2026 年 13 巻 p. 477-483
Iatrogenic cerebral amyloid angiopathy is a rare condition thought to result from prion-like transmission of amyloid-β, typically after childhood neurosurgery or exposure to cadaveric dura mater, with a latency period of several decades. We report the case of a 60-year-old woman who underwent a right frontal craniotomy for traumatic head injury at the age of 12 years, although it was unknown whether cadaveric dura mater was used. Approximately 48 years later, she developed a right frontal lobar hemorrhage near the prior craniotomy site and experienced more than 15 recurrent intracerebral hemorrhages, predominantly involving lobar regions over 3 months. Histopathology of hematoma-adjacent tissue revealed marked amyloid-β deposition without significant tau pathology or inflammation, and apolipoprotein E genotype was ε3/ε3. Based on clinical, pathological, and epidemiological findings, the case met the Queen Square criteria for possible iatrogenic cerebral amyloid angiopathy. The exceptionally long latency and unusually frequent recurrent hemorrhages were the most striking features of this case. Iatrogenic cerebral amyloid angiopathy should be considered even in older adults, as remote neurosurgical history remains an important diagnostic clue. Because amyloid-β is resistant to conventional sterilization, special attention to the sterilization of surgical instruments is also warranted when patients with a remote neurosurgical history undergo further neurosurgery.
Cerebral amyloid angiopathy (CAA) is classified according to the type of amyloid protein deposited, with amyloid-β being the most representative subtype.1) Amyloid-β predominantly accumulates in the walls of small- to medium-sized arteries and veins in the cerebral cortex and leptomeninges, leading to vascular injury such as intracerebral hemorrhage.1)
CAA caused by amyloid-β deposition is clinically divided into 3 forms. The most common is sporadic CAA, which is associated with aging and frequently coexists with Alzheimer's disease. In contrast, hereditary CAA often presents with a family history and is linked to specific genetic mutations. More recently, iatrogenic CAA, resulting from exogenous transmission of amyloid-β, has gained attention.2,3)
We encountered an older patient who experienced very frequent recurrent lobar hemorrhages over a short period. Given the high likelihood that childhood neurosurgical intervention contributed to disease onset, we considered this case to have important implications for future clinical practice. We also reviewed previously reported cases of iatrogenic CAA following neurosurgical procedures identified through PubMed and compared them with our case.
A 60-year-old woman with no family history of dementia had undergone a right frontal craniotomy at approximately 12 years of age for a traumatic head injury caused by a baseball bat. Information regarding the use of cadaveric human dura mater during that surgery was unavailable. The patient had no prior history of hypertension or malignant neoplasm and was not taking any antiplatelet or anticoagulant medications. Coagulation studies were within normal limits.
She experienced sudden-onset headache and visited a local clinic, where mild left hemiparesis was noted. Computed tomography revealed a right frontal lobar hemorrhage, and she was transferred emergently to our hospital (Figure 1A). The hemorrhage was located very close to the previous childhood craniotomy site.
At initial presentation, computed tomography (A) demonstrated a right frontal lobar hemorrhage located directly beneath the posterior margin of the childhood craniotomy site. On day 3, susceptibility-weighted images (B, C) showed hypointense lesions in the medial right parietal lobe and posterior left temporal lobe, consistent with possible cortical microhemorrhages.
On day 2, she developed a decreased level of consciousness. Cerebral angiography showed no vascular abnormalities. Emergency craniotomy and hematoma evacuation were performed, during which a small sample of damaged perihematomal cortex was obtained for pathological examination. On day 3, magnetic resonance imaging with susceptibility-weighted imaging demonstrated additional hypointense lesions suspicious for lobar hemorrhage in the medial right parietal lobe and posterior left temporal lobe (Figure 1B and C).
During the subsequent 3 months, she experienced more than 15 recurrent intracerebral hemorrhages (Figure 2). Most hemorrhages were lobar, although some involved the thalamus and brainstem. Four months after the initial hemorrhage, she remained in a state of severe consciousness disturbance and persistent quadriparesis.
Representative serial computed tomography images during the clinical course. Multiple intracerebral hemorrhages recurred over a 3-month period. Most hemorrhages were lobar in distribution.
