Clinical Profiles and Treatment Outcomes of 51 Cases of Carotid Cavernous Fistula: A Retrospective Observational Study

Article information

Abstract

Materials and Methods

Results

A total of 61 patients diagnosed with CCF were documented. Of which, 10 patients were excluded due to incomplete medical record data, leaving 51 patients in this study. Patient demographics and clinical characteristics are shown in Table 1. Traumatic CCF (34 patients, 66.7%) was more common than spontaneous CCF (17 patients, 33.3%). The mean age of patients with traumatic CCF was 36.8 ± 17.08 years, while that of patients with spontaneous CCF was 54.8 ± 11.6 years, though the difference was not statistically significant. Spontaneous CCF was more commonly found in female patients, while traumatic CCF was predominant in male patients. (p = 0.005). The majority of CCFs were unilateral, both in traumatic and spontaneous CCF. In traumatic CCF, the median duration between trauma and the onset of symptoms was 4.5 weeks (range, 0–516 weeks).

Table 1

Patient demographic and clinical characteristics (n = 51)

In traumatic CCF, the most common symptom was proptosis (73.5%), followed by blurred vision and red eyes (both 11.8%). Proptosis was also the most common symptom in spontaneous CCF (47.1%), but red eyes (29.4%) and diplopia (17.6%) were more prevalent compared to traumatic CCF. However, the differences were not statistically significant (p > 0.05). Proptosis and lagophthalmos were more commonly found in traumatic CCF than in spontaneous CCF; however, there were no statistically significant difference (p > 0.05). Conjunctival abnormalities were highly prevalent in both groups. Eye movement restriction and bruit were significantly more common in traumatic CCF than in spontaneous CCF (p = 0.006 and p = 0.008, respectively). BCVA and IOP were not statistically different in both groups. Antiglaucoma treatment was given to 24 patients in traumatic CCF and 8 patients in spontaneous CCF (p = 0.101).

Various neuroimaging techniques were used to diagnose CCF, including computed tomography angiography (CTA) in most cases (29 cases), followed by CT (11 cases), DSA (8 cases), magnetic resonance imaging (2 cases), and magnetic resonance angiography (1 case). The neuroimaging findings in CCF are shown in Table 2. Superior ophthalmic vein and cavernous sinus dilatation were found in both groups equally. Other veins dilatation had a higher proportion in traumatic CCF but were not statistically significant. Of the 51 patients, 35 underwent DSA, while the remaining 16 had not undergone DSA at the time of data collection. CCF was not found in 5 patients, and an incomplete DSA report was found in 3 patients; thus, 27 patients could be classified under Barrow classification. Most patients in traumatic CCF had Barrow type A classification and direct classification (both p < 0.001) (Table 2). The etiology of all direct CCF were trauma. In spontaneous CCF, all patients had indirect anatomical classification and Barrow type B, C, or D (both p < 0.001). In traumatic CCF, one patient had Barrow type B, and the other had type D. In spontaneous CCF, six patients had type B, and the other had type C. Among the nine patients with indirect CCF, the identified systemic risk factors were hyperlipidemia (four patients), women over 50 years of age (four patients), diabetes mellitus (three patients), hypertension (three patients), and a history of trauma (two patients).

Table 2

Neuroimaging findings, Barrow classification, and anatomical classification

Treatment modalities were divided into conservative and endovascular treatments (Table 3). Of the 46 patients, 27 were managed with conservative treatment, whereas 19 received endovascular treatment. Notably, 16 of the 27 patients undergoing conservative treatment had not undergone DSA. In most patients, endovascular treatment was performed via embolization in 19 patients. Two patients had been embolized in this group, but the fistula could not be closed because of its sharp angulation and large fistula size. In traumatic and spontaneous CCF, 17 (53.1%) and 2 (14.3%) patients, respectively, underwent endovascular treatment. The median time from diagnosis of CCF to DSA and/or endovascular treatment was 29 weeks (range, 2–118). Based on the anatomical classification (direct and indirect), in patients who underwent DSA (27 patients), conservative treatment was performed in the majority of indirect CCF (5 out of 8), while endovascular treatment was performed in the majority of direct CCF (15 out of 19).

