Risk of Clinical Events in Presymptomatic Familial Cerebral Cavernous Malformations

ABSTRACT

Background and Purpose

Familial cerebral cavernous malformations (fCCM) may cause focal neurological deficits (FND) due to hemorrhagic or non‐hemorrhagic events and epileptic seizures (ES). With the widespread use of magnetic resonance imaging (MRI) and genetic testing, diagnosis is now frequently established at a presymptomatic stage, before any clinical manifestation. However, the subsequent risk of a first clinical event in these individuals remains poorly defined.

Methods

We conducted a cohort‐based analysis of retrospectively collected longitudinal data from 79 patients with fCCM who were asymptomatic at diagnosis (median age, 41.7 years) and managed in a national referral center.

Results

Over 649.6 patient‐years (median follow‐up, 4.0 years), 23 patients (29%) experienced a first clinical event. Post‐diagnostic event‐free survival was approximately 97% at 1 year, 79% at 5 years, and 69% at 10 years. Incidence rates for a first event were 35.4 per 1000 person‐years for any first event (FND or ES), 23.1 per 1000 person‐years for a first FND, and 15.4 per 1000 person‐years for a first ES. Male sex was independently associated with higher risk (HR 4.20; 95% CI, 1.60–11.04; p = 0.0036).

Conclusions

These data provide pragmatic post‐diagnostic risk estimates to guide counseling and surveillance in asymptomatic fCCM.

1. Introduction

Familial cerebral cavernous malformation (fCCM) is an autosomal dominant condition characterized by the progressive development of multiple cavernous malformations (CCM) in the central nervous system. CCM lesions may remain silent or cause neurological complications, including symptomatic intracerebral hemorrhage, focal neurological deficits of non‐hemorrhagic origin, and epilepsy [1, 2]. Although some patients are diagnosed at the presymptomatic stage, either after incidental imaging or through presymptomatic genetic testing, the exact risk of a first clinical event in these individuals has not been well established. We therefore aimed to estimate event‐free survival among presymptomatic subjects following fCCM diagnosis.

2. Patients and Methods

Retrospective data were obtained from a cohort study of fCCM patients evaluated between 2010 and 2025 at the national referral center for rare vascular diseases of brain and eyes in France (CERVCO; www.cervco.fr). The fCCM patients included in the present study were selected based on the following criteria: (1) presence of multiple non‐contiguous cerebral CCM on magnetic resonance imaging (MRI) and/or evidence of a pathogenic variant in CCM1 (KRIT1), CCM2 (MGC4607), or CCM3 (PDCD10), (2) no history of focal neurological deficit (FND) or epileptic seizure (ES) at time of diagnosis. The date of diagnosis corresponded to the earliest documented evidence of disease, defined by either the first MRI suggestive of CCM or confirmation by genetic testing. All patients gave informed consent to participate in this cohort study, approved by the Inserm Institutional Review Board (IRB No. 0003888).

In each patient, the following data were systematically assessed: sex, genotype, age at diagnosis, age at last visit, occurrence and age at first clinical event (FND or ES) presumably related to fCCM. Only FND (related to hemorrhage or mass effect) and ES were considered, as they constitute objective and clinically significant neurological events, unlike more nonspecific symptoms such as headache. Patients without follow‐up data since 2020 were contacted by phone using standardized interviews to update outcome status and age at last evaluation (25 contacted; 8 responded). The primary outcome was the occurrence of a first clinical event (FND or ES) after diagnosis. Follow‐up was calculated from the date of diagnosis (T0) to the first event; patients who remained event‐free were censored at their last available follow‐up.

Event‐free survival was estimated using the Kaplan–Meier method, and reported at 1, 5, and 10 years after diagnosis (95% CI, log transformation). To assess the potential impact of event‐driven referral bias, we performed a sensitivity analysis excluding patients whose first event occurred before their first evaluation at the referral center and who were referred within 2 years after that first event. Incidence rates of clinical events within the cohort were computed for the composite endpoint and separately for first FND and first ES. Rates were expressed per 1000 person‐years with 95% confidence intervals based on a Poisson approximation. Kaplan–Meier curves were compared across predefined subgroups, including sex, genotype (CCM1 versus CCM2), and age at diagnosis (dichotomized at the median). Differences between groups were assessed using the log‐rank test. In addition, an exploratory multivariable Cox proportional hazards model was constructed including sex, genotype, and age at diagnosis as covariates. All analyses were performed using R version 4.5.2.

3. Results

Seventy‐nine patients with fCCM who were asymptomatic (i.e., without any history of FND or ES) at diagnosis were included in the present study. Diagnosis of fCCM on brain MRI was incidental in 48 patients (61%) and made in context of presymptomatic screening (brain MRI and/or genetic screening) in 31 patients (39%). CCM were identified on MRI in all patients except one, who was genetically diagnosed through presymptomatic family screening and had not undergone MRI. Median age at diagnosis was 41.7 years (interquartile range [IQR], 30.9–53.5). Genotype was available for 75 patients; 49 (62%) carried a pathogenic variant in CCM1, 25 (32%) in CCM2, and 1 (1%) in CCM3. In one patient, genetic diagnosis had been established outside our center, but the genotype was unavailable. In three patients, genetic testing was not available, but the diagnosis was considered definite based on evidence of multiple CCM on MRI and a consistent family history.

