Prognosis of septic cavernous sinus thrombosis remarkably improved: a case series of 12 patients and literature review

The most common presenting signs are orbital symptoms (proptosis, chemosis, impaired ocular motility) and cranial nerve impairment [12]. All cranial nerves passing through the cavernous sinus can be affected (III, IV, V1, V2, VI). In our series, the VIth cranial nerve is most frequently impaired. This can be explained by the fact that the VIth nerve runs medially through the sinus, while the other nerves pass through the lateral wall of the cavernous sinus, protected by a thick layer of dura [11‐13]. Symptoms of fever and headache also occur in the majority of patients [14].

CST can be difficult to distinguish from orbital cellulitis. Both can cause eyelid swelling, chemosis, proptosis, double vision, and reduced visual acuity. In orbital cellulitis, the symptoms are usually unilateral. In CST, there is often a unilateral onset with the development of bilateral complaints and symptoms in later stages. Additionally, abnormal pupil responses and papilledema are seen more frequently in cases of CST. Orbital cellulitis usually arises as a complication of paranasal sinusitis, whereas CST can be a complication from any infection in the head and neck area.

Other causes can mimic the symptoms of CST, including tumors, idiopathic orbital inflammation, Tolosa–Hunt Syndrome (inflammatory disease characterized by severe and unilateral headaches with orbital pain and ophthalmoplegia), internal carotid artery aneurysm and carotid-cavernous fistula [15, 16]. However, differentiation with septic CST is normally clear due to the absence of septic signs (such as fever) and the subacute or chronic course.

The diagnosis of CST is made by imaging. We used CT in 3 cases, MRI in 6, and the combination of both in 3 patients. Both direct and indirect signs on CT and MR can be seen in cases of CST. Direct signs are expansion of and filling defects in the cavernous sinus. Indirect signs, caused by venous obstruction, are dilatation of the superior ophthalmic vein, exophthalmos, and increased dural enhancement of the border of the cavernous sinus [17]. Associated findings are secondary thrombosis (for example of the superior ophthalmic vein, petrosal sinus, sphenoparietal sinus and sigmoid sinus) and narrowing of the internal carotid artery [17, 18]. These findings were also frequently found in our series. Besides narrowing of the internal carotid artery, occlusion and aneurysmal formation are also described [19]. We did not find these complications in our series.

Up to date literature about imaging modalities for CST is scarce. Contrast-enhanced CT and MRI are both used to diagnose CST. Eustis et al. reviewed all orbital complications of acute rhinosinusitis, including CST. They prefer MR imaging over CT for CST and describe that false negative CT scans are more common until late in the course of the disease. Their only reference, however, is a single case report which describes an early diagnosis of CST assessed by MR imaging [20].

Schuknecht et al. reviewed 8 patients with CST and compared CT and MR images [17]. They preferred contrast-enhanced CT over contrast-enhanced MR and use CT as primary imaging technique [17].

For cerebral venous thrombosis, T2 weighted MR with gradient echo (GRE) sequence (a sequence more sensitive to tissue susceptibility differences than regular T1 and T2 sequences) is found to be most sensitive for detection of thrombosis. However, in these series of 11, 17, and 45 patients, no patients with CST are included [21‐23].

We prefer contrast-enhanced MR imaging to diagnose CST since we found the imaging quality superior to CT imaging in patients with both modalities performed. In addition, potential other intracranial complications can be assessed on MR accurately as well. However, based on the data of this series we cannot draw any conclusion about the sensitivity of MR versus CT imaging.

The main component of the management of CST is treatment of the underlying infection with adequate intravenous antibiotics. Patients in our series received a cephalosporin or penicillin combined with metronidazole, to cover Staphlycoccus and Streptococcus species as well as Gram-negative and anaerobic species which are frequently involved [15].

Antifungal therapy is indicated in patients at risk for fungal infection; for example, immunocompromised patients or patients with poorly controlled diabetes mellitus.

Surgical treatment also focuses on the underlying infection, and may consist of drainage of sinus, mastoid, pharynx, etc., depending on the source of the infection.

In the last decades, AT is recommended for patients with CST by many authors [9, 11, 15, 24]. One should bear in mind however, that this recommendation is based on retrospective series and reviews, in the absence of clinical trials addressing this subject. In addition, many authors refer to older literature, namely Southwick [14] and Levine [25].

