Introduction

Among all subtypes of stroke, subarachnoid hemorrhage (SAH) resulting from a ruptured cerebral aneurysm (RCA) is the most serious. Despite recent diagnostic, medical, and surgical advances, the prognosis of aneurysmal SAH remains poor [34, 36, 39, 48, 49, 52, 53, 62, 68]. Previously, autopsy was an important tool for investigating cerebral aneurysms (CAs). However, advances in imaging technologies such as magnetic resonance angiography (MRA), computed tomography angiography (CTA), and digital subtraction angiography (DSA) have significantly reduced the opportunity for autopsy. Nevertheless, some of the key data from autopsy studies are still invaluable and should be kept in mind. The present review of previous autopsy studies was conducted to reaffirm the important knowledge that can be gained from them. The prevalence of CAs in autopsy studies was investigated in terms of age, gender and aneurysm site, size, and multiplicity and compared with clinical data for CAs. One of the main purposes was to provide clues as to the natural course of CA prevalence, i.e., from development of an unruptured cerebral aneurysm (UCA) to enlargement and finally to CA rupture.

Materials and methods

A comprehensive literature review of CAs investigated in autopsy studies was performed using standard search engines, including PubMed, Web of Science, and Cochrane Library. The search terms “autopsy”, “cerebral aneurysm,” “formation,” “prevalence,” and “subarachnoid hemorrhage” and their corresponding synonyms were combined with the “AND” operator. References were reviewed to identify further articles for inclusion. There were no limits with regard to dates. In this study, autopsy data for UCAs and RCAs were compared with clinical data obtained mainly from hospital-based and community-based studies (CBSs), respectively. Hospitalized patients with CAs are not truly representative of patients in the community because of inevitable case selection, while CBSs include all incident cases of disease, regardless of the patient’s age or clinical condition. With regard to CBSs, if a large geographic area is studied, a number of patients, especially those in outlying areas, may be missed [47, 48, 82]. In such populations, many patients in poor condition, especially the elderly, may not be picked up, with the consequence that the observed incidence might be considerably lower than the actual incidence. The best understanding of CAs can be obtained from CBSs of small geographical areas that include all incident cases of disease, regardless of the age and clinical condition of the patients. Among CBSs cited in this study, the geographic areas covered by 13 of them were small with populations of less than 100,000 [34, 36,37,38,39,40, 42, 43, 47,48,49, 53, 64], and 5 covered large areas with populations of over 100,000 [41, 52, 55, 62, 68]. The Izumo Study was conducted in a small city in Japan with an area of approximately 175 km2 and a population of 80,000 [34, 36,37,38,39,40, 42, 43, 47,48,49].

