Prevalence of celiac artery compression by median arcuate ligament in patients with splanchnic artery aneurysms/pseudoaneurysms submitted to endovascular embolization

Median arcuate ligament (MAL) is a fibrous arch that joins the right and left crura of the diaphragm on either side of the aortic hiatus, usually at the level of the 12th thoracic vertebra, crossing the aorta superior to the celiac artery (CA) origin [1].

Celiac artery compression by MAL is increasingly being detected due to the widespread use of imaging studies [2]. In fact, it is suggested in the literature that in up to a third of people, MAL has a lower insertion and may compress the proximal portion of the CA, which is usually more severe in expiration, given the cranial displacement of the aortic branches [3]. It is considered to constitute one of the main causes of CA stenosis [4]. Still, it is mostly asymptomatic [2].

When this anatomy is associated with symptoms, it is named median arcuate ligament syndrome (MALS), whose prevalence is extremely low (0.002% of the population), and more frequently detected in female patients. Clinical presentation is variable and includes weight loss, postprandial epigastric pain, nausea, and vomiting, most frequently manifesting at the ages of 20 to 50 years [1, 2]. The diagnosis is usually only established after the exclusion of other more common causes of symptoms [5].

Two main mechanisms have been attributed to MALS, namely compression of vascular and nervous structures. Vascular compression alone does not seem to be enough explanation, given that adequate blood supply to the digestive organs is provided by the abundant collateral circulation via the mesenteric arteries. As so, it is hypothesized that the myoelectric activity of the smooth muscles may also be altered by the compression of the nervous splanchnic plexus fibers in the digestive tract. Either splanchnic vasoconstriction or direct stimulation of pain neurons may result in neurogenic-related pain [1].

On computed tomography angiography (CTA), the diagnosis is confirmed by the visualization of a focal narrowing at the origin of the CA with a hooked appearance, better depicted at arterial phase imaging in the sagittal plane [6].

The coexistence between CA stenosis or occlusion and aneurysms in its visceral arterial collaterals was first described in 1973 by Sutton and Lawton [1, 7]. Since then, several studies have reported an association between MALS and pancreaticoduodenal arcades (PDAs) aneurysm formation [2]. Hypothesized explanations for this phenomenon include the increased blood flow and hypertension in the small collateral arteries connecting the CA and the superior mesenteric artery (SMA) beds (enlarging them and eventually weakening its wall), and the increased arterial wall shear stress [1, 2, 8]. Besides, the dynamic nature of the CA compression may result in alternating anterograde and retrograde blood flow in these collaterals, with consequent turbulent flow [3]. Pancreaticoduodenal arcades aneurysms not related to MALS are mostly secondary to atherosclerosis in CA and its branches [9].

Endovascular treatment with CA angioplasty and stent placement may increase the blood flow to the CA, but surgical MAL incision (either laparoscopically or by open surgery) should be the standard therapy for MALS, given that the cause of the stenosis is the extrinsic compression by this ligament [1, 2, 10].

Despite the reported cases of PDAs aneurysm regression or stability following CA decompression in patients with MALS, the treatment of these aneurysms is recommended regardless of their size, given their significant risk of rupture, which is associated with a relatively high mortality rate. When possible, endovascular embolization (often with metallic coils) is currently considered the preferred option for both ruptured and unruptured aneurysms, given its less invasiveness and increased safety in recent years [1, 2, 6, 11]. Surgery is recommended in unstable patients and cases of endovascular treatment failure [1]. It is technically feasible but associated with a mortality rate of up to 30% in ruptured aneurysms [11]. A potential complication of PDAs aneurysm embolization in patients with CA stenosis is the ischemic damage to organs such as the liver or the pancreas, due to interruption of the collateral circulation [1].

However, being a rare disease, no therapeutic guidelines have been established for patients with splanchnic artery aneurysms and MALS. Celiac artery decompression to correct the hemodynamic imbalance in the PDAs is often recommended to prevent ischemia and aneurysm recurrence [1, 2, 10, 12‐18], whereas others reported favorable outcomes with aneurysm embolization alone [1, 16, 19‐27].

The main goal of this article was to determine the prevalence of median arcuate ligament compression (MALC) of celiac artery in patients with splanchnic artery aneurysms/pseudoaneuryms (SAAPs) submitted to transcatheter arterial embolization, and compare demographic data and clinical outcomes between patients with and without MALC. As a secondary objective, we analyzed other causes of CA stenosis and compared patient characteristics and outcomes between patients with different causes of CA stenosis.