Over a median follow‐up of 4.0 years (IQR, 2.4–13.2) from diagnosis, 23 patients (29%) experienced a first clinical event (isolated FND in 13, isolated ES in 8, combination of FND‐ES in 2). Among them, the first event occurred before the first visit at the CERVCO referral center in 12 patients, and after the first visit in 11 patients. Fifty‐six patients (71%) remained asymptomatic during follow‐up period, with a high heterogeneous duration of follow‐up across patients (median duration of 3.7 years; IQR, 2.0–12.2; Table 1). Event‐free survival after diagnosis was 97.4% (95% CI, 89.9–99.3) at 1 year, 78.6% (95% CI, 65.1–87.4) at 5 years, and 68.7% (95% CI, 53.5–79.9) at 10 years (Figure 1A).

TABLE 1.

Patient characteristics and clinical outcomes.

	Total (n = 79)	Remained asymptomatic (n = 56)	Developed first clinical event (n = 23)
Sex, n (%)
Female	48 (61%)	41 (73%)	7 (30%)
Male	31 (39%)	15 (27%)	16 (70%)
Age at diagnosis, years, median (IQR)	41.7 (30.9–53.5)	42.1 (33.2–54.2)	39.3 (23.7–50.8)
Age at event or last follow‐up, years, median (IQR)	50.4 (39.9–58.2)	50.7 (41.5–63.8)	46.8 (36.7–56.8)
Time from diagnosis to first event or censoring, median (IQR)	4.0 (2.4–13.2)	3.7 (2.0–12.2)	4.4 (3.4–14.9)
Genotype, n (%)
CCM1	49 (62%)	35 (63%)	14 (61%)
CCM2	25 (32%)	16 (29%)	9 (39%)
CCM3	1 (1%)	1 (2%)	0
Not available	4 (5%)	4 (7%)	0
Mode of diagnosis, n (%)
Incidental	48 (61%)	36 (64%)	12 (52%)
Presymptomatic	31 (39%)	20 (36%)	11 (48%)
Type of first clinical event, n (%)	—	—
FND	—	—	15 (65%)
ES	—	—	10 (43%)

Note: Data are presented as median (interquartile range) or number (%). Analyses comparing groups are descriptive. Age at event or last follow‐up corresponds to age at first clinical event for symptomatic patients and age at last available follow‐up for asymptomatic patients.

Abbreviations: ES, Epileptic seizure; FND, Focal neurological deficit; IQR, Interquartile range.

FIGURE 1.

Event‐free survival after diagnosis in patients with asymptomatic familial cerebral cavernous malformations. Kaplan–Meier estimates of event‐free survival in patients with familial cerebral cavernous malformations who were asymptomatic at diagnosis. Panel A shows the main analysis including all 79 patients. Panel B shows a sensitivity analysis excluding patients referred to the tertiary center within ≤ 2 years after their first clinical event (n = 73). Panels C–E present Kaplan–Meier curves stratified by sex (C), genotype (CCM1 vs. CCM2) (D), and age at diagnosis dichotomized at the median (E). Follow‐up time is shown in years after diagnosis; time zero corresponds to the time of diagnosis. The event of interest was the first clinical event. Patients without events were censored at last follow‐up. Shaded areas represent 95% confidence intervals. Numbers at risk are shown below each plot.

Estimates were similar in the sensitivity analysis excluding 6 patients referred to CERVCO within 2 years after their first clinical event. In this analysis, event‐free survival after diagnosis was similar to the main analysis with 97.2% (95% CI, 89.1–99.3) at 1 year, 80.0% (95% CI, 65.6–88.9) at 5 years, and 68.5% (95% CI, 51.8–80.4) at 10 years (Figure 1B).

Across 649.6 patient‐years, the incidence rate for any first event (FND and/or ES) was 35.4 per 1000 patient‐years (95% CI, 22.4–53.1). The incidence rate was 23.1 per 1000 patient‐years for first FND (95% CI, 12.9–38.1) and 15.4 per 1000 patient‐years for first ES (95% CI, 7.4–28.3).

Kaplan–Meier curves stratified by sex, genotype and age at diagnosis are presented in Figure 1C–E. Event‐free survival appeared lower in males compared with females. In contrast, no clear differences were observed according to genotype or age at diagnosis, with overlapping confidence intervals between groups. In a multivariable Cox model, male sex was independently associated with a higher risk of a first clinical event after accounting for age at diagnosis and genotype (HR, 4.20; 95% CI, 1.60–11.04; p = 0.0036).

