Eur Stroke J. 2026 Oct 5;11(10):aakag130. doi: 10.1093/esj/aakag130.
ABSTRACT
INTRODUCTION: Cerebral cavernous malformations (CCMs) are lesions that may be incidental or present with haemorrhage or focal neurological deficits. Some authors have reported apparent lesion enlargement during pregnancy and hypothesised that hormonal fluctuations may promote CCM growth or haemorrhage. Whether pregnancy increases the risk of symptomatic haemorrhage (SH) from CCMs therefore represents an important clinical concern.
PATIENTS AND METHODS: We collected prospective data from patients with CCMs. A case-crossover framework was used to compare the risk of SH during pregnancy-related and non-pregnant observation periods. Pregnancy-related observation periods included pregnancy and the 6-week puerperium, while all other follow-up time after CCM diagnosis was considered non-pregnant. Each woman served as her own control, with the risk of SH during pregnancy-related periods compared with that during her non-pregnant follow-up periods.
RESULTS: After screening, the final cohort included 42 women diagnosed with a CCM at 46 years of age or younger. Of these, 23 women experienced at least 1 pregnancy (30.4% familial CCM; 21.7% brainstem CCM) at or after CCM diagnosis, accounting for 40 pregnancies. Overall, 2 SHs occurred during 25.6 pregnancy patient-years, corresponding to an annual SH rate of 7.8% (95% CI, 0.9-28.3), compared with 42 SHs during 420.7 non-pregnancy patient-years, corresponding to an annual SH rate of 10.0% (95% CI, 7.2-13.5). No statistically significant difference was found in the haemorrhagic rates between pregnant vs non-pregnant periods (P = .751). When observation periods were stratified according to a history of prior haemorrhage, no SHs occurred during 10.3 pregnancy patient-years among periods with prior haemorrhage, whereas 8 SHs occurred during 12.7 non-pregnancy patient-years, corresponding to an annual SH rate of 63.0% (95% CI, 27.2-124.2). Among periods without prior haemorrhage, 2 SHs occurred during 15.2 pregnancy patient-years, corresponding to an annual SH rate of 13.1% (95% CI, 1.6-47.5), compared with 34 SHs during 408.0 non-pregnancy patient-years, corresponding to an annual SH rate of 8.3% (95% CI, 5.8-11.6). In this subgroup, generalised estimating equation (GEE)-Poisson analysis showed no significant difference in haemorrhage risk between pregnancy and non-pregnancy periods (incidence rate ratio [IRR] 1.64, 95% CI, 0.44-6.13; P = .462). When analysis was restricted to brainstem CCMs, GEE-Poisson analysis showed no significant difference between pregnancy and non-pregnancy periods (IRR 2.09; 95% CI, 0.36-12.14; P = .411). One SH occurred during 8.2 pregnancy patient-years, corresponding to an annual SH rate of 12.2% (95% CI, 0.3-67.9), vs 6 SHs during 91.6 non-pregnancy patient-years, corresponding to 6.6% (95% CI, 2.4-14.3). Among periods with prior haemorrhage, no SHs occurred during 1.5 pregnancy patient-years (0.0%; 95% CI, 0.0-194.7), vs 2 SHs during 4.4 non-pregnancy patient-years (45.2%; 95% CI, 5.5-163.4). Among periods without prior haemorrhage, 1 SH occurred during 6.7 pregnancy patient-years (15.0%; 95% CI, 0.4-83.6), vs 4 SHs during 87.2 non-pregnancy patient-years (4.6%; 95% CI, 1.3-11.7), with no significant difference on GEE-Poisson analysis (IRR 3.50; 95% CI, 0.62-19.78; P = .156).
CONCLUSION: Pregnancy was not associated with a significantly increased rate of SH in this cohort of women with CCMs. However, the limited number of events during pregnancy and the limited pregnant follow-up require cautious interpretation. Counselling should remain individualised according to lesion location, prior haemorrhagic history, neurological status and treatment history.
PMID:42949441 | DOI:10.1093/esj/aakag130