Results

Aneurysm formation and rupture

The pathogenesis of CAs is unclear and has been debated for many years. In previous autopsy studies, two basic theories concerning the etiology of aneurysm formation have been discussed: the congenital or medial defect theory [20, 26], and the acquired degenerative theory [13, 24, 72, 73]. In 1859, Gul [26] stated that aneurysms were due to a congenital developmental error or an inborn mural defect. This congenital theory became widespread. In 1930, Forbus [20] first described defects in the media of cerebral arteries. He assumed that such defects were congenital and that aneurysms arose at such sites. However, the tunica media defect theory has since been refuted by many studies. In 1940, Glynn [24] considered that tunica media defects have doubtful etiological significance in the formation of CAs, and that instead, degenerative changes in the internal elastic lamina play an essential role. It is well known that tunica media defects are prevalent in the posterior part of the normal cerebral circulation system and at the bifurcations of external and internal cranial arteries in different animals, whereas aneurysms are much less common in those areas [16, 70]. In 1963, Hassler [28] suggested that gaps in the media — or so-called medial defects — were important for aneurysm formation, whereas gaps were frequently observed even in the normal bifurcations of cerebral arteries [70]. In 1963, Stehbens [72] reported that histological changes in the aneurysm wall were similar to those in the arterial wall, and that the architectural differences between the arterial wall and the sac wall could be explained by the enormous wall expansion that occurs during aneurysm formation. On this basis, Stehbens [72, 73] concluded that CAs were acquired degenerative lesions induced by the effect of hemodynamic stress. Some authors believed that aneurysm formation could involve a combination of both congenital and acquired factors [7, 13, 80]. In 1950, Carmichael [7] theorized that defects of both the tunica media and the internal elastic lamina must be present before an aneurysm forms. Tunica media defects were usually congenital whereas internal elastic lamina defects might be due to atheroma formation, although both could occur independently. In 1954, Walker and Allegre [80] reported that atheromatous lesions were found at the sites of CAs, and that such degenerative vascular changes precipitated vascular dilatation. In 1959, Crawford [13] described three main factors that contributed to the pathogenesis of CAs; developmental faults in the media, atherosclerosis, and hypertension were considered to play roles of varying importance according to the age at which an aneurysm developed. In 1966, Crompton [15, 16] suggested that infiltration of leukocytes and fibrin into the aneurysm wall preceded expansion or rupture of the aneurysm, and that this was due to endothelial damage arising through pulsation, turbulent flow, or ischemia as a result of mural thrombosis. In 1985, an autopsy study by Anim [2] showed that hypertension increased the formation and development of CAs, whereas severe atherosclerosis of cerebral vessels was not the main pathogenetic factor in CA formation. In 1994, Kosierkiewiz [54] performed immunocytochemical studies of human saccular aneurysms and concluded that progression of atherosclerosis within the aneurysmal sac was correlated positively with aneurysm growth and might contribute to rupture. In 1999, Chyatte [11] stated that extensive inflammatory and immunological reactions were common in UCAs and might be related to aneurysm formation and rupture. These aforementioned hypotheses were based primarily on pathological findings in human autopsy cases or during surgery.

Recent clinical, experimental, and animal studies have suggested that disruption of the internal elastic lamina is the most important factor in the creation of CAs, and that hemodynamic stress and hypertension may be auxiliary factors [22, 86].

In a postmortem series of 163 RCAs by Crawford [13] in 1959, the rupture occurred at the fundus or apex of the sac in 105 cases (64%), the lateral segment in 17 (10%), the neck or proximal third of the sac in 3 (2%), and was undetermined in 38 (24%). In 1966, Crompton [15] examined the brains of 289 patients who had died of RCA, and among 271 of them, the rupture had occurred through the apex in 227 (84%), the body in 38 (14%), and the neck in only 6 (2%). When an aneurysm ruptures, hemorrhage is arrested by thrombus formation within the sac [35, 72]. The critical size for rupture of UCAs based on autopsy studies has been estimated to be 4 mm [15, 75], 5 mm [59, 71], and 6–15 mm [13]. In any event, the critical size may differ according to patient age and aneurysm site.

Prevalence of cerebral aneurysms at autopsy

Table 1 shows the percentages of brains with aneurysms based on the number of brain examinations and the percentages of brain examinations relative to total autopsy series. In 1918, Turnbull [76] found 42 brains (0.9%) having 48 aneurysms among 4547 patients subjected to head opening. The reported prevalence of UCAs varied from 0.3 to 4.0% of total brain examinations across all age groups [3, 10, 17, 45, 58]. In a planned study of 2786 consecutive autopsies across all age groups by Chason and Hindman [10] in 1958, 137 patients (4.9%) were found to have berry aneurysms, comprising 80 (2.9%) UCAs and 57 (2.0%) RCAs. According to Berry et al. [3] in 1966, incidental asymptomatic UCAs were found in 39 (1.0%) of 3871 brains examined among a series of 6686 autopsies across all age groups. In a prospective and consecutive autopsy series performed by McCormick [58] in 1973, patients with UCAs accounted for 4.0% (114) of 2831 patients of all ages, 5.0% (114) of 2276 patients aged ≥ 1 year, and 5.9% (114) of 1945 patients aged ≥ 20 years, but no aneurysms were found in 331 patients aged 1–19 years. The reported prevalence of CAs including UCAs and RCAs has varied from 1.3 to 7.6% of total brain examinations across all age groups [3, 10, 17, 29, 45, 46, 51, 58, 71], the highest reported prevalences being 4.6% [51], 4.9% [10], 6.5% [58], and 7.6% [71]. One important aspect of CA prevalence was that women were affected more frequently than men. This trend was noted in patients with UCAs [45], RCAs [17, 25, 46], and CAs including UCAs and RCAs [10, 45, 46, 51]. The only exception was a more recent forensic autopsy study by Ronkainen [69] in 1998, in which the prevalence of UCAs in men relative to women was 27:2. The autopsy rate for men was four times higher than that for women. One explanation is that forensic autopsy material differs from the general population in having a male preponderance and a higher age, the number of autopsies increasing in older age groups.

