Microhemorrhages: Causes, Risks, and Their Meaning

Microhemorrhages are tiny bleeds in the brain, usually just a few millimeters across, that leave behind deposits of iron-containing hemosiderin visible on certain types of MRI. They are surprisingly common, especially in older adults, and their presence can signal anything from normal aging to serious vascular disease. What makes them clinically interesting is that their location in the brain, their number, and the conditions that produced them all carry different implications for future stroke risk, cognitive decline, and treatment decisions around blood-thinning medications.

What a Microhemorrhage Actually Is

When a tiny blood vessel in the brain leaks, red blood cells escape into the surrounding tissue. The body’s cleanup crew, mainly immune cells called macrophages, breaks down the hemoglobin from those cells and converts the iron into hemosiderin, a brownish storage form. These hemosiderin deposits are what radiologists see on specialized MRI sequences as small dark spots. A pathological study of 200 people aged 65 and older found hemosiderin deposits in the putamen (a deep brain structure) in 99% of cases, with significantly more deposits clustered around small arteries and arterioles than around capillaries.

1PubMed Central. Brain haemosiderin in older people: pathological evidence for an ischaemic origin of magnetic resonance imaging MRI microbleeds

The cellular details have been confirmed through autopsy studies using iron stains and immunohistochemistry, showing that iron gets trapped inside macrophages and pericytes (the cells that wrap around small blood vessels).

2PubMed Central. Cerebral microbleeds in the elderly: a pathological analysis

Dysfunction of the blood-brain barrier, the tightly sealed lining of brain blood vessels that normally keeps blood components out of brain tissue, is considered the initiating event. Risk factors like high blood pressure and aging can weaken this barrier’s structure and function, allowing red blood cells to leak through.

3PubMed Central. Dysfunction of the Blood-brain Barrier in Cerebral Microbleeds: from Bedside to Bench

How They Are Found

Microhemorrhages do not show up well on standard MRI. They require specialized sequences that are highly sensitive to iron’s magnetic properties. The two main techniques are gradient-recalled echo (GRE) imaging and susceptibility-weighted imaging (SWI). SWI is the more sensitive of the two because it incorporates additional magnetic information into the image, making tiny hemosiderin deposits stand out more clearly.

4PubMed Central. Detection of traumatic cerebral microbleeds by susceptibility-weighted image of MRI

The clinical difference between these techniques is substantial. In one study, the prevalence of microbleeds jumped from 23% when using GRE to 40% when using SWI, and the total count of detected microbleeds increased from 219 to 284.

5PubMed. Clinical relevance of improved microbleed detection by susceptibility-weighted magnetic resonance imaging

This means two patients scanned at different hospitals using different MRI protocols could receive very different microbleed counts, even if their actual burden is identical. SWI provides greater precision in identifying microbleeds compared to GRE, though modern automated detection algorithms can work across both modalities with comparable sensitivity.

6Scientific Reports. Automated detection of cerebral microbleeds on T2*-weighted MRI

How Common They Are

In the general population, microhemorrhages become more common with age, and they can appear even in people without any obvious vascular risk factors. A longitudinal study following healthy adults over several years found that about 10% developed at least one microbleed during the study period, with most new cases appearing in people who were 40 or older at the start.

7PubMed Central. Incident Risk and Progression of Cerebral Microbleeds in Healthy Adults: A Multi-Occasion Longitudinal Study

In elderly Japanese adults, roughly one in five had microbleeds on MRI, with men affected more than women (23% versus about 16%). The prevalence climbed steadily with age in both sexes.

8PubMed Central. Prevalence of and risk factors for cerebral microbleeds in a general Japanese elderly community

A separate study of healthy subjects without major cerebrovascular risk factors confirmed that microbleeds increased with aging and were linked to other markers of brain aging, such as white matter lesions and hippocampal shrinkage.

9PubMed. Age-related changes in white matter lesions, hippocampal atrophy, and cerebral microbleeds in healthy subjects without major cerebrovascular risk factors

So finding one or two microbleeds on a brain scan of someone over 60 is fairly ordinary. The clinical alarm bells start ringing when the number is high or when the location follows a pattern that suggests a specific disease process.

