NF1 is one of the most frequent autosomal-dominant inherited disorders, with an incidence of 1 in 3000 births (Falcone et al. 2010). There is a great variety of manifestations in NF1, involving mesenchymal tissue in musculoskeletal and cardiovascular systems (Hinsch et al. 2008). Although the estimated prevalence of NF-V has been reported as 0.4–6.4%, the actual frequency is unknown, as most patients with NF-V are asymptomatic throughout life. According to an autopsy series of patients with NF1 who died of other causes, vasculopathy was reported in 44% of cases (Salyer and Salyer 1974). Therefore, with the inclusion of asymptomatic NF-V, the frequency of vasculopathy in patients with NF1 may be higher than previously reported.
Patients with NF1 have a decreased life expectancy, by ∼15 years, compared to that of the general population. Malignant neoplasms are the most common cause of death in NF1, followed by vasculopathy (Rasmussen et al. 2001). Cardiovascular disease, haemorrhage, and embolism associated with vasculopathy are frequent causes of death in adult patients with NF1 (Zoller et al. 1995). NF-V includes aneurysms, aneurysmal dilatation, stenosis, rupture, and arteriovenous malformations (D’Errico et al. 2018; Hinsch et al. 2008; Falcone et al. 2010). Among these, aneurysmal dilatation is the most common form of vasculopathy (D’Errico et al. 2018). Additionally, the involvement of multiple vessels is a common characteristic of NF-V (Friedman et al. 2002). In fact, in the present case, multiple aneurysmal dilatations and multiple splenic aneurysms were observed on CT.
Fatal haemorrhaging due to the spontaneous rupture of extracranial arteries in NF1 has been reported, with thoracic vessels in most cases (Ishizu et al. 2006; Bargiela et al. 2018). In contrast, there are only 5 reported cases of ruptured visceral arteries, with SMA, renal artery, and gastroduodenal artery involvement (Huffman et al. 1996; Hinsch et al. 2008; Bargiela et al. 2018); none involved the spontaneous rupture of the PAA.
NF-V is usually described as mesodermal dysplasia or fibromuscular hyperplasia (Hinsch et al. 2008) (Raborn et al. 2020). Mutations in the NF1 gene result in decreased neurofibromin, with subsequent proliferation of endothelial and smooth muscle cells in the arteries and veins (Hung et al. 2012). As a result, the vascular tissue is fragile in patients with NF1 (Aizawa et al. 2010). Moreover, the following have been reported as causes of rupture in patients with NF1: 1) the direct invasion of neurofibromas; 2) compression of the vasa vasorum by neurofibromas; 3) physical movement (e.g. orthopaedic traction); and 4) pregnancy (Ishizu et al. 2006; Leier et al. 1972).
In the present case, in addition to the aforementioned fragile nature of the arterial vessel, we speculated that the spontaneous rupture of the PAA, rather than an existing aneurysm associated with NF1, might have resulted from a direct invasion of neurofibromas. The pseudoaneurysm involved multiple arteries (ASPDA, AIPDA, and PIPDA), and selective angiography of the AIPDA showed early visualization of the portal vein at the AIPDA. We believe that a minor arterioportal shunt was created when the pseudoaneurysm ruptured, and an arteriovenous malformation did not exist. These findings suggest a sudden rupture due to a direct invasion of neurofibromas, rather than the rupture of an existing aneurysm. In fact, a previous CT scan (performed 7 years before TAE at another hospital; images not shown) showed a soft-tissue density around the SMA and no aneurysm in the PAA. Additionally, coronal short-TI inversion recovery magnetic resonance imaging (MRI) performed before TAE (as follow-up for a meningocele from the lumbar spine to the sacrum at another hospital; images not shown), showed a high-intensity area around the SMA, indicating the presence of neurofibromas. We believe that these findings support our speculation. Additionally, although the possibility that an existing PAA aneurysm had ruptured cannot be completely ruled out, this was considered unlikely, as there was no median arcuate ligament stenosis on celiac angiography and no aneurysm on previous CT images. Even if the present case involved a ruptured PAA aneurysm, it is still of interest, as there are only 2 reported cases of ruptured pancreaticoduodenal artery aneurysm associated with NF1 (Serleth et al. 1998; Fukushima et al. 2020).
To our best knowledge, the present case is first reported case of intraperitoneal bleeding due to spontaneous PAA rupture associated with NF1, successfully treated by TAE combined with stent-graft placement and IABO. Endovascular management is safe and effective at all ages, even in haemodynamically unstable patients with NF1 (Bargiela et al. 2018). Vital signs in the present case were unstable; therefore, we used IABO, which is effective in non-traumatic, as well as traumatic, intra-abdominal haemorrhage (Hoehn et al. 2019). However, her vital signs remained unstable under IABO after ASPDA embolization. Therefore, we controlled the flow of the SMA at the area of major extravasation with a compliant balloon. After this, her vital signs became stable, and endovascular treatment of the SMA system (AIPDA and PIPDA) could be performed.
In cases of neurofibromatosis-associated aneurysms, endovascular treatment with coiling is the most commonly used technique (55/66 cases; 83.3%), followed by stent-graft placement (10/66 cases; 15.2%) (Bargiela et al. 2018). Furthermore, in half of the latter cases, as well as in the present case, coil embolization was performed in addition to stent-graft placement. Thus, this treatment method should be considered in cases of rupture. We used a steerable microcatheter due to the steep angles of both the AIPDA and PIPDA, resulting in a successful catheterization. This device, which has a remote-controlled flexible tip, manipulated using a dial in the handgrip, is effective in the quick catheterization of such challenging vessel cases (Soyama et al. 2017).
Surgical control of haemorrhages in NF1 is reportedly difficult, due to the high fragility of the involved vessels with this disease (Hung et al. 2012; Aizawa et al. 2010). Therefore, if an arterial rupture is suspected, endovascular therapy should be considered (Hung et al. 2012).
Currently, there are no standard recommendations for routine follow up in NF-V. Vascular complications, including spontaneous rupture, can occur over time (Hung et al. 2012). A spontaneous arterial rupture is rare, but can sometimes be fatal (Roberts et al. 2019). Therefore, careful observation with imaging is necessary (Bargiela et al. 2018). In the present case, we planned to perform regular CT examinations once a year. Additionally, as she underwent stent-graft placement, there is a risk of SMA thrombotic occlusion; thus, lifelong antiplatelet therapy with aspirin (100 mg/day) is necessary. Good long-term stent patency was reported in 9 of 11 cases (mean follow-up, 28 months), including 1 case with stent placement in the SMA and antiplatelet therapy (100 mg/day aspirin) (Kunzle et al. 2013).
Based on the present case, PAA rupture should be considered as one of the potential conditions of NF-V. Furthermore, spontaneous PAA rupture associated with NF1 can be successfully treated by TAE combined with stent-graft placement and partial IABO.