Table 1 Percentages of brains with cerebral aneurysms based on the number of brain examinations
Full size table

In clinical studies, the reported prevalence of UCAs, mostly diagnosed by conventional imaging techniques, was estimated to be 1–2%, with higher detection rates for women [1, 78, 84]. The prevalence of UCAs also varies according to the assessment modality employed, i.e., MRA, CTA, or DSA. Two recent studies using 3-T MRA have reported higher prevalences of 4.3% [31] and 8.4% [30] than those reported previously. This can be explained by an increased detection rate of UCAs smaller than 3 mm due to the higher resolution of the 3-T MRA system. A systematic review of the literature on autopsy and angiography studies showed that the prevalence of UCAs in adults of the general population was approximately 2–3% [67, 79]. In a review analysis based on 51 CBSs, the overall incidence of SAH was approximately 9 per 100,000 people per year [18]. Some CBSs reported that the annual incidence of SAH was higher in men, but most other studies reported the incidence to be higher in women [36, 49, 53, 55, 64]. The incidence of aneurysmal SAH was higher in Japan than in Western countries [18, 34, 36, 47, 49, 53, 55, 64]. Nevertheless, a recent review analysis showed that the prevalence of UCAs did not differ between Japan and other Western countries [79]. In the Izumo Study [36], the crude and age- and sex-adjusted incidence rates of aneurysmal SAH were both 23 per 100,000 people per year for all ages. Furthermore, when including in the incidence calculations patients whose death certificates indicated that they might have died of aneurysmal SAH, but who had not been formally examined, the rates increased to 32 and 29 per 100,000 people per year for all ages, respectively [34]. In Izumo, patients with suspected SAH were immediately referred to a hospital irrespective of age or condition, and CT scans were performed as far as practicable to confirm SAH. Some patients who died before arrival at hospital still underwent a CT scan. One of the reasons for the high incidence of SAH in Izumo is a more aggressive diagnostic strategy employed for SAH in this small city, even in older patients with poor clinical status. These factors make it unlikely that many patients have been missed in Japanese studies, even if they are elderly. In addition, Japan has the longest life expectancy in the world, and the number of elderly patients is increasing. From CBSs [18, 36, 47], assuming that the incidence of aneurysmal SAH is 10–20 per 100,000 people per year, RCAs will occur in 0.1–0.2% of the general population in 10 years.

Age distribution of patients with cerebral aneurysms

Table 2 shows the age distributions of patients with CAs in autopsy studies. The age distributions of UCAs and RCAs were different. Namely, patients with UCAs were generally older than those with RCAs [10, 17, 45, 46, 51, 58]. Six studies have reported the age distributions of patients with UCAs and RCAs. Patients aged ≥ 70 years who had UCAs and RCAs accounted for 29% and 4% (33 of 114 and 3 of 71) [58], 39% and 13% (33 of 84, and 17 of 133) [17], 23% and 5% (9 of 39, and 7 of 131) [45, 46], 81% and 34% (22 of 27, and 10 of 29) [51], and 26% and 4% (21 of 80, and 2 of 57) [10], respectively. In autopsy studies, the peak prevalence of incidental UCAs was observed in the seventh decade [10, 17, 45, 58], whereas that of RCAs was recorded in the sixth decade [14, 17, 46, 58, 74]. According to Iwamoto et al. [51], the prevalence of UCAs increased with age, whereas that of RCAs decreased with age; the mean age of patients with unruptured CAs (80.0 years) was higher than that of patients with ruptured CAs (63.5 years). The age difference between patients with UCAs and those with RCAs suggests that middle age is an important risk factor for aneurysm rupture and that most UCAs are less likely to rupture.