Table 3

Treatment modalities

Treatment outcomes in patients with traumatic CCF who underwent either conservative or endovascular treatment are shown in Table 4. Before treatment, there was no statistically significant difference in the proportion of clinical signs between the conservative and endovascular groups. After treatment, however, the proportions of proptosis, conjunctival abnormalities, and bruit were significantly lower in the endovascular treatment group compared to the conservative group (p < 0.05). Table 5 shows the treatment outcomes in patients with spontaneous CCF who underwent conservative or endovascular treatment. There was no statistically significant difference in clinical signs between the conservative and endovascular treatment groups in pretreatment and posttreatment. Table 6 shows conservative treatment outcomes in traumatic and spontaneous CCF. We found no statistically significant difference in the proportion of clinical signs between spontaneous and traumatic CCF (p > 0.05).

Table 4

Treatment outcomes in traumatic carotid cavernous fistula (n = 32)

Table 5

Treatment outcomes in spontaneous carotid cavernous fistula (n = 14)

Table 6

Conservative treatment outcomes in traumatic and spontaneous CCF (n = 27)

Discussion

Traumatic CCF accounts for 75% to 80% of patients [6]. Yuneta et al. [10] also reported that 91.3% of 23 patients had trauma, which corresponds with our results. Traumatic CCF occurs in younger patients, although the difference was not statistically significant. This corresponds to the results of studies by Yuneta et al. [10] and Holland et al. [11]. Trauma commonly causes CCF in young adults owing to its high activity and mobility. Spontaneous CCF commonly occurs in elderly patients. Alexander et al. [12] reported that indirect CCF averaged 60.9 years old. Spontaneous CCF is related to hypertension, diabetes, and atherosclerosis [4,6]. Most spontaneous CCF were women, while traumatic CCF mainly occurred in men (p < 0.05), which corresponds with Oishi et al. [13] who reported 81.1% of spontaneous CCF patients were women. Pregnancy, peripartum, and menopause are the risk factors for spontaneous CCF. Reduced estrogen levels may produce dura mater’s small blood vessel atherosclerosis and collateral circulation from arteries to veins [14].

Most CCF symptoms and clinical signs originate from the arteriovenous blood flow shunts. Eye movement and bruit were significantly higher in traumatic CCF than in spontaneous CCF, whereas conjunctival abnormalities were high in both groups. Similar results were reported by Chi et al. [15], who found proptosis, cranial nerve palsy, and red eyes in 80.8%, 55.2%, and 55% of 172 traumatic CCF cases, respectively, with 98.8% having bruits [2]. Venous congestion in the cavernous sinuses causes hypertension in the surrounding veins, including the superior ophthalmic vein. Reduced venous flow from the eye and surrounding tissues causes proptosis and arterialization of conjunctival blood vessels. The cranial nerves are compressed by excessive cavernous sinus pressure, causing ophthalmoplegia and hypoesthesia. Chronic venous congestion can increase extraocular muscles, restricting eye movement [4,16,17].

Neuroimaging was performed to diagnose CCF. CTA was performed in most cases in this study. CTA and DSA share similarities; they have high sensitivity, specificity, and accuracy and can accurately detect fistula size [18]. Superior ophthalmic veins and cavernous sinus dilatation were equally found in both groups. Holland et al. [11] reported cavernous sinus dilatation and superior vein dilatation in direct and indirect CCF. Although DSA is the gold standard for the diagnosis of CCF, it was only conducted in 68.6% of patients during data collection due to the lengthy waiting period for DSA and/or embolization procedures. Most patients with traumatic CCF were classified as Barrow type A (p < 0.05). In contrast, all spontaneous CCF cases had indirect anatomical classifications and Barrow types B, C, or D (p < 0.05). This corresponds with Holland et al. [11], who reported that direct CCF was 100% Barrow type A, while indirect CCF (23 patients) was Barrow type B, C, or D.

