic brain injury. Trends Neurosci, 2024; 47(9):677–692; doi: 10.1016/j.tins.2024.07.003

22.

Duckworth

, Azor

, Wischmann

, et al. A finite element model of cerebral vascular injury for predicting microbleeds location. Front Bioeng Biotechnol, 2022; 10:860112; doi: 10.3389/fbioe.2022.860112

23.

Low

, McKiernan

, Prats-Sedano

, et al.; PREVENT Dementia Investigators. Neuroimaging and clinical findings in healthy middle-aged adults with mild traumatic brain injury in the PREVENT dementia study. JAMA Netw Open, 2024; 7(8):e2426774; doi: 10.1001/jamanetworkopen.2024.26774