Table 2 Age distribution of patients with cerebral aneurysms in autopsy series
Full size table

In clinical studies, UCAs were predominantly diagnosed in the middle-aged and older general population [1, 31, 78, 84]. CBSs showed that the incidence of aneurysmal SAH increased almost linearly with age [34, 36, 49, 53, 55, 64]. In Izumo Study, although hypertension and cigarette smoking were independent positive risk factors for aneurysmal SAH, age was not associated with an increased risk of aneurysmal SAH [37, 38, 41].

Site of cerebral aneurysms

Table 3 shows the sites of CAs reported in autopsy studies. This review study revealed that the sites of UCA in patients without SAH were the anterior cerebral artery (ACA), including the anterior communicating artery (ACoA) and distal ACA, in 27–50% of cases; the internal carotid artery (ICA) in 9–28%; the middle cerebral artery (MCA) in 25–45%; and the vertebrobasilar artery (VBA) in 0–8% [9, 45, 51, 63]. Similarly, excluding the forensic data [19, 21], RCA occurred in the ACA in 22–55% of cases, the ICA in 14–49%, the MCA in 10–38%, and the VBA in 0–21% [3, 9, 13, 14, 46, 51, 57, 83], whereas UCAs and RCAs occurred at CA sites in 16–45%, 11–50%, 15–44%, and 8–22% of cases, respectively [3, 9, 10, 23, 29, 45, 46, 51, 58, 60, 66, 71, 74, 83]. Comparing the sites of RCAs and UCAs in four studies of SAH patients with multiple aneurysms, the proportion of ACA aneurysms was higher in the RCA series (22–37%) than in the UCA series (18–23%), and the proportion of ACoA aneurysms was 30–34% for RCAs and 13–20% for UCAs [3, 9, 46, 83]. These findings suggest that ACoA aneurysms are prone to rupture [15, 32, 61]. In general, MCA aneurysms are usually encountered more often at autopsy than in clinical studies [45]. McKissock et al. [61] found that 27 UCAs had not been diagnosed during the lives of 20 of 100 complete necropsy patients who had SAH, and that the most common site for these was the MCA. Therefore, the rate of rupture of UCAs seems to be lower for MCA aneurysms than for ACA or ICA aneurysms [45]. The frequency of ruptured VBA aneurysms is higher in autopsy studies than in CBSs [42]. The difference is attributable to a higher case of fatality rate in patients with VBA aneurysms compared to patients with aneurysms at other sites. In both of two forensic studies, the proportion of ruptured VBA aneurysms was 14% [4, 21], and in one study of patients who had died immediately, the figure reached 18% (27 of 149) [21].

Table 3 Sites of cerebral aneurysms in autopsy series
Full size table

In a systematic review of UCAs found at autopsy and by angiography by Vlak et al. [79], those in the ACA accounted for 18%, those in the ICA 42%, those in the MCA 35%, and those in the VBA 5%. In the Unruptured Cerebral Aneurysm Study of Japan (UCAS Japan) [77], which was based on a large number of UCAs, the ACoA accounted for 16% of UCA sites, the ICA for 34%, the MCA for 36%, the VBA for 9%, and other sites for 6%. In CBSs, the frequency of RCAs was highest in the ACA (24–43%) and lowest in the VBA (5–14%) [42]. The frequencies of ICA aneurysms (19–29%) and MCA aneurysms (21–28%) were intermediate between them [42]. Inagawa [42] roughly estimated that the proportions of RCA sites based on CBSs were 40% for the ACA, including the ACoA and distal ACA, 25% for the ICA, 25% for the MCA, and 10% for the VBA. Therefore, in clinical studies, ACA aneurysms appear to be more common in RCA series than in UCA series, whereas ICA and MCA aneurysms are more common in UCA series.