Southwick et al. reviewed case reports of patients with CST treated between 1940 and 1984. Of 86 reported cases, 32% were treated with heparin. Mortality was lower among patients treated with heparin. Of the 8 cases they treated in their own hospital, one received AT. This patient developed a large intracranial hemorrhage due to an overdose of anticoagulants [14].

Levine et al. reviewed their seven cases and literature cases between 1941 and 1988, and found no conclusive evidence for reduced mortality in patients treated with anticoagulation, but they did find a reduced morbidity [25].

In a more recent review of Weerashinge et al. from 2016, 47% of the 88 cases were treated with anticoagulation. The treatment varied in terms of medication and in duration of treatment. Comparing patients with and without anticoagulation, patients treated with anticoagulants had more chance on a full recovery and fewer patients died [9].

Trials performed to investigate if anticoagulation improves mortality and morbidity have been done in patients with cerebral venous thrombosis. In none of these trials, patients had septic thrombosis, let alone CST [26]. In our center, we treat CST with anticoagulants, based on a meta-analysis of two trials that showed a clear benefit of anticoagulants without bleeding complications [26]. We thus have generalized the results of the trials about cerebral venous thrombosis for CST. Although there are differences between these entities, there is considerable overlap in pathophysiology. We, therefore, use data on cerebral venous thrombosis in our considerations for the treatment of CST, as suggested by Desa et al. [27] In patients with meningitis, we usually refrain from anticoagulation, given the increased risk of intracerebral bleeding [28]. The decision whether or not to start with AT is preceded by consultation with a neurologist in every case. The optimal duration of AT is not known, but we normally treat patients for with anticoagulation for 3–6 months, which is a usual term internationally [29].

In our series of twelve patients, 1 died and 11 survived, of whom 9 without complications. CST has always been notorious for its bad prognosis. Most articles still refer to Southwick, describing a mortality of 50% of their 8 cases and 29% in their literature review of cases treated between 1940 and 1984 [14]. Yarington published a literature review of 878 patients (treated between 1821 and 1960), with a mortality of 80% and a morbidity of 75% [30]. Later, the same author published a series of 28 patients, with a mortality of 13.6% and morbidity of 22.7% [31]. These patients were treated up to 1977. Since then, the mortality rate in many articles has been described as 20–30% in the antibiotic era [11, 12, 15].

For more recent data we have a comprehensive review of case reports, with a mortality of 10/88 (11%) [9]. This review also reveals a high morbidity; of the 78 patients who survived 12/78 (15%) have residual symptoms, of which 8 patients had permanent blindness, 1 patient had double vision, 2 patients had hemiparesis, 1 patient had VIth nerve palsy. In a large multi-center study on cerebral venous thrombosis, the mortality is 8%, however, only a small proportion of the 624 patients had CST [3]. A review of cerebral venous thrombosis describes a mortality of 0–28% [32]. A higher mortality rate (> 5%) was associated with poor clinical condition upon admission (coma, cranial nerve deficits and seizures). The explanation of the decline in mortality of CST in the last decades is sought in improved treatment and identification of less serious case by better quality and availability of imaging techniques. Our study thus adds another contemporary series that shows that mortality is not so dramatically anymore, like several recent articles suggested [24, 33, 34]. Lizé et al. performed a retrospective study of patients with CST as a complication of sinusitis and reported no deaths among their seven patients [24]. Smith et al. published a case series of twelve children with CST and a literature review of pediatric cases. They report a mortality of 0% in their own series and 8% in pediatric cases published between 1994 and 2014. Morbidity rates were 25% in both their own patients and in the selected literature [33]. Wang et al. published a case report and literature review on CST caused by sinusitis between 2002 and 2015. They found a mortality rate of 0% and a morbidity rate of 37.5% [34]. Whether or not AT plays a role in the outcomes of CST is hard to assess by our small series. In the 7 patients treated with AT, we found complete or near complete recovery. However, 3 out of 5 patients not treated with anticoagulation also completely recovered. The one patient who died in the group not treated with anticoagulants was in very bad condition and can be regarded as an outlier. Since we did not find any hemorrhage complications of AT, and considering the findings in literature, we do favor AT in absence of contra-indications.