What Location Tells You About the Cause

One of the most useful things about microhemorrhages is that where they sit in the brain often points to why they formed. The two dominant causes split geographically in a remarkably clean way.

Microbleeds in deep brain structures (the basal ganglia, thalamus, brainstem, and cerebellum) are strongly associated with hypertensive small vessel disease. High blood pressure damages tiny penetrating arteries in these regions over time, making them prone to leaking. By contrast, microbleeds restricted to the lobar regions, the outer cortex of the brain, are the hallmark of cerebral amyloid angiopathy (CAA), a condition in which abnormal amyloid protein builds up in the walls of surface brain vessels.

10PubMed Central. Hypertension-Related Cerebral Microbleeds

CAA is a major cause of lobar brain hemorrhage and cognitive impairment in older adults, and it is closely tied to the same amyloid deposits seen in Alzheimer’s disease.

11PubMed Central. Cerebral amyloid angiopathy in the elderly

Research has confirmed that strictly lobar microbleeds in both the cerebral hemispheres and the cerebellum are related to amyloid angiopathy, while any combination of lobar and deep microbleeds together suggests hypertensive vascular damage.

12PubMed. Strictly Lobar Microbleeds Reflect Amyloid Angiopathy Regardless of Cerebral and Cerebellar Compartments

This location-based distinction matters because the two conditions carry different risks and require different management strategies. CAA-related microbleeds, for instance, raise particular concern about future lobar hemorrhage, while hypertensive microbleeds are more tightly linked to deep brain bleeds and certain patterns of cognitive decline.

Links to Dementia and Cognitive Decline

Having a handful of microbleeds does not necessarily mean your thinking skills will suffer. In a large population study of over 3,200 people followed for an average of about six years, having any single microbleed was not associated with cognitive decline. But having more than four microbleeds was a different story: those individuals performed significantly worse on tests of processing speed, word fluency, memory, and fine motor skills. Microbleeds were also associated with roughly double the risk of developing dementia, including Alzheimer’s disease.

13JAMA Neurology. Association of Cerebral Microbleeds With Cognitive Decline and Dementia

A meta-analysis that pooled results from seven prospective studies found that the risk of dementia climbed more steeply once someone had two or more microbleeds, with more than double the risk compared to people without them. The location mattered here too: deep and mixed-location microbleeds were significantly associated with dementia, while strictly lobar microbleeds alone were not.

14Brain Disorders. Association of cerebral microbleeds with risks of cognitive impairment and dementia: A systematic review and meta-analysis of prospective studies

In a memory clinic population, deep microbleeds were linked to executive dysfunction (trouble with planning, attention, and mental flexibility), though this connection faded after accounting for other markers of small vessel disease. Among people with mild cognitive impairment specifically, lobar microbleeds were independently tied to worse global cognition and memory.

15PubMed Central. Association between cerebral microbleeds and cognition in a memory clinic population

The takeaway is that microbleeds are markers of broader brain damage. They rarely cause cognitive problems on their own but tend to travel alongside other forms of vascular and neurodegenerative injury that do.

Stroke Risk and Blood-Thinning Medications

This is where microhemorrhages create the trickiest clinical dilemmas. A comprehensive meta-analysis found that microbleeds were associated with roughly double the risk of future ischemic stroke, but the relative increase in future brain hemorrhage risk was even greater, roughly quadrupled.

16PubMed Central. Clinical significance of cerebral microbleeds on MRI: A comprehensive meta-analysis of risk of intracerebral hemorrhage, ischemic stroke, mortality, and dementia in cohort studies

This creates a genuine bind for people who need blood thinners, especially those with atrial fibrillation who take anticoagulants to prevent clot-caused strokes. A meta-analysis of stroke patients with atrial fibrillation showed that the annual rate of brain hemorrhage was about 0.3% in people without microbleeds, jumped to about 0.8% in those with microbleeds, and rose to about 2.5% in those with five or more.

17PubMed. Brain microbleeds, anticoagulation, and hemorrhage risk: Meta-analysis in stroke patients with AF

A separate observational study found that the rate of symptomatic brain hemorrhage in anticoagulated patients with microbleeds was about 9.8 per 1,000 patient-years, compared to 2.6 per 1,000 patient-years in those without, and that adding microbleed status to standard bleeding risk scores significantly improved the prediction of who would bleed.