Two autopsy studies have analyzed the site of aneurysms separately for men and women: one focused on patients with UCAs [63] and the other on patients with UCAs and RCAs [74]. Both studies found that ACoA aneurysms were more frequent in men, whereas ICA aneurysms were more common in women.

Similarly, two CBSs showed that regardless of age, ruptured ACoA aneurysms were found more frequently in men, whereas ICA aneurysms were encountered more commonly in women [42, 55].

Size of cerebral aneurysms

At autopsy, it is often extremely difficult to find a small aneurysm located within a blood clot [23]. For this reason, the presence of small aneurysms is often overlooked, which can reduce the apparent prevalence. McCormick and Acosta-Rua [59] demonstrated that the unfixed size of UCAs was 30–60% greater in maximal diameter during infusion of the cerebral arteries with a 0.9% saline solution under 70 mmHg pressure. Accordingly, they considered that an aneurysm measuring 1 cm in a living patient would measure only about 6–7 mm at autopsy. The same authors also demonstrated that there was a decrease in size of 6–12% following formalin fixation. Allowing for all these factors, the size of aneurysms 10 mm in diameter during life would decrease to 5.5–7.2 mm in autopsy material [45]. Therefore, even in autopsy examinations, CAs may be overlooked, and there is a possibility that the actual prevalence would be higher than those reported so far. Careful examination of the brain is very important to determine the exact prevalence of CAs. In clinical studies, the potential accuracy of angiography in demonstrating CAs has been investigated in a number of studies. However, in a significant number of patients who show clinical symptoms and signs of SAH, the causative lesion cannot be found. In patients in whom angiography is negative, the cause may be a small aneurysm that undergoes thrombosis within the sac, which is sometimes quite extensive [35, 72].

Table 4 shows the size of CAs in autopsy studies. Smaller aneurysms tended to be more common in UCAs than in RCAs, while larger aneurysms showed an opposite tendency [21, 45, 46, 58]. Whether or not the size of CAs differs with patient age has been a controversial issue. In autopsy studies, McCormick and Acosta-Rua [59] reported that small UCAs were more common in patients over 45 years of age, whereas larger RCAs occurred more often in patients under 50. In contrast, dealing primarily with RCAs, Sugai and Shoji [74] suggested that large aneurysms were more common in older age groups. However, Crompton [14] found no tendency for the diameter of 278 RCAs to increase with age, and Inagawa and Hirano [45] found no relationship between UCA size and patient age.

Table 4 Size distribution of cerebral aneurysms in autopsy series
Full size table

In two large clinical studies of UCAs, older patients had larger UCAs, and the size of the UCAs increased with age [8, 77]. For RCAs, some hospital-based studies have shown that the size was not related to patient age [8, 44, 50, 65, 81]. However, in the Izumo Study [43], RCAs tended to be larger in older age groups, and age (≥ 60 years of age) and cigarette smoking were independently associated with aneurysms ≥ 5 mm in diameter. Considering that the incidence of aneurysmal SAH increases with age [34, 36, 49], and atherosclerosis of the circle of Willis is more prominent in older patients [46], UCAs in the elderly may not rupture as easily as those in younger patients [43]. This may be one of the reasons why a relationship between aneurysm size and age has been suggested [43].

Regarding the relationship between aneurysm site and size, ACoA aneurysms were smaller than ICA and MCA aneurysms (Table 5) [45, 74], similar to the results of the Izumo Study [42, 43].

Table 5 Sites and sizes of cerebral aneurysms in autopsy series
Full size table

Multiple cerebral aneurysms

Table 6 gives details of multiple CAs found at autopsy. In this study, the proportion of multiple aneurysms was 14–19% in patients without a history of SAH [45, 63], 6–31% in patients with SAH [9, 13, 15, 19, 21, 25, 46, 57, 74, 83], and 12–36% in patients with and without SAH [3, 10, 12, 29, 45, 46, 51, 58, 60, 66, 71]. In 1958, Housepian and Pool [29] found that arteriosclerosis was more frequent in patients with multiple aneurysms than in those with single aneurysms. Regarding the relationship between age and the multiplicity of aneurysms, Sugai and Shoji [74] showed that all patients aged ≤ 29 years had a solitary aneurysm, and that after the fourth decade, the frequency of multiple aneurysms increased with age. In 1966, Crompton [15] reported that in 90 patients with multiple RCAs, the largest aneurysm had ruptured in 79 (88%), and that when the combination of ICA and MCA aneurysms was on the same side, proximal aneurysms ruptured more often.