The histopathological findings are summarised in Figure 3. Histopathological examination of the surgical specimen revealed amyloid-β deposition in leptomeningeal and cortical small arteries, with adventitia-predominant involvement. The vascular lumina remained patent. In the vicinity of the hematoma, arterial walls showed multilayering of the internal elastic lamina, a double-barrel appearance, atrophy and loss of smooth muscle cells, and severe amyloid-β deposition extending from the adventitia to the media, without evidence of vasculitis (Vonsattel grade 3 features).
Histopathological features of the cerebral cortex and hematoma in a case of suspected iatrogenic cerebral amyloid angiopathy. (A) Elastica-Masson stain, low-magnification overview (scale bar: 500 μm). (B-1 to B-4) Amyloid-β immunohistochemistry: (B-1) high-density diffuse plaques distributed from the cortical middle to superficial layers (low magnification); (B-2) adventitia-predominant amyloid-β deposition in a leptomeningeal/cortical small artery with extension to superficial cortical layers; (B-3) sparse dense-core plaques in the deep cortical layers; (B-4) perivascular amyloid-β deposition (dyshoric changes) in the subcortical white matter (scale bars: 500 μm for B-1; 100 μm for B-2, B-3, B-4). (C) Elastica-Masson stain of a cortical artery (scale bar: 100 μm). (D) Phosphorylated tau (p-tau) immunohistochemistry showing a single p-tau-positive neuron with sparse neuropil threads (scale bar: 20 μm). (E) Elastica-Masson stain of an artery near the hematoma demonstrating multilayering of the internal elastic lamina, double-barrel appearance, and smooth muscle atrophy (scale bar: 50 μm). (F) Amyloid-β immunohistochemistry of the same vessel as in (E) showing severe amyloid-β deposition from adventitia to media (scale bar: 50 μm). (G) Elastica-Masson stain of the hematoma interior revealing multiple bleeding foci with concentric (onion-skin-like) fibrin deposits (scale bar: 100 μm). (H) Amyloid-β immunohistochemistry within the hematoma showing amyloid-β-deposited arterioles adjacent to bleeding points, presumed ruptured vessels (scale bar: 100 μm).
Parenchymal amyloid-β pathology showed high-density diffuse plaques distributed from the cortical middle layers (layers III-IV) to the superficial layers (layers I-II), with slight predominance in the middle layers, and sparse dense-core plaques in the deep layers (layers V-VI). Perivascular amyloid-β deposition (dyshoric changes) was noted in the subcortical white matter immediately beneath the cortex. Phosphorylated tau (p-tau) pathology was minimal, with only a single p-tau-positive neuron identified across the entire specimen and sparse neuropil threads in the surrounding area.
Within the hematoma, multiple bleeding foci were present, accompanied by concentric onion-skin-like fibrin deposits. In the vicinity of these bleeding points, arterioles with amyloid-β deposition were observed and presumed to represent sites of rupture.
Apolipoprotein E (ApoE) genotyping performed using the polymerase chain reaction-invader method demonstrated an ε3/ε3 genotype.
This case report was approved by the Institutional Review Board of Dokkyo Medical University (approval no. 161). Written informed consent for publication was obtained from the patient's family because the patient had severe disturbance of consciousness.
A literature search was performed using the PubMed database for articles published up to March 31, 2026. The following search string was applied:
("Cerebral Amyloid Angiopathy" [MeSH] OR "cerebral amyloid angiopathy" OR "amyloid angiopathy") AND ("Neurosurgical Procedures" [MeSH] OR "neurosurgery" OR "brain surgery" OR "craniotomy") AND ("Postoperative Complications" [MeSH] OR "iatrogenic" OR "surgery-induced") AND (1900/01/01:2026/03/31[dp])
Iatrogenic CAA was first recognized when extensive amyloid-β deposition and severe CAA were identified in young adults who had received cadaver-derived human pituitary growth hormone during childhood and later died of iatrogenic Creutzfeldt–Jakob disease.4) Subsequent reports described cases of individuals developing CAA many years after exposure to cadaveric dura mater. These findings established that not only prions but also amyloid-β can be transmitted.