4. Discussion

In our study, we observed that at the presymptomatic stage of fCCM, post‐diagnostic event‐free survival was approximately 97% at 1 year, 79% at 5 years, and 69% at 10 years in a cohort study of individuals recruited at a national referral center for rare cerebrovascular diseases. Incidence rates for a first event were also estimated at 35.4 per 1000 person‐years for any first event, with 23.1 per 1000 person‐years for a first FND and 15.4 per 1000 person‐years for a first ES. These results are not comparable to those obtained in most longitudinal CCM studies that included both symptomatic and asymptomatic patients or considered together both first and recurrent events [3, 4, 5, 6]. Clinical outcomes were previously reported in only 33 asymptomatic patients with fCCM among whom two experienced a first clinical event (one symptomatic hemorrhage and one ES) [7]. Both the sample size and the mean follow‐up were, however, substantially lower than in our study.

Interestingly, a recent large international multicenter retrospective hospital‐based cohort study, including 1592 fCCM patients, evaluated the lifetime risk of clinical events in a broad multicenter fCCM cohort, regardless of whether the diagnosis of fCCM was made before or following a hemorrhagic or epileptic event. They showed incidence rates of 11.8 events per 1000 patient‐years for a first symptomatic hemorrhage and 6.8 events per 1000 patient‐years for a first ES [6]. However, this hospital‐based study, incorporating neurosurgical series (32% of patients having undergone interventional treatment), was likely enriched for patients diagnosed following a clinical event, thereby potentially inflating estimates of event risk in the broader fCCM population. In contrast, our cohort included a substantial proportion of patients referred for genetic diagnosis or genetic counseling rather than for therapeutic decisions related to clinical events. Moreover, our study design, which restricted inclusion to patients diagnosed prior to any clinical event, was intended to limit event‐driven diagnosis bias, another potential source of overestimation of event risk in fCCM. Although methodological differences preclude direct comparison with the study by Dammann et al. [6], our results suggest that first‐event risk among presymptomatically diagnosed patients may not substantially differ from that observed in unselected tertiary‐care fCCM cohorts including patients diagnosed either presymptomatically or in the context of clinical symptoms. This observation remained consistent in a sensitivity analysis excluding patients referred shortly after a first event, suggesting that the impact of event‐driven referral bias on our estimates was limited.

Interestingly, an exploratory multivariable analysis in our cohort identified male sex as independently associated with a higher risk of a first post‐diagnostic clinical event. To our knowledge, this association has not been reported in longitudinal studies of fCCM. These findings may suggest that sex could influence the onset of clinical symptoms in fCCM. However, given the monocentric design, the modest sample size and the limited number of events, these results should be interpreted with caution and do not allow definitive conclusions.

Importantly, our findings may not be generalizable to all asymptomatic fCCM carriers at the population level. Recruitment through a referral center may have introduced selection bias, favoring patients with characteristics influencing prognosis, such as a more severe family history or higher lesion burden. In addition, the predominantly adult population and the very limited representation of PDCD10 carriers (n = 1), which precluded genotype‐specific analyses for this subgroup, further restrict extrapolation. Accordingly, our estimates should be interpreted as pragmatic post‐diagnostic risks in CCM1 or CCM2 mutation carriers managed in a specialized adult care setting, rather than as population‐based penetrance. Moreover, the impact of MRI lesion burden, which may potentially influence clinical outcomes, was not addressed in the present analysis and should be considered as a limitation.

Interpretation of our findings requires consideration of additional limitations. Post‐diagnostic follow‐up duration was overall limited, as expected in an asymptomatic population more likely to discontinue tertiary‐center follow‐up. Notably, 17 patients were lost to follow‐up since 2020 and could not be reached by phone (with a median follow‐up of 3.11 years, IQR 0.91–3.89), leading to right‐censored observations and potentially affecting event detection.

In addition, event‐driven referral may have influenced our estimates, as 12 patients were referred to CERVCO after the first event. However, estimates were similar in a sensitivity analysis excluding the 6 patients referred to CERVCO within 2 years of their first event, suggesting that the impact of event‐driven referral bias was limited.

In conclusion, this study provides clinically relevant estimates of first‐event risk after presymptomatic fCCM diagnosis. These results offer a foundation for patient counseling and individualized follow‐up, to be refined by future large prospective studies.

Author Contributions

Estelle Dubus: data curation. Nassira Alili: investigation, resources. Constantina Rodica Popa: data curation, resources, investigation, writing – original draft. Abbas Taleb: data curation. Hugues Chabriat: investigation, methodology, resources, project administration, supervision, writing – review and editing. Fanny Fernandes: data curation. Dominique Hervé: investigation, resources. Stéphanie Guey: conceptualization, data curation, resources, investigation, methodology, project administration, supervision, writing – original draft, writing – review and editing, validation. Elisabeth Tournier Lasserve: resources. Florence Riant: resources. Louis Lambert: formal analysis, software, validation.

Funding

The authors have nothing to report.

Disclosure

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Anonymized data are available from the corresponding author upon reasonable request.