Table 6 Multiple cerebral aneurysms found at autopsy
Full size table

In 1964, Wood [85] performed an angiographic study of multiple aneurysms and concluded that the largest aneurysm was the one most prone to rupture.

In the Izumo Study [40], the average annual incidence rates of SAH in patients with multiple aneurysms and in those with single aneurysms were higher in women than in men, and the age-specific incidence rates of SAH in both groups tended to increase with age. In the Izumo Study for multiple aneurysms, the RCAs were mostly larger than UCAs, and the largest were most likely to rupture [40]. Multivariate analyses demonstrated that only the size of RCAs was a significant positive predictor of multiplicity [40, 43]. There is a possibility that the greater the aneurysm size at rupture, the longer the interval from development. Furthermore, the longer the interval from development to rupture, the higher the possibility that further new aneurysms will develop, resulting in an increase in both the number and/or size of the aneurysms [40]. Aneurysm multiplicity is related mainly to aneurysm size at rupture; in other words, whether newly developed aneurysms rupture easily when they are small [40]. In CBSs, the frequency of multiple aneurysms has ranged from 18 to 34% [40, 52, 62, 68], depending mainly on the completeness of the diagnostic procedures. Inagawa [40] roughly estimated that the frequency of multiple aneurysms based on CBSs and autopsy studies was approximately 30%.

Minute cerebral aneurysms

In autopsy studies, aneurysms of microscopic size and some larger ones constitute a definite group that is seldom discovered without the aid of special measures. In 1941, Richardson and Hyland [66] reported two patients with minute unruptured berry aneurysms; one was at the MCA bifurcation and the other was at the angle between the posterior cerebral and posterior communicating arteries. Each was a minute outpouching about 2 mm in diameter, which they thought could be a very early stage of aneurysm formation. In 1958, Brolin and Hassler [5] examined 35 randomly selected brains, and although macroscopic examination did not reveal any aneurysms, four small aneurysms were discovered by microscopic examination. Hassler [27] divided CAs into two groups according to their maximum diameters in fixed preparations: minute aneurysms with a maximum diameter of less than 2 mm and major aneurysms with a maximum diameter of at least 2 mm. Sugai and Shoji [74] adopted Hassler’s definition of minute and major aneurysms.

In clinical studies, UCAs a few millimeters in diameter, which were not found by preoperative angiography, were occasionally encountered during surgery for known aneurysms [6, 33]. However, even in the large UCAS Japan survey [77], all 6697 registered UCAs were ≥ 3 mm in size, and none were < 3 mm.

Table 7 shows the age distribution of patients with minute UCAs. In 1961, Hassler [27] autopsied a normal series of 250 individuals whose clinical signs suggested that there had been no direct cause of death in the brain and found 32 minute UCAs in 27 (11%). All of these minute UCAs were observed in 27 (19%) of 144 individuals aged ≥ 31 years, and 18 (67%) of them were ≥ 61 years of age. However, no minute aneurysms were found among 106 individuals aged ≤ 30 years. Minute UCAs were observed more frequently in women (15 of 123, 12%) than in men (12 of 127, 9%). Multiple minute aneurysms were found in 5 (19%) of 27 individuals. In addition, 8 minute UCAs were encountered in 5 of 10 patients with major RCAs. In another two patients, SAH was suspected to have been caused by rupture of minute aneurysms; one was a 46-year-old woman and the other a 46-year-old man. Sugai and Shoji [74] reported that the occurrence of minute UCAs was very frequent in SAH patients as compared with the remaining disease groups. In these two studies, minute UCAs were found in about half of patients with SAH. Minute aneurysms were observed at the junction of the cerebral arteries, as was the case for major aneurysms, and in patients with multiple aneurysms, minute aneurysms were located mostly distal to the major aneurysms, especially in the MCA [27, 74]. About 60% of minute aneurysms were located at the MCA (Table 8) [27, 74].