During the study period, we found a prevalence of CA compression by MAL in 7 out of 57 patients submitted to endovascular embolization of SAAPs at our institution (12.3%), after the exclusion of 7 patients, due to lack of enough clinical and imaging data. Celiac artery stenosis was found in 3 other patients, caused by atherosclerotic plaques (Fig. 2). Minimal atherosclerotic plaques were found in the CA of three patients with MALC, but they were not occlusive.

Statistical comparison between patients with and without MALC is shown in Table 1. The mean age of patients with and without MALC was 59.7 years (± 18.4) and 61.6 years (± 15.2), respectively (P = 0.763). The male gender was predominant in both groups (57.1% and 68%, P = 0.675). High blood pressure (HBP) and diabetes mellitus (DM) were present in 42.9% and 71.4% of patients with MALC, respectively, with lower prevalence in patients without MALC (32% and 30%, P = 0.675 and 0.084). Smoking and alcohol consumption were not clarified in more than three thirds of patients. The main indication for embolization was rupture with active bleeding in both groups, namely 71.4% and 54% of patients with and without MALC (P = 0.450), respectively (Figs. 3 and 4). Notably, a “string-of-beads” appearance was noted in the ileocolic artery of a patient with MALC and PDA aneurysm rupture, indicative of segmental arterial mediolysis or fibromuscular dysplasia (not confirmed), suggesting an exacerbating and not only causative role of MALC.

The location of SAAPs in the PDAs (Fig. 5) was more prevalent in patients with MALC (57.1%), whereas it occurred in 10% of patients without MALC. This difference was statistically significant (P = 0.009), with a 5.71 higher chance of this location in the first group (95% confidence interval (CI) of [1.99–16.33]). The only other variable that showed a statistically significant association with the presence or absence of MALC was the type of vascular dilatation, with a greater proportion of aneurysms in patients with MALC (71.4%) when compared with those without MALC (24%), whereas pseudoaneurysms were less common (28.6% vs 76%). The risk of aneurysm instead of pseudoaneurysm was 2.98 times higher in patients with MALC (P = 0.020, 95% CI [1.51–5.88]).

Coils and microcoils were the most frequent embolization materials used in both patients with and without MALC (85.7% and 82%, respectively). The procedure was successful in most cases, namely in 85.7% and 90% of patients with and without MALC, respectively. Few complications occurred after the procedure, including 5 immediate (in 28.6% and 6% of patients with and without MALC) and 14 non-immediate (in 28.6% and 24% of patients with and without MALC). Specifically, one patient with MALC presenting with a 50 mm splenic artery aneurysm suffered arterial rupture during the procedure and partial permeabilization 3 months later. Other complications in this group included coil embolization (immediate) and recurrent bleeding within 6 months. The mortality rate was 0% in patients with MALC, whereas patients without MALC had 14% and 24% mortality rates at 30 and 90 days, respectively.

During a mean follow-up of 1967 days (ranging from 401 to 3116 days), only 1 out of 7 patients with MALC performed surgical MAL incision, 16 months after the endovascular treatment of the SA aneurysm. Follow-up CTA at 3 months showed CA residual stenosis, although less severe (Fig. 6). The remainder patients were kept under clinical and imaging surveillance, with no register of aneurysm recurrence or other complications.

Descriptive analysis of cases with CA stenosis caused by MALC and atherosclerosis is presented in Table 2. Patients with atherosclerosis were older (mean age of 76.7 years ± 8.6) than those with MALC (59.7 years ± 18.4, P = 0.170). Diabetes mellitus and HBP were frequent in both groups, namely in 42.9% and 71.4% of patients with MALC, and in 100% and 66.7% of patients with atherosclerosis, respectively. Patients with atherosclerotic CA stenosis had pseudoaneurysms caused by gastric ulcer and neoplasm perforation, chronic pancreatitis, and percutaneous transhepatic cholangiography. One patient with MALC had a pseudoaneurysm secondary to biliodigestive surgery. Active bleeding was the greater indication for embolization in both groups (71.4% and 100% of patients with MALC and atherosclerosis, respectively, P = 0.1.000). The only exceptions were two cases of prophylactic embolization in the MALC group.

The mean size of SAAPs was larger in patients with MALC (19.4 mm) than in those with atherosclerotic stenosis (12.3 mm, P = 0.729). The mean grade of CA stenosis was similar in both groups (70.9% in the MALC group and 68.0% in the atherosclerosis group, P = 1.000), although the range was higher in patients with MALC (50% to 100% vs. 62.5% to 71.4%). Median arcuate ligament thickness in patients with MALC ranged from 4 to 10 mm, with a mean of 6.7 mm. In general, the mean caliber of the evaluated splanchnic arteries was greater in patients with MALC, especially in GDA (4.6 vs. 3.3 mm, P = 0.059) and PDAs (3.6 vs. 2.1 mm, P = 0.153).

