Patient data were gathered from 17 institutions. This retrospective study was approved by the review board of each institution. Patients with a pathologic or clinical diagnosis of visceral artery aneurysm due to SAM between 2005 and 2015 were enrolled in this study. The inclusion criteria for clinical diagnosis of SAM were defined according to previous literature [9] as follows: aneurysm located at the main trunk or branch of the celiac artery, superior mesenteric artery (SMA), or inferior mesenteric artery; no calcification; patient age older than 20 years; no clinical symptoms, such as constitutional, cutaneous, arthritic, ocular, pulmonary, or renal symptoms; and no celiac artery stenosis. Patients were excluded if they had isolated renal artery aneurysms, to exclude fibromuscular dysplasia. Patients with a renal artery aneurysm who also had an aneurysm of another visceral artery were included in the study.

Images of all patients with a clinical diagnosis of SAM were collected and assessed by central radiologic review. Radiologic diagnosis of SAM was defined as a string-of-beads appearance or the presence of multiple fusiform aneurysms, according to the literature [5, 10]. In the central radiologic review, all images were interpreted by two experienced diagnostic and interventional radiologists, and any discrepancies were resolved through consensus. We previously reported data on eight aneurysms in four patients [6]. We divided aneurysms due to SAM into unruptured and ruptured cases. To clarify the natural history of unruptured aneurysms, we assessed their morphologic changes. To evaluate the period between the first examination and the appearance of morphologic changes, we reviewed all available images. We also evaluated the efficacy and safety of TAE for visceral artery aneurysms due to SAM.

Overall, records of 153 aneurysms in 70 patients were gathered; five in two patients were diagnosed pathologically, and 148 in 68 patients were diagnosed clinically. The images of the 148 clinically diagnosed aneurysms were assessed in central radiologic review. As a result, 118 aneurysms in 43 patients were diagnosed radiologically. Therefore, a total of 123 aneurysms due to SAM in 45 patients (33 men, 12 women; median age, 60 years; range, 30–82 years) were included in the study (Fig. 1). Thirty-six patients had clinical symptoms, and all 36 reported abdominal pain. The other nine patients had no symptoms, and their aneurysms were found incidentally. The median follow-up period was 12 months (range, 0.5–99 months).

The aneurysm characteristics and clinical course are summarized in Table 1 and Figure 2. The median size of the aneurysms was 8 mm (range, 3–63 mm). Among the 123 aneurysms analyzed in this study, 92 (75%) were unruptured and 31 (25%) were ruptured. Among the 92 unruptured aneurysms, 21 were treated, and 71 were observed. Nine of 31 (29%) patients with 71 observed aneurysms received medications to control hypertension. No patients received antiplatelet agents, anticoagulants, steroids, or immunosuppressants. One of the 71 aneurysms could not be evaluated because of a metallic artifact from coils placed in another aneurysm. Therefore, a total of 70 of 123 (57%) aneurysms were evaluated for natural history. Of 70 aneurysms, 45 (64%) had no morphologic changes at follow-up. Among the other 25 aneurysms, nine (13% of the 70) were reduced in size, 13 (19%) had disappeared (Fig. 3), and three (4%) were newly present at follow-up.

The 70 unruptured aneurysms were found to be unruptured at initial diagnosis of SAM. The 31 ruptured aneurysms included two aneurysms that newly appeared in another artery during follow-up. In these two cases, the periods between the prior imaging and rupture were 99 and 10 months; TAE was performed successfully (Fig. 4). Thus, in total, five new aneurysms in four patients appeared during follow-up in other arteries within unaffected segments confirmed at prior imaging.

The median period between the first examination and the appearance of any morphologic changes of observed aneurysms was 3 months (range, 2 days to 99 months). The median period between the first examination and the appearance of morphologic changes was 10 months (range, 2 days to 99 months), disappearance of morphologic changes was 6 months (range, 4 days to 9 months), and reduction in morphologic changes was 2 months (range, 1–99 months) (Fig. 5). Twenty-nine of the 31 ruptured aneurysms were treated, 25 by TAE and four by surgical resection. For 21 of the 92 unruptured aneurysms, treatments included 20 TAEs and one surgical resection. Eleven of the 20 TAEs were performed in the same TAE session as for ruptured aneurysms. The indication for TAE was at the discretion of the operators.

To summarize the interventions, 50 of the 123 aneurysms were treated (five surgical resections, 45 TAEs). TAE was successful in all cases, and there was no recurrence during a median follow-up of 16 months (range, 0.25–99 months). The following materials were used for TAE: coils in 37 aneurysms, N-butyl-2 cyanoacry-late (NBCA) in four, both NBCA and gelatin sponge in three, and both coils and NBCA in one. Complications occurred in three aneurysms (6.7%) and involved slight arterial wall injury, which occurred in abnormal-appearing segments in all cases. No severe complications, such as bleeding or organ infarction, occurred.

The morphology of aneurysms associated with SAM likely changes quickly. Thus, we assessed the natural history of aneurysms due to SAM in this study. Because TAE is an important treatment of ruptured aneurysms due to SAM, we also assessed the efficacy of TAE in a large sample.

We assessed the natural history of 70 unruptured aneurysms due to SAM. The results showed no change in the natural history in 64%, size reduction in 13%, and disappearance in 19% of the aneurysms. In a previous study [10], the natural history of SAM in 10 patients with aneurysm, dissection, occlusion, or wall thickening was found to stabilize or resolve in 50% of cases within a median follow-up period of 33 months. In another study [5], four patients with aneurysms due to SAM underwent follow-up, and all aneurysms remained stable or resolved. Naidu et al. [8] reported on the disease course of SAM, including aneurysm, dissection, occlusion, and wall thickening in 97 patients; 80% of the aneurysms were stable or resolved within a median follow-up period of 3.1 months.