18The Lancet Neurology. Cerebral microbleeds and risk of symptomatic intracranial haemorrhage in patients with small-vessel disease and atrial fibrillation taking anticoagulants (CROMIS-2): a multicentre observational cohort study

Despite these numbers, current evidence does not support automatically stopping blood thinners when microbleeds are found. Studies have shown that antiplatelet therapy appears safe for ischemic stroke patients who have microbleeds, and that withholding treatment carries its own serious risks from clot-caused strokes.

19PubMed Central. Antiplatelet therapy may be safe in ischemic stroke patients with cerebral microbleed

Similarly, a study of patients with atrial fibrillation and microbleeds found that antithrombotic therapy did not significantly increase brain hemorrhage events or death, though patients with a history of high blood pressure or prior brain hemorrhage alongside their microbleeds warranted closer monitoring.

20PubMed Central. The safety of antithrombotic therapy in patients with cerebral microbleeds and cardiogenic cerebral embolism due to nonvalvular atrial fibrillation

Where the count gets very high, the calculus shifts. A retrospective study found no general association between the mere presence of microbleeds and symptomatic hemorrhage after acute stroke treatment, but patients with more than ten microbleeds had dramatically higher odds of bleeding complications.

21PubMed Central. Cerebral microbleeds shouldn’t dictate treatment of acute stroke: a retrospective cohort study evaluating risk of intracerebral hemorrhage

A pooled analysis of patients receiving clot-busting therapy for stroke confirmed that those with microbleeds had a higher risk of post-treatment hemorrhage and that increasing microbleed burden was associated with progressively higher risk of symptomatic bleeding.

22PubMed. Microbleeds, Cerebral Hemorrhage, and Functional Outcome After Stroke Thrombolysis

In practice, most clinicians treat the number of microbleeds as a dial rather than an on-off switch. A few microbleeds may be noted and monitored; a heavy burden, especially ten or more, prompts serious conversation about the risks and benefits of anticoagulation.

Less Common Causes

While hypertension and amyloid angiopathy account for the large majority of microhemorrhages, a wide range of other conditions can produce them. Traumatic brain injury is one of the most recognized, where microbleeds serve as an important imaging marker for diffuse axonal injury, the shearing of nerve fibers from mechanical forces.

23PubMed Central. Diffuse axonal injury after traumatic cerebral microbleeds: an evaluation of imaging techniques

However, the relationship is not one-to-one. A study examining traumatic microbleeds under the microscope found that about 64% were associated with axonal injury, meaning roughly a third occurred without nerve damage at the same spot.

24medRxiv. TRAUMATIC MICROHEMORHAGES ARE NOT SYNONYMOUS WITH AXONAL INJURY

Infective endocarditis, a bacterial infection of the heart valves, produces microbleeds at strikingly high rates. In one case-control study, 57% of endocarditis patients had microbleeds compared to 15% of matched controls, and the odds of having endocarditis climbed sharply with each additional microbleed found.

25PubMed. Cerebral microbleeds are frequent in infective endocarditis: a case-control study

A broader review of less common causes includes cerebral embolism, vasculitis (blood vessel inflammation), hemorrhagic metastases from cancer, radiation therapy to the brain, and rare inherited conditions.

26PubMed. MRI of cerebral microhemorrhages

Genetic Conditions That Cause Microbleeds in Younger People

When microhemorrhages appear in someone younger than expected, or in someone without high blood pressure, genetic small vessel diseases deserve consideration. Several inherited conditions are known to damage brain blood vessels and produce microbleeds. The best-studied is CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), caused by mutations in the NOTCH3 gene. In symptomatic CADASIL patients, about 31% had microbleeds on MRI, predominantly in the thalamus, and standard vascular risk factors like hypertension did not explain them.

27PubMed. Cerebral microbleeds in CADASIL

Other genetic conditions linked to cerebral small vessel disease include CARASIL (a recessive form), COL4A1 mutations that weaken vessel collagen, and Fabry disease, a storage disorder affecting blood vessel linings.

28PubMed Central. Genetics of cerebral small vessel disease

These are rare, but recognizing microbleeds as a possible clue can lead to a genetic diagnosis that changes management for the patient and their family members.