Most indirect or dural CCF arise spontaneously when the thin-walled dural artery wall crossing the cavernous sinuses ruptures. This rupture dilates the dural artery anastomosis and provides collateral blood. Spontaneous venous thrombosis can induce collateral venous flow in the cavernous sinus. This study found that women over the age of 50 years, hyperlipidemia, diabetes, hypertension, and trauma were indirect CCF patients. Prasad et al. [19] identified hypertension, diabetes, and postmenopause as indirect CCF risk factors.

Treatment depends on the fistula type, anatomy, size, and operator or institution preferences. [17]. This study divided treatment into conservative and endovascular treatments. Endovsascular treatment aims to close the fistula while maintaining the internal carotid artery blood flow through transarterial or transvenous obliteration with a balloon, a coil, or embolic material [17]. This study showed that the proportion of proptosis, conjunctival abnormalities, and bruit in the endovascular treatment of the traumatic CCF group was significantly lower than that of the conservative group (p < 0.05). Permana et al. [20] found that 28 traumatic CCF patients who received balloon or coil endovascular treatment had fewer proptosis, chemosis, bruit, and third and sixth cranial nerve palsies. Holland et al. [11] observed a reduction in the number of patients exhibiting proptosis, restricted eye movement, pulsatile tinnitus, and elevated IOP at 6 months following endovascular treatment. The vision improved significantly after endovascular treatment. Yuneta et al. [10] discovered that endovascular treatment improved 11 of 12 CCF patients. A closed fistula reduces or stops carotid artery-cavernous sinus blood flow, causing a reduction in venous congestion, which results in the improvement of clinical symptoms and signs [20].

Patients waiting for or have undergone DSA procedures are treated conservatively. Conservative treatment is generally used for low-flow fistulas. Medical treatment includes reduced IOP and lubricants to maintain ocular integrity. External carotid-jugular manual compression with the contralateral hand can be performed on indirect CCF for 10 seconds, four to six times per hour. The goal is to reduce arteriovenous shunts, decrease arterial inflow, and increase venous outlet pressure to promote fistula thrombosis. Manual compression treats 30% of spontaneous CCF cases. Compression treatment fails in the high-flow CCF, which requires endovascular treatment [17]. Manual compression data compliance was infrequently recorded in medical records. At the final follow-up, there was no significant difference in the proportion of clinical signs between the conservative and endovascular treatment. This study has a variable follow-up time, which may affect conservative treatment outcomes. Iampreechakul et al. [21] reported 46 spontaneously resolved CCF cases, of which 33 (71.3%) were traumatic, 11 (24.0%) were spontaneous, and 2 (4.3%) were iatrogenic cases. Rarely, venous stasis and tissue damage may cause cavernous sinus thrombosis and venous drainage, causing spontaneous direct CCF regression. Conservative treatment is performed only in patients with normal or mild visual impairment, small fistulas, low flow, cavernous sinus thrombosis, or venous drainage, which usually resolves spontaneously.

This study has limitations because it was a retrospective study that obtained data from the medical records. The completeness of the data in terms of ophthalmological examination and neuroimaging results varied during the data collection. Different follow-up durations may affect clinical profiles and treatment results.

In conclusion, spontaneous CCF is less prevalent than traumatic CCF. Traumatic and spontaneous CCF may have different clinical characteristics. Eye movement restriction and bruit are more prevalent in traumatic CCF. Most traumatic CCF patients exhibited Barrow classification type A, while all spontaneous CCF were type B, C, or D Barrow classification. Endovascular treatment improved clinical signs during posttreatment follow-up. After treatment of traumatic CCF, the proportion of proptosis, conjunctival abnormalities, and bruit in the endovascular treatment group was significantly lower than that in the conservative group.

Supplementary Materials

Article information Continued