Diagnostic criteria for iatrogenic CAA, known as the Queen Square diagnostic criteria, were proposed by Banerjee et al.5) They classify iatrogenic CAA into "probable" and "possible" categories based on age at onset, exposure history, clinical and radiological features, evidence of amyloid-β accumulation, and exclusion of hereditary amyloid-β disorders. Confirmation of amyloid-β deposition and genetic testing are required for the diagnosis of probable iatrogenic CAA.
In the present case, the exposure history included childhood neurosurgery performed around 1975, which coincided with the period in Japan when cadaveric dura mater was widely used. Histopathology confirmed amyloid-β deposition. Although the family history of dementia was negative, hereditary disorders could not be fully excluded. Therefore, this case fulfilled the criteria for possible iatrogenic CAA according to the Queen Square diagnostic criteria.
Clinical and radiological featuresThe clinical course in the present case was marked by exceptionally frequent hemorrhagic recurrences, with more than 15 intracerebral hemorrhages occurring within 3 months. Although previous reports have also described frequent hemorrhagic recurrences in iatrogenic CAA, the number of episodes observed in our patient was particularly remarkable.6,7)
We collected 74 cases of iatrogenic CAA related to neurosurgical procedures (Table 1, Supplementary Table 1). The age at diagnosis ranged from 29 to 84 years, with a median of 42 years. The latency period ranged from 23 to 52 years, with a median of 36 years. Thirty-three cases (44.6%) involved exposure to cadaveric dura mater. The vast majority (61 cases; 82.4%) presented with intracerebral hemorrhage, and lobar hemorrhage was observed in 58 cases (78.4%). The most frequent ApoE genotype was ε3/ε3 (24/43 cases; 55.8%), and ε4/ε4 was not reported. Among the 74 cases, the first hemorrhage was ipsilateral in 38 (51.4%), contralateral in 9 (12.2%), distant in 16 (21.6%), and unclear or absent in 11 (14.9%).
Summary of Previously Reported Cases of Iatrogenic Cerebral Amyloid Angiopathy Associated with Neurosurgical Procedures
| n = 74 | ||
|---|---|---|
| Distant refers to hemorrhage occurring across the tentorium (supratentorial vs infratentorial) or between the cerebrum and spinal cord. N/A: not available | ||
| Age at diagnosis | 29-84 (median 42) | |
| Sex | ||
| man | 51 (68.9%) | |
| woman | 20 (27.0%) | |
| unknown | 3 (4.1%) | |
| Latency period (year) | 23-52 (median 36.0) | |
| The underlying disease | ||
| Trauma | 32 (43.2%) | |
| Usage of cadaveric dura mater | ||
| yes | 33 (44.6%) | |
| no | 8 (10.8%) | |
| N/A | 33 (44.6%) | |
| Mode of onset | ||
| Intracerebral hemorrhage | 61 (82.4%) | |
| Other | 13 (17.6%) | |
| Lobar type hemorrhage | 58 (78.4%) | |
| Familial history of dementia | ||
| Present | 2 (2.7%) | |
| Absent | 38 (51.4%) | |
| Unknown | 34 (45.9%) | |
| Apolipoprotein E genotype | [n = 43] | |
| ε2/ε3 | 15 (34.9%) | |
| ε3/ε3 | 24 (55.8%) | |
| ε2/ε4 or ε3/ε4 | 4 (9.3%) | |
| N/A | 31 | |
| Relationship between initial surgical site and first haemorrhage site | ||
| Ipsilateral | 38 (51.4%) | |
| Contralateral | 9 (12.2%) | |
| Distant | 16 (21.6%) | |
| Unclear/No hemorrhage | 11 (14.9%) | |
Previous studies have demonstrated a spatial relationship between the site of prior neurosurgery and the first hemorrhage in iatrogenic CAA.8) In their series, 69% of initial hemorrhages occurred on the same side as the original operation, suggesting that iatrogenic CAA may originate near the surgical site. In our review, ipsilateral hemorrhage was more than 4 times as common as contralateral hemorrhage (Table 1). In the present case, the first hemorrhage occurred extremely close to the previous craniotomy site, supporting the possibility of amyloid-β propagation.