High Altitude and Environmental Triggers

One of the more surprising contexts for microhemorrhages is extreme altitude. People who develop high-altitude cerebral edema (HACE), a dangerous swelling of the brain that occurs at very high elevations, often have microbleeds that persist long after recovery. MRI scans of HACE survivors have revealed hemosiderin deposits concentrated in the corpus callosum, the thick band of fibers connecting the brain’s two hemispheres. These deposits can be detected months after the episode and may serve as a diagnostic footprint of past HACE even when the person has recovered clinically.

29PubMed. Microhemorrhages in nonfatal high-altitude cerebral edema

Using higher-strength MRI magnets (3 Tesla rather than 1.5 Tesla), researchers found that the microbleeds in HACE were more extensive than previously realized, extending beyond the areas of visible swelling and persisting over time, eventually coalescing into larger deposits.

30American Journal of Neuroradiology. Acute and Evolving MRI of High-Altitude Cerebral Edema: Microbleeds, Edema, and Pathophysiology

Whether these persistent microbleeds contribute to any lasting cognitive effects in HACE survivors is an open question and an area of active research.

Telling Microbleeds Apart from Mimics

Not every dark dot on an MRI is a microhemorrhage, and the false-positive rate is not trivial. Studies estimate that somewhere between 11% and 24% of suspected microbleeds turn out to be mimics.

31PubMed. Cerebral Microbleeds: Imaging and Clinical Significance

The most common imposters include small calcium deposits, cross-sections of blood vessels, and tiny cavernous malformations (tangles of abnormal small vessels). Calcium deposits can look identical to hemosiderin on standard sequences, but the two can be distinguished using phase images from SWI, because calcium and iron distort the magnetic signal in opposite directions.

32Journal of Neurosonology and Neuroimaging. Cerebral Microbleeds: Incidence, Imaging Characteristics, Common and Uncommon Causes

Blood vessels caught in cross-section can be told apart from microbleeds by tracing them across multiple image slices, where they reveal their tubular, continuous structure rather than ending as an isolated dot. Cavernous malformations tend to show a characteristic ring of hemosiderin with stagnant blood at various stages of breakdown visible on other MRI sequences, making them distinguishable from a simple microbleed if the radiologist looks carefully.

33PubMed Central. Cerebral Microbleeds: A Field Guide to their Detection and Interpretation

Bone artifacts near the base of the skull and air in the sinuses add another layer of difficulty, sometimes obscuring real microbleeds in the temporal and frontal lobes or producing dark spots that look like them.

Automated Detection and the Future of Counting

Manual counting of microbleeds is tedious, time-consuming, and unreliable. Two trained radiologists can disagree substantially on how many microbleeds a scan contains, especially when the count is high or the scan quality is imperfect. This has pushed researchers toward automated detection using deep learning.

Several algorithms now exist that combine object-detection neural networks with three-dimensional analysis to first flag potential microbleed candidates and then filter out false positives.

34PubMed Central. Automated detection of cerebral microbleeds in MR images: A two-stage deep learning approach

One validated algorithm achieved over 92% sensitivity in detecting microbleeds in a community-based cohort, with fewer than 10 false positives per scan on SWI images. It also outperformed human raters at catching new microbleeds on follow-up scans, correctly identifying 50% of new longitudinal microbleeds while manual raters caught 30% or less, and reduced visual rating time by 43%.

6Scientific Reports. Automated detection of cerebral microbleeds on T2*-weighted MRI

Newer approaches use knowledge distillation techniques that work across different MRI modalities, meaning the same algorithm can handle GRE, SWI, or quantitative susceptibility mapping inputs without retraining.

35Frontiers in Neuroinformatics. Automated detection of cerebral microbleeds on MR images using knowledge distillation framework

As these tools become more standardized, microbleed counts should become more reproducible across hospitals and research centers, which would help clarify the thresholds at which a given count changes a patient’s risk profile or treatment plan. Right now, there is no universally agreed-upon cutoff. The field has settled loosely on groupings of “few” versus “many,” but the specific numbers (more than 4, more than 5, more than 10) vary across studies, partly because counting methods do too.