Within a period of 11 years, we found a prevalence of celiac artery compression by MAL in 12.3% (n = 7) of patients submitted to endovascular embolization of SAAPs at our institution (n = 57). Considering aneurysms only, it was present in 35.2% (greater than the rate reported by Regus et al., although the number of visceral aneurysms in their sample was almost twice the size of ours) [29]. Boll et al. found a much greater prevalence of CA stenosis or occlusion (> 60%) in patients treated with coil embolization, but they only included aneurysms of the GDA or PDA [19]. In their retrospective study of 37 patients with MALS diagnosed at CTA, Heo et al. found SA aneurysms in 24.3% of cases, which were more common in patients with collateral circulation. However, only 2 patients underwent embolization, one due to aneurysm rupture and the other due to complete CA obstruction [5]. These findings suggest that the number of patients with co-existent MALC and SA aneurysms is probably higher than we estimated, given that we only analyzed SAAPs submitted to endovascular treatment, which is still more often performed in the presence of complications.

A high prevalence of MALS has been reported in patients with PDAs aneurysms (33–100%), most probably due to the increased blood flow through these collaterals [1‐3, 7, 16]. In our study sample, SAAPs in patients with MALC were also more frequent in PDAs (57.1%, n = 4), whereas they occurred in 10% of patients without MALC. This difference was statistically significant (P = 0.009), with a 5.71 higher chance of this location in the first group (95% CI [1.99–16.33]). Besides PDAs, some reports have described aneurysms in other locations in patients with MALS [13, 19, 21, 23, 26, 30‐37]. In the present sample of patients with MALC, we also found them in the splenic and gastroepiploic arteries.

The only other statistically significant difference we found between patients with and without MALC, was a greater proportion of aneurysms in patients with MALC (71.4%) when compared with those without MALC (24%), whereas pseudoaneurysms were less common (28.6% vs 76%). The risk of aneurysm instead of pseudoaneurysm was 2.98 times higher in patients with MALC (P = 0.020, 95% CI [1.51–5.88]). This is an expected finding, as the hypothesized physiopathological explanation for the development of splanchnic artery aneurysms in patients with MALC is the weakening of the wall of small collateral arteries connecting the CA and the SMA beds, due to increased blood flow, hypertension, and shear stress, whereas pseudoaneurysms are usually secondary to wall disruption from other causes [1, 2, 8]. In fact, the only pseudoaneurysm we found in patients with MALC was secondary to biliodigestive surgery, although it is possible that the weakness of the vessel walls induced by MALC may still have been contributory.

Caruana et al.found that most PDAs aneurysms (7 out of 8) were associated with severe CA stenosis, defined as 80–100% stenosis or ≥ 8 mm in length [3]. Bonardelli et al.also reported a preponderance of severe stenosis (56%) in their study sample [38]. In our group of patients with MALC, 50% was the lower value of stenosis, in a patient submitted to prophylactic embolization of a splenic artery aneurysm. In the other cases, the grade of stenosis was higher, ranging from 55 to 100% (mean 70.86%). This may suggest that a greater CA stenosis predisposes to the formation of SA aneurysms. Also, the diameter of collaterals was greater in cases of more severe stenosis. These findings are consistent with those reported by others [39]. However, the grade of CA stenosis may have been underestimated, since CTA scans are usually obtained on inspiration, with partial release of the compression.

In their study of PDAs aneurysms, Antoniak et al.also graded the thickness of MAL into severe or nonsevere, with a threshold value of 4 mm. Most cases (10 out of 15) were severe [37]. Adopting the same grading, all cases in our study were severe, and the mean thickness was 6.3 mm, ranging from 4 to 10 mm. This could also suggest that SAAPs are more prevalent in patients with greater CA stenosis. However, no correlation was found between that measure and the grade of stenosis, but both are influenced by the inspiration status, which was not controlled in this study.

