We conducted a retrospective cross-sectional study of adult (age >18 years) patients admitted to our hospital with CVST from 2010 to 2022 inclusively on patients with CVST who were managed in our hospital. Deidentified data on the patient's clinical presentation, imaging findings, treatment given, and clinical outcomes were directly entered into REDCap^®.

The Aga Khan University Institutional Scientific and Ethics Review Committee (Ref: 2022-ISERC-126), and the National Committee for Science, Technology and Innovation (NACOSTI, Ref No: 238026) granted ethical approval before the study commenced.

The majority of our CVST cohort was female and relatively young, comparable to other studies from within and outside SSA (Baduro and Ferro, 2021). More than 80% were of Black African origin, providing adequate representation of the local population.

The top three symptoms at presentation were headache, focal neurological deficits, and seizures, which are, again, comparable to case series from SSA (Baduro and Ferro, 2021). However, we did not find any cases of cavernous sinus syndrome (CSS), which has been more often described, in up to 13.8% of cases, in other SSA cohorts. This could be because CSS is usually infective in nature, and infections were less implicated in our cohort (18% vs. 63.1% for other case series from SSA).

When comparing our cohort to other SSA cohorts, frequencies of contraceptive use (8.5% vs. 7.3%), pregnancy- and puerperium-related complications (9.8% vs. 6.2%), and traumatic and iatrogenic (5.6% vs. 0.8%) were not significantly different (Baduro and Ferro, 2021). However, we had more cases attributed to underlying systemic inflammatory/connective tissue diseases (9.0% vs. 3.7%) and malignancies (8.2% vs. 1.5%). Our center is a regional referral hospital for rheumatological and oncological conditions and has international-standard diagnostic services, which may explain this higher level of diagnosis (Pathology Laboratory Medicine, 2015). As has been described (Zhang et al., 2023), CVST with underlying malignancy, especially hematological, was associated with a poorer outcome in our cohort.

We found blood-borne viral infections, including HIV and hepatitis, to be the main infective risk factor for CVST in our cohort, similar to what has been found in other parts of the continent (Baduro and Ferro, 2021; Gaye et al., 2021; Napon et al., 2010) but different from the findings of the other Kenyan study (Onyambu et al., 2013), which was CVST being more common with direct brain infections such as tuberculosis. This within-country difference in the type of infection associated with CVST is likely due to the other Kenyan cohort being from a large public hospital where tuberculosis and other HIV-related opportunistic infections are more common (Chepkondol et al., 2020).

Of considerable significance was the number of patients newly diagnosed with thrombophilia in our cohort (30.3%), with protein S deficiency being the most common diagnosis (16.4%). This is an important finding brought about by routine thrombophilia testing in all CVST patients, contrary to what has been suggested by the latest European guidelines (Ferro et al., 2017). In comparison, the SSA cohorts had thrombophilia detected in 2.2%, with protein S deficiency being slightly less common than anti-phospholipid syndrome (Baduro and Ferro, 2021). The larger North African and Mexican cohorts had similar low findings of thrombophilia (Sassi et al., 2017; Ruiz-Sandoval et al., 2012); our cohort is similar to what has been found in, for example, European studies (Ferro et al., 2004).

A large proportion of our patients (83.6%) had contrast-enhanced MRI brain scan with venography and angiography, compared to 20.5% having contrast-enhanced CT head scans with or without venography. This contrasts significantly with other local studies, where the most common modality was a head CT scan with contrast, rarely with venography, and no MRI scans (Onyambu et al., 2013). This reflects the variability of MR diagnostic services available in private hospitals vs. public hospitals in SSA, including in Kenya (Hasford et al., 2022). The superior sagittal sinus was the dominant sinus involved (51.6%), as is the case with other cohorts (Baduro and Ferro, 2021; Ruiz-Sandoval et al., 2012; Ferro et al., 2004; Duman et al., 2017; Khealani et al., 2008). The most common secondary parenchymal brain lesions were hemorrhage (60.7%) followed by what has usually been reported in other studies: venous infarction (51.8%), brain oedema (26.8%), and herniation (7.1%) (Gaye et al., 2021; Wasay et al., 2019). The higher rates of hemorrhage are probably due to the main non-communicable disease comorbidity being hypertension in our patients (16.4%). Interestingly, patients with hypertension were significantly less likely to have complete resolution of CVST at follow-up imaging (p = 0.003), a finding worth exploring in larger future studies to identify the nature behind this relationship.

In the acute phase, all our patients were started on low-molecular-weight heparin (LMWH); one was treated with fondaparinux due to associated comorbidities. Slightly more than half transitioned to warfarin, and the remainder were treated with direct oral anticoagulants (DOACs), even though the evidence for DOAC use in CVST has only been published in recent years (Ferro et al., 2019; Field et al., 2023). DOAC preference was probably because the infrastructure for monitoring warfarin treatment in our setting is poor, a systemic problem across SSA (Mouton et al., 2021; Tadesse et al., 2022). Patients from our hospital are more likely to be prescribed DOACs due to a number of factors, including the ability to afford them compared to other hospitals (Mucyo et al., 2022). It is noteworthy that the patients in our study who were managed with warfarin had slightly better clot resolution rates, but this was not statistically significant (p = 0.175).

The majority (82.5%, 33/40) of patients who experienced seizures were treated with levetiracetam for a median duration of 16.0 (10.0, 24.0) months, in keeping with international guidelines (Ferro et al., 2017), chiefly to avoid interactions with anticoagulation using other anti-seizure medications. Only 20% of these patients had seizures beyond the first 3 months, as is expected in acute symptomatic seizures. However, headaches remained a significant problem for most patients throughout their treatment (77.1%), even though only a small majority had premorbid primary headache conditions. This is an increasingly recognized phenomenon globally (Dias et al., 2024).

We had a higher proportion of patients having good outcomes compared to other studies from the continent (Baduro and Ferro, 2021; Gaye et al., 2021; Sassi et al., 2017): 92.7% had a modified Rankin Scale (mRS) score of 0–2, of which 81.2% had no disability (mRS score: 0–1) at last follow-up. We suspect this could be because of the ability of our center to reach an early diagnosis compared to other SSA cohorts:

There are inherent limitations in our study: