The Incidence of Cerebral Venous Thrombosis: A Cross-Sectional Study

The incidence of cerebral venous thrombosis (CVT) is estimated at 0.2 to 0.5 per 100 000 per year.^1,2 One estimate was derived from an extrapolation of mortality and autopsy data from studies performed several decades ago.³ Kalbag and Woolf⁴ provide mortality figures from the British Registrar General. From 1952 to 1961 an average of 21.7 deaths from CVT were reported annually in a population of 56 million (0.39 deaths per million). The mortality of CVT probably varied between 20% and 50%. This gives an incidence of approximately 0.1 to 0.2 cases per 100 000. These indirectly calculated incidence rates are inaccurate, especially because improved neuroimaging has shown that, unlike previously believed, CVT often has a benign course. Other studies give estimates of 0.22,⁵ 0.34,⁶ and 1.23⁷ per 100 000, but in most of these studies, determination of the incidence was not the primary objective. We determined the incidence of CVT among adults in 2 provinces in The Netherlands over a 3-year period by hand-searching the medical records of all 19 hospitals in this region.

We identified 9270 potential CVT cases (Figure). After hand-searching all medical records, we identified 147 patients diagnosed with CVT. Fifty-three patients did not meet the eligibility criteria; therefore, 94 patients were included in the analysis. The combined adult population of the 2 provinces was 2 353 429 in 2008, 2 379 236 in 2009, and 2 405 611 in 2010. The overall annual incidence of CVT among adults was 1.32 per 100 000 person-years (95% CI, 1.06–1.61; Table 1). As expected, the incidence was significantly higher in women than men (1.86 versus 0.75) and higher among patients aged 31 to 50 years (1.71).

The median age of patients was 41 years (Table 2). Fifty-two percent of female patients used oral contraceptives and 18% were pregnant or had recently given birth. The transverse and sigmoid sinuses were thrombosed most often (70% and 53%, respectively). Parenchymal lesions at baseline occurred in 44%. Nearly all patients (96%) received anticoagulant treatment. Mortality was 1% at discharge and 3% at follow-up.

We found an incidence of adult cerebral venous thrombosis of 1.32 per 100 000 person-years, much higher than previously published. This indicates that CVT in adults has an incidence comparable to bacterial meningitis.⁸ A possible explanation is that due to an increased awareness of CVT and improved imaging techniques, CVT is more frequently diagnosed. One of the strengths of our study is that we hand-searched the medical records of almost 10 000 potential CVT cases and used strict inclusion and diagnostic criteria. Because exact figures on population distribution are available in The Netherlands, we believe that our data provide a reliable estimate of the true incidence of CVT.

Several well-designed studies have examined the incidence of CVT among children. The Canadian Pediatric Ischemic Stroke Registry found an incidence of 0.67 per 100 000 per year with a peak among neonates.⁹ As a result, CVT was always believed to be more common among children than adults. In contrast, our data suggest that CVT is twice as common in adults as in children.

One previous study found an incidence similar to our study. Janghorbani et al⁷ performed a case registry in 2 hospitals in Iran and reported an incidence of 1.23 per 100 000. Their study, however, was performed in tertiary care hospitals, which may have caused referral bias. Furthermore, because recent data on population were lacking, they calculated their denominator based on an older census, which may be less accurate. Finally, it is unclear if they checked for duplicate counts or referrals from outside the caption area.

The mortality in our study, 1% at discharge and 3% at follow-up, is lower than in previous studies. In the International Study on Cerebral Vein and Dural Sinus Thrombosis (ISCVT), mortality was 4% at discharge and 8% at follow-up.¹⁰ This is in agreement with the idea that improved diagnosis has identified not only more cases, but also less severe cases with a benign course. Possibly, the almost universal application of anticoagulant treatment (96%) may have contributed to the low mortality.

Our study has several limitations. First, despite our extensive search, we cannot exclude that some cases have been missed. For instance, patients who were miscoded, who were admitted to a hospital outside the search region, or who were not treated by a neurologist would all have been missed. Thus, the incidence we found may be an underestimate of the true incidence. Another limitation is the retrospective design of our study, but because we manually checked the original medical records and imaging, we feel it is unlikely that this introduced bias.

We thank the following people for their help in the acquisition of data: M. C. Visser (VU Medical Centre); H. Weinstein (St. Lucas-Andreas Hospital); N. D. Kruyt (Slotervaart Hospital); V. I. H. Kwa (Onze Lieve Vrouwe Gasthuis); M. Janmaat and M. M. van der Graaff (Bovenij Hospital); P. Lanting (Amstelland Hospital); W. van der Meulen (Rode Kruis Hospital); T. C. van der Ree (Westfries Gasthuis); E. Wouda (Zaans Medical Center); I. Merkies (Spaarne Hospital); S. Heckenberg (Kennemergasthuis); P. M. Laboyrie (Tergooi Hospital); F. Scholtes (Gemini Hospital); R. Tans (MC Group); M. Limburg (Flevo Hospital); D. Molenaar (Amstelland Hospital) and C. Zwetsloot (Waterland Hospital); and D. Brandsma (Antoni van Leeuwenhoek Hospital). We thank K. Stronks for her helpful comments on the article.