Although some authors reported that most patients with SAAPs and MALC were asymptomatic [37], in our 7 patients with MALC, only two of them were, and rupture was the main indication for embolization (71.4%). This is probably explained by the patient selection, given that we analyzed SAAPs submitted to embolization, which is mostly done following complications, namely rupture. In their review of five SA aneurysms associated with MALS, Sugiyama and Takehara found rupture in 60% of them (2 in PDAs and 1 in epiploic artery), successfully treated with transarterial coil embolization [4]. Boll et al.also found a large proportion of symptomatic patients (85%) in their study of embolized PDA and GDA aneurysms, including 45% ruptured cases [19]. These findings reinforce the potential clinical relevance of SA aneurysms in this group of patients. It is important to be aware of the presence of CA stenosis before the endovascular treatment of PDA aneurysm rupture, as it can influence the vascular approach (CA cannulation may be difficult in some patients) [40]. Also, PDA embolization in patients with CA stenosis can cause ischemia of the upper abdominal organs [1, 40]. Simultaneous blood pressure monitoring of the common hepatic artery has been proposed as a safety measure in patients with MALC, to prevent liver ischemia [41]. We found no reported signs of ischemic damage to the liver or other organs, presumably due to the presence of a rich collateral circulation in all cases.

Interestingly, Tétreau et al.found that ruptured aneurysms due to CA stenosis were significantly smaller than other ruptured aneurysms (9 mm vs. 26 mm) [42]. Although the mean size of ruptured SAAPs in our study was also smaller in cases of MALC (12.9 mm) compared with the others (15.5 mm), we found no statistically significant difference (P = 0.538).

Besides CA stenosis secondary to MALC, atherosclerotic disease has also been associated with the presence of PDAs aneurysms. At their article, Bonardelli et al.described 23 patients with concomitant CA stenosis/occlusion and visceral aneurysms, mostly secondary to MALS (44%), followed by atherosclerotic lesions (32%) [38]. Like those authors, we also found that mean age was higher in the atherosclerotic group (76.7 years) compared with MALC group (59.7 years). As expected, the mortality rate was higher in the atherosclerotic group, as these patients have more comorbities and are older.

Management of SA aneurysms remains controversial, given the unpredictability of PDAs aneurysms to rupture, which is neither correlated with the size and multiplicity of the aneurysms, nor the patient’s age [37]. In fact, PDAs aneurysms have a high propensity for spontaneous rupture, regardless of size (as opposed to the other SA aneurysms, in which the risk increases with sizes greater than 2 to 2.5 cm). Also, it as an associated mortality rate of 30% to 50%. Therefore, prophylactic endovascular treatment is recommended to prevent such events [3‐5].

However, there is no established consensual approach in terms of when and in whom CA stenosis should be addressed in patients with SA aneurysms and MALC [1, 10, 15, 19‐21]. Some authors have recommended CA decompression to prevent organ ischemia and reduce the risk of aneurysm recurrence, either with endovascular procedures (angioplasty and/or stent placement), or surgery (MAL release and/or bypass grafting) (1,2,10,13,15–18]. Splanchnic artery aneurysms and CA stenosis can be addressed simultaneously (one-stage treatment) or separated (two-stage treatment), although there are no established recommendations regarding the optimal order of and duration of time between the two procedures. Also, it depends on the patient’s condition, grade of stenosis, operator’s preference, and availability of a hybrid operation room [15].

Resolution or stability of collateral circulation and aneurysms after treating CA stenosis alone have been reported, presumably due to reduced flow and thrombosis [12, 14, 43‐45], but there are reported cases of aneurysm growth after CA decompression [32, 46]. On the other hand, others have obtained good outcomes using aneurysm embolization only [1, 16, 19‐27]. However, Takase et al.noted a higher mean age in patients with MALS and PDA aneurysms compared to those without aneurysms (56.9 years vs. 44.6 years), suggesting that longer follow-ups (10 years or longer) may be necessary to detect aneurysm recurrence [27].

In the present study, only 1 out of 7 patients with MALC performed surgical MAL incision 16 months after the endovascular treatment of the SA aneurysm, showing less severe residual CA stenosis in the follow-up CTA 3 months later. The remainder patients had no register of aneurysm recurrence or other complications during a mean follow-up of 1967 days (ranging from 401 to 3116 days). Boll et al.also reported no recurrent PDA or GDA aneurysms after coil embolization alone, although their mean follow-up was shorter (12 months). Still, our study sample is small and further studies with more patients and longer follow-up are necessary to provide evidence for the creation of management recommendations.

In patients with SAAPs submitted to endovascular embolization, the prevalence of CA compression by MAL is not uncommon. The most frequent location for aneurysms in patients with MALC is in the PDAs. Endovascular management of SAAPs is very effective in patients with MALC, with low complications, even in symptomatic ruptured aneurysms.