Our results are similar to those reported in the literature. From a pathologic point of view, lesions due to SAM have been reported to undergo changes from the injurious phase to the reparative phase [11–14]. After the initial arterial injury, the reparative phase begins with the growth of granulation tissue in the arterial gaps. Fibrosis replaces the granulation tissue to heal the arterial wall and help restore its shape [12, 13]. Thus, size reduction or disappearance of an aneurysm may reflect this reparative phase.

In this study, two ruptured aneurysms newly appeared during follow-up at other arteries within the unaffected segments confirmed at prior imaging; however, none of the aneurysms unruptured at initial diagnosis of SAM ruptured during the follow-up period. Considering this natural course, conservative management seems appropriate for unruptured aneurysms due to SAM. To decrease the risk of rupture, controlling hypertension with antihypertensive medications is considered important [15] and may be useful for observation of aneurysms. In our study, nine of 31 (29%) patients with unruptured aneurysms received antihypertensive medications. The role of antiplatelet agents and anticoagulants has not been established for aneurysms, but these agents are generally used for dissection with a thrombosed false lumen causing luminal compression, end-organ ischemia, or infarction [8]. In addition to the two ruptured aneurysms, there were three newly appearing unruptured aneurysms. Therefore, observation of patients by vascular imaging is necessary, but the duration and timing of imaging follow-up are unclear.

Development of new aneurysms or dissections 10 months [6] and up to 1.5 years [16] after treatment of the initial lesion has been reported. Kalva et al. [10] recommended imaging follow-up at least once per year. Among the five newly appearing aneurysms in our study, however, the period between the first examination and the appearance of the new aneurysm varied from 2 days to 99 months. Thus, it is not easy to determine the timing of imaging follow-up, even considering our results. Because all patients with ruptured aneurysms in this study had abdominal pain, it may be important to advise patients to seek emergency treatment immediately after experiencing any abdominal symptom, because it might indicate a ruptured aneurysm. Furthermore, among the five newly appearing aneurysms in this study, two (40%) were ruptured when they were found. Because such aneurysms may be in the injurious phase, we think it is important to pay attention to rupture at aneurysm appearance.

TAE was performed on 45 aneurysms due to SAM in this study and was successful in all cases; no recurrence or bowel ischemia was found during the follow-up period. Therefore, TAE should be considered a useful treatment of aneurysms due to SAM. However, slight arterial wall injury was found in 6.7% of cases, occurring in abnormal-appearing segments in all cases. It has been reported that the arterial wall in SAM might be prone to dissection from mediolysis [10, 17]. Thus, careful manipulation should be undertaken in the performance of TAE. Furthermore, given the risk of arterial injury for unruptured aneurysms due to SAM, TAE can be avoided, and observation is recommended. It is also important to monitor not only identified unruptured aneurysms but also arteries without aneurysms to detect any further changes.

The locations of the aneurysms in this study varied; 17 were found in the celiac artery, 15 in the main duct of the SMA, 11 in the middle colic artery (MCA), nine in the left or right hepatic artery, nine in the splenic artery, eight in the left gastric artery, eight in the right gastroepiploic artery, and 46 in other arteries. Ruptured aneurysms occurred in the MCA in 10 cases, splenic artery in three, right gastroepiploic artery in three, and other arteries in 15 cases. Aneurysms of the MCA were ruptured in 10 of 11 (91%) cases.

It is unclear why aneurysms of the MCA easily rupture. We assume that temporal stenosis of the visceral artery may appear and disappear in SAM, and hemodynamic changes can occur. Thus, the MCA is likely to be exposed to such hemodynamic changes because it is connected to both the SMA and the inferior mesenteric artery. Therefore, we suspect that the risk of rupture is high for aneurysms in the MCA, as is the case for pancreaticoduodenal artery aneurysms [18–20]. Given the high risk of rupture, aneurysms of the MCA should probably be treated even if they are unruptured. In addition, the MCA in particular should be checked during follow-up examinations.

In our study, we did not evaluate visceral artery dissection related to SAM. However, dissection was reported to be the most common arterial imaging finding of SAM [8]. Visceral artery dissection can cause bowel infarction and necrosis, and there is risk of aneurysm degeneration and rupture [21]. Ruptured dissection of the SMA related to SAM has also been reported [22]. Therefore, management of dissection due to SAM should be considered an important issue. Further studies are necessary to investigate dissection related to SAM to clarify optimal management.

Our study had several limitations, the primary one being its retrospective design. Most of our patients had no pathologic evaluation of the involved arteries, and diagnosis was based on clinical and radiologic findings. Furthermore, 21 of 92 unruptured aneurysms were treated, and the indication for treatment was at the discretion of the operators. Hence, there was selection bias in the evaluation of the natural history of the unruptured aneurysms. In addition, regarding the results of TAE, multiple operators at 17 institutions were involved in this study. Thus, details of the procedure were not completely standardized owing to variability in the techniques of individual operators.

Unruptured aneurysms due to SAM at initial SAM diagnosis are likely to be stable or resolved; thus, conservative management seems appropriate. Because the risk of rupture of aneurysms in the MCA appears to be high, such aneurysms should be treated even if unruptured. TAE is a useful treatment of aneurysms due to SAM, but it is important to be cautious of arterial injury because the arterial wall in SAM is prone to dissection; thus, we recommend it only for ruptured aneurysms.

We thank the members of the study group: Masanori Inoue, Department of Radiology, Keio University; Seiji Kamei, Department of Radiology, Kainan Hospital; Yasutaka Baba, Department of Radiology, Hiroshima University; and Kenji Murakami, St. Mari-anna University School of Medicine.