Table 7 Age distribution of patients with minute unruptured cerebral aneurysms measuring < 2 mm in autopsy series
Full size table
Table 8 Sites of minute unruptured cerebral aneurysms measuring < 2 mm in autopsy series
Full size table

With increasing patient age, these minute aneurysms increased in size accompanied by arteriosclerotic changes and developed into saccular aneurysms [74]. A few of the minute aneurysms became atherosclerotic without any increase in size. Due to the fact that both minute and major aneurysms were dome-shaped and had similar histological structures, minute aneurysms were considered to be in the very early stages of formation, and major aneurysms probably originated from them [5, 27, 66, 74]. If major aneurysms develop from minute aneurysms, then the abundance of the latter would suggest that only a few minute aneurysms undergo this change [27]. The chances of a minute aneurysm growing into a major aneurysm presumably depend to a great extent upon the size of the media defect [27]. These autopsy studies suggest that CAs occur more frequently than indicated in the literature, and that many minute aneurysms are overlooked. As for the age distribution, most minute aneurysms appear from the fourth decade [27].

Discussion

To discuss the prevalence of CAs, it is more understandable to divide aneurysms into two groups according to maximum diameter: minute aneurysms measuring < 2 mm, and major aneurysms measuring ≥ 2 mm [27]. In autopsy studies involving total brain examinations across all age groups, the prevalence of CAs was 0.3–4.0% for UCAs alone [3, 10, 17, 45, 58] and 1.3–7.6% for CAs including UCAs and RCAs [3, 10, 17, 29, 45, 46, 51, 58, 71]. In patients with UCAs [45], RCAs [17, 25, 46], and CAs including UCAs and RCAs [10, 45, 46, 51], the prevalence of CAs was higher in women. Patients with UCAs were generally older than those with RCAs [10, 17, 45, 46, 51, 58]. UCAs tended to be smaller than RCAs, whereas the latter tended to be larger [21, 45, 46, 58]. MCA aneurysms usually predominated in autopsy material in comparison with clinical studies [45]. In autopsy studies, ACoA aneurysms were more frequent in men, but ICA aneurysms were more common in women [63, 74], and ACoA aneurysms were smaller than ICA and MCA aneurysms [45, 74]. Similar findings have also been suggested in CBSs [42, 43, 55]. In autopsy studies of SAH patients with multiple aneurysms, the proportion of ACA aneurysms was higher in RCA series than in UCA series [3, 9, 46, 83]. That is, ACoA aneurysms were prone to rupture [61]. However, most of these data were derived from patients with major aneurysms.

Minute aneurysms were found by microscopic examination at autopsy [5, 27, 66, 74] or encountered occasionally during surgery for known aneurysms [6, 33], but it is difficult to detect them using conventional imaging techniques, such as MRA, CTA, and DSA. In recent studies, 3-T MRA has improved the ability to diagnose small aneurysms 2–3 mm in diameter, but a clear diagnosis is still difficult for aneurysms smaller than 2 mm [30, 31, 56]. Autopsy studies, especially microscopic examinations, are more sensitive for detection of CAs than imaging techniques [5]. However, even autopsy can miss minute aneurysms due to post-mortem shrinkage because of blood flow arrest and formalin fixation [45, 59]. Minute aneurysms are thought to be in the very early stages of formation because of the similarity of their histological structure to major aneurysms [5, 27, 66, 74]. In terms of age distribution, minute aneurysms appear from the fourth decade and are not detected in neonates or younger age groups [27], suggesting that aneurysms are acquired lesions. The high prevalence of UCAs and RCAs in the elderly and in women can be attributed to the more frequent development of minute aneurysms in these populations. With regard to the sites of minute aneurysms and major aneurysms, MCA aneurysms accounted for a much higher proportion of the former [27, 74]. In patients with ipsilateral ICA and MCA aneurysms, proximal aneurysms tended to rupture more easily [15]. Therefore, hemodynamic stress on dilation from minute aneurysms to major aneurysms may be weaker in MCA aneurysms than at other aneurysm sites, resulting in a high frequency of minute MCA aneurysms [45]. If so, hemodynamic stress may not contribute much to aneurysm formation, as the frequency of minute MCA aneurysms is high. Recent clinical studies have discussed risk factors for the development and rupture of CAs based on the prevalence of major aneurysms [38, 41]. However, risk factors for the development of CAs that have been discussed previously appear to be risk factors for the enlargement of minute aneurysms to major aneurysms rather than the formation of minute aneurysms. The true risk factors for the development of CAs should be studied in relation to minute aneurysms. After minute aneurysms develop, most of them are stable, but some become unstable as they enlarge to major aneurysms. Some major UCAs remain unruptured, and the size of UCAs increases with age. In this situation, new minute aneurysms may develop, resulting in multiple aneurysms. However, some UCAs will eventually rupture as UCAs grow to a critical size that varies with age and aneurysm site [13, 15, 59, 71, 75].