Neuropathological featuresPathologically, iatrogenic CAA is characterized by amyloid-β deposition in the vessel walls of the leptomeninges and cortex, with relatively mild tau pathology.5) Although tau deposition within neurons has been described in some cases, it typically does not meet the criteria for Alzheimer's disease (Table 1). This may reflect the slower development of tau pathology following amyloid-β seeding compared with the earlier vascular amyloid deposition.9) A recent systematic review of iatrogenic CAA cases with neuropathological data similarly reported that concomitant parenchymal amyloid-β deposition is frequent, whereas tau pathology is absent or minimal in most cases in which it was examined, in contrast to the substantial tau burden seen in moderate-to-severe sporadic CAA.10) In our case, the discrepancy between pronounced vascular amyloid deposition and minimal neuronal tau accumulation was consistent with iatrogenic CAA.
Conversely, when an inflammatory response predominates in vessels with amyloid-β deposition, the condition is termed CAA-related inflammation. Although historically regarded as a distinct subtype, CAA-related inflammation is now considered part of a spectrum of pathological processes shared across multiple forms of CAA.11-13) Recent reports indicate that iatrogenic CAA may also be accompanied by CAA-related inflammation.13) A recent case report further described an inflammatory phenotype of iatrogenic CAA with leptomeningeal and arterial wall enhancement and a favorable response to immunosuppressive therapy, indicating that inflammation may be a treatable contributor in selected patients.14) No pathological features suggestive of CAA-related inflammation were present in our patient.
ApoE genotyping is a useful adjunct in differentiating causes of CAA.5,15) Most reported cases of iatrogenic CAA have not shown the ε4/ε4 genotype typically associated with sporadic CAA; instead, ε3/ε3 is the most common (Table 1). The absence of a family history of dementia, the marked amyloid-β deposition despite the relatively young age of 60 years, and the ε3/ε3 genotype in this patient supported the diagnosis of iatrogenic CAA.
Age at onset and latencyAlthough this patient was 60 years old, which is slightly older than the classical age range for iatrogenic CAA, this does not preclude the diagnosis. Indeed, cases of iatrogenic CAA in older adults have been reported,16,17) and a multicenter European study suggested that the conventional age threshold of <55 years may require reconsideration.15) This case highlights the need to consider iatrogenic CAA even in older individuals.
In Japan, cadaveric dura mater was used between 1973 and 1997, and several cases of iatrogenic CAA have also been reported domestically.18,19) The present patient underwent initial surgery during this period, making exposure possible, although definitive evidence was unavailable. Assuming a latency of 40 years, cases may continue to occur until around 2037, and considering latencies approaching 50 years, iatrogenic CAA could potentially appear until approximately 2047.
Amyloid-β is resistant to standard sterilization methods, including those involving formaldehyde, raising the possibility of transmission via contaminated neurosurgical instruments. Indeed, iatrogenic CAA has been reported in patients without exposure to cadaveric dura mater but with prior neurosurgery.15,20) A recent epidemiological analysis reported an increased risk of spontaneous intracerebral hemorrhage in recipients of red blood cell transfusions from donors who subsequently developed multiple spontaneous intracerebral hemorrhages, suggesting that even transfusion-mediated transmission of amyloid-β may be possible, although a causal relationship has not yet been established.21) Decontamination of surgical instruments likely requires prion-level sterilization, including high-temperature alkaline treatment followed by autoclaving.22) For patients with a history of neurosurgery during the era of cadaveric dura use, special attention to sterilization procedures is warranted when considering re-operation.
We would like to express our sincere appreciation to the pathology laboratory technologists for their meticulous work in tissue processing and histological preparation, which greatly contributed to the accuracy of our pathological evaluations.
All authors have no conflict of interest. Among the authors, Takeo Uzuka, Takuma Sumi, Shunsuke Shibao, Miki Sakuma, Yoshihiro Abe, and Hiroyoshi Akutsu, who are members of the Japan Neurosurgical Society, have submitted the self-reported Conflicts of Interest Disclosure Statement Forms online.
Artificial intelligence-assisted tools were used to assist with language editing and pre-submission quality checking. The authors reviewed and verified all content and take full responsibility for the accuracy, integrity, and originality of the manuscript.