As the prevalence of minute aneurysms is much higher than that of major aneurysms, only a small number of the former develop to the latter [27, 74]. To date, most autopsy and clinical studies have discussed major aneurysms, and the exact prevalence of minute aneurysms is unknown. In the present study, due to the poor autopsy data for minute aneurysms, their prevalence was tentatively estimated by taking into account the prevalence of minute and major UCAs at autopsy. As a result, it can be hypothesized from this review analysis that, in the very early stages, minute aneurysms will develop in approximately 10–20% of the general population mostly after middle age [27, 74]. Minute aneurysms are difficult to detect by conventional imaging techniques, but some enlarge to major aneurysms that can be diagnosed in this way. In clinical studies, the reported prevalence of UCAs, mostly diagnosed by conventional imaging techniques, is estimated to be 1–2% among the general population, mainly in middle-aged and older individuals, with higher detection rates in women [1, 78, 84]. In CBSs, CAs will rupture in 0.1–0.2% of the general population within 10 years [18, 36, 47]. The prevalence of RCAs increases with age and is higher in women [34, 36]. As a rough estimate, the following assumptions are possible. In a general population of 10,000, minute UCAs will develop in 1000–2000 individuals (10–20%), minute UCAs will enlarge to major UCAs in 100–200 (1–2%), and major UCAs will eventually rupture in 10–20 (0.1–0.2%) within 10 years. In other words, minute UCAs will develop in 10–20% of the general population, 10% of minute UCAs will enlarge to major UCAs, and 10% of major UCAs will eventually rupture within 10 years. At present, risk factors for the development and enlargement of minute aneurysms are unknown and difficult to investigate. In the future, advances in imaging technology such as MRA, CTA, and DSA may enable more accurate diagnosis of minute aneurysms. If so, risk factors for the development and enlargement of minute aneurysms may become clearer.

Conclusions

Despite dramatic improvements in imaging techniques, such as MRA, CTA, and DSA, the diagnosis of small CAs is still limited. In autopsy studies, UCAs tended to be smaller than RCAs, and patients with UCAs were generally older than those with RCAs. Microscopic examination at autopsy revealed the existence of minute CAs (< 2 mm), which are considered to be in a very early stage of aneurysm formation. Minute UCAs were occasionally seen during surgery on known aneurysms, but it is difficult to detect them using conventional imaging techniques. Most of the minute CAs were located at the MCA. Based on autopsy and clinical studies, it can be hypothesized that minute UCAs develop in 10–20% of the general population, that 10% of minute UCAs enlarge to major UCAs (≥ 2 mm) detectable by conventional imaging techniques, and that 10% of major UCAs will eventually rupture within 10 years. The high prevalence of UCAs and RCAs in the elderly and in women can be attributed to the more frequent development of minute aneurysms in these populations. However, a potential limitation of this study was that data on minute CAs has been scant to date, and it is not expected that more such autopsy data will be available in the future. Further advances in diagnostic technology may lead to the detection of minute CAs, and if so, the natural course of CA prevalence